Detects are now reliable before downloading, correct? You're probably now hitting ground bounce problems because ground tracks are too thin. Next board layout, fill empty board space with ground pours.
1: I Connect FTDI USB to PC
-- the FTDI 2 Green LEDs light Brightly.
2: I plug power into P2 VIn, 5V from FTDI -or- External like 9V Battery.
-- the FTDI 2 Green LEDs light dimly.
(-- if I leave it like this - then the LEDs brighten up after about 60 seconds)
(-- if I Disconnect P2 Power only - then the LEDs brighten up again)
3: If I run Detect before 60 seconds - then I usually get DETECT Success repeatedly.
4: If I click UPLOAD - then I get "Uploading" then "Verifying" - then the Popup Window closes after a few Seconds - and the FTDI LEDs stay Dim (Failing ?)
-- Upload / Detect does not work - unless I repower the P2 again.
You could try to measure the impedance of the P2 receive and transmit pins between power and ground rails and compare with other pins you've not yet connected to see if anything has burned in the chip...? Hopefully nothing got shorted out on the serial pins inside the P2 when you were feeding it 5V via the 220 ohms for a while. Maybe the protection diode circuit has gone open circuit which could indicate it burned out or may have fed parts of the P2 with a higher voltage for a while. Perhaps do a diode check between these pins and and the VIO pin for that upper group with your multimeter (VIO_60_63) while unpowered with nothing else connected and see what you get vs the same test for some other as yet unused pins which could not have been damaged in this way. Result should hopefully give some confidence that the P2 might still work.
Also if you go scope TX and RX pins while you reset/detect/download in the software you should see activity on both P2 pins. If you only see your transmit activity into the P2 then it indicates the P2 is not responding which could be a bad receiver pin or bad transmitter pin or both - or some other power related problem. Also check that the reset signal is being driven high in time before the serial activity begins.
EDIT: okay I posted before reading your latest. So you have at least one board that is partially working and not completely dead? Are all your boards basically behaving the same now? Earlier posts indicated something had stopped working completely after being flaky for a while.
The P2 chips just seem flaky.
I hope that I am not wasting my time on these chips.
I need reliable chips in my robots to work in the real world, inside, outside, motorized, etc .
If this is not reliable, I have other 240MHz chips that run at about 50mA that i can just run in parallel using I2C, instead of multicores in 1 chip ...
Ok that's good, so it seems like you have a semi-working setup. Are you planning to scope the serial pins, reset line and input voltages while the chip is being detected/downloaded to determine what you see? Do you have an oscilloscope you can use?
2v at the FTDI TX isn't right. Either the IO pins are being fed 1v8 (and body diodes are conducting to clamp it at 2v) or the FTDI is in a very weak drive mode. Or its a fake?
@jlsilicon said:
The P2 chips just seem flaky.
I hope that I am not wasting my time on these chips.
I need reliable chips in my robots to work in the real world, inside, outside, motorized, etc .
You probably should be starting with a reference board. There's nothing flaky about the Prop2 chips when wired right. Just they do need more board layout consideration because they are faster and bigger than the Prop1.
That's the reason why there is already a bunch of mini Prop2 modules with bare minimum support circuits on the market.
I repeat it again what I said here: It doesn't work well to start with a "quick and dirty" design and try to improve it step by step. There's a high probability that when you fix one thing there's still something else broken. So it still doesn't work and you don't know why.
It's much better to start with an exemplary design where everything is perfect and all recommendations are implemented. Don't care if it costs some $ more, it will pay back by saving lots of trouble. If you think you desperately need to save cost for series production, later, you can then step by step leave out decoupling caps and pullup resistors and try if it still works. If it doesn't you immediately know why.
So please either buy a ready-to-use board or carefully study the existing boards. The P2ME and the KISS boards are quite simple. Every component and copper trace are there for a reason.
I have Designed and ordered/manufactured boards many times for different Microcontrollers, etc.
-- I have assorted projects lying around from them.
Maybe, but you definitely haven't designed a board with high speed digital circuitry on it that passed any EMI test. I apologize if I tell you something you don't want to hear. But it doesn't get better if you ignore it. Google "ground bounce" and "PCB ground plane". You'll surely find some tutorials on how to do PCB layout.
Actually I think I have a good idea what the problem is.
I still have not yet received the AP2114D-1.8V for the 1.8V VReg.
I am still using the MCP1700T-1.8V for the 1.8V VReg. It is probably not enough to support the entire chip - and is failing as the chip comes up.
That would be a good guess.
Jeepers, yes MCP1700 is woefully underpowered for the task.
That said... If you were suppling that with 3.3V (and not 5V -- hmm, but maybe even 5V, just about), then you should get sufficient current out of it to be able to detect the P2 and upload some simple sign-of-life code. So I suspect that regulator is not your only issue with your circuit.
