@Cluso - I can understand, and empathize, if you think this project trespasses into your multi-prop domain. In the same situation, I'd be angry too. But I think it's a common enough notion that nobody could truly think to have exclusive rights to it forever. I support mctrivia's efforts because, frankly, your ideas disappoint me. In your critique of McT's design, you don't see a use for it, and that's what makes me disinterested in your "multi-blade" boards... you have already designed in all the functionality *you* think they need. What I need is a mega-multiprocessor clean slate, like a petri dish of stem cells, that I can configure in whatever fashion I choose. Also, last time I looked, you don't offer an assembled unit, even as an option. So I applaud your skill as an engineer, your multi-blades are pretty cool, but they're not what I want... try not to be too annoyed when someone else comes up with the design that I (and some others) do want.
Thank you, very much, for pointing out some areas in McT's design that may need further scrutiny. After all, that's what a community like this forum is for, and your contributions to the Zicog project are tremendous... but that's not a multi-prop project is it?
@McT - As stated above, I can really understand Cluso not wanting to give aid to "the competition", but he was nice enough to point toward some things. In electronics, I'm strictly a technician, not an engineer, but I'd suggest looking into the need for shielding traces between the buss traces. I'm not saying they are needed, I'm saying you should ask if they are.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
"They may have computers, and other weapons of mass destruction." - Janet Reno
it would be preferable to be able to shield traces but this would require a 4 later board to effectively do adding another $30. better shielding can be done still by grounding the via to every other pin on the bus but then you lose pins. I am still thinking about switching to a 4 layer board if I can eat most of the cost myself. I do not want to charge More then $180
as for competition I have recommended his tri blade to several people even here. I am not trying to compete. My design is smaller and does not have all the built in ram and other features.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Need to make your prop design easier or secure? Get a PropMod has crystal, eeprom, and programing header in a 40 pin dip 0.7" pitch module with uSD reader, and RTC options.
I don't have a problem with the fact he said there was anything wrong it was that he makes fun of my education and is very vague on what he thinks is wrong initially just saying it will not work
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Need to make your prop design easier or secure? Get a PropMod has crystal, eeprom, and programing header in a 40 pin dip 0.7" pitch module with uSD reader, and RTC options.
I'm not suggesting you should put bus shielding in because I mentioned it. I don't want this thing to cost any more than it has to. What I'm suggesting is that you find out if shielding the bus lines is a necessary evil.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
"They may have computers, and other weapons of mass destruction." - Janet Reno
Unfortunately us being human, we are not perfect.. Public forums are people to publicly come and say "look what I have done, or what I can do".. This is good and bad, because for people like myself that don't have the proper education to know all of the technical aspects, I rely on others for help, and also for inspiration. That begin said, I assume that the people with these skill sets come to places like this to not only show what they have but help others to get to that level with their experience. I also believe that if you show what you have to others, it should be expected that that will inspire others to do the same if not more. If one doesn't like this fact, one should not post their work in a public sharing forum.
I joined this forum a few months ago, and have had nothing but positive responses from everyone.. When I make mistakes in my logic or even just stupid mistakes. I post and expect to be corrected, but in a reasonable fashion. The posts above remind me of the Apple BBS scene flame wars when I was 12 ( a long time ago ).. If help has been asked for, I was raised to give that help if I was able, or at least try to help in a POSITIVE manor, not a NEGATIVE one..
Since everyone has an opinion, and is entitled to express it, (within reason and morals ), Mine is simple, I am here to learn and to help out when I can, without malice, I may not always be right, but I expect to be told I am wrong with respect, and have an explanation of what I am doing/saying wrong so I can learn and not continue making that mistake, but the explanation is needed.
Lets all play nice.
KPR
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
New ICON coming, gotta wait for the INK to heal, now we have colour!
WNed shielding is not needed for low frequency communication. at 8 inch long traces at the upper limits of what the prop can do it may cause a few problems without but error rate would still be prety low. not sure on the exact ammount would have to dig out my old books
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Need to make your prop design easier or secure? Get a PropMod has crystal, eeprom, and programing header in a 40 pin dip 0.7" pitch module with uSD reader, and RTC options.
