Phil Pilgrim (PhiPi) said...
Cluso,
However, if this is just some Aussie/British rivalry that, as an American, I can never hope to understand, please carry on. I shall have to content myself with being entertained by it, I suppose.
-Phil
Of course you can hope to understand, Phil. Fortunately we have a dispute resolution process for such oz/uk quarrels. Unfortunately it only comes around every 4 years and is called "the ashes". If you want to understand it better, i'd start here with Warnie's "ball of the century" in 1993 www.youtube.com/watch?v=XNIUjI_LINE
Back to the topic at hand,
* The lack of code protection, no matter how cosmetic it is in other devices, instantly removes a large share of the available market. The best way of winning back some of that share might be to partner with other MCUs that do have some protection
* The video capabilities are too hard to grasp at first glance. Someone needs to round up all the available modes and have a demo where these are selected from a menu to highlight their capabilities literally at a glance. Back in the late 80's I remember a vesa test utility that cycled through all the non standard SVGA text modes one after another, we need something similar but for text+graphics.
* Parallax need a small product product with integrated uSD right next to/on top of the prop, and three video outputs (S-video, VGA, TV). The uSD is a sensible way to counter other products out there which have more flash onboard.
* One of the best aspects of using the prop is this forum. I would like to some way of linking OBEX to forums posts inside he Prop Tool. So when I right click on FullDuplexSerial I can hyperlink to latest version, and key thread posts. I realize this may require a bit of maintenance by Parallax, but no other software tool I use could do this and it would be a great point of differentiation
That'll do me. Agreed with many things others have said. I don't see price as an issue because the video capabilities lift the prop into another category of "system on a chip" and $5 difference is nothing in that game.
Phil Pilgrim (PhiPi) said...
....I do not find Leon's posts to be either off-topic or inappropriate to the subject at hand.....
Yes, I, too, find it's unfortunate that Leon and Cluso are in a quarrel of some sort. I know One-liner Leon's responses to forum questions can sometimes be so compact and so terse as to feel abrasive, but I'm always interested to hear what he's got to say. So far nothing Leon has informed us about has caused me to jump ship and move away from the Prop. I often look into the other chips Leon brings up, but so far I've not seen anything as user-friendly as the Prop. But maybe that's just because I'm a non-EE. In any case, what Leon's inputs have done for me is to provide a sense of confidence that, for my particular applications, the Prop is a pretty good choice.
So c'mon, Cluso. C'mon, Leon. You two shake hands and agree to disagree or whatever. Kiss and make up. Or whatever.
Thanks Andrew. I will be interested in the response you get.
Yes there is a lot of bias. It is in everything. We are much happier with what we know than what we don't.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔ Links to other interesting threads:
I will say one more thing, though. Parallax must do something to curb this "Arduino" craze. Arduino is being recognized everywhere as the official starter micro.
Over on the "ProDuino" thread we are now facing the cold hard factual details of what it takes to do Arduino stuff on a Propeller. Turns out that whilst the Prop is "obviously superior" to a dinky 8 bit ATMEGAxxx it is not so easy going for the Propeller in that space. Taking into account the features, the cost, the IDE etc.
That comparison is an interesting exercise and I don't think we are done with it yet.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
For me, the past is not over yet.
I have to say that, I agree with Phil on the reactions to Leon. Especially from Cluso99. While I would prefer that Leon not bring up other MCUs as often as he does, I think the open hostility in the forums reacting to him are worse. It creates a very unwelcome place to converse.
I avoided contributing to this thread in part because of it. I think it's better if you ignore Leon's posts that bother you, or address him in PMs with your displeasure.
Anyway, one thing that I think might contribute to more people giving the prop a shot is more proto board configurations. I would venture to say that most small embedded applications have no use for keyboard, mouse, or tv/vga displays. How about a proto board with an ADC built in. using I2C so it's just hanging off the eeprom pins. Perhaps one with a level translator for a bank of I/O pins? Something that makes it more trivial to connect to sensors and signals other stuff.
On another thread people are discussing a ProDruino or similar. If we had a Prop Proto board with a 4/8 channel ADC and a bank of 5v compatible I/Os (one level translator chip) then you gain pretty much all of the spirit of the Arduino. You just use PropTool with Spin/PASM and interface it to anything you want just as easily as Ardruino. I don't think we need to make it shield compatible or arduino IDE compatible at all. A lot of people use the Arduino because you can just poke a few wires into the I/O pins and connect them to switches or whatever nad make it do something tangible.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Check out the Propeller Wiki·and contribute if you can.
Roy is correct. People keep raising the video aspects of the Propeller on this thread. Nice as they are for the hobbyist, they are irrelevant to most embedded applications, and hence won't cut any ice with design engineers out there in the real world. In fact, they might be a turn-off - why waste silicon on something one will never use? Similarly with the built-in Spin interpreter.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Leon Heller
Amateur radio callsign: G1HSM
So if you must. I can make the prop do things your favourite Xmos chip cannot do. I can do things with a prop cheaper than an Intel i5. Is this really relevant? Specifically it is not to THIS thread.
Your comments have not gone unnoticed and they are not appreciated by others too! Please leave the forum.
I had a PM from Dave, I think it was, at Parallax, the other day, thanking me for my contributions!
Cluso99 said...
Here is another real life design example. I designed a centronics interface to an ICL minicomputer (~1981) and sold to ICL. It was a tiny pcb ~2"x3" and used a pair of the newly sampled MC68705P3S at $150 each !!! There were also 2 74xx and 2 transistors and a Coil on the pcb. My pcb plugged inside the printer and it replaced a ginormous pcb which came in it's own box ~24"x18"x6" and cost >$2,000. It used a cheap micro maybe worth ~$30. BTW a centronics printer was worth ~$5,000 or ICL's ~$10,000. Assembly was extremely simple, failures almost non-existent. I was still selling them for the same price 7 years later, even though (hope no ICL'ers reading this) the motorola chips came down to $8 ea. A reverse bell curve because it was a closed market.
Another thing here is that I could have done it much cheaper with a single conventional micro (motorola 6802) with more parts but still fit it into the space allowed. I chose to use two expensive single chip micros (internal 1.7KB eeprom) because I could seperate the code into two simple and distinct threads, one for the printer, one for the ICL. It made coding simpler and quicker time to market.
I did something similar where I once worked for a comms hub - I used a large PIC with four little AVRs interfaced to it via SPI, each acting as an intelligent, buffered, high-speed UART. We couldn't find a single chip device that would do the job, and my solution was cheap, and easily expanded if the need arose. I could have used those nice Maxim SPI UARTs, but they were very expensive.
Cluso99 said...
Leon said...
I don't know about that. I once worked on a Propeller-based project which was an ideal fit for the device. Unfortunately, we were only given about two weeks to deliver a working prototype, and all I was able to do in that time was show all the various functions running on their own cogs. They still needed integrating, and the company involved pulled the plug, costing my client a lot of money.
So now the truth comes out... You lost a client so you blame Parallax! Either you should have used another chip or the task was impossible. Either way, it's your fault, not the propeller chip & Parallax. Back to topic (well it's off topic now anyway)...
Where did I say I blamed Parallax? The blame really lies with the company my client was working for spending too long (I think they'd spent six months on it) trying to roll their own solution, giving up, and going to my client (who used to work for them) with an impossible time-scale when their customer was jumping up and down about no working system being forthcoming. I'd have had a fully working prototype in another week, thanks to the Propeller! Orders would have been into the thousands - one in every refrigerated lorry owned in Europe by the customer and replacing a very unreliable unit that was much more expensive, with fewer features. For instance, we were able to incorporate video at virtually no additional cost. Most of the cogs were used, it was a very neat design.
Leon
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Leon Heller
Amateur radio callsign: G1HSM
The Propeller is a brilliant chip but not perfect. Comparisons with other chips, positive and negative, will arise and can be expected in a thread which is about competing with the competition. Leon is not out of line in that respect. One can hide inconvenient truths but potential users will still see them.
What seems to be the case is that the answer to "can I do <this> using a Prop?" ( where a one-chip solution using something else will do ), has the answer, "Yes, if you do <this>, add <that>, jump through some hoops".
