I'm not sure why more people don't use them. I've built my company around the propeller chip and find it extremly easy to use and flexible enough to tackle some of the biggest problems we've had. (Every one of our products has at least one propeller chip in it!)
I've seen people complain about two major points, IP protection and chip cost. We've found both of those points to be moot. In a product that costs as much as our, the $6 price per unit is nothing, especially considering the flexibility and code base we base it on, that cost savings alone easily makes up for the unit price difference.
As for IP protection, we've found that we don't need it as long as our code is well documented and the appropriate copyright measures are in place. Who cares if someone can see our code? If they copy it and try to use it, we'll sue... simple as that. Protecting the code isn't going to get you any more protection... it'll simply buy you a small delay before they can actually use your code. Trust me, even the best protection schemes can be defeated... and then you're right back at square one. Seriously, if Apple/Sony can't protect their hardware/software from invasion, what makes you think an AVR or PIC can?
BTW... we announced our "Wingman" product at Oshkosh and have already sold over 250 units... with commitments for more than 100 per month ongoing.
BTW... we announced our "Wingman" product at Oshkosh and have already sold over 250 units... with commitments for more than 100 per month ongoing.
Woo hoo! Glad to see more successful products.
I was thumbing through my new Mouser catalog today. I find the listings of AVR and Freescale, and ARM variant chips pretty amazing. There are only a few micros with more SRAM than Propeller and none with similar flexibility, but there are plenty with more or less program memory.
I have an ARM LPC2148 which has oodles of program memory. The funny thing is getting a program loaded on it is pretty painful. If the LPC2148 was as easy to program as Propeller I might use it more often. Of course part of the pain comes from having to wait for a larger image to load
As far as copy protection goes. How much defensible legal protection do you get from just putting "Copyright (c) year, name All rights reserved." on your source? Seems like fraudulent claims would be easy without requiring proof.
I've found the LPC2148 very easy to use. I developed a large program using one to control a medical instrument without any problems. A Propeller wouldn't have been suitable - not enough memory or I/O, and no ADCs. Loading memory during development was very fast, via the JTAG interface.
I've found the LPC2148 very easy to use. I developed a large program using one to control a medical instrument without any problems. A Propeller wouldn't have been suitable - not enough memory or I/O, and no ADCs. Loading memory during development was very fast, via the JTAG interface.
Leon, what amount of memory did you need? Why not just add a memory chip to the Prop? How many I/Os did you need? How about simply adding another prop chip for 64 ports? There's ADC function too with a few simple components or you can read the pins with the internal rc capability. (there's some detail in the tester thread about the latter)
The Propeller is used in several high-volume commercial products. These customers aren't open with their product designs, as you can well imagine. The commercial use is increasing and we support them through their design cycle lead times, too.
Having worked with several venture-backed companies who also create semiconductors for specific purposes, and having experienced their short-term need to satisfy managers, investors and board of directors, as a customer I'd find the Propeller and Parallax an ideal partner to use for a commercial design. You're dealing with a company who promises to hold the course and not reinvent ourselves every year. Customers who've chosen the Propeller for commercial products have been very satisfied with their choice. Their decision to use our chip is based on capabilities as much as our company's track record.
You'll be seeing more Parallax evolution for commercial customers this Autumn. A few of our Propeller forum members are even included in these efforts as a focus group.
We assure you that the Propeller 1 and future versions will be around long enough to meet your needs. This is something that isn't easy to come by in this industry: stability, consistency, and support. Just take a look at the BS1 - it's been around for nearly 20 years.
To hobbyists, educators and engineers alike - we're your company. Everybody counts as a key customer in our company regardless of how much you buy. You can hear this statement elsewhere but we practice it every day.
I needed two UARTs, 10 (accurate) analogue inputs, one SPI, one I2C, two PWMs, one DAC, four outputs for pumps and valves, two LED outputs and two switch inputs. I had to process a large amount of data from a four-quadrant photo-diode sensor, thermistors and conductivity sensors, and send it to a PC every second, with a CRC, whilst checking for commands from the PC, which also had a CRC. The customer wanted the software written in C. I needed a couple of large circular buffers, as well as quite a lot of storage for arrays.
The instrument is actually an "ektacytometer", for measuring red blood cell deformation when subjected to a shear force. Blood samples mixed with a buffer are fed into the space between a fixed perspex cylinder and a concentric rotating perspex cylinder. A laser illuminates the blood cells and the diffraction pattern is detected by the photo-diode sensors and used to determine the deformation, using von Hulst's anomalous diffraction technique. Various rare blood disorders can be diagnosed using this instrument, and the prototype is being used in a Paris hospital. My customer (he works at a university in France) is applying for further funding to take the design to the next stage.