We are all a community of human beings. Given that, the diversity of views and approaches will be different. Being helpful is a good goal for contributors. Sometimes we say well meaning things that may be regrettable later; I have done it on this forum too often to varying degrees of discomfort.
That being said, engineering can be tough. All the theory in the world helps to achieve specific resullts but is no way a guarantee of perfection. Knowing enough about something to achieve desirable results is what is important. How much "is enough" is hard to quantify and specify when varying degrees of knowledge and ability are present. I tend to trust people with vast experience because I know how getting painted into a corner failing for various things tastes ... it ain't Carmel.
I remember once working with a hardware fellow I'll call Hank who "knew everything" on a project. For 6 months I was writing code to help him understand the behaviour of his board. The first problem encountered was unusable JTAG because Hank didn't really study how or why the cpu connection requested by the software team was to be used. He redesigned the JTAG breakout for the vendor a few times [noparse]:)[/noparse] The second problem was the land-pattern for the SDRAM was wrong and various bits popped everytime the board was moved ugh. From there on, it got tougher. Everytime we got past one problem and added a new feature, things would mysteriously break. A big issue turned out Hank did not ensure enough ground-plane in the board (another was his lack of FPGA programming experience - I've worked with top notch teams and it was obvious he was green there). Another hardware engineer took over and had a product ready to ship soon thereafter.
Regarding multiple Propeller loads: The input capacitance per propeller pin is very small ... typically 6pf according to the spec. Loading for some number of propellers would not be too bad. The little stampish 1x1 propmod boards are attractive for finding out what limits exist in this and other areas of a design. Signal integrity would need to be checked with a high quality scope. The main concerns for AC signal integrity as I know it would be ringing and slew rate. My scope is too low bandwidth to get an true picture of propeller pin ringing and overshoot/undershoot. Does anyone have a 1GHz scope? Verification for large systems can be rough and takes lots of effort and money; small projects should not require so much. Question is: When does a multi-propeller become a large system?
Whatever happens I hope it doesn't cost too much blood, sweat, tears, and money to make something worthwhile.
I have a lot of experience with things like bus loading, and not having to deal with that kind of bupkis is one of the things I love about the Prop.
First of all, the Prop has extremely strong drive transistors; each and every one of its 32 I/O pins can source or sink enough milliamps to light up the room with a white LED. This is not true of the chips which have input capacitance issues. People who are used to building conventional CPU buses where fan-out is a critical PITA are not in the habit of shrugging and ignoring this, but you can do that with the Prop.
Second, there are circuits such as firewire and USB 2.0 drivers and those pesky bus circuits where line length is critical, but those circuits are being operated at such high frequencies that typical Prop comms look like DC by comparison. Operating in PASM, the fastest you could possibly toggle a line in software is around 10 MHz. The highest frequencies you could achieve even with the counter PLL's are around 160 MHz. At those frequencies, line length differences of 4-5 inches are no problem; at the software driven frequencies line length has no meaningful effect on a single circuit board.
That said, there is an argument to be made for not directly linking prop pins, since if you drive one high and one low those strong drive transistors I mentioned will try to crowbar the 3.3V supply. I don't think it's likely the Props would be damaged, especially by a short duration short of this type, but it could create havoc with the power supply especially if more than a couple of pinsets did this at once. There is a real danger of a software bug manifesting as "the thing resets at random" because you're noise hashing the power supply. Current limiting resistors on such connections would be very advisable. Putting a 100 ohm resistor in series with all such pin-pin connections eliminates the chance of either shorting the power supply or damaging the drive transistors, while leaving plenty of ooomph to drive high frequencies against whatever capacitance accumulates. However, there are other arguments for leaving the crowabar-ability in place, not least of which is that you might really want to use those lines to power an output. In a truly general-purpose board that's not unlikely (and one of the things I had in mind when I suggested routing unused pins to vias for outside hookup).
My experience so far with the Prop is that it is an extraordinarily robust chip, as one might expect from a company that sells so much of its output to children, and it puts up with stuff that would make most computer chips fall on their back with their legs in the air. Just producing glitch-free video in software for hours on end is an amazing thing; it's easy to lose awareness of just how _unusual_ that is. Going from the Prop back to a chip that uses external bussed memory and I/O and an interrupt system is truly painful.
