openrcforums

It is always worth going over the solder joints and checking each one out. From the picture it does look like pin 36 is starved of solder along with pin 34.

Is the chip the correct one? It should be a Atmega 128. Not the Atmega 128L .... I take it that the orientation of the chip is correct?

Usually things like this come down to something very simple, especially when all you have done is change the chip.

ok, resoldered all the pins, checked for shorts, checked continuity..
With the board out of the radio I powered it from the programmer and tried to read the fuses with -B 100 option. Nothing, the same error..
even with the board out and powered from the programmer, this should work right??

So, what you guys think? could it be the chip? I ran out of ideas.. well, didn't have many anyway.

Thanks..

the chip is the atmega128 16AU

Flaps 30 wrote: Usually things like this come down to something very simple, especially when all you have done is change the chip.

Yeah, I agree, nothing that a hammer wouldn't solve

I take it that the chip is orientated correctly?

Yes a hammer would help resolve the issue, along with giving you a lovely set of spare parts..

Of course the chip might be faulty, as the proper precautions were not taken regarding the packaging of the chip. I take it that you did have the earŧhed wrist band fitted, along with all the other standard anti static measures in place when you fitted the chip? FYI - I didn't bother with any of that when I changed my chips.

So what to do? ... I'm stuck at this point, as I assume that you have been thorough with your checks, along with taking note of past posts concerning this subject..

You could purchase a couple of new chips from a respected supplier and not some far east supplier. You might have another chip with you and you could change the one you have with that, The other option would be to go back to the old M64 you pulled out and forget the whole thing. Of course it would be a shame to take that route, as we know that the M128 works and a number of us have done that with no pain at all.

I would tend to stay put for a while, and hopefully with the help of others, this problem can be resolved to make it work or at the very least pinpoint what is at fault.

Check each of the programmer connections. Power up the board, possibly by disconnecting them from the programmer, just leaving power and ground connected, then powering the programmer up. Check you have 5 volts to the chip.
Now, use a resistor (say 1k) and connect each of the programming pads alternately to 5v and ground through the resistor. Check the pad goes to 5v and ground in each case.

Make sure you have cleaned off all flux, I find a "cotton bud" with cellulose thinner works quite well, use one end 'wet', then dry off with the other end.

Use a needle, and lightly scrape the board between each pad, pushing through under the plastic body to make sure there are no shorts under the chip. Sometimes you can get a high resistance short due to flux etc.

Mike.

Flaps 30 wrote:I take it that the chip is orientated correctly?

I think so, the little dot is pin 1, right?

Yes a hammer would help resolve the issue, along with giving you a lovely set of spare parts..

I will skip it for now

Of course the chip might be faulty, as the proper precautions were not taken regarding the packaging of the chip. I take it that you did have the earŧhed wrist band fitted, along with all the other standard anti static measures in place when you fitted the chip? FYI - I didn't bother with any of that when I changed my chips.

No, no precautions

You could purchase a couple of new chips from a respected supplier and not some far east supplier. You might have another chip with you and you could change the one you have with that

I do have another one, from the same supplier.. will probably buy another one from the online shop here in Germany.. but it cost €9.50

The other option would be to go back to the old M64 you pulled out and forget the whole thing.

I think the m64 might not be good anymore.. with all that trouble to remove it..

I would tend to stay put for a while, and hopefully with the help of others, this problem can be resolved to make it work or at the very least pinpoint what is at fault.

Yeah, I will wait for some more ideas.. thank you all for your help..

João

MikeB wrote:Check each of the programmer connections. Power up the board, possibly by disconnecting them from the programmer, just leaving power and ground connected, then powering the programmer up. Check you have 5 volts to the chip.
Now, use a resistor (say 1k) and connect each of the programming pads alternately to 5v and ground through the resistor. Check the pad goes to 5v and ground in each case.

Make sure you have cleaned off all flux, I find a "cotton bud" with cellulose thinner works quite well, use one end 'wet', then dry off with the other end.

Use a needle, and lightly scrape the board between each pad, pushing through under the plastic body to make sure there are no shorts under the chip. Sometimes you can get a high resistance short due to flux etc.

Mike.

Thanks Mike.. will try all that later, starting by cleaning the flux (I only have alcohol) and using the needle..
Thanks

Yes it is the little circle/dot that is pin 1. From your picture it wasn't clear if the chip was fitted correctly.

Ouch! That is a high price for the chip.. Surely there are cheaper suppliers in Germany?

Listen to what Mike has said. Go over things again, remove all the flux residues and look with a high magification glass of some sort at each pin and the spaces between pins to ensure that each joint is correct and that you don't have a dry or no contact joint. Yes I know you have been around it a few times. No harm in looking again. There isn't much else that can be wrong. Do the tests that Mike has suggested after another check.

I will later.. I did check with the multimeter. no shorts and there is continuity to the places where they should connect.. pads, resistors, etc.. Did the neddle test.. they all seemed to sing the same song, so the pins are probably all soldered..
I'm very inclined to replace the chip, but I'm a bit afraid because the other one is from the same supplier and came also in styrofoam.. I'm also a bit concerned because of the pads.. don't know how much more abuse they can take..

João

Is there any test I can do with the scope to see if at least the m128 is working, or if it is completely dead?
because if it is dead it's not worth it wasting more time on it..

anyway bought a couple more m128's just in case.. I hope this ones are at least originals.. and I hope I bought the right ones..
I'm just thinking here.. maybe the seller bought them from the same chinese person and is reselling them.. that'd be funny.. not..

http://www.ebay.de/itm/300774498765?ssP ... 1439.l2649

jhsa wrote:If i power the programmer, do i also have to connect the back of the radio?