The LPC2148 costs about $10 (one off) and I used a prototyping board costing about $100. We just had to add a few connectors to it.
We have been offered the commercial rights to the design for nothing by the university, but it won't make us rich. The diseases are so rare that only one machine will be needed per country.
How much defensible legal protection do you get from just putting "Copyright (c) year, name All rights reserved." on your source?
Steve, legally, all you need is your copyright notice to have full protection. However, a good attorney will also have several other items in place. For example, you might have a 3rd party that can hold on to the source code and design documents in escrow just for such an occasion. You can also, like we do, actually use all three protection mechanisms. One, being copyright, two, being patents and the third being trade secrets. All three are enforcable in different ways to protect different areas of your business.
Our business model, however, is not dependant on code protection.. Instead, we're a little different in that we plan on attacking our competitors with unique products that are superior in major ways than from the competition. (Think about how Apple approaches their business model.) Our Wingman is a great example... it's the first and only portable attitude display unit on the market and we went right after the most expensive competitors with a unit that is 1/10 of their price but offers a lot more functionallity. Our heads up displays are just a unique in that regard.
@Leon: My suggestion: order it somewere. A lot of those chips are available. If you need 50,000,000 a year and what you need is not off the shelf, I suggest to have Chip as development partner. ;-)
I needed two UARTs, 10 (accurate) analogue inputs, one SPI, one I2C, two PWMs, one DAC, four outputs for pumps and valves, two LED outputs and two switch inputs. I had to process a large amount of data from a four-quadrant photo-diode sensor, thermistors and conductivity sensors, and send it to a PC every second, with a CRC, whilst checking for commands from the PC, which also had a CRC. The customer wanted the software written in C.
Hello Leon,
I'll take the bait, come on - what's this all about, anyway? If you want us to help you with the project, we will, but your mind is already made up. And I know you are totally capable of doing the project without our help, too.
Maybe we could squeeze most of your application into the Propeller, perhaps even built with Catalina if C is a requirement. If you had a need for using the Propeller in such an application we'd work with you, but if you truly don't want to use it, that's perfectly okay. Nobody at Parallax is going to attempt you to do something you don't want to do. Please use the tools and processors that you find appropriate for the job.
As I mentioned above, the choice to use the Propeller isn't always based on specifications or cost alone. I recognize from your posts that these are not important factors in your decision making process. Leon, you are also part of our team, too. I know that from positive posts you have made in other forums about the Propeller.
Now, how about that project! Who's the customer? How many chips do they need (a joke)!
Sincerely,
Ken Gracey
Note: Leon's post was submitted in two pieces. I replied based on his first quote. The second part of his post was made after he completed the first part. Leon, I wouldn't give away so many details about my customer's project.
It's a university project, so the details don't matter. I've deleted them, anyway. I can't see it being feasible to use a Propeller for that lot, especially as the software had to be written in C.
This is the time critical code that *has* to be executed within 20 ms, and transmitted to the PC after 50 interrupts:
//---------------------- Timer 0 ISR every 20 ms -----------------//
static void timer0ISR(void)
{
char temp[10];
char s[44];
u_long crc;
unsigned int ptr;
unsigned int C, A, B, E, F;
unsigned int i;
static unsigned int n;
int R;
ticks++; // increment general delay ticks
Timer0Count++; // increment 20 ms ticks
IO0PIN ^= ACTIVITY_LED; // toggle activity LED
C_total = A_total = B_total = E_total = F_total = 0;
if (!(IO0PIN & PB)) // check if Panic button has been pressed
{
#ifndef NODEBUG
debug_printf("Panic button pressed!!!\n");
#endif
cancel(); // cancel function resets processor using watchdog
}
// Skip if bridge mode or wait for '1001' from WinEkta
if ((flag == 1) && (Send_data == 1))
{
for (i = 0; i < 16; i++)
{
C_total += getAnalogueInput_AD1(0); // P0.6
A_total += getAnalogueInput_AD0(2); // P0.29
B_total += getAnalogueInput_AD0(3); // P0.30
E_total += getAnalogueInput_AD0(7); // P0.5
F_total += getAnalogueInput_AD0(1); // P0.28
}
C_temp += (C_total >> 4);
A_temp += (A_total >> 4);
B_temp += (B_total >> 4);
E_temp += (E_total >> 4);
F_temp += (F_total >> 4);
// every 50 interrupts (1 s) average the values and transmit them to the PC
while (n++ < 49)
{
T0IR = 0xFF; // Clear the timer 0 interrupt
return;
}
C = C_temp/50;
A = A_temp/50;
B = B_temp/50;
E = E_temp/50;
F = F_temp/50;
// get status bits
get_status();
// get motor speed
R = get_motor_speed();
// increment secs counter
secs++;
// construct output string
my_itoa_4(C, temp);