You might make your life easier by adding a PropPlug header, and using that to get things setup (instead of your custom FTxxx interface). At least you'd have less unknowns to start with. You can always bring on more custom circuitry later as you build on your confidence in each circuit section.
Also, one other thing might be that if the FTDI board expects 5V IO then the logic level of the P2 might not be sufficient to drive it reliably at high speeds, especially if there is additional stray capacitance, underpowered regulators and other general breadboard or ground bounce flakiness also contributing to the problem. Your earlier 2V level comment for Tx sounded somewhat confusing, especially so if that voltage level was ever also observed from the P2 transmit driving out data to the FTDI as well. Weak 2V signals feeding 5V input device logic thresholds isn't good, for CMOS inputs particularly.
You probably will need a scope to get to the bottom of all this bringup mess using the parts you have already fitted, or wait for the better regulator to see if that helps. But as VonSzarvas has suggested, a real PropPlug is a good investment if you intend to work with a Propeller 2.
@Rayman said:
I’d imagine a 5v io rigged ftdi chip can be made to work, I’m just not aware of anybody having done it…
You could likely achieve that with a series 10K in parallel with something like a 1-10nF to punch the logic level transition through, on each of the tx and rx lines. For that matter, on DTR, just a series cap with the P2's 10K pullup would be good to pulse a reset. The BJT is not really essential. Pretty sure the P2-EVAL board circuit works like that, as one example. Probably a couple of more recent P1 boards too.
I can only re-iterate the good advice above: Getting a scope and looking at the impact of experiments is pretty crucial with logic play (and for sure, when trying to fix things with resistor ladders and such, especially those that pull to an opposite rail - more often than not they will mangle the logic level and switching speed).
@jlsilicon You could try putting a second MCP1700T-1.8V on top of the first for now...
That won't be perfect, but should provide more current to 1.8V rail...
Ha, I should have thought of that. The core supply needs at least 2.0 Amps to handle higher clock rates. It won't need all that normally but I've managed to force it over even that with crafted overclocking.
Comments
Detects are now reliable before downloading, correct? You're probably now hitting ground bounce problems because ground tracks are too thin. Next board layout, fill empty board space with ground pours.
Ok,
1: I Connect FTDI USB to PC
-- the FTDI 2 Green LEDs light Brightly.
2: I plug power into P2 VIn, 5V from FTDI -or- External like 9V Battery.
-- the FTDI 2 Green LEDs light dimly.
(-- if I leave it like this - then the LEDs brighten up after about 60 seconds)
(-- if I Disconnect P2 Power only - then the LEDs brighten up again)
3: If I run Detect before 60 seconds - then I usually get DETECT Success repeatedly.
4: If I click UPLOAD - then I get "Uploading" then "Verifying" - then the Popup Window closes after a few Seconds - and the FTDI LEDs stay Dim (Failing ?)
-- Upload / Detect does not work - unless I repower the P2 again.
You could try to measure the impedance of the P2 receive and transmit pins between power and ground rails and compare with other pins you've not yet connected to see if anything has burned in the chip...? Hopefully nothing got shorted out on the serial pins inside the P2 when you were feeding it 5V via the 220 ohms for a while. Maybe the protection diode circuit has gone open circuit which could indicate it burned out or may have fed parts of the P2 with a higher voltage for a while. Perhaps do a diode check between these pins and and the VIO pin for that upper group with your multimeter (VIO_60_63) while unpowered with nothing else connected and see what you get vs the same test for some other as yet unused pins which could not have been damaged in this way. Result should hopefully give some confidence that the P2 might still work.
Also if you go scope TX and RX pins while you reset/detect/download in the software you should see activity on both P2 pins. If you only see your transmit activity into the P2 then it indicates the P2 is not responding which could be a bad receiver pin or bad transmitter pin or both - or some other power related problem. Also check that the reset signal is being driven high in time before the serial activity begins.
EDIT: okay I posted before reading your latest. So you have at least one board that is partially working and not completely dead? Are all your boards basically behaving the same now? Earlier posts indicated something had stopped working completely after being flaky for a while.
i never fed 5V to the chip pins.
The P2 chips just seem flaky.
I hope that I am not wasting my time on these chips.
I need reliable chips in my robots to work in the real world, inside, outside, motorized, etc .
If this is not reliable, I have other 240MHz chips that run at about 50mA that i can just run in parallel using I2C, instead of multicores in 1 chip ...
Did a 5V level signal get fed via a 220 ohm resistor, pushing 5mA or so through the protection diodes?
No, we assumed that the FTDI was driving 5V to 220R to Tx, but all I read was about 2V on the FTDI:Tx pin - on a previous board.
I am now using the 3rd Board assembled, just yesterday.
Ok that's good, so it seems like you have a semi-working setup. Are you planning to scope the serial pins, reset line and input voltages while the chip is being detected/downloaded to determine what you see? Do you have an oscilloscope you can use?
