... but PropTool can't find the P1V.
Two different P123 boards.
That means you could sanity check this with a borrowed laptop or another PC or tablet ..
I think you just need a serial port ? (maybe even a terminal ?)
I went to the unconditioned little hole that houses my cnc machine... which is hosted by Windows 7. Initially, I had the same problem. I got busy sweating in the other room ... took a nap and before leaving, decided to give it one more try... AND to my pleasant surprise, it worked and it worked repeatedly, supporting serial rates of 230400... didn't try anything higher:)
So, I came home, plugged in the P123 board and it failed.... so, I took another nap and to my pleasant surprise, when I woke up... it was working on my Windows 8.1 machine.
I have noticed that occasionally, when starting cold, the P1V doesn't show the Cog0 LED lit. That's how it was on first try today. After napping the LED was lit. It is as though there is a chip on the board that has to warm up before it works correctly.
I don't care, I'll just leave it plugged in from now on.
Many thanks to everyone who spent valuable time and effort trying to help with this.
Do you have a can of freeze ?
Could be useful to know which part is marginal, as Parallax are making some A9 boards as we type...
The power supply is OK I presume ?
I usually preface my remarks by reminding everyone that I barely know what I am talking about:)
I think the issue is somewhere in the the reset circuitry. Somewhere, between when the green LED's all light up in Mindrobot/Jac's P123 port... and then the Cog0 LED lights up. Getting to step one always happens... but until something in the system transitions, the P1V is stuck in place. Even after this, I have had problems ... so it looks like even though Cog0 seems to be running it can be stuck in place.
I think Ozpropdev has a better handle on it than I do, and I am sure Ken and Chip are more than aware of it by now. I'm thinking it could be as small an issue as a single resistor. It took me a couple of months to serendipity my way around it, but I think its small change.
But we are talking about a P2 that inherits deeply from the P1... and the real news is that I think Spin->counters is broken in an odd way. I just got up and running again. So I really don't want to say much until I play some more. I saw the issue on the DE2-115 and thought maybe it was my code. But the problem is there in spades in Jac's initial port, so either it is in something Jac did or it is somewhere else in the P1V. I need to nail it down and I am not trying to tease it. I just don't want to be wrong.... but right now, I would bet good money on it.
On the P123 issue, I would like to time this out. I wonder if either Mindrobots or Jac could make it so that when Cog0 is up an operating... the LED cycles on and off slowly enough so that noise on the reset line doesn't cause it.
It is beyond me, but worth doing, I think.
... and the real news is that I think Spin->counters is broken in an odd way.
I don't know if this is what you're referring to, but there was a bug in the 2014 code drop by Parallax that had an "&" instead of an "|" in the part that simulates the Propeller PLLs. This was fixed here: https://github.com/jacgoudsmit/P1V/commit/feada86ade7a5a0dcabaf90a990483c243f7f4f9. Or are you talking about the startup problem with your specific board?
On the P123 issue, I would like to time this out. I wonder if either Mindrobots or Jac could make it so that when Cog0 is up an operating... the LED cycles on and off slowly enough so that noise on the reset line doesn't cause it.
It is beyond me, but worth doing, I think.
It's easy to generate a once-per-second clock. For example you could add/modify the fpga123.v module as follows (NOTE: I didn't test this):
reg [24:0] blinkcount;
reg blink;
always @(posedge clock_50)
begin
blinkcount <= blinkcount + 1; // EDITED
if (blinkcount == 25'd25000000)
begin
blinkcount <= 0;
blink <= ~blink;
end;
end
// replace the line: assign led[7:0] = ~cogled[8:1]; // lit when COG is inactive
assign led[7:0] = ~(blink ? cogled[8:1] : 8'b0); // blink when COG is inactive
Many, many thanks. Have the rug rat today... so I have to wait til tomorrow to give it a try.
Jac said:
It's easy to generate a once-per-second clock. For example you could add/modify the fpga123.v module as follows (NOTE: I didn't test this):
Everything is easy, when you know how to do it:)
On the other issue... sounds close. But I downloaded your zip in the last week or so. So I think that fix would be there?
I am fairly sure there is an issue, but I don't want to mis-state or mistake it... So, I have a little work to do.
And all of a sudden my free time is evaporating... looking bleak til Tuesday.
On the other issue... sounds close. But I downloaded your zip in the last week or so. So I think that fix would be there?
I am fairly sure there is an issue, but I don't want to mis-state or mistake it... So, I have a little work to do.
And all of a sudden my free time is evaporating... looking bleak til Tuesday.