Hey, anyone can tell me I'm full of poop. Just include that it's nice poop, even if you don't want it.
I expect to be corrected, even if I think I'm right. I just hope the person who does it, chooses the nicest way he/she can. It may not work the first time, so be pre-warned.
I come and go on this forum, mainly from the challenges of life, but I've been around for a long time. I just want people to have some consideration when dealing with others. It is easy to get annoyed, or aggravated, especially dealing with me. I'm slow to view things as others do, but I will try as long as you will.
I hold no malice to anyone here. I hope everyone can get past any issues, and move on to do bigger and better projects.
James Long said...
Hey, anyone can tell me I'm full of poop. Just include that it's nice poop, even if you don't want it.
I expect to be corrected, even if I think I'm right. I just hope the person who does it, chooses the nicest way he/she can. It may not work the first time, so be pre-warned.
I come and go on this forum, mainly from the challenges of life, but I've been around for a long time. I just want people to have some consideration when dealing with others. It is easy to get annoyed, or aggravated, especially dealing with me. I'm slow to view things as others do, but I will try as long as you will.
I hold no malice to anyone here. I hope everyone can get past any issues, and move on to do bigger and better projects.
James L
Ditto..
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
New ICON coming, gotta wait for the INK to heal, now we have colour!
localroger my thinking to prevent the crow bar effect was to pull the lines up with resistors and tell everyone to use like i2c change to input for high
as for the pins you can cut the traces on the back of the board for the lines you don't need and solder to the via.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Need to make your prop design easier or secure? Get a PropMod has crystal, eeprom, and programing header in a 40 pin dip 0.7" pitch module with uSD reader, and RTC options.
I've been following this thread, but have been pretty busy without much to add.· I had an "oh Wow!" moment reading about the tremendous amount of experience you people have and are sharing with each other.· Mostly we're hearing about someone trying to do something way beyond their personal experience and expertise and wanting help from someone with a different skill set, but it's still impressive to think about the accumulated years of experience in digital design and implementation, analog design and implementation, etc. available here in these forums (fora?)
I have actually done this before with motorola 68hc11 to limited use. I came on here looking for what people would want in a system like this.
yes it has been a few years since I have had time to play with micro controllers like I did in school. And it is an order of magnitude more complex to connect up to 64 props together compared to 8 68hc11 with a 3 wire bus.
it is impressive though how much talent is here. I have great respect for several people here whom seem to be jeneiouses at the prop. You, old bit collector, and heater to name a few. Hippy and jazzed i have herd less from but also seem to know a lot about what they are doing. I pale in my knowledge in comparison to any of you.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Need to make your prop design easier or secure? Get a PropMod has crystal, eeprom, and programing header in a 40 pin dip 0.7" pitch module with uSD reader, and RTC options.
OK Back on Topic but first I hope I have not offended anyone. If so I am sorry.
localroger said...
Current limiting resistors on such connections would be very advisable. Putting a 100 ohm resistor in series with all such pin-pin connections eliminates the chance of either shorting the power supply or damaging the drive transistors, while leaving plenty of ooomph to drive high frequencies against whatever capacitance accumulates. However, there are other arguments for leaving the crowabar-ability in place, not least of which is that you might really want to use those lines to power an output. In a truly general-purpose board that's not unlikely (and one of the things I had in mind when I suggested routing unused pins to vias for outside hookup).
There is enough room length wise down the board to put in 0603 resisters now that I have spaced each props header 0.1" apart so daughter boards could have a little space. But there is only 15mil center to center between the traces and this I am pretty sure is less then an 0402 quad resister is packed. Also the assembler does not want to deal with components smaller then 0603.
Originally I could not figure out how to utilize P8 and P9 but have just realized how I could connect P9 of each chip to the P8 of the next chip. I can also connect this ring trace to a pin so it could be used as a ring or as a way to regain 1 more pin from each chip.
jazzed said...
The main concerns for AC signal integrity as I know it would be ringing and slew rate. My scope is too low bandwidth to get an true picture of propeller pin ringing and overshoot/undershoot. Does anyone have a 1GHz scope?