Or as that's received, "Yes it can be done, may result in a better product through doing so, may be more difficult, needs more, and is consequently more expensive". Some may see that as okay, but take away the necessity for it to be any better than an alternative, simply creating an equal, and one is left with, "Yes it can be done, may be more difficult, needs more, and is consequently more expensive".
One can come up with all sorts to offset that balance but what bottom line are potential users likely to see ? What are they likely to decide to do ?
What is the compelling reason to bang a square peg through a round hole when there's already a number of round pegs available ?
For what it's worth, it was *me* that hired Leon to write the software for the refrigerated-truck project.
We knew that the desired timescale was ridiculous - and told my customer this clearly and regularly, while
developing madly through night and day in order to complete the project as quickly as possible.
We prepared hardware/software demo's every couple of days to show progress.
Leon's mixed hardware and software background was fantastic. At our first meeting, he confirmed that the
Prop was a great fit for the task; we were helped by extensive use of the Obex and with answers from
the 'Forum that never sleeps'.
When I read his description 'most of the cogs were used, it was a very neat design' I was really thrilled. It
also featured an advanced 9-Amp buck-boost power supply because the customer's printer was so demanding.
What we didn't know is that there was a two-horse race going on, and an alternative outfit had been developing an
8051-based board for nine months. The last time I saw this, it was encased in glue, with numerous strap-modifications
and six huge capacitors hanging off of it to compensate for the tiny inductors. I'll never understand what was their
decision-making process, which seemed irrational/impulsive/unscientific.
It is ironic to recount here, but I intended to use Cluso's excellent RAM-blade for the successor requirement, which
demanded even more dataprocessing on the fly.
Another irony: I used Chip's 'Singing Seven' demo to illustrate the Prop's capabilities. The company's name has
the word 'Seven' in its title and what really sealed the choice of Prop for the controller was the thought that their
product could sing the name of the company on exhibition stands!
In business and even engineering, purchases are made with regard to "is it a solution to the problem" and emotional issues are generally back seat issues.· So, while we all love the propeller for what ever reasons, most of us are emotionaly attached and driven to the propeller.
Cost is a HUGE factor for electronic products - especially anything mass produced.· The financial decision between a $2.00 ATINY chip versus an $8.00 propeller isn't even a choice of consideration.· Bottom line keeps a company going, and in the world of mass produced electronics, little costs add up very fast.
The right tool for the job is another huge factor in a decision like this.· Personally I don't think the Propeller is the end all cure all of micro controller needs.· It is a very good solution for projects that require muti-tasking.· Very few projects actually utilize this feature.· Most people use the various cores to act as native, hardware based peripheral solutions found on other micros.· This methodology works and works fine, but an existing hardware solution on another micro is tough to beat in performance and reliability. I cannot say that those hardware solutions are as easy to use though.
Just for the sake of a thought experiment, think about all the projects posted here on the forums and in the contests. Consider how many of them actually needed muti-tasking and how many just used the extra cogs to handle some other task that could have been handled by hardware peripheral in another micro.·
Then there is the issue of programming language.· I have been working with ATMEGA chips for a couple years and I use BASCOM to program them.· It didn't take me long at all to learn how to program these chips even though I had no microcontroller experience.· I suspect that a programmer used to programming in C for 20 years isn't all that excited about working in another language, especially when his boss is complaining he has already spent too much time on the project.
Anyway, back to solutions... A business purchase is solution driven -you have a problem and you·need an ecomical solution.·A micro is the solution - which micro, the one that provides the cheapest and most efficient means to solve the problem.· Cheapest, that one is pretty easy to identify.· The most efficient means to solve the·problem - no so easy to identify.· The most efficient·from a programmer's view point is the one that he knows the best and if he (or she) has been in the industry a long time already has experience with some brand.· Having that experience,and probably a supporting library of·software tools usually directs you to the most efficient solution.·
Don't get me wrong, I like the propeller chip, I like spin, I love PASM.· But I would never use it for·projects where I have already used BASCOM and ATMEL chips.··Those chips can be cheaper and I already have a decent library of solutions (I2C, SPI,·LCD, PWM, etc. etc.).· I don't use just one model of Atmel chips either, I use the one·with the·features that match the problem, for instance, I don't use an ATMEGA2560 for everything·- it is too expensive!· However, it·is the ideal choice for some projects I have done.
So, frankly, I don't think you are going to see the Propeller be in the same choice category as Atmega, Atiny's, or the PIC micros.· The decision to go with a Propeller for me was all about the peformance advantages of the multi-tasking capabilies of the 8 cogs.· That is what is unique and that is what has to be marketed as the solution.
Is a real shame that you have locked horns with Leon. You both have fabulous knowledge about many different processors etc. What strikes me is that these threads are being read by for example, people in technical education that might wish to use the Propeller as a teaching tool. For these little spatts to continue for much longer would see this to be of damage to Parallax and the excellent goals they have worked hard and continue to reach for.
Should we count on our fingers the talanted people that no longer post here? Life is often about dealing with abrasive people in the most appropriate place.
heater said...
As you know I do like this Prop as "System On a Chip" approach to it's promotion.
As I have said before we should drop the language of COGs and HUB etc. Use terms the uninitiated would immediately recognize and feel comfortable with.
Talk about "peripheral blocks" and flexibility not COG's. HUB RAM is just "RAM"
Such a collection of peripheral drivers, FullDuplexSerial, SD card, I2C, SPI etc define a variant of the Prop "System On a Chip". Same chip, just comes out of the box with that selection of drivers ready to go. Wish I could find my post suggesting the Prop as "System On a Chip" again.
I'm actively working on a solution for this.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
There are 10 types of people in the world: those who understand binary, and those who don't
Apologies for another post on the secondary theme of this thread--the tension between Cluso and Leon--but I feel compelled to "chip-in" in defense and admiration of them both.·
A few months back, Leon seemed to be heavily pushing the XMOS, even seemingly ramming it together with a Prop to justify talking about it (an over-simplification, I'm sure), which resulted in Cluso understandably "blowing a gasket" (understandable because I think there were others that felt the same way and, after all, this is principally a Prop forum)·and calling him out on it.· At the time, I appreciated Cluso's honesty for speaking out...because I, too, can often get frustrated when I think people aren't trying to be cooperative or whatever, and it made me feel good to realize that I'm not the only one with a temper.· However,·I don't think that Cluso has a hair-trigger by any means, but rather had reached a breaking point over a period of time, causing him, if I recall correctly,·to even·suggest that the Forum Administrator take action to suppress Leon.
But what happened?· Well, clearly Leon was not banned.· Parallax in their wisdom did not see fit to take such a harsh action against a contributor.· But there was a change!· Leon stopped pushing the XMOS so hard and barely ever mentioned it afterwards until recently.· That means that either [noparse][[/noparse]1] LEON reconsidered his approach and voluntarily limited himself so as to avoid giving offense (whether real or imagined), or, [noparse][[/noparse]2] that Parallax--though not banning Leon--asked him if there was a way for him to kind of keep-the-peace out of their respect for Cluso.· One or the other happened, I believe, or both,·but either way, Leon adjusted his behavior because he values this forum.· He may have gotten a little bolder recently in his enthusiasm to educate forum users on where the Prop stands in comparison to other chips, but he definitely (based on my perception) curtailed his mentioning of the XMOS after Cluso's protest.· That means that a sort of compromise was already indirectly reached between Leon and Cluso.· But now, it appears, that some of that·"bad blood" (maybe only felt on one side, I don't know) from the past has entered in again into recent threads.· But I think it's important to recognize that Leon already clearly changed his behavior, even if he didn't feel he was in the wrong, and that is commendable.· In summary, [noparse][[/noparse]A] it was understandable that Cluso got riled, and, [noparse][[/noparse]B]·Leon made an effort to change.· Well, that concludes my defense remarks, though neither retained me as arbitrator.
So on to the admiration, starting with Cluso.· Regular forum readers can probably readily cite 5 to 10 names of frequent posters to this forum that make outstanding contributions to the forum, both in terms of comments as well as code and hardware.· Cluso is definitely in that group and no doubt widely admired and appreciated.· He constantly shows his willingness to make suggestions when people have problems or ideas, creates innovative ideas and applications for the Prop (both in H/W and S/W), and is a kind of all-around evangelist for the Propeller, not to mention that he brings a vast array of professional experience to this forum (as do many others).· Cluso would be sorely missed if absent from this forum.· I think we can tolerate an atypical outburst or two from him, and I think that it would be madness for anyone to purposefully agitate him, comparable to shooting oneself in the foot.