ptr = 0;
s[ptr++] = 'C';
for (i = 0; i < 4; i++)
s[i+ptr] = temp[i];
my_itoa_4(A, temp);
ptr = 5;
s[ptr++] = 'A';
for (i = 0; i < 4; i++)
s[i+ptr] = temp[i];
my_itoa_4(B, temp);
ptr = 10;
s[ptr++] = 'B';
for (i = 0; i < 4; i++)
s[i+ptr] = temp[i];
my_itoa_4(E, temp);
ptr = 15;
s[ptr++] = 'E';
for (i = 0; i < 4; i++)
s[i+ptr] = temp[i];
my_itoa_4(F, temp);
ptr = 20;
s[ptr++] = 'F';
for (i = 0; i < 4; i++)
s[i+ptr] = temp[i];
my_itoa_4(R, temp);
ptr = 25;
s[ptr++] = 'R';
for (i = 0; i < 4; i++)
s[i+ptr] = temp[i];
my_itoa_4(secs, temp);
ptr = 30;
s[30] = 'T';
s[31] = '0';
s[32] = '0';
ptr = 33;
for (i = 0; i < 4; i++)
s[i+ptr] = temp[i];
ptr = 37;
sprintf(temp, "S%0.2x", status);
for (i = 0; i < 3; i++)
s[i+ptr] = temp[i];
crc = crc32(s, 39);
sprintf(temp, "%0.8x", crc);
ptr = 40;
for (i = 0; i < 8; i++)
s[i+ptr] = temp[i];
s[48] = '\0';
_puts0(s); // transmit string to PC via UART0
IO0PIN ^= TX_LED; // toggle transmit LED
C_temp = A_temp = B_temp = E_temp = F_temp = 0;
n = 0;
}
// feed watchdog
WDFEED = 0xAA;
WDFEED = 0x55;
T0IR = 0xFF; // Clear the timer 0 interrupt
}
It's a university project, so the details don't matter. I can't see it being feasible to use a Propeller for that lot, especially as the software had to be written in C.
Leon, in a prior post you mentioned that the project is an "ektacytometer" for measuring red blood cell deformation when subjected to a shear force. This seems like an instrument that would really benefit from having a visual display to couple whatever datalogging/communication requirements. Since the device is being developed for hospital use you could simplify the setup and feedback by providing something that humans relate to so readily - a display. And the Propeller offers convenient graphics/text via composite or SVGA. How about if you talk to the customer and sell them on the many benefits of the Propeller's display capabilities?
Perhaps if you demonstrate the visual abilities of the Propeller they'll come right back to you for Phase 2 of the product development. You're one of the few engineers who knows what's possible with the Propeller in this regard. Everybody else will be thinking about whatever off-the-shelf solution they can find, won't they?
Just tell me where to send the engineer if you need help. . . unless your mind is already made up!
There is a large program running on the PC that displays all the data, logs data for each patient, prints out the results, etc. The machine is actually controlled by the PC via my ARM software, all that the operator has to do is put the sample into the machine and select the type of test. My ARM program controls all aspects of the test, depending on what has been selected using the PC program.
... hmmm.. new forum appears to be missing the the key "User Ignore" button ...
Such a button would have to have "baby" and "bathwater" sub-buttons before I would ever use such a thing, and a user-programmable way to make the distinction. Oh, wait! [slaps forehead] I can do that in my head as I skim over the posts!
This is a thread that would attract a novice interested in the Propeller. Perhaps here for the first time and looking around. From a that point of view they are just collecting information and forming opinions.
One opinion would be obvious: "If the pros or industry isn't using the Propeller, why should I use it" and Ken answered that quite well.
The other opinion would be based on how Leon was treated. From their point of view, Leon is a Senior Member with an incredible 3500+ postings. That affords him some credibility. Leon tried to make a point, a point that someone else with Pic or Avr experience may relate too. But he didn't do a very good job of it, and so he got pee'd on.
Thank you, Capt. Quirk. A thread like this can't be just about the Propeller, because, if it were, it would lack the critical, wider context necessary to support the discussion. Leon brings us news from beyond the frontier of our own Propeller-centric universe, and for that he deserves our appreciation and attention. Those of us who confuse factual reportage with mere advocacy do so to our own impoverishment. It's a big world beyond our little Eden, and it pays to stay abreast of what lurks outside our borders.
I think of the Prop as ideal for instrumentation. That's not a consumer market, so this is off topic, but the thread has jumped. Instrumentation is a professional market, where the cost of the Prop itself is negligible. There is need for display, user interface, and communication in addition to measurement, control and computation tasks. All of that is so neatly partitioned on a Propeller, and if one won't do it, why not use two or more as clever people here have done with amazing results? A lot of companies doing this kind of work are small and fill very specific niches. The requirements vary tremendously, so one built in ADC or one set of built-in peripherals will not cover the field by any means, but the symmetry of the Propeller, the cogs, the pins, the counters, not to mention the open support and code library, well, it allows so many possibilities. It will take a while to catch on. The support and stability of a company like Parallax, as Ken has said, are indeed worth a lot. The education market may be a key to dissemination.