I only have a 100MHz scope but when I get the first prototype running I will take several readings and post the images anyways. a 1GHz scope is probably needed to see the slew rate but if the 100MHz can't show anything then there should be no problems for communication. If anyone has a 1GHz scope and is willing to run some tests for me I would be very appreciated.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Need to make your prop design easier or secure? Get a PropMod has crystal, eeprom, and programing header in a 40 pin dip 0.7" pitch module with uSD reader, and RTC options.
Ok i will be sending this board to be made this weekend so last chance to put in any requests.
I have included the schematic.
There is one error on this diagram that needs to be fixed. The ring should have a header pin attached to it so you can use it to reclaim an io.
The blue lines mean that the same pin on each chip is connected.
The resistor and switch in the top left are physically overlaping. only 1 or the other can be populated. By default a 0 ohm will be placed there unless otherwise requested.
Not shown in close up but there is by the header a 10k resistor tying each bus line high. can be removed if you don't want.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Need to make your prop design easier or secure? Get a PropMod has crystal, eeprom, and programing header in a 40 pin dip 0.7" pitch module with uSD reader, and RTC options.
Mac,
If you have followed the ...6.6MB/s thread, you will understand why pins P0..7 are better for shared Propeller byte data. Perhaps a special ribbon cable could be used for the headers. Not sure why you need the Schottky diode on the reset line but I do find it very odd indeed that you don't have a common pull-up on RESN.
shotkey diode lets you reset each prop seperatly by reset line on propplug header or all at once through resetn.
I have been following the 6.6MB thread and I am thinking the added cost to make the board wider to suport this is worth it. Sucks it will be hours of work for me to change but I will do it. And I will eat the cost difference so the price will still be the same(at least for initial offering).
I am also going to widen the bus to 12 bits instead of 10 making the later 2 optional with 0 ohm resistor.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Need to make your prop design easier or secure? Get a PropMod has crystal, eeprom, and programing header in a 40 pin dip 0.7" pitch module with uSD reader, and RTC options.
By all means save your work! [noparse]:)[/noparse] I forgot about the individual resets ... sorry. I think 10 bits will be fine for parallel transactions between propellers, but it is good to hedge your bets. I noticed you opted out of having series resistors between propeller pins ... must be a space issue.
If you want drive the buslines only Low and let the Pullups bring it to High, then a value of 10 kOhm is to high. Especailly if you want the 6.6MHz or 10 MHz datarates. 8 parallel Port pins have a capacity of around 50pF and with 10k you get a rise time of several 100 ns.
I think a value of 2.2k should be fast enough and draws not to much current.
Go with 2.2K ... I've never seen a 1K pull-up on TTL level pins.
Thing is the data bits "should" always be driven.
Someone might use drive low pull high on a control pin though which might be any pin.
the problem wth drive high is if you are not perfect in your programing you may crowbar the traces. If you drive low only pull high like i2c does then you can't have that problem. Just as fast because you just use dira to send data instead of outa.
edit: woups you are thinking the same. i just miss read you. i have both values so i will try both and send back the results. should i make the pull ups quad packs or individuals. quad takes less space individual gives more hacking flexibility.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
My new unsecure propmod both 1x1 and full size arriving soon.
But that's where you need series resistors. No way in heck am I going to use dira to drive bytes onto a bus. Having the right "realestate" or land-pattern is important for·the FAB ... values can be changed easy.
well i am now including series 100 ohm resistor so it will not crow bar instead you will have 1.6v at inputs if you do. i can place 0 ohms in if anyone wants instead. bus width can be 4,5,8,9, or 12 depending on number and type of resistors put in.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
My new unsecure propmod both 1x1 and full size arriving soon.
well the new design is coming along. the power switch has been banished to the bottom though so anyone that wants to use it will have to flip the board upside down.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
My new unsecure propmod both 1x1 and full size arriving soon.
is there any reason to keep the pull up resistors on the bus lines now that there are series resisters in place? if 1 chip goes high and another low that would cause 16.5mA draw across the 2 100ohm resisters 27mW well within tolerances of the resisters and the prop.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
My new unsecure propmod both 1x1 and full size arriving soon.
Looks like I don't need to make the board much bigger. I am almost done replacing all the parts and it looks like each prop will take up 0.8x0.9" area.
Images are not finalized but i have uploaded schematic and a quick pcb pinout.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
My new unsecure propmod both 1x1 and full size arriving soon.