But Leon also brings value, often of another kind, to this forum.· In particular, he frequently helps us put the Prop in perspective amongst the world of chips in which it resides.· Personally, I don't feel right about choosing a solution if I haven't made some effort to consider alternatives.· But I, for one, lack professional experience in this field.· Sure, I can search the Internet quite well, but that's no substitute for the experience of others (which can also be had by searching).· For me, Leon provides·some of the experience and perspective·that I would otherwise be at least partially missing.· I really value that.· Furthermore, Leon typcially packs as much into a two-line post as is semantically possible.· I've learned a lot from him.· He's even set me straight a time or two when I've been wrong or one-sided.· Sure, his remarks are often terse, but they are filled with gems that come from experience.· In that I'm not a professional in the microcontroller field, this forum is a big part of my "education," and Leon's posts are a significant part of that.· I'd miss his insights if he were gone, both in terms of his general electronics knowledge and in terms of his insights into other microcontrollers.··Once again, he just packs so much information into a two-line·(or so)·post.·
Lastly, the Propeller can't avoid being compared to other chips out there.· And the Propeller has no reason to avoid comparison.· It's a fabulous chip!· But it's just like job applicants:· some are right for some jobs but not for others.· However, it's really only by comparison that we can make informed choices.· I kind of feel bad that I don't often read the forums for other chips, I like our beloved Propeller enough as it is.· But I'm glad tha·I can learn about the strengths and, yes, weaknesses of the Propeller for various tasks via some comparisons to other chips on this forum.· Do I want this forum to only do comparisons and not focus on the Propeller applications?· No.· But I like·this forum·basically the way it is:· fresh and dynamic, with comparisons where appropriate.· What we are lucky to have in Cluso and Leon are two experienced contributors that make it easier for us to know "how" and "when" to apply the Prop.· They are really on the same side.· And Leon's above-mentioned change-of-behaviour speaks volumes to me·(as it is truly difficult to change), showing that he values this forum but doesn't blindly use the Propeller.· I click on the links that he occassionally provides to learn more about other possible solutions.· Like another poster said, his links haven't disuaded me from using the Propeller, but I think they help me to use it more intelligently.· And Cluso's contributions are simply invaluable and too numerous to detail.· We need them both!·
I am going to answer the Leon problem once and for all. You do not see me getting upset with anyone else who mentions another chip legitimately - and there haven't been any to my knowledge, despite some thinking they may get my rage if they mention another multicore cpu. Leon mentions another chip at every opportunity,·adds links, etc·and puts the prop down also at every opportunity. I am absolutely not alone here, and I can say that Parallax is not amused either, no matter what Dave may have said.
I am one who can put it behind me immediately after I have said my piece. I do acknowledge that there are times when comparisons will be made and there are times when they are not called for. This thread did not call for it. I asked what would entice design engineers, not what was wrong with the prop or Parallax, but this is where the thread evolved.
Ok I have said my piece, lets move on. We are so off-topic we may as well continue along the lines the thread has taken.
The prop chip is ~$6 per 1,000. Since the prop is not a high moving line at the distributors like DigiKey, etc, they probably have to be given a higher margin than say a PIC. Notice how quickly the prop reduces in volume to say a PIC. Now if you compare a PIC for $2 then we are talking apples and oranges and in most circumstances and even I would use a PIC, but there are some·instances where·I would still use a Prop (BTW I am). So lets say the margin was 30% (not real). So the distributor would buy for $1.33 and sell for $2.00 and the prop would buy for $4.00 and sell for $6.00. Parallax cannot undercut it's dealers, but you are really comparing a $1.33 chip to a $4.00 chip if you see what I mean. BTW we have to add an eeprom which IIRC a 32KB can be had for ~50c in quantity. I am sure you would get a great deal if you went to Parallax with a firm order for 10,000 pcs you would get a great price. An order for 100,000 and Chip would deliver them to you personally (No Chip - get back to the Prop II!).
I raised an idea on another thread about booting from·SD (or microSD) without an eeprom. How this would be done is that a new ROM mask would be required. I am unsure of it's cost and I am presuming it is a traditional mask and not a hardwired thing. However, some piece of ROM must go and that was not liked on the other thread. Everyone is using something. Just to refresh what's in the ROM, 16KB for the character font, 8KB for the Log tables, $4KB+2 for the Sine and 4KB-2 for the Interpreter and Booter. The Booter has to a have the option to boot from SD card added which would include some low level SPI access which would probably take 1KB or maybe even less if it was in spin (it could be). There would·be room for the faster Interpreter which would then have space for the LMM extensions. We may even have room for 2KB 8x8 character font too, depending on what was removed. Or maybe we could get a Basic Interpreter in there, not that this is any use for volume sales.
If someone had the killer app, they could have security with their own mask without eeprom. You will not get that from a big company these days. No chip is totally secure though. BTW I have not confirmed Parallax could/would do it, but it's an idea for those with volume.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔ Links to other interesting threads:
I am not trying to hijack this forum anymore than it already has....Sorry Cluso. But i think Heater and i have a valid point. The propeller is bending the knee and trying to please the Arduino lovers. This is kinda like Americans reverting to communism(Okay, not really, haha) I think Parallax should very carefully consider their competition. The chess game players are starting to switch sides. Parallax is now taking fire from Arduino on the education front and they can't even compete in the manufacturing business. I think Parallax has to decide what is their primary goal. Hobbyists or Business's. I think a strong C++ programmable Prop would greatly help the company, regardless of it's target audience......
I'd really value the ability to boot from SD w/o the EEPROM, even though the EEPROM is cheap and small. It's just more elegant, assuming that one's design needs SD for other purposes, I mean. I look forward to such in any follow-on chip (i.e., the Prop II). Also, it's too bad that other process technologies apparently aren't readily available for the Prop. Is it really only because of the design tools that Parallax uses? Could they readily use or switch to others? Maybe bring on other expertise and/or out-source that. Are "translators" available to convert designs between various tools? Also, if FLASH could be used in a future Propeller design (which the current process technology doesn't permit, of course, as we've been told), could such FLASH work as fast as mask ROM? Is there something special in terms of connections to 8 cores that alters things? I assume such memory just goes through some kind of "multiplexer" (selector) to connect it to the currently selected cog (but I don't know). What I'm getting at is, wouldn't it be nice if a future Prop could be flashed with various versions of firmware! But I suppose something might prevent that, just don't know the details. At any rate, being able to conveniently and quickly access an SD card would relieve some of the need for different firmware as, for example, alternate character sets (data in this case rather than code, though perhaps code could be delivered that way, too) could likely be stored on SD (assuming that they could be read in fast enough for the driver that was using them). Anyway, if any of what Cluso suggested is possible, I think it's worth considering. I think the advantages of having just a single Prop that everyone can target have passed. That might have been important in the beginning, but the Prop is much more "established" now. Anyway, if indeed firmware has to be frozen in mask ROM from the manufacturer (in addition to just basic boot-up code), then that's always going to be a limiting factor, isn't it, unless one has RAM to burn to load in a variant? Well, in the case of the Prop II, having a larger address space and much more ROM at least partially alleviates some of the concern. Anyway, Parallax has to live in the real world where things need to be practical, so of course some mix of compromises will always be called for. I'm just curious about the particulars, not being schooled on the details of process technology and what not.
Post Edited (JRetSapDoog) : 7/16/2010 5:49:40 PM GMT
It's more a factor of the semiconductor process than the tools. The developers of the-chip-which-must-not-be-named have stated that they could include flash on the chip, but they would have to use the TSMC 125 nm process, IIRC, instead of the 90 nm and 45 nm processes they are using. They would do it if there was sufficient demand, though.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Leon Heller
Amateur radio callsign: G1HSM
Perhaps you should hyphenate that, Leon: the-chip-which-must-not-be-named. Or perhaps you could use an acronym: TCWMNBN, lol, as opposed to, you know, its real 4-letter name. Anyway, when I first read your comment, I thought that you meant the creators of the Prop or future Prop shouldn't be named, rather than the name of a competing chip.