Not all instruments are all that complicated, compared to the one that Leon described. As a case in point, I just finished a project for a University group where a $15000 instrument they purchased did not have enough battery life or data storage capacity or the correct timing for what they wanted to do. I made them an add-on box to add those features, but to do so I had to hack into the controls of the instrument and was surprised to find that it was built around a Ticket processor, which is kind of like a BASIC Stamp. How much more it could do and more easily with a Propeller at its core. It included DOS software for Windows 3.1, so that gives you an idea that the instrument could be successful for years without supercharging its interface or its innards. The same is true of so many single purpose instruments.
I have stated my disgust at Leon's posts in the past. Everything is negative. It's a shame because Leon does have some good information. As you can see, I am not alone here.
Anyone can find something else that will do the job, for better of for worse. But can you find the SAME chip that will do all of these things, for better or worse. The answer is a simple and definitive NO.
I have run my own companies since 1976 and designed and programmed them using many different micros. I have my favourites - the Motorola 68705 family (now known as Fairchild MC9S08 family) was one, but now it is the Propeller - and it wins hands down.
One of the problems when you have a range of designs is to minimise the inventory. This is extremely important. The main reason, is not what you would expect, that being inventory cost. It is actually a little thing called consistency of supply. Every few+ years as the economy goes into recession the building of the next generation of factories slows down. As we come out of recession and the demand starts to outstrip supply. The new factories cannot handle the new demand. The lead times increase. Then the vultures come in and buy up all the stock to sit on and scalp the market when it gets desperate. I have seen chips go from ex-stock to 1 year lead times overnight. DRAM used to be a classic of this. We got caught with modem chips where Rockwell went to allocation and we would only get 20-30% of our projected orders, without increaded demand.
So, what does this mean. Well, if I have fewer different chips used in my designs, I can hold a buffer stock more easily. I can then determine which products to produce with a limited supply of parts. Am I going to make 1,000 of item A that I make $10 profit, or am I going to make 100 of item B where I make $100 profit. It doesn't take an accountant to figure this out.
And notice I haven't addressed the other advantages of using the one chip for lots of products - things like code, familiarity, layout, etc.
Price of the Prop is really not an issue. We would all like a cheaper prop. If the other commercial products use a lot of Prop chips, the price will come down. As Ken says, some of those products are out there, but they are often confidential and cannot be revealed here. Wjsteele has showed his product off for us - congratulations Bill. We just need to use enough features of the prop to ensure it is price competitive.
As for security, did you see the other thread with a link to the person who revealed the internals of the prop die? He etched the plastic packaging off the chip, then the first interconnect layer. Pitty Parallax had already released this info years ago I suspect from this method you could find the internal Flash code within any chip.
My apologies that much of this is Off Topic, but we do need a reality check from time to time.
1) I agree that there is no ultimate code security. But it would be nice to have a way to prevent wholesale copying of commercial code with just 5 minutes of effort from the perp.
2) If Parallax granted me one wish, I'd wish for a PASM quick reference. I'm certain I'd use the Prop more often if I didn't have to wade through a thick and wordy manual.
A perfect example of a quick reference is Herbert Schildt's "C the Pocket Reference." I'd love to have such a short and concise summary of PASM! I like the fact that he includes well-considered one or two line code snippets of almost everything. An example is far more valuable than a thousand words, if one is in a hurry. And we always are in a hurry.
In any event, PASM is the only way to get real speed out of the Prop. Otherwise, it's meaningless to talk about the Prop being a 160 or 200 MIPS machine.
There has been a quick reference in the Propeller Tool for as long as I can remember.
Humm... That is a very nice command summary. Particularly welcome is the operator summary on the last page.
Given that the Propeller is a bit odd, perhaps some selected examples and a sentence or two could be devoted to such topics as: Cog architecture, conditional execution, the choice of 2's compliment arithmetic, the composition of an instruction word, the two-bit condition code register, the unusual TRUE and FALSE, cog synchronization, etc. I raise the Schildt text again as a pretty good model of what I wish we had for the Prop.
Blaise Pascal: "I would have written a shorter letter, but I did not have the time."
Pity that our two choices are 1) a list of commands and 2) a 300 page tutorial that starts with the creation of the World...but leaves out all the good stuff.
XMOS chips are used on lots of those gigantic displays used at sports grounds and so on - one XMOS chip per LED tile...............................The Propeller obviously can't compete in those markets.