Comments
Thank you, very much, for pointing out some areas in McT's design that may need further scrutiny. After all, that's what a community like this forum is for, and your contributions to the Zicog project are tremendous... but that's not a multi-prop project is it?
@McT - As stated above, I can really understand Cluso not wanting to give aid to "the competition", but he was nice enough to point toward some things. In electronics, I'm strictly a technician, not an engineer, but I'd suggest looking into the need for shielding traces between the buss traces. I'm not saying they are needed, I'm saying you should ask if they are.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
"They may have computers, and other weapons of mass destruction." - Janet Reno
as for competition I have recommended his tri blade to several people even here. I am not trying to compete. My design is smaller and does not have all the built in ram and other features.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Need to make your prop design easier or secure? Get a PropMod has crystal, eeprom, and programing header in a 40 pin dip 0.7" pitch module with uSD reader, and RTC options.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Need to make your prop design easier or secure? Get a PropMod has crystal, eeprom, and programing header in a 40 pin dip 0.7" pitch module with uSD reader, and RTC options.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
"They may have computers, and other weapons of mass destruction." - Janet Reno
I joined this forum a few months ago, and have had nothing but positive responses from everyone.. When I make mistakes in my logic or even just stupid mistakes. I post and expect to be corrected, but in a reasonable fashion. The posts above remind me of the Apple BBS scene flame wars when I was 12 ( a long time ago ).. If help has been asked for, I was raised to give that help if I was able, or at least try to help in a POSITIVE manor, not a NEGATIVE one..
Since everyone has an opinion, and is entitled to express it, (within reason and morals ), Mine is simple, I am here to learn and to help out when I can, without malice, I may not always be right, but I expect to be told I am wrong with respect, and have an explanation of what I am doing/saying wrong so I can learn and not continue making that mistake, but the explanation is needed.
Lets all play nice.
KPR
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
New ICON coming, gotta wait for the INK to heal, now we have colour!
WNed shielding is not needed for low frequency communication. at 8 inch long traces at the upper limits of what the prop can do it may cause a few problems without but error rate would still be prety low. not sure on the exact ammount would have to dig out my old books
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Need to make your prop design easier or secure? Get a PropMod has crystal, eeprom, and programing header in a 40 pin dip 0.7" pitch module with uSD reader, and RTC options.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
"They may have computers, and other weapons of mass destruction." - Janet Reno
That being said, engineering can be tough. All the theory in the world helps to achieve specific resullts but is no way a guarantee of perfection. Knowing enough about something to achieve desirable results is what is important. How much "is enough" is hard to quantify and specify when varying degrees of knowledge and ability are present. I tend to trust people with vast experience because I know how getting painted into a corner failing for various things tastes ... it ain't Carmel.
I remember once working with a hardware fellow I'll call Hank who "knew everything" on a project. For 6 months I was writing code to help him understand the behaviour of his board. The first problem encountered was unusable JTAG because Hank didn't really study how or why the cpu connection requested by the software team was to be used. He redesigned the JTAG breakout for the vendor a few times [noparse]:)[/noparse] The second problem was the land-pattern for the SDRAM was wrong and various bits popped everytime the board was moved ugh. From there on, it got tougher. Everytime we got past one problem and added a new feature, things would mysteriously break. A big issue turned out Hank did not ensure enough ground-plane in the board (another was his lack of FPGA programming experience - I've worked with top notch teams and it was obvious he was green there). Another hardware engineer took over and had a product ready to ship soon thereafter.
Regarding multiple Propeller loads: The input capacitance per propeller pin is very small ... typically 6pf according to the spec. Loading for some number of propellers would not be too bad. The little stampish 1x1 propmod boards are attractive for finding out what limits exist in this and other areas of a design. Signal integrity would need to be checked with a high quality scope. The main concerns for AC signal integrity as I know it would be ringing and slew rate. My scope is too low bandwidth to get an true picture of propeller pin ringing and overshoot/undershoot. Does anyone have a 1GHz scope? Verification for large systems can be rough and takes lots of effort and money; small projects should not require so much. Question is: When does a multi-propeller become a large system?