So, that competing chip doesn't have FLASH? I've read about it, but forgotten the details. But it could have, you're saying, if they switched to the TSMC process, and so perhaps a future Prop could also. Seems like I recall that the next Prop is still several process sizes larger than any of the processes you've mentioned (though my memory is fuzzy on that). Perhaps that's to reduce leakage current, though they've said the new chip will be much more power-hungry than the current Prop, making it another kind of beast. For my intended usage, I don't need days, weeks or months of operation on battery power, so I welcome the more powerful chip.
Also, I've read that die-size is a concern with the new design, so I don't understand why smaller process technology isn't being used, but I've kind of just assumed that it has to do with either power consumption or design tools. BTW, does anyone know which IC fab Parallax uses for the Prop? I know it's in Taiwan (or at least their packager is), but don't know which.· Anyway, that still leaves open my question about whether FLASH-based firmware could run as fast as ROM-based firmware and/or be compatible with the Prop's unique architecture, if that is at all a factor.
Regarding the process tools Parallax uses, I can certainly understand how a small company must use the tools they have and know. That's what I'd do if I had a unique design, even if there existed "better" tools or process technology. I'd do it that way because I'd know that I'd just be lucky to get a working design out the door, and I'd hope that its unique features would carry it beyond any limitations that the process technology introduced. However, this time around is not Parallax's first chip effort. And I think it's pretty clear that they are on the edge of greatness with the new one. We're all chomping at the bit for it, and professionals will be, too, once they see how the unique abilities of the current Prop have scaled up so much. As such, perhaps Parallax can't just default to taking the known route of working with the tools they have and know. That is, perhaps it is worth considering other process technologies/tools. I don't mean to suggest that they haven't done that. Of course they have, just that it might be worth taking another look at.
Of course, I don't want to see anything significantly slow down the release of a new chip, but I hope it can be as great as it can be. RAM memory still seems to be one of the biggest limitations. It's a real pity that that seems to be unavoidable. But anyway, I'll sure be happy to get anywhere near the 256KB they're striving for, even if 4X would be a killer in terms of one-chip designs for certain applications. Or perhaps a POP chip could be done, though I wouldn't especially welcome losing a lot of I/O's to accommodate that.
Even if Parallax is widely successful with their next chip, I don't see them wanting to become mostly a chip company, per se, in such a competitive market. I still think they should be the "ease-of-use" company. That leads me to wonder if they could partner with an established chip-design company (one like Microchip or the maker (XXXX, one of those X's being correct) of the-chip-which-must-not-be-named (TCWMNBN); after all, there are some simularities. Of course, if they did that, they'd have to do it in such a way that they maintained management and creative control...because Parallax's mission includes being a user-friendly, socially-responsible and not-purely-focused-on-profit company, one that encourages the creativity of others in a win-win kind of way (can't really say that for the other companies; it's only about profit for them). But one has to partner these days on some level. For example, very few companies can have their own fabs. Therefore, it might be something to consider and could bring significantly more resources to bear...and bring us a line of truly great products.
Having said that, what's coming down the pipe from their own resources looks pretty incredible in and of itself. And, it seems, Parallax is even more open to a Prop 1.5 or similar. So maybe various kinds of actions will be happening in the future. That 3C mini-board (of another thread) sounds quite interesting and practical, for example.
Post Edited (JRetSapDoog) : 7/16/2010 7:00:13 PM GMT
I don't think that Parallax has mentioned whose fab they use, but it doesn't sound like TSMC. They are the biggest, and one of the few that were making a profit. The others are probably coining it at the moment, because of the chip famine. There have been rumours that some of their customers have been paying extra to jump the queue.
I think that the on-chip flash would just load programs into RAM, for speed. There are tricks that speed up execution from flash, like the wide interface that NXP uses on their ARM chips, but execution from RAM will always be fastest.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Leon Heller
Amateur radio callsign: G1HSM
RossH said...
But of course the tools should really be offered (or at least backed) by Parallax themselves.
Exactly so, and this is why IMO Parallax should pay you something so
you can continue to work on Catalina. In this way the upgrades to Catalina
could be offered to users for free. The optimizing version of Catalina you
mentioned could then be freely downloaded. You can't be expected to toil
endlessly on Catalina without some $ support.
Tubular said...
The video capabilities are too hard to grasp at first glance. Someone needs to round up all the available modes and have a demo where these are selected from a menu
Tubular said...
a small product product with integrated uSD right next to/on top of the prop, and three video outputs (S-video, VGA, TV)
Great ideas
It seems to me it should be possible to repackage a sm prop
inside a plastic module that would look like a normal ic.
Put a sm prop, a small 5 or 6.25 xtal, an eeprom, flash memory
(uSD or whatever), sm resistors..etc. Bring out to pins the video
signals for easy use.
I think many here could build this simple little board, all we
need is a way to package it in a plastic module with pins on it
I for one have a dislike for the look of those little dil boards
where you can see all the parts....it just looks more professional
to encapsulate in plastic.... (and leave off the cute propeller beanie image)
Flash and ROM would work the same way. But Flash takes a much larger die size (more transistors) and therefore more expensive, and the fab house used cannot do it anyway. Added to this, Parallax do not have the resources to do this.
A ROM change may be easy though. What I envisage is a new ROM part that can boot from SD if the eeprom does not exist. So it is backwards compatible except that it would have something in the ROM removed. That is the real stumbling block - what to remove... Character font 16x32 and replace with multiple smaller 8x8 fonts and other things too (because we gain half the ROM space back); Log tables; Sin tables. They are the only choices available. Anythig else is currently off the table.
Unfortunately if we ditch the font, then it's gone because there is not enough space in hub to replace it. However, the 4K for Log or 4K for Sin users may have enough hub space. Alternately, Parallax could keep both chips with an inventory cost penalty. There are a few other things that can suport this concept and they have been communicated to the powers to be. It is a shame that we cannot just add more ROM but it isnt going to happen.
BTW This is not for an Arduino replacement although with the extra SD, VGA/TV, Kbd, etc it may be able to compete in an upmarket way. It is more for the volume users. The prop is an upmarket chip and so SD in very likely to be included in that killer app (IMHO anyway).
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔ Links to other interesting threads:
on page 3, I said...
Being that I interact with product engineers and design engineers almost on a daily basis, I can be open and honest with many. I'll throw some questions out to a few and see what I can get back to be posted here.
I have some feedback already:
Company A (market specific robotics manufacturer, field proven product lines, 4 product lines/~100 per month)
Most of their products are made with Atmel microcontrollers. Cost is not considered in their component selection process because the electronics (PCBs) account for about 5% of the total cost of the robot. Their main selection criteria was based upon using a single family of microcontrollers to cover all of their needs. Some of their boards are extremely simple (IE: read a sensor and relay data via serial I/O line back to the main baord) and some are very complex (IE: host system managing the entire system with various protocols such as I2c, one-wire, serial, Ethernet, USB, etc. It also drives a human interface on an LCD/keypad that is duplicated via a proprietary RF connection). Being able to use a single family of microcontrollers makes it easier to manage all programming efforts with staff that works as a well oiled machine.
Company B (very small company with project specific marketing products such as store displays, 8 product lines ranging from 50 to 4,000 per month)
Products are heavily burdened with a "low-cost" push, so the cheapest microcontroller that can do the job is selected. The company typically use Cypress PSoC when possible because of the low cost and the head engineer has the most experience with the PSoC which means the work can be done quickly and in house. For more complex projects, the PSoC coding is outsourced to a long time used programming partner. Other jobs have had their "code" portion outsourced at the beginning, and the vendor is allowed to use whatever microcontroller they choose to get the job done as long as it fits a cost target.
Company C (wireless sensing products for a specific target market, a few year old startup with some very experienced engineers, 18 product lines, a couple around 20 per quarter, but most around 300-500 per month)
This companies product line is directly affected by power consumption as the units are battery operated, but transmit data every 5 minutes. They chose to pair TIs MSP430 controller with the TIs CC2420 2.4 GHZ radio chip for their product. The product has performed very well in the market for the past 2 years, but other new designs from them are branching out into other microcontrollers to meet new needs such as Ethernet to serial, power outlet monitoring, etc. Cost has not been a major factor, but definitely considered. For low power usage, it seems that TIs MSP430 has the top spot.