Correct, but I also don't see why anyone would think that it would. From what I can tell from the information on the XMOS website, I think the target market is different than that of the prop, so comparing the two chips doesn't really say much. The XMOS site pushes USB audio, LED tile controllers, iPOD docks, and AVB ethernet as the main markets. I don't foresee the prop in the selection process of any of those markets (of course I could be wrong).
As a side note, of the two customers (at work) that do LED based products, one uses XILINX on their entire suite of LED control boards and the other uses various chips, but seems to favor Atmel. My guess is that the Xilinx user could accomplish many of their products with XMOS if they chose.
In regards to code security, your code is only as safe as the desire of someone else to steal it. Even devices with internal flash can have the code harvested.
We build several products that boot from external flash devices and the security risk never seems to be an issue. Some of these products are in very competitive markets, so I am sure code security was discussed during design.
My opinion is that if you have a product that relies on code security that heavily, any microcontroller choice will still leave you at risk, so you need hardware methods for security, such as potting. My last employer built modems for ATM machines and the last step of the process was to encapsulate the PCB in a potting compound that turned into a crystallized looking form of tar. Any attempt to dis-assemble the device would destroy it. It also had about 10 "hardware key" connections made with enameled wire that was suspended in the potting material so that if someone did manage to remove the potting and leave the board itself intact, it would corrupt itself repeatedly when powered up due to the missing connections that would have been removed with the potting.
Another method is embedding connections into cases. Many POS terminals contain connection pads in the cases, PCB guides, and other mechanical parts that prevent the product from working properly unless properly assembled into the case with all of the individual hardware components.
In the main, it is a completely different market, but there is some overlap at the bottom end. I provided that link because you said that you hadn't seen any commercial applications for XMOS devices. I do feel that Parallax needs to take a similar professional approach to XMOS, if they want professionals to use the device.
Somebody just mentioned the need for a PC to display the data that they got with some other microcontroller. And, that's really, to me, the best part about the Propeller, is that I can make autonomous systems with keyboard and mouse input and VGA display that aren't dependent on a PC. Especially, since you need to always keep PCs up to date with fixes and updates, that's just a major pain when you just want something that you can turn on and have it work. And sometimes, you want systems that are on all the time. PCs aren't too good at that...
Yes - a professional approach geared towards the specific customer. I agree with you, Leon, but in our own flavor and style of course. It's coming.
Try to get in touch with David May and Chip Gracey. Then you know the difference in professionalism and what it means. If you plan to compete a giant, be sure to have a slingshot ;-)
Comments
I've seen people complain about two major points, IP protection and chip cost. We've found both of those points to be moot. In a product that costs as much as our, the $6 price per unit is nothing, especially considering the flexibility and code base we base it on, that cost savings alone easily makes up for the unit price difference.
As for IP protection, we've found that we don't need it as long as our code is well documented and the appropriate copyright measures are in place. Who cares if someone can see our code? If they copy it and try to use it, we'll sue... simple as that. Protecting the code isn't going to get you any more protection... it'll simply buy you a small delay before they can actually use your code. Trust me, even the best protection schemes can be defeated... and then you're right back at square one. Seriously, if Apple/Sony can't protect their hardware/software from invasion, what makes you think an AVR or PIC can?
BTW... we announced our "Wingman" product at Oshkosh and have already sold over 250 units... with commitments for more than 100 per month ongoing.
Bill
I was thumbing through my new Mouser catalog today. I find the listings of AVR and Freescale, and ARM variant chips pretty amazing. There are only a few micros with more SRAM than Propeller and none with similar flexibility, but there are plenty with more or less program memory.
I have an ARM LPC2148 which has oodles of program memory. The funny thing is getting a program loaded on it is pretty painful. If the LPC2148 was as easy to program as Propeller I might use it more often. Of course part of the pain comes from having to wait for a larger image to load
As far as copy protection goes. How much defensible legal protection do you get from just putting "Copyright (c) year, name All rights reserved." on your source? Seems like fraudulent claims would be easy without requiring proof.
Cheers,
--Steve
Humanoido
The Propeller is used in several high-volume commercial products. These customers aren't open with their product designs, as you can well imagine. The commercial use is increasing and we support them through their design cycle lead times, too.
Having worked with several venture-backed companies who also create semiconductors for specific purposes, and having experienced their short-term need to satisfy managers, investors and board of directors, as a customer I'd find the Propeller and Parallax an ideal partner to use for a commercial design. You're dealing with a company who promises to hold the course and not reinvent ourselves every year. Customers who've chosen the Propeller for commercial products have been very satisfied with their choice. Their decision to use our chip is based on capabilities as much as our company's track record.
You'll be seeing more Parallax evolution for commercial customers this Autumn. A few of our Propeller forum members are even included in these efforts as a focus group.