Whatever happens I hope it doesn't cost too much blood, sweat, tears, and money to make something worthwhile.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
--Steve
Propalyzer: Propeller PC Logic Analyzer
http://forums.parallax.com/showthread.php?p=788230
First of all, the Prop has extremely strong drive transistors; each and every one of its 32 I/O pins can source or sink enough milliamps to light up the room with a white LED. This is not true of the chips which have input capacitance issues. People who are used to building conventional CPU buses where fan-out is a critical PITA are not in the habit of shrugging and ignoring this, but you can do that with the Prop.
Second, there are circuits such as firewire and USB 2.0 drivers and those pesky bus circuits where line length is critical, but those circuits are being operated at such high frequencies that typical Prop comms look like DC by comparison. Operating in PASM, the fastest you could possibly toggle a line in software is around 10 MHz. The highest frequencies you could achieve even with the counter PLL's are around 160 MHz. At those frequencies, line length differences of 4-5 inches are no problem; at the software driven frequencies line length has no meaningful effect on a single circuit board.
That said, there is an argument to be made for not directly linking prop pins, since if you drive one high and one low those strong drive transistors I mentioned will try to crowbar the 3.3V supply. I don't think it's likely the Props would be damaged, especially by a short duration short of this type, but it could create havoc with the power supply especially if more than a couple of pinsets did this at once. There is a real danger of a software bug manifesting as "the thing resets at random" because you're noise hashing the power supply. Current limiting resistors on such connections would be very advisable. Putting a 100 ohm resistor in series with all such pin-pin connections eliminates the chance of either shorting the power supply or damaging the drive transistors, while leaving plenty of ooomph to drive high frequencies against whatever capacitance accumulates. However, there are other arguments for leaving the crowabar-ability in place, not least of which is that you might really want to use those lines to power an output. In a truly general-purpose board that's not unlikely (and one of the things I had in mind when I suggested routing unused pins to vias for outside hookup).
My experience so far with the Prop is that it is an extraordinarily robust chip, as one might expect from a company that sells so much of its output to children, and it puts up with stuff that would make most computer chips fall on their back with their legs in the air. Just producing glitch-free video in software for hours on end is an amazing thing; it's easy to lose awareness of just how _unusual_ that is. Going from the Prop back to a chip that uses external bussed memory and I/O and an interrupt system is truly painful.
I expect to be corrected, even if I think I'm right. I just hope the person who does it, chooses the nicest way he/she can. It may not work the first time, so be pre-warned.
I come and go on this forum, mainly from the challenges of life, but I've been around for a long time. I just want people to have some consideration when dealing with others. It is easy to get annoyed, or aggravated, especially dealing with me. I'm slow to view things as others do, but I will try as long as you will.
I hold no malice to anyone here. I hope everyone can get past any issues, and move on to do bigger and better projects.
James L
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
James L
Partner/Designer
Lil Brother SMT Assembly Services
Ditto..
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
New ICON coming, gotta wait for the INK to heal, now we have colour!
as for the pins you can cut the traces on the back of the board for the lines you don't need and solder to the via.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Need to make your prop design easier or secure? Get a PropMod has crystal, eeprom, and programing header in a 40 pin dip 0.7" pitch module with uSD reader, and RTC options.
I've been following this thread, but have been pretty busy without much to add.· I had an "oh Wow!" moment reading about the tremendous amount of experience you people have and are sharing with each other.· Mostly we're hearing about someone trying to do something way beyond their personal experience and expertise and wanting help from someone with a different skill set, but it's still impressive to think about the accumulated years of experience in digital design and implementation, analog design and implementation, etc. available here in these forums (fora?)
yes it has been a few years since I have had time to play with micro controllers like I did in school. And it is an order of magnitude more complex to connect up to 64 props together compared to 8 68hc11 with a 3 wire bus.
it is impressive though how much talent is here. I have great respect for several people here whom seem to be jeneiouses at the prop. You, old bit collector, and heater to name a few. Hippy and jazzed i have herd less from but also seem to know a lot about what they are doing. I pale in my knowledge in comparison to any of you.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Need to make your prop design easier or secure? Get a PropMod has crystal, eeprom, and programing header in a 40 pin dip 0.7" pitch module with uSD reader, and RTC options.