2v at the FTDI TX isn't right. Either the IO pins are being fed 1v8 (and body diodes are conducting to clamp it at 2v) or the FTDI is in a very weak drive mode. Or its a fake?
You probably should be starting with a reference board. There's nothing flaky about the Prop2 chips when wired right. Just they do need more board layout consideration because they are faster and bigger than the Prop1.
That's the reason why there is already a bunch of mini Prop2 modules with bare minimum support circuits on the market.
I repeat it again what I said here: It doesn't work well to start with a "quick and dirty" design and try to improve it step by step. There's a high probability that when you fix one thing there's still something else broken. So it still doesn't work and you don't know why.
It's much better to start with an exemplary design where everything is perfect and all recommendations are implemented. Don't care if it costs some $ more, it will pay back by saving lots of trouble. If you think you desperately need to save cost for series production, later, you can then step by step leave out decoupling caps and pullup resistors and try if it still works. If it doesn't you immediately know why.
So please either buy a ready-to-use board or carefully study the existing boards. The P2ME and the KISS boards are quite simple. Every component and copper trace are there for a reason.
Maybe, but you definitely haven't designed a board with high speed digital circuitry on it that passed any EMI test. I apologize if I tell you something you don't want to hear. But it doesn't get better if you ignore it. Google "ground bounce" and "PCB ground plane". You'll surely find some tutorials on how to do PCB layout.
Actually I think I have a good idea what the problem is.
I still have not yet received the AP2114D-1.8V for the 1.8V VReg.
I am still using the MCP1700T-1.8V for the 1.8V VReg. It is probably not enough to support the entire chip - and is failing as the chip comes up.
That would be a good guess.
The 3.3V VReg is the AP2114D-3.3V though.
That might be it.
Also would be good to figure out if you have ftdi chip set for 3.3v or 5v io.
If it is 3.3 v, then don’t need resistors. If 5v probably need bigger resistor value on tx from ftdi…
Jeepers, yes MCP1700 is woefully underpowered for the task.
That said... If you were suppling that with 3.3V (and not 5V -- hmm, but maybe even 5V, just about), then you should get sufficient current out of it to be able to detect the P2 and upload some simple sign-of-life code. So I suspect that regulator is not your only issue with your circuit.
You might make your life easier by adding a PropPlug header, and using that to get things setup (instead of your custom FTxxx interface). At least you'd have less unknowns to start with. You can always bring on more custom circuitry later as you build on your confidence in each circuit section.
Also, one other thing might be that if the FTDI board expects 5V IO then the logic level of the P2 might not be sufficient to drive it reliably at high speeds, especially if there is additional stray capacitance, underpowered regulators and other general breadboard or ground bounce flakiness also contributing to the problem. Your earlier 2V level comment for Tx sounded somewhat confusing, especially so if that voltage level was ever also observed from the P2 transmit driving out data to the FTDI as well. Weak 2V signals feeding 5V input device logic thresholds isn't good, for CMOS inputs particularly.
You probably will need a scope to get to the bottom of all this bringup mess using the parts you have already fitted, or wait for the better regulator to see if that helps. But as VonSzarvas has suggested, a real PropPlug is a good investment if you intend to work with a Propeller 2.
Or, if you can get that ftdi board to work with P1, it should also work with P2.
Think all you need is a multimeter on the TX output of ftdi board, when not connected to P2, to see if 5v or 3.3v io.
I’d imagine a 5v io rigged ftdi chip can be made to work, I’m just not aware of anybody having done it…
You could likely achieve that with a series 10K in parallel with something like a 1-10nF to punch the logic level transition through, on each of the tx and rx lines. For that matter, on DTR, just a series cap with the P2's 10K pullup would be good to pulse a reset. The BJT is not really essential. Pretty sure the P2-EVAL board circuit works like that, as one example. Probably a couple of more recent P1 boards too.
I can only re-iterate the good advice above: Getting a scope and looking at the impact of experiments is pretty crucial with logic play (and for sure, when trying to fix things with resistor ladders and such, especially those that pull to an opposite rail - more often than not they will mangle the logic level and switching speed).
"How to" is an often asked question.
@jlsilicon You could try putting a second MCP1700T-1.8V on top of the first for now...
That won't be perfect, but should provide more current to 1.8V rail...
I was thinking of that. Seems like a delicate task.
But, I think the AP2114H-1.8V is arriving today.
-- Guess not. Should arrve tomorrow then.
Ha, I should have thought of that. The core supply needs at least 2.0 Amps to handle higher clock rates. It won't need all that normally but I've managed to force it over even that with crafted overclocking.
Yeah it was a temp solution, but I forgot about it - that it could cause a problem.
Now I have to wait for the VReg Chips to arrive.