We assure you that the Propeller 1 and future versions will be around long enough to meet your needs. This is something that isn't easy to come by in this industry: stability, consistency, and support. Just take a look at the BS1 - it's been around for nearly 20 years.
To hobbyists, educators and engineers alike - we're your company. Everybody counts as a key customer in our company regardless of how much you buy. You can hear this statement elsewhere but we practice it every day.
Sincerely,
Ken Gracey
The instrument is actually an "ektacytometer", for measuring red blood cell deformation when subjected to a shear force. Blood samples mixed with a buffer are fed into the space between a fixed perspex cylinder and a concentric rotating perspex cylinder. A laser illuminates the blood cells and the diffraction pattern is detected by the photo-diode sensors and used to determine the deformation, using von Hulst's anomalous diffraction technique. Various rare blood disorders can be diagnosed using this instrument, and the prototype is being used in a Paris hospital. My customer (he works at a university in France) is applying for further funding to take the design to the next stage.
The LPC2148 costs about $10 (one off) and I used a prototyping board costing about $100. We just had to add a few connectors to it.
We have been offered the commercial rights to the design for nothing by the university, but it won't make us rich. The diseases are so rare that only one machine will be needed per country.
Steve, legally, all you need is your copyright notice to have full protection. However, a good attorney will also have several other items in place. For example, you might have a 3rd party that can hold on to the source code and design documents in escrow just for such an occasion. You can also, like we do, actually use all three protection mechanisms. One, being copyright, two, being patents and the third being trade secrets. All three are enforcable in different ways to protect different areas of your business.
Our business model, however, is not dependant on code protection.. Instead, we're a little different in that we plan on attacking our competitors with unique products that are superior in major ways than from the competition. (Think about how Apple approaches their business model.) Our Wingman is a great example... it's the first and only portable attitude display unit on the market and we went right after the most expensive competitors with a unit that is 1/10 of their price but offers a lot more functionallity. Our heads up displays are just a unique in that regard.
Bill
Hello Leon,
I'll take the bait, come on - what's this all about, anyway? If you want us to help you with the project, we will, but your mind is already made up. And I know you are totally capable of doing the project without our help, too.
Maybe we could squeeze most of your application into the Propeller, perhaps even built with Catalina if C is a requirement. If you had a need for using the Propeller in such an application we'd work with you, but if you truly don't want to use it, that's perfectly okay. Nobody at Parallax is going to attempt you to do something you don't want to do. Please use the tools and processors that you find appropriate for the job.
As I mentioned above, the choice to use the Propeller isn't always based on specifications or cost alone. I recognize from your posts that these are not important factors in your decision making process. Leon, you are also part of our team, too. I know that from positive posts you have made in other forums about the Propeller.
Now, how about that project! Who's the customer? How many chips do they need (a joke)!
Sincerely,
Ken Gracey
Note: Leon's post was submitted in two pieces. I replied based on his first quote. The second part of his post was made after he completed the first part. Leon, I wouldn't give away so many details about my customer's project.
This is the time critical code that *has* to be executed within 20 ms, and transmitted to the PC after 50 interrupts:
//---------------------- Timer 0 ISR every 20 ms -----------------// static void timer0ISR(void) { char temp[10]; char s[44]; u_long crc; unsigned int ptr; unsigned int C, A, B, E, F; unsigned int i; static unsigned int n; int R; ticks++; // increment general delay ticks Timer0Count++; // increment 20 ms ticks IO0PIN ^= ACTIVITY_LED; // toggle activity LED C_total = A_total = B_total = E_total = F_total = 0; if (!(IO0PIN & PB)) // check if Panic button has been pressed { #ifndef NODEBUG debug_printf("Panic button pressed!!!\n"); #endif cancel(); // cancel function resets processor using watchdog } // Skip if bridge mode or wait for '1001' from WinEkta if ((flag == 1) && (Send_data == 1)) { for (i = 0; i < 16; i++) { C_total += getAnalogueInput_AD1(0); // P0.6 A_total += getAnalogueInput_AD0(2); // P0.29 B_total += getAnalogueInput_AD0(3); // P0.30 E_total += getAnalogueInput_AD0(7); // P0.5 F_total += getAnalogueInput_AD0(1); // P0.28 } C_temp += (C_total >> 4); A_temp += (A_total >> 4); B_temp += (B_total >> 4); E_temp += (E_total >> 4); F_temp += (F_total >> 4); // every 50 interrupts (1 s) average the values and transmit them to the PC while (n++ < 49) { T0IR = 0xFF; // Clear the timer 0 interrupt return; } C = C_temp/50; A = A_temp/50; B = B_temp/50; E = E_temp/50; F = F_temp/50; // get status bits get_status(); // get