Post Edited (mctrivia) : 4/29/2009 1:26:00 AM GMT
There is enough room length wise down the board to put in 0603 resisters now that I have spaced each props header 0.1" apart so daughter boards could have a little space. But there is only 15mil center to center between the traces and this I am pretty sure is less then an 0402 quad resister is packed. Also the assembler does not want to deal with components smaller then 0603.
Originally I could not figure out how to utilize P8 and P9 but have just realized how I could connect P9 of each chip to the P8 of the next chip. I can also connect this ring trace to a pin so it could be used as a ring or as a way to regain 1 more pin from each chip.
I only have a 100MHz scope but when I get the first prototype running I will take several readings and post the images anyways. a 1GHz scope is probably needed to see the slew rate but if the 100MHz can't show anything then there should be no problems for communication. If anyone has a 1GHz scope and is willing to run some tests for me I would be very appreciated.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Need to make your prop design easier or secure? Get a PropMod has crystal, eeprom, and programing header in a 40 pin dip 0.7" pitch module with uSD reader, and RTC options.
I have included the schematic.
There is one error on this diagram that needs to be fixed. The ring should have a header pin attached to it so you can use it to reclaim an io.
The blue lines mean that the same pin on each chip is connected.
The resistor and switch in the top left are physically overlaping. only 1 or the other can be populated. By default a 0 ohm will be placed there unless otherwise requested.
Not shown in close up but there is by the header a 10k resistor tying each bus line high. can be removed if you don't want.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Need to make your prop design easier or secure? Get a PropMod has crystal, eeprom, and programing header in a 40 pin dip 0.7" pitch module with uSD reader, and RTC options.
Post Edited (mctrivia) : 5/6/2009 10:40:37 PM GMT
If you have followed the ...6.6MB/s thread, you will understand why pins P0..7 are better for shared Propeller byte data. Perhaps a special ribbon cable could be used for the headers. Not sure why you need the Schottky diode on the reset line but I do find it very odd indeed that you don't have a common pull-up on RESN.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
--Steve
Propalyzer: Propeller PC Logic Analyzer
http://forums.parallax.com/showthread.php?p=788230
shotkey diode lets you reset each prop seperatly by reset line on propplug header or all at once through resetn.
I have been following the 6.6MB thread and I am thinking the added cost to make the board wider to suport this is worth it. Sucks it will be hours of work for me to change but I will do it. And I will eat the cost difference so the price will still be the same(at least for initial offering).
I am also going to widen the bus to 12 bits instead of 10 making the later 2 optional with 0 ohm resistor.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Need to make your prop design easier or secure? Get a PropMod has crystal, eeprom, and programing header in a 40 pin dip 0.7" pitch module with uSD reader, and RTC options.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
--Steve
Propalyzer: Propeller PC Logic Analyzer
http://forums.parallax.com/showthread.php?p=788230
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
My new unsecure propmod both 1x1 and full size arriving soon.
If you want drive the buslines only Low and let the Pullups bring it to High, then a value of 10 kOhm is to high. Especailly if you want the 6.6MHz or 10 MHz datarates. 8 parallel Port pins have a capacity of around 50pF and with 10k you get a rise time of several 100 ns.
I think a value of 2.2k should be fast enough and draws not to much current.
Andy
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
My new unsecure propmod both 1x1 and full size arriving soon.
Thing is the data bits "should" always be driven.
Someone might use drive low pull high on a control pin though which might be any pin.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
--Steve
Propalyzer: Propeller PC Logic Analyzer
http://forums.parallax.com/showthread.php?p=788230
edit: woups you are thinking the same. i just miss read you. i have both values so i will try both and send back the results. should i make the pull ups quad packs or individuals. quad takes less space individual gives more hacking flexibility.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
My new unsecure propmod both 1x1 and full size arriving soon.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
--Steve
Propalyzer: Propeller PC Logic Analyzer
http://forums.parallax.com/showthread.php?p=788230
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
My new unsecure propmod both 1x1 and full size arriving soon.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
My new unsecure propmod both 1x1 and full size arriving soon.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
My new unsecure propmod both 1x1 and full size arriving soon.
Images are not finalized but i have uploaded schematic and a quick pcb pinout.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
My new unsecure propmod both 1x1 and full size arriving soon.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
My new unsecure propmod both 1x1 and full size arriving soon.