Company D (legacy modem technology for industrial applications, very old company/product line, 6 product lines, 120-200 per quarter)
Even though this company could easily re-design their products using different microcontrollers, their products have a field proven design and very low profit margin. Most of their products built are for field repair of installs that are 15+ years old. Their products use various microcontrollers and the "why" is anyone's guess. (most of the original design engineers have passed on)
Company E (retail beverage equipment products, solid company, new products each year, 7 product lines, 300-1000 per month)
Microcontroller choices are left to the design engineers that are assigned product lines. They have products with Atmel AVR and Microchip PICs. They have a few products that use Qtouch chips and those all use the Atmel Atmega128.
So, it seems the choice is very much driven by the needs of product/project more so than something that can be said that encompasses all of them into a group so that one item can be hailed as the largest concern. All but Company B, have the appropriate software packages and tools for the products because that is a factor of bringing a product to market and sustaining it. Having to pay for these tools is expected, so even though it's a plus by getting free tools, some have purchased other software tools to help them anyways. In all cases except where the PSoC is used, programming in C appears to be the champion. On the PSoC, it is split between the free PSoC designer software and manually coding in C. I asked a few about using a chip that required an external EEPROM and most did not care other than that meant more PCB real estate when compared to a part that had internal flash. The only place that I see that the Propeller could not be used is with Company A that chose a microcontroller family to address duties that the Propeller cannot perform such as full-blown ethernet.
With all that said, my formal response to the thread title's question would be this:
Having a C compiler that is equivalent to those being used for the microcontrollers listed above will increase the Propeller's ability to gain marketshare.
Any chance of getting volumes (order of magnitude approximations would be fine) for each of your companies A - E? It would be interesting to see how the volume affects the decision making process.
Ross.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Catalina - a FREE C compiler for the Propeller - see Catalina
I added volumes to the descriptions next to the company. By "product lines", I mean either a PCB or a completed assembly that uses at least one PCB that has a microcontroller. We do have some assemblies here that have more than one PCB that gets assembled into an enclosure.
Here's comparable information for my company - let's call it "company F":
Company F (distributed telemetry equipment for a specific industrial market, about 10 years experience in this market with some very experienced engineers and 4 product lines - one around 10 per quarter, but the others around 10,000 - 20,000 units per annum each (we tend to get large annual orders and do one or two production runs each year). The main product line is directly affected by power consumption as the units are battery operated, and expected to last in the field for up to 5 years with no maintenance. They sample data every 6 minutes and transmit by ISM band radio, PSTN or GSM dialer periodically (typically once per day). Company policy means I can't list you the exact chips we use, but the processors are mostly ATmega. We use C and buy professional compilers and other tools - but we use the manufacturer's "freebie" tools when evaluating new components, and also for various non-critical development tasks.
Ross.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Catalina - a FREE C compiler for the Propeller - see Catalina
Comments
Of course you can hope to understand, Phil. Fortunately we have a dispute resolution process for such oz/uk quarrels. Unfortunately it only comes around every 4 years and is called "the ashes". If you want to understand it better, i'd start here with Warnie's "ball of the century" in 1993 www.youtube.com/watch?v=XNIUjI_LINE
Back to the topic at hand,
* The lack of code protection, no matter how cosmetic it is in other devices, instantly removes a large share of the available market. The best way of winning back some of that share might be to partner with other MCUs that do have some protection
* The video capabilities are too hard to grasp at first glance. Someone needs to round up all the available modes and have a demo where these are selected from a menu to highlight their capabilities literally at a glance. Back in the late 80's I remember a vesa test utility that cycled through all the non standard SVGA text modes one after another, we need something similar but for text+graphics.
* Parallax need a small product product with integrated uSD right next to/on top of the prop, and three video outputs (S-video, VGA, TV). The uSD is a sensible way to counter other products out there which have more flash onboard.
* One of the best aspects of using the prop is this forum. I would like to some way of linking OBEX to forums posts inside he Prop Tool. So when I right click on FullDuplexSerial I can hyperlink to latest version, and key thread posts. I realize this may require a bit of maintenance by Parallax, but no other software tool I use could do this and it would be a great point of differentiation
That'll do me. Agreed with many things others have said. I don't see price as an issue because the video capabilities lift the prop into another category of "system on a chip" and $5 difference is nothing in that game.
tubular
Yes, I, too, find it's unfortunate that Leon and Cluso are in a quarrel of some sort. I know One-liner Leon's responses to forum questions can sometimes be so compact and so terse as to feel abrasive, but I'm always interested to hear what he's got to say. So far nothing Leon has informed us about has caused me to jump ship and move away from the Prop. I often look into the other chips Leon brings up, but so far I've not seen anything as user-friendly as the Prop. But maybe that's just because I'm a non-EE. In any case, what Leon's inputs have done for me is to provide a sense of confidence that, for my particular applications, the Prop is a pretty good choice.
So c'mon, Cluso. C'mon, Leon. You two shake hands and agree to disagree or whatever. Kiss and make up. Or whatever.
Yes there is a lot of bias. It is in everything. We are much happier with what we know than what we don't.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Links to other interesting threads:
· Home of the MultiBladeProps: TriBlade,·RamBlade,·SixBlade, website
· Single Board Computer:·3 Propeller ICs·and a·TriBladeProp board (ZiCog Z80 Emulator)
· Prop Tools under Development or Completed (Index)
· Emulators: CPUs Z80 etc; Micros Altair etc;· Terminals·VT100 etc; (Index) ZiCog (Z80) , MoCog (6809)·
· Prop OS: SphinxOS·, PropDos , PropCmd··· Search the Propeller forums·(uses advanced Google search)
My cruising website is: ·www.bluemagic.biz·· MultiBlade Props: www.cluso.bluemagic.biz
Over on the "ProDuino" thread we are now facing the cold hard factual details of what it takes to do Arduino stuff on a Propeller. Turns out that whilst the Prop is "obviously superior" to a dinky 8 bit ATMEGAxxx it is not so easy going for the Propeller in that space. Taking into account the features, the cost, the IDE etc.
That comparison is an interesting exercise and I don't think we are done with it yet.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
For me, the past is not over yet.
I avoided contributing to this thread in part because of it. I think it's better if you ignore Leon's posts that bother you, or address him in PMs with your displeasure.
Anyway, one thing that I think might contribute to more people giving the prop a shot is more proto board configurations. I would venture to say that most small embedded applications have no use for keyboard, mouse, or tv/vga displays. How about a proto board with an ADC built in. using I2C so it's just hanging off the eeprom pins. Perhaps one with a level translator for a bank of I/O pins? Something that makes it more trivial to connect to sensors and signals other stuff.
On another thread people are discussing a ProDruino or similar. If we had a Prop Proto board with a 4/8 channel ADC and a bank of 5v compatible I/Os (one level translator chip) then you gain pretty much all of the spirit of the Arduino. You just use PropTool with Spin/PASM and interface it to anything you want just as easily as Ardruino. I don't think we need to make it shield compatible or arduino IDE compatible at all. A lot of people use the Arduino because you can just poke a few wires into the I/O pins and connect them to switches or whatever nad make it do something tangible.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Check out the Propeller Wiki·and contribute if you can.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Leon Heller
Amateur radio callsign: G1HSM
Post Edited (Leon) : 7/16/2010 8:37:26 AM GMT
I had a PM from Dave, I think it was, at Parallax, the other day, thanking me for my contributions!
I did something similar where I once worked for a comms hub - I used a large PIC with four little AVRs interfaced to it via SPI, each acting as an intelligent, buffered, high-speed UART. We couldn't find a single chip device that would do the job, and my solution was cheap, and easily expanded if the need arose. I could have used those nice Maxim SPI UARTs, but they were very expensive.
Where did I say I blamed Parallax? The blame really lies with the company my client was working for spending too long (I think they'd spent six months on it) trying to roll their own solution, giving up, and going to my client (who used to work for them) with an impossible time-scale when their customer was jumping up and down about no working system being forthcoming. I'd have had a fully working prototype in another week, thanks to the Propeller! Orders would have been into the thousands - one in every refrigerated lorry owned in Europe by the customer and replacing a very unreliable unit that was much more expensive, with fewer features. For instance, we were able to incorporate video at virtually no additional cost. Most of the cogs were used, it was a very neat design.
