| Author |
 Topic  |
|
|
dranore
USA
4 Posts |
 Posted - 09/30/2012 : 02:52:44
|
So I have 3x AM39F032's soldered to 36pin adapters with a custom adapter for those boards to interface with the GQ-4X. I the first one I did never reported anything correctly. Probably a bad chip or a bad soldering job. Probably the latter. :P
The second chip has worked successfully for me in the past, but it was giving me the, seemingly common, verification issues. However, when tested the chip behaved correctly.
Today, after a long break, I soldered up a new board to see if I could address the verification issues. The details are fuzzy now, but at one point today I could ID the previously working chip. Then when I blanked one of the chips soon after, things started to get weird.
Now I can't ID any of my chips with 5.03B, 6.13C, or 6.16. The chips report 0xFFFF for their ID and read all FF. The "good" chips will write at -1/-2, but verification fails and the chips do not work when tested. Nothing seems to have changed on the programmer hardware side, so I'm very confused. Any suggestions? |
|
| Reply #1
dranore
USA
4 Posts |
Posted - 09/30/2012 : 13:05:12
|
So this morning, as a means of isolating the issue I've done a few things:
First, I installed the software and driver on a different PC. Like before, with the first chip I put in, I got a positive ID on a chip with 5.03 (which I tried first). Immediately after the positive ID, the ID checks have consistently returned "FFFF" as if there were no chip in the socket at all. This is without having touched the device at all. It just stopped ID'ing the chip. I went on to test my other chips with similar results. I upgraded to 6.16 and I still can't ID, erase, or write.
Next I tested every pin on my adapter from the the extended address adapter to my 36-pin socket. The A20 line was a *little* finicky, so I resoldered the headers on my adapter socket that connect to the extended address adapter. Otherwise every pin on my adapter reads cleanly for continuity. This was all precautionary anyways as this exact setup has previously worked for me.
I have also soldered a new connector between the extended address board and the 36-pin socket. I also verified continuity for all the pins on my AM29F032B and the 36-pin adapter board's pins. Again, no improvement.
I'm somewhat at my whit's end here. I've now spent almost my entire weekenend on this problem:
- tested three versions of your software on two different machine
- tested three different chips (one that's known to work)
- verified continuity on all components
- replaced one link in the chain as a precaution
Is there anything I can do to ensure a chip will ID properly? Why would a chip ID once, and then immediately stop? How can I rule out an issue with the GQ-4X itself? |
Edited by - dranore on 09/30/2012 16:43:49 |
 |
|
| Reply #2
ZLM
2945 Posts |
Posted - 10/02/2012 : 20:42:43
|
The AM29F032B TSOP40 chip is verified on programmer GQ-4X with adapter ADP-082 with top ZIF module ADP-077.
The ADP-082 has on board signal driver to match the signal between GQ-4X and chip AM29F032B.
If you are using self-made adapter, then the software may not generate the signal correctly. The software is adjusted with ADP-082 configuratin.
I suggest you to get a ADP-082 at least for this chip. You still can use your soldering top board I think. |
|
|
| Reply #3
dranore
USA
4 Posts |
Posted - 12/31/2012 : 00:18:56
|
As for using the ADP-082, I don't believe that will work for me because I would then need to design another adapter to connect my board to your adapter, which would get very expensive in both time and money. I am also aware of others having success with similar setups.
It seems that I have somehow damaged my existing boards. I soldered a new adapter board and it shows the correct id (0104).
Now I have a different problem. When I write to the chip at speeds of 0 to +2, it errors with:
"Write failed" at 0x000000 (even if the chip is verified as erased)
If I write with a -1 or -2, it writes but there are a number of errors introduced. This has happened with the USBPrg 6.17. I will be trying with earlier software next.
Edit:
Testing with 5.03B, and I get similar results. I'm noticing that I'm getting errors at the same spot when writing at speed 0. It seems to be erroring at address 0x000009.
|
Edited by - dranore on 12/31/2012 00:41:45 |
|
|
| Reply #4
dranore
USA
4 Posts |
Posted - 12/31/2012 : 11:18:00
|
I found that the source of my problem with an almost invisible bridge between two pins. Now I can successfully write, but I'm noticing that I always fail on verify with either version of the software at any speed (with and without double write).
There are exactly 64 (coincidence?) errors when I diff the read data from the chip and the source data. 63 of the errors were suppose to be FF and were read as a different value, only one was an FB. There doesn't seem to be an overarching pattern of addresses.
It might still be a soldering issue, but with so few errors I have a hard time believing that. I'm going to look into this more, but any additional suggestions are welcome.
|
Edited by - dranore on 12/31/2012 11:23:37 |
|
|
| Reply #5
ZLM
2945 Posts |
|
| |
Topic |
|
|
|
Suddenly they've stopped working... AM29F032B