motor speed R = get_motor_speed(); // increment secs counter secs++; // construct output string my_itoa_4(C, temp); ptr = 0; s[ptr++] = 'C'; for (i = 0; i < 4; i++) s[i+ptr] = temp[i]; my_itoa_4(A, temp); ptr = 5; s[ptr++] = 'A'; for (i = 0; i < 4; i++) s[i+ptr] = temp[i]; my_itoa_4(B, temp); ptr = 10; s[ptr++] = 'B'; for (i = 0; i < 4; i++) s[i+ptr] = temp[i]; my_itoa_4(E, temp); ptr = 15; s[ptr++] = 'E'; for (i = 0; i < 4; i++) s[i+ptr] = temp[i]; my_itoa_4(F, temp); ptr = 20; s[ptr++] = 'F'; for (i = 0; i < 4; i++) s[i+ptr] = temp[i]; my_itoa_4(R, temp); ptr = 25; s[ptr++] = 'R'; for (i = 0; i < 4; i++) s[i+ptr] = temp[i]; my_itoa_4(secs, temp); ptr = 30; s[30] = 'T'; s[31] = '0'; s[32] = '0'; ptr = 33; for (i = 0; i < 4; i++) s[i+ptr] = temp[i]; ptr = 37; sprintf(temp, "S%0.2x", status); for (i = 0; i < 3; i++) s[i+ptr] = temp[i]; crc = crc32(s, 39); sprintf(temp, "%0.8x", crc); ptr = 40; for (i = 0; i < 8; i++) s[i+ptr] = temp[i]; s[48] = '\0'; _puts0(s); // transmit string to PC via UART0 IO0PIN ^= TX_LED; // toggle transmit LED C_temp = A_temp = B_temp = E_temp = F_temp = 0; n = 0; } // feed watchdog WDFEED = 0xAA; WDFEED = 0x55; T0IR = 0xFF; // Clear the timer 0 interrupt }Leon, in a prior post you mentioned that the project is an "ektacytometer" for measuring red blood cell deformation when subjected to a shear force. This seems like an instrument that would really benefit from having a visual display to couple whatever datalogging/communication requirements. Since the device is being developed for hospital use you could simplify the setup and feedback by providing something that humans relate to so readily - a display. And the Propeller offers convenient graphics/text via composite or SVGA. How about if you talk to the customer and sell them on the many benefits of the Propeller's display capabilities?
Perhaps if you demonstrate the visual abilities of the Propeller they'll come right back to you for Phase 2 of the product development. You're one of the few engineers who knows what's possible with the Propeller in this regard. Everybody else will be thinking about whatever off-the-shelf solution they can find, won't they?
Just tell me where to send the engineer if you need help. . . unless your mind is already made up!
There is a large program running on the PC that displays all the data, logs data for each patient, prints out the results, etc. The machine is actually controlled by the PC via my ARM software, all that the operator has to do is put the sample into the machine and select the type of test. My ARM program controls all aspects of the test, depending on what has been selected using the PC program.
Is Marijuana legal in the UK?
-Phil
One opinion would be obvious: "If the pros or industry isn't using the Propeller, why should I use it" and Ken answered that quite well.
The other opinion would be based on how Leon was treated. From their point of view, Leon is a Senior Member with an incredible 3500+ postings. That affords him some credibility. Leon tried to make a point, a point that someone else with Pic or Avr experience may relate too. But he didn't do a very good job of it, and so he got pee'd on.
Good on you for contributing it.
Statement retracted, and message deleted. My apologies.
OBC
-Phil
Not all instruments are all that complicated, compared to the one that Leon described. As a case in point, I just finished a project for a University group where a $15000 instrument they purchased did not have enough battery life or data storage capacity or the correct timing for what they wanted to do. I made them an add-on box to add those features, but to do so I had to hack into the controls of the instrument and was surprised to find that it was built around a Ticket processor, which is kind of like a BASIC Stamp. How much more it could do and more easily with a Propeller at its core. It included DOS software for Windows 3.1, so that gives you an idea that the instrument could be successful for years without supercharging its interface or its innards. The same is true of so many single purpose instruments.
Anyone can find something else that will do the job, for better of for worse. But can you find the SAME chip that will do all of these things, for better or worse. The answer is a simple and definitive NO.
I have run my own companies since 1976 and designed and programmed them using many different micros. I have my favourites - the Motorola 68705 family (now known as Fairchild MC9S08 family) was one, but now it is the Propeller - and it wins hands down.
One of the problems when you have a range of designs is to minimise the inventory. This is extremely important. The main reason, is not what you would expect, that being inventory cost. It is actually a little thing called consistency of supply. Every few+ years as the economy goes into recession the building of the next generation of factories slows down. As we come out of recession and the demand starts to outstrip supply. The new factories cannot handle the new demand. The lead times increase. Then the vultures come in and buy up all the stock to sit on and scalp the market when it gets desperate. I have seen chips go from ex-stock to 1 year lead times overnight. DRAM used to be a classic of this. We got caught with modem chips where Rockwell went to allocation and we would only get 20-30% of our projected orders, without increaded demand.