Leon
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Leon Heller
Amateur radio callsign: G1HSM
Post Edited (Leon) : 7/16/2010 9:40:27 AM GMT
What seems to be the case is that the answer to "can I do <this> using a Prop?" ( where a one-chip solution using something else will do ), has the answer, "Yes, if you do <this>, add <that>, jump through some hoops".
Or as that's received, "Yes it can be done, may result in a better product through doing so, may be more difficult, needs more, and is consequently more expensive". Some may see that as okay, but take away the necessity for it to be any better than an alternative, simply creating an equal, and one is left with, "Yes it can be done, may be more difficult, needs more, and is consequently more expensive".
One can come up with all sorts to offset that balance but what bottom line are potential users likely to see ? What are they likely to decide to do ?
What is the compelling reason to bang a square peg through a round hole when there's already a number of round pegs available ?
We knew that the desired timescale was ridiculous - and told my customer this clearly and regularly, while
developing madly through night and day in order to complete the project as quickly as possible.
We prepared hardware/software demo's every couple of days to show progress.
Leon's mixed hardware and software background was fantastic. At our first meeting, he confirmed that the
Prop was a great fit for the task; we were helped by extensive use of the Obex and with answers from
the 'Forum that never sleeps'.
When I read his description 'most of the cogs were used, it was a very neat design' I was really thrilled. It
also featured an advanced 9-Amp buck-boost power supply because the customer's printer was so demanding.
What we didn't know is that there was a two-horse race going on, and an alternative outfit had been developing an
8051-based board for nine months. The last time I saw this, it was encased in glue, with numerous strap-modifications
and six huge capacitors hanging off of it to compensate for the tiny inductors. I'll never understand what was their
decision-making process, which seemed irrational/impulsive/unscientific.
It is ironic to recount here, but I intended to use Cluso's excellent RAM-blade for the successor requirement, which
demanded even more dataprocessing on the fly.
Another irony: I used Chip's 'Singing Seven' demo to illustrate the Prop's capabilities. The company's name has
the word 'Seven' in its title and what really sealed the choice of Prop for the controller was the thought that their
product could sing the name of the company on exhibition stands!
I kid you not, this was the 'killer app'!
Regards,
T o n y
Cost is a HUGE factor for electronic products - especially anything mass produced.· The financial decision between a $2.00 ATINY chip versus an $8.00 propeller isn't even a choice of consideration.· Bottom line keeps a company going, and in the world of mass produced electronics, little costs add up very fast.
The right tool for the job is another huge factor in a decision like this.· Personally I don't think the Propeller is the end all cure all of micro controller needs.· It is a very good solution for projects that require muti-tasking.· Very few projects actually utilize this feature.· Most people use the various cores to act as native, hardware based peripheral solutions found on other micros.· This methodology works and works fine, but an existing hardware solution on another micro is tough to beat in performance and reliability. I cannot say that those hardware solutions are as easy to use though.
Just for the sake of a thought experiment, think about all the projects posted here on the forums and in the contests. Consider how many of them actually needed muti-tasking and how many just used the extra cogs to handle some other task that could have been handled by hardware peripheral in another micro.·
Then there is the issue of programming language.· I have been working with ATMEGA chips for a couple years and I use BASCOM to program them.· It didn't take me long at all to learn how to program these chips even though I had no microcontroller experience.· I suspect that a programmer used to programming in C for 20 years isn't all that excited about working in another language, especially when his boss is complaining he has already spent too much time on the project.
Anyway, back to solutions... A business purchase is solution driven -you have a problem and you·need an ecomical solution.·A micro is the solution - which micro, the one that provides the cheapest and most efficient means to solve the problem.· Cheapest, that one is pretty easy to identify.· The most efficient means to solve the·problem - no so easy to identify.· The most efficient·from a programmer's view point is the one that he knows the best and if he (or she) has been in the industry a long time already has experience with some brand.· Having that experience,and probably a supporting library of·software tools usually directs you to the most efficient solution.·
Don't get me wrong, I like the propeller chip, I like spin, I love PASM.· But I would never use it for·projects where I have already used BASCOM and ATMEL chips.··Those chips can be cheaper and I already have a decent library of solutions (I2C, SPI,·LCD, PWM, etc. etc.).· I don't use just one model of Atmel chips either, I use the one·with the·features that match the problem, for instance, I don't use an ATMEGA2560 for everything·- it is too expensive!· However, it·is the ideal choice for some projects I have done.
So, frankly, I don't think you are going to see the Propeller be in the same choice category as Atmega, Atiny's, or the PIC micros.· The decision to go with a Propeller for me was all about the peformance advantages of the multi-tasking capabilies of the 8 cogs.· That is what is unique and that is what has to be marketed as the solution.
Chris
Is a real shame that you have locked horns with Leon. You both have fabulous knowledge about many different processors etc. What strikes me is that these threads are being read by for example, people in technical education that might wish to use the Propeller as a teaching tool. For these little spatts to continue for much longer would see this to be of damage to Parallax and the excellent goals they have worked hard and continue to reach for.
Should we count on our fingers the talanted people that no longer post here? Life is often about dealing with abrasive people in the most appropriate place.
This is still a very good thread however.
Regards
Mike.
I'm actively working on a solution for this.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
There are 10 types of people in the world: those who understand binary, and those who don't
A few months back, Leon seemed to be heavily pushing the XMOS, even seemingly ramming it together with a Prop to justify talking about it (an over-simplification, I'm sure), which resulted in Cluso understandably "blowing a gasket" (understandable because I think there were others that felt the same way and, after all, this is principally a Prop forum)·and calling him out on it.· At the time, I appreciated Cluso's honesty for speaking out...because I, too, can often get frustrated when I think people aren't trying to be cooperative or whatever, and it made me feel good to realize that I'm not the only one with a temper.· However,·I don't think that Cluso has a hair-trigger by any means, but rather had reached a breaking point over a period of time, causing him, if I recall correctly,·to even·suggest that the Forum Administrator take action to suppress Leon.
But what happened?· Well, clearly Leon was not banned.· Parallax in their wisdom did not see fit to take such a harsh action against a contributor.· But there was a change!· Leon stopped pushing the XMOS so hard and barely ever mentioned it afterwards until recently.· That means that either [noparse][[/noparse]1] LEON reconsidered his approach and voluntarily limited himself so as to avoid giving offense (whether real or imagined), or, [noparse][[/noparse]2] that Parallax--though not banning Leon--asked him if there was a way for him to kind of keep-the-peace out of their respect for Cluso.· One or the other happened, I believe, or both,·but either way, Leon adjusted his behavior because he values this forum.· He may have gotten a little bolder recently in his enthusiasm to educate forum users on where the Prop stands in comparison to other chips, but he definitely (based on my perception) curtailed his mentioning of the XMOS after Cluso's protest.· That means that a sort of compromise was already indirectly reached between Leon and Cluso.· But now, it appears, that some of that·"bad blood" (maybe only felt on one side, I don't know) from the past has entered in again into recent threads.· But I think it's important to recognize that Leon already clearly changed his behavior, even if he didn't feel he was in the wrong, and that is commendable.· In summary, [noparse][[/noparse]A] it was understandable that Cluso got riled, and, [noparse][[/noparse]B]·Leon made an effort to change.· Well, that concludes my defense remarks, though neither retained me as arbitrator.
So on to the admiration, starting with Cluso.· Regular forum readers can probably readily cite 5 to 10 names of frequent posters to this forum that make outstanding contributions to the forum, both in terms of comments as well as code and hardware.· Cluso is definitely in that group and no doubt widely admired and appreciated.· He constantly shows his willingness to make suggestions when people have problems or ideas, creates innovative ideas and applications for the Prop (both in H/W and S/W), and is a kind of all-around evangelist for the Propeller, not to mention that he brings a vast array of professional experience to this forum (as do many others).· Cluso would be sorely missed if absent from this forum.· I think we can tolerate an atypical outburst or two from him, and I think that it would be madness for anyone to purposefully agitate him, comparable to shooting oneself in the foot.