The bugfix that I linked to above has definitely been implemented in the P1V repo.
My free time isn't looking much better, I'm afraid. Hopefully I'll have some time this weekend to finish the QuickP1V PCB design but it's not looking good, in spite of the national holiday on Monday.
I changed the clock to clock_160, but haven't checked it yet. I am well aware that I shouldn't put a P123 on a conductive surface... but that didn't stop me:)... One down, one to go:)
In my defense, I was trying to keep a four year old from knocking over my microscope on his way to my soldering gun.
Comments
I think you just need a serial port ? (maybe even a terminal ?)
That's the next step. Problem is I am betwixt and between. Have to travel to my XP... sometime today, which is all set up for the P2Hot:)
I went to the unconditioned little hole that houses my cnc machine... which is hosted by Windows 7. Initially, I had the same problem. I got busy sweating in the other room ... took a nap and before leaving, decided to give it one more try... AND to my pleasant surprise, it worked and it worked repeatedly, supporting serial rates of 230400... didn't try anything higher:)
So, I came home, plugged in the P123 board and it failed.... so, I took another nap and to my pleasant surprise, when I woke up... it was working on my Windows 8.1 machine.
I have noticed that occasionally, when starting cold, the P1V doesn't show the Cog0 LED lit. That's how it was on first try today. After napping the LED was lit. It is as though there is a chip on the board that has to warm up before it works correctly.
I don't care, I'll just leave it plugged in from now on.
Many thanks to everyone who spent valuable time and effort trying to help with this.
Now, where is that P2;)
Could be useful to know which part is marginal, as Parallax are making some A9 boards as we type...
The power supply is OK I presume ?
I think the issue is somewhere in the the reset circuitry. Somewhere, between when the green LED's all light up in Mindrobot/Jac's P123 port... and then the Cog0 LED lights up. Getting to step one always happens... but until something in the system transitions, the P1V is stuck in place. Even after this, I have had problems ... so it looks like even though Cog0 seems to be running it can be stuck in place.
I think Ozpropdev has a better handle on it than I do, and I am sure Ken and Chip are more than aware of it by now. I'm thinking it could be as small an issue as a single resistor. It took me a couple of months to serendipity my way around it, but I think its small change.
But we are talking about a P2 that inherits deeply from the P1... and the real news is that I think Spin->counters is broken in an odd way. I just got up and running again. So I really don't want to say much until I play some more. I saw the issue on the DE2-115 and thought maybe it was my code. But the problem is there in spades in Jac's initial port, so either it is in something Jac did or it is somewhere else in the P1V. I need to nail it down and I am not trying to tease it. I just don't want to be wrong.... but right now, I would bet good money on it.
It is beyond me, but worth doing, I think.
I don't know if this is what you're referring to, but there was a bug in the 2014 code drop by Parallax that had an "&" instead of an "|" in the part that simulates the Propeller PLLs. This was fixed here: https://github.com/jacgoudsmit/P1V/commit/feada86ade7a5a0dcabaf90a990483c243f7f4f9. Or are you talking about the startup problem with your specific board?
It's easy to generate a once-per-second clock. For example you could add/modify the fpga123.v module as follows (NOTE: I didn't test this):
reg [24:0] blinkcount; reg blink; always @(posedge clock_50) begin blinkcount <= blinkcount + 1; // EDITED if (blinkcount == 25'd25000000) begin blinkcount <= 0; blink <= ~blink; end; end // replace the line: assign led[7:0] = ~cogled[8:1]; // lit when COG is inactive assign led[7:0] = ~(blink ? cogled[8:1] : 8'b0); // blink when COG is inactive===Jac
Jac said:
Everything is easy, when you know how to do it:)
On the other issue... sounds close. But I downloaded your zip in the last week or so. So I think that fix would be there?
I am fairly sure there is an issue, but I don't want to mis-state or mistake it... So, I have a little work to do.
And all of a sudden my free time is evaporating... looking bleak til Tuesday.
The bugfix that I linked to above has definitely been implemented in the P1V repo.
My free time isn't looking much better, I'm afraid. Hopefully I'll have some time this weekend to finish the QuickP1V PCB design but it's not looking good, in spite of the national holiday on Monday.
===Jac
I implemented your "heart beat" function but ...
For anyone else...in Jac's code the line : blinkcount <= 1; should read blinkcount<= blinkcount + 1;
Oops! I edited the code above. Thanks!
===Jac
In my defense, I was trying to keep a four year old from knocking over my microscope on his way to my soldering gun.