So, what does this mean. Well, if I have fewer different chips used in my designs, I can hold a buffer stock more easily. I can then determine which products to produce with a limited supply of parts. Am I going to make 1,000 of item A that I make $10 profit, or am I going to make 100 of item B where I make $100 profit. It doesn't take an accountant to figure this out.
And notice I haven't addressed the other advantages of using the one chip for lots of products - things like code, familiarity, layout, etc.
Price of the Prop is really not an issue. We would all like a cheaper prop. If the other commercial products use a lot of Prop chips, the price will come down. As Ken says, some of those products are out there, but they are often confidential and cannot be revealed here. Wjsteele has showed his product off for us - congratulations Bill. We just need to use enough features of the prop to ensure it is price competitive.
As for security, did you see the other thread with a link to the person who revealed the internals of the prop die? He etched the plastic packaging off the chip, then the first interconnect layer. Pitty Parallax had already released this info years ago
My apologies that much of this is Off Topic, but we do need a reality check from time to time.
1) I agree that there is no ultimate code security. But it would be nice to have a way to prevent wholesale copying of commercial code with just 5 minutes of effort from the perp.
2) If Parallax granted me one wish, I'd wish for a PASM quick reference. I'm certain I'd use the Prop more often if I didn't have to wade through a thick and wordy manual.
A perfect example of a quick reference is Herbert Schildt's "C the Pocket Reference." I'd love to have such a short and concise summary of PASM! I like the fact that he includes well-considered one or two line code snippets of almost everything. An example is far more valuable than a thousand words, if one is in a hurry. And we always are in a hurry.
In any event, PASM is the only way to get real speed out of the Prop. Otherwise, it's meaningless to talk about the Prop being a 160 or 200 MIPS machine.
There has been a quick reference in the Propeller Tool for as long as I can remember. Simply click on the Help menu and select "Quick Reference."
The PASM is below the Spin reference.
Bill
Humm... That is a very nice command summary. Particularly welcome is the operator summary on the last page.
Given that the Propeller is a bit odd, perhaps some selected examples and a sentence or two could be devoted to such topics as: Cog architecture, conditional execution, the choice of 2's compliment arithmetic, the composition of an instruction word, the two-bit condition code register, the unusual TRUE and FALSE, cog synchronization, etc. I raise the Schildt text again as a pretty good model of what I wish we had for the Prop.
Blaise Pascal: "I would have written a shorter letter, but I did not have the time."
Pity that our two choices are 1) a list of commands and 2) a 300 page tutorial that starts with the creation of the World...but leaves out all the good stuff.
Correct, but I also don't see why anyone would think that it would. From what I can tell from the information on the XMOS website, I think the target market is different than that of the prop, so comparing the two chips doesn't really say much. The XMOS site pushes USB audio, LED tile controllers, iPOD docks, and AVB ethernet as the main markets. I don't foresee the prop in the selection process of any of those markets (of course I could be wrong).
As a side note, of the two customers (at work) that do LED based products, one uses XILINX on their entire suite of LED control boards and the other uses various chips, but seems to favor Atmel. My guess is that the Xilinx user could accomplish many of their products with XMOS if they chose.
In regards to code security, your code is only as safe as the desire of someone else to steal it. Even devices with internal flash can have the code harvested.
We build several products that boot from external flash devices and the security risk never seems to be an issue. Some of these products are in very competitive markets, so I am sure code security was discussed during design.
My opinion is that if you have a product that relies on code security that heavily, any microcontroller choice will still leave you at risk, so you need hardware methods for security, such as potting. My last employer built modems for ATM machines and the last step of the process was to encapsulate the PCB in a potting compound that turned into a crystallized looking form of tar. Any attempt to dis-assemble the device would destroy it. It also had about 10 "hardware key" connections made with enameled wire that was suspended in the potting material so that if someone did manage to remove the potting and leave the board itself intact, it would corrupt itself repeatedly when powered up due to the missing connections that would have been removed with the potting.
Another method is embedding connections into cases. Many POS terminals contain connection pads in the cases, PCB guides, and other mechanical parts that prevent the product from working properly unless properly assembled into the case with all of the individual hardware components.
Yes - a professional approach geared towards the specific customer. I agree with you, Leon, but in our own flavor and style of course. It's coming.
Try to get in touch with David May and Chip Gracey. Then you know the difference in professionalism and what it means. If you plan to compete a giant, be sure to have a slingshot ;-)
Go the galactic way, go parallax