If you could program the radio before with the power off, then I see no reason why the back needs to be in place. The +5 Volts comes from the USBASP. So you can check that volts are getting to where they should, and let you scope things out if required, whilst the USBASP is connected.

I have a mySmartUSB Light programmer.. it's a STK500.. at the moment the board is no installed in the radio.. I'm using the 5V from the programmer to power the board, so I guess it should work??!!

The atmega has 2 other pins called MISO and MOSI. is it worth it trying to program it using those 2 pins?

NO, they are not proramming pins.
You could try removing the capacitor on the reset signal, the big one (47uF) just below the crystal.

Mike.

if I try to read the fuses with the programmer disconnected from the pcb, I get the exact same error.. so I guess the programmer is not seeing the atmega whatsoever

I'm out of ideas also, everything I would have tried has been suggested already. Do you have another programmer, maybe a USBASP to try?

no, I only have this one..

This would be the second failure out of a fair number that have gone in without a hitch. IMO if the programmer worked fine with the original chip, then there is no reason that I can see as to why it shouldn't be able to to so with the new one.

There is so little that can go wrong, other than poor solder joins/short circuits or a faulty new chip, whhich isn't unknown but not that common unless they are abused. Hence why the emphasis on checking and checking again the work done.

I found this document that goes some way to understanding how to program/flash the chip ----> In-System Programming

Don't give up João. I feel that a lot of others here would like to see the outcome of your adventure. The answer is likely to be something fairly basic. Have you another TX to check to see if your programmer is still healthy?

yes, I have my other radio..
I will only give up if the board gets damaged.. Or maybe not if someone has a stock board that I could buy not so expensive in case it is necessary..

Just tried my programmer with my other radio.. it works very well. Normally I don't supply the radio from the programmer when I flash but Tried just to test, and it works also..
Just cleaned all spaces between pins with a needle, reflowed the solder in all of them, nothing. I think this chip is dead..

João

Apparantly, there is a way to force the chip into serial programming mode. One of the fuses can disable serial programming mode.
What you need to do is the following:
Connect pin 1 to ground, and also connect the SCK signal from the programmer to ground.
Apply power to the M128, this will cause a power on reset with pin 1 (PEN - programming enable) active low (and SCK low). This forces serial programming mode on.
WIth power still connected, connect the programmer, and then attempt to program the chip.
This is documented in the M128 (and M64) datasheets.

Mike.

2 questions Mike..
1- just connect the SCK pad to ground, so the programmer and SCK pin from the 128 are grounded at the same time, and keep them grounded all the time?

2- if the board is powered then I don't send power from the programmer, right?

And if this fails the chip can go out of the window I suppose.. not without being hammered a few times of course

1. Yes, as long as the programmer has a resistor in the SCK line, otherwise you would short out the output from the programmer processor. If in doubt, add a 100 ohm resistor in the line.
2. Correct.

Mike.

ok, just to make sure I understood, I will add a 100 ohm resistor from the programmer to the SCK pad on the board.. then ground the pad and keep it grounded all the time.

Another question, can I power the board using the wire that powers the sound module with battery voltage taken from the capacitor (Mine is yellow) next to the regulator? I would use the battery pack to power it..

Thanks,

João

Thought I would jump in here, as Mike might have gone to bed. Much of what Mike has mentioned is in the document that I gave a link to..

jhsa wrote:ok, just to make sure I understood, I will add a 100 ohm resistor from the programmer to the SCK pad on the board.. then ground the pad and keep it grounded all the time.

Yup.. Programmer SCK line is grounded via the 100 ohm(ish) resistor to 0 Volts. Pin 1 of the M128 is connected to 0 Volts while you try to program the chip.. Still using the -B 100 as before

jhsa wrote:Another question, can I power the board using the wire that powers the sound module with battery voltage taken from the capacitor (Mine is yellow) next to the regulator? I would use the battery pack to power it..

I think that you will find that the M128 is powered via your USBASP device.. You should see +5 Volts on pin 21 of the M128. So no need to do what you propose. Just check to see if you have +5 Volts on the chip.. You never know.

Yup.. Programmer SCK line is grounded via the 100 ohm(ish) resistor to 0 Volts. Pin 1 of the M128 is connected to 0 Volts while you try to program the chip.. Still using the -B 100 as before

And the SCK pin of the 128 is also grounded, right?

The programmer only gives power when doing something.. so, it's just a few seconds for reading the fuses for example.

If your programmer only outputs +5 volts when it is doing something, then it is different to the ones I use. The ones I have output 5 Volts all of the time. I would double check that as being the case with yours. I would doubt that the chip would like the 5 Volt supply to just come on during the programming time, that sounds a bit suspect.

As for the SCK (Serial Clock) pin of the M128.. I was just following what Mike has written.. Yes he has mentioned the SCK pin being held low..

yeah, my programmer just supply the 5V when is doing something.. and it works to pogramm the other radio.. I never actually use it like this but just tried to make sure it works..
But that is not the problem with this board.. will try Mikes idea in the morning..

Thanks

João

Okay... Yes I hear what you are saying about the 5volt supply.. It was just that when I did my two chips (one with the Smartieparts board and another with a Chinese USBASP reprogrammed to remove the SCK error) both of them gave the chip a continual supply.. Maybe giving chips a 'first birthday' with programmers that have a switched supply that only comes on just before sending data, might be part or all of the problem..

Have a strong coffee before playing later today...

i also tried with the radio powered. The same result. The programmer doesn't see the chip...