But Leon also brings value, often of another kind, to this forum.· In particular, he frequently helps us put the Prop in perspective amongst the world of chips in which it resides.· Personally, I don't feel right about choosing a solution if I haven't made some effort to consider alternatives.· But I, for one, lack professional experience in this field.· Sure, I can search the Internet quite well, but that's no substitute for the experience of others (which can also be had by searching).· For me, Leon provides·some of the experience and perspective·that I would otherwise be at least partially missing.· I really value that.· Furthermore, Leon typcially packs as much into a two-line post as is semantically possible.· I've learned a lot from him.· He's even set me straight a time or two when I've been wrong or one-sided.· Sure, his remarks are often terse, but they are filled with gems that come from experience.· In that I'm not a professional in the microcontroller field, this forum is a big part of my "education," and Leon's posts are a significant part of that.· I'd miss his insights if he were gone, both in terms of his general electronics knowledge and in terms of his insights into other microcontrollers.··Once again, he just packs so much information into a two-line·(or so)·post.·
Lastly, the Propeller can't avoid being compared to other chips out there.· And the Propeller has no reason to avoid comparison.· It's a fabulous chip!· But it's just like job applicants:· some are right for some jobs but not for others.· However, it's really only by comparison that we can make informed choices.· I kind of feel bad that I don't often read the forums for other chips, I like our beloved Propeller enough as it is.· But I'm glad tha·I can learn about the strengths and, yes, weaknesses of the Propeller for various tasks via some comparisons to other chips on this forum.· Do I want this forum to only do comparisons and not focus on the Propeller applications?· No.· But I like·this forum·basically the way it is:· fresh and dynamic, with comparisons where appropriate.· What we are lucky to have in Cluso and Leon are two experienced contributors that make it easier for us to know "how" and "when" to apply the Prop.· They are really on the same side.· And Leon's above-mentioned change-of-behaviour speaks volumes to me·(as it is truly difficult to change), showing that he values this forum but doesn't blindly use the Propeller.· I click on the links that he occassionally provides to learn more about other possible solutions.· Like another poster said, his links haven't disuaded me from using the Propeller, but I think they help me to use it more intelligently.· And Cluso's contributions are simply invaluable and too numerous to detail.· We need them both!·
·
I am one who can put it behind me immediately after I have said my piece. I do acknowledge that there are times when comparisons will be made and there are times when they are not called for. This thread did not call for it. I asked what would entice design engineers, not what was wrong with the prop or Parallax, but this is where the thread evolved.
Ok I have said my piece, lets move on. We are so off-topic we may as well continue along the lines the thread has taken.
The prop chip is ~$6 per 1,000. Since the prop is not a high moving line at the distributors like DigiKey, etc, they probably have to be given a higher margin than say a PIC. Notice how quickly the prop reduces in volume to say a PIC. Now if you compare a PIC for $2 then we are talking apples and oranges and in most circumstances and even I would use a PIC, but there are some·instances where·I would still use a Prop (BTW I am). So lets say the margin was 30% (not real). So the distributor would buy for $1.33 and sell for $2.00 and the prop would buy for $4.00 and sell for $6.00. Parallax cannot undercut it's dealers, but you are really comparing a $1.33 chip to a $4.00 chip if you see what I mean. BTW we have to add an eeprom which IIRC a 32KB can be had for ~50c in quantity. I am sure you would get a great deal if you went to Parallax with a firm order for 10,000 pcs you would get a great price. An order for 100,000 and Chip would deliver them to you personally (No Chip - get back to the Prop II!).
I raised an idea on another thread about booting from·SD (or microSD) without an eeprom. How this would be done is that a new ROM mask would be required. I am unsure of it's cost and I am presuming it is a traditional mask and not a hardwired thing. However, some piece of ROM must go and that was not liked on the other thread. Everyone is using something. Just to refresh what's in the ROM, 16KB for the character font, 8KB for the Log tables, $4KB+2 for the Sine and 4KB-2 for the Interpreter and Booter. The Booter has to a have the option to boot from SD card added which would include some low level SPI access which would probably take 1KB or maybe even less if it was in spin (it could be). There would·be room for the faster Interpreter which would then have space for the LMM extensions. We may even have room for 2KB 8x8 character font too, depending on what was removed. Or maybe we could get a Basic Interpreter in there, not that this is any use for volume sales.
If someone had the killer app, they could have security with their own mask without eeprom. You will not get that from a big company these days. No chip is totally secure though. BTW I have not confirmed Parallax could/would do it, but it's an idea for those with volume.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Links to other interesting threads:
· Home of the MultiBladeProps: TriBlade,·RamBlade,·SixBlade, website
· Single Board Computer:·3 Propeller ICs·and a·TriBladeProp board (ZiCog Z80 Emulator)
· Prop Tools under Development or Completed (Index)
· Emulators: CPUs Z80 etc; Micros Altair etc;· Terminals·VT100 etc; (Index) ZiCog (Z80) , MoCog (6809)·
· Prop OS: SphinxOS·, PropDos , PropCmd··· Search the Propeller forums·(uses advanced Google search)
My cruising website is: ·www.bluemagic.biz·· MultiBlade Props: www.cluso.bluemagic.biz
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Propeller + Picaxe = Romeo & Juliet
Post Edited (JRetSapDoog) : 7/16/2010 5:49:40 PM GMT
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Leon Heller
Amateur radio callsign: G1HSM
Post Edited (Leon) : 7/16/2010 6:52:42 PM GMT
So, that competing chip doesn't have FLASH? I've read about it, but forgotten the details. But it could have, you're saying, if they switched to the TSMC process, and so perhaps a future Prop could also. Seems like I recall that the next Prop is still several process sizes larger than any of the processes you've mentioned (though my memory is fuzzy on that). Perhaps that's to reduce leakage current, though they've said the new chip will be much more power-hungry than the current Prop, making it another kind of beast. For my intended usage, I don't need days, weeks or months of operation on battery power, so I welcome the more powerful chip.
Also, I've read that die-size is a concern with the new design, so I don't understand why smaller process technology isn't being used, but I've kind of just assumed that it has to do with either power consumption or design tools. BTW, does anyone know which IC fab Parallax uses for the Prop? I know it's in Taiwan (or at least their packager is), but don't know which.· Anyway, that still leaves open my question about whether FLASH-based firmware could run as fast as ROM-based firmware and/or be compatible with the Prop's unique architecture, if that is at all a factor.
Regarding the process tools Parallax uses, I can certainly understand how a small company must use the tools they have and know. That's what I'd do if I had a unique design, even if there existed "better" tools or process technology. I'd do it that way because I'd know that I'd just be lucky to get a working design out the door, and I'd hope that its unique features would carry it beyond any limitations that the process technology introduced. However, this time around is not Parallax's first chip effort. And I think it's pretty clear that they are on the edge of greatness with the new one. We're all chomping at the bit for it, and professionals will be, too, once they see how the unique abilities of the current Prop have scaled up so much. As such, perhaps Parallax can't just default to taking the known route of working with the tools they have and know. That is, perhaps it is worth considering other process technologies/tools. I don't mean to suggest that they haven't done that. Of course they have, just that it might be worth taking another look at.
Of course, I don't want to see anything significantly slow down the release of a new chip, but I hope it can be as great as it can be. RAM memory still seems to be one of the biggest limitations. It's a real pity that that seems to be unavoidable. But anyway, I'll sure be happy to get anywhere near the 256KB they're striving for, even if 4X would be a killer in terms of one-chip designs for certain applications. Or perhaps a POP chip could be done, though I wouldn't especially welcome losing a lot of I/O's to accommodate that.
Even if Parallax is widely successful with their next chip, I don't see them wanting to become mostly a chip company, per se, in such a competitive market. I still think they should be the "ease-of-use" company. That leads me to wonder if they could partner with an established chip-design company (one like Microchip or the maker (XXXX, one of those X's being correct) of the-chip-which-must-not-be-named (TCWMNBN); after all, there are some simularities. Of course, if they did that, they'd have to do it in such a way that they maintained management and creative control...because Parallax's mission includes being a user-friendly, socially-responsible and not-purely-focused-on-profit company, one that encourages the creativity of others in a win-win kind of way (can't really say that for the other companies; it's only about profit for them). But one has to partner these days on some level. For example, very few companies can have their own fabs. Therefore, it might be something to consider and could bring significantly more resources to bear...and bring us a line of truly great products.
Having said that, what's coming down the pipe from their own resources looks pretty incredible in and of itself. And, it seems, Parallax is even more open to a Prop 1.5 or similar. So maybe various kinds of actions will be happening in the future. That 3C mini-board (of another thread) sounds quite interesting and practical, for example.
Post Edited (JRetSapDoog) : 7/16/2010 7:00:13 PM GMT
some white space please! [noparse];)[/noparse]
I think that the on-chip flash would just load programs into RAM, for speed. There are tricks that speed up execution from flash, like the wide interface that NXP uses on their ARM chips, but execution from RAM will always be fastest.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Leon Heller
Amateur radio callsign: G1HSM
Post Edited (Leon) : 7/16/2010 7:25:39 PM GMT
That's the same reason I use the prop.
Video and timing are the jewels in the prop.
Exactly so, and this is why IMO Parallax should pay you something so
you can continue to work on Catalina. In this way the upgrades to Catalina
could be offered to users for free. The optimizing version of Catalina you
mentioned could then be freely downloaded. You can't be expected to toil
endlessly on Catalina without some $ support.
Great ideas
It seems to me it should be possible to repackage a sm prop
inside a plastic module that would look like a normal ic.
Put a sm prop, a small 5 or 6.25 xtal, an eeprom, flash memory
(uSD or whatever), sm resistors..etc. Bring out to pins the video
signals for easy use.
I think many here could build this simple little board, all we
need is a way to package it in a plastic module with pins on it
I for one have a dislike for the look of those little dil boards
where you can see all the parts....it just looks more professional
to encapsulate in plastic.... (and leave off the cute propeller beanie image)
A ROM change may be easy though. What I envisage is a new ROM part that can boot from SD if the eeprom does not exist. So it is backwards compatible except that it would have something in the ROM removed. That is the real stumbling block - what to remove... Character font 16x32 and replace with multiple smaller 8x8 fonts and other things too (because we gain half the ROM space back); Log tables; Sin tables. They are the only choices available. Anythig else is currently off the table.
Unfortunately if we ditch the font, then it's gone because there is not enough space in hub to replace it. However, the 4K for Log or 4K for Sin users may have enough hub space. Alternately, Parallax could keep both chips with an inventory cost penalty. There are a few other things that can suport this concept and they have been communicated to the powers to be. It is a shame that we cannot just add more ROM but it isnt going to happen.
BTW This is not for an Arduino replacement although with the extra SD, VGA/TV, Kbd, etc it may be able to compete in an upmarket way. It is more for the volume users. The prop is an upmarket chip and so SD in very likely to be included in that killer app (IMHO anyway).
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Links to other interesting threads:
· Home of the MultiBladeProps: TriBlade,·RamBlade,·SixBlade, website
· Single Board Computer:·3 Propeller ICs·and a·TriBladeProp board (ZiCog Z80 Emulator)
· Prop Tools under Development or Completed (Index)
· Emulators: CPUs Z80 etc; Micros Altair etc;· Terminals·VT100 etc; (Index) ZiCog (Z80) , MoCog (6809)·
· Prop OS: SphinxOS·, PropDos , PropCmd··· Search the Propeller forums·(uses advanced Google search)
My cruising website is: ·www.bluemagic.biz·· MultiBlade Props: www.cluso.bluemagic.biz
Thanks for the vote of support. The problem is that the Propeller is so addictive that Parallax probably realizes I'd keep toiling away in any case
Ross.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Catalina - a FREE C compiler for the Propeller - see Catalina
I have some feedback already:
Company A (market specific robotics manufacturer, field proven product lines, 4 product lines/~100 per month)
Most of their products are made with Atmel microcontrollers. Cost is not considered in their component selection process because the electronics (PCBs) account for about 5% of the total cost of the robot. Their main selection criteria was based upon using a single family of microcontrollers to cover all of their needs. Some of their boards are extremely simple (IE: read a sensor and relay data via serial I/O line back to the main baord) and some are very complex (IE: host system managing the entire system with various protocols such as I2c, one-wire, serial, Ethernet, USB, etc. It also drives a human interface on an LCD/keypad that is duplicated via a proprietary RF connection). Being able to use a single family of microcontrollers makes it easier to manage all programming efforts with staff that works as a well oiled machine.
Company B (very small company with project specific marketing products such as store displays, 8 product lines ranging from 50 to 4,000 per month)
Products are heavily burdened with a "low-cost" push, so the cheapest microcontroller that can do the job is selected. The company typically use Cypress PSoC when possible because of the low cost and the head engineer has the most experience with the PSoC which means the work can be done quickly and in house. For more complex projects, the PSoC coding is outsourced to a long time used programming partner. Other jobs have had their "code" portion outsourced at the beginning, and the vendor is allowed to use whatever microcontroller they choose to get the job done as long as it fits a cost target.
Company C (wireless sensing products for a specific target market, a few year old startup with some very experienced engineers, 18 product lines, a couple around 20 per quarter, but most around 300-500 per month)
This companies product line is directly affected by power consumption as the units are battery operated, but transmit data every 5 minutes. They chose to pair TIs MSP430 controller with the TIs CC2420 2.4 GHZ radio chip for their product. The product has performed very well in the market for the past 2 years, but other new designs from them are branching out into other microcontrollers to meet new needs such as Ethernet to serial, power outlet monitoring, etc. Cost has not been a major factor, but definitely considered. For low power usage, it seems that TIs MSP430 has the top spot.
Company D (legacy modem technology for industrial applications, very old company/product line, 6 product lines, 120-200 per quarter)
Even though this company could easily re-design their products using different microcontrollers, their products have a field proven design and very low profit margin. Most of their products built are for field repair of installs that are 15+ years old. Their products use various microcontrollers and the "why" is anyone's guess. (most of the original design engineers have passed on)
Company E (retail beverage equipment products, solid company, new products each year, 7 product lines, 300-1000 per month)
Microcontroller choices are left to the design engineers that are assigned product lines. They have products with Atmel AVR and Microchip PICs. They have a few products that use Qtouch chips and those all use the Atmel Atmega128.
So, it seems the choice is very much driven by the needs of product/project more so than something that can be said that encompasses all of them into a group so that one item can be hailed as the largest concern. All but Company B, have the appropriate software packages and tools for the products because that is a factor of bringing a product to market and sustaining it. Having to pay for these tools is expected, so even though it's a plus by getting free tools, some have purchased other software tools to help them anyways. In all cases except where the PSoC is used, programming in C appears to be the champion. On the PSoC, it is split between the free PSoC designer software and manually coding in C. I asked a few about using a chip that required an external EEPROM and most did not care other than that meant more PCB real estate when compared to a part that had internal flash. The only place that I see that the Propeller could not be used is with Company A that chose a microcontroller family to address duties that the Propeller cannot perform such as full-blown ethernet.
With all that said, my formal response to the thread title's question would be this:
Having a C compiler that is equivalent to those being used for the microcontrollers listed above will increase the Propeller's ability to gain marketshare.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Andrew Williams
WBA Consulting
PowerTwig Dual Output Power Supply Module
My Prop projects: Reverse Geo-Cache Box, Custom Metronome, Micro Plunge Logger
Post Edited (WBA Consulting) : 7/19/2010 11:51:26 PM GMT
Thanks for the information - very useful.
Any chance of getting volumes (order of magnitude approximations would be fine) for each of your companies A - E? It would be interesting to see how the volume affects the decision making process.
Ross.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Catalina - a FREE C compiler for the Propeller - see Catalina
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Andrew Williams
WBA Consulting
PowerTwig Dual Output Power Supply Module
My Prop projects: Reverse Geo-Cache Box, Custom Metronome, Micro Plunge Logger
Here's comparable information for my company - let's call it "company F":
Company F (distributed telemetry equipment for a specific industrial market, about 10 years experience in this market with some very experienced engineers and 4 product lines - one around 10 per quarter, but the others around 10,000 - 20,000 units per annum each (we tend to get large annual orders and do one or two production runs each year). The main product line is directly affected by power consumption as the units are battery operated, and expected to last in the field for up to 5 years with no maintenance. They sample data every 6 minutes and transmit by ISM band radio, PSTN or GSM dialer periodically (typically once per day). Company policy means I can't list you the exact chips we use, but the processors are mostly ATmega. We use C and buy professional compilers and other tools - but we use the manufacturer's "freebie" tools when evaluating new components, and also for various non-critical development tasks.
Ross.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Catalina - a FREE C compiler for the Propeller - see Catalina