PCB 5.0 and 27C400

By: Damion - 2009-01-19 06:17:53

Hi!

First, I would like to give a huge thanks in advance for any thoughts on my dilemma, any info is appreciated much.

Seems I can't burn 27C400 (AmigaOS ROM). I can read ROMs OK, and verify them as working in WinUAE (Amiga emulator).

OS: WinXP, polling diabled (DisableWarmPoll reg setting, "plug and play" disabled in Device Manager)
HW: Willem dual-power, "latest" PCB 5.0 + 16-bit adapter from Sivava, 40-pin ROM inserted into the 42-pin socket (top two socket pins empty), external PSU
Software: Tried both 0.97ja (verified by others as working to burn these babies, at least with older PCB revision), jumpers at PCB3B, also latest 0.98D9, with jumpers at PCB45, software set to burn 27C400
VCC = 12.5v
VPP = 6.2v

Reading only seems to work reliably with 0.98D9/PCB45 settings.

Here's what happens when writing:

The first few addresses or so may be written OK, then it will crap out with something like "Error at 000003, Chip = FF, Buffer = 6B" or equivalent. Sometimes, it will erroneously write "00" instead of dying with the usual blank (FF). There was one single occasion where it miraculously did write, however, I had to set tWP to max in order for it to burn without randomly failing through the process. After resetting that particular computer, it never worked again.

I've tried about 10 different eproms (working fine prior to 10-15 min UV erase), and make sure to blank check before burning. I've also tried 3 different computers so far (along with all the bios settings, jumpers, software revisions, etc) with the same results on each one.

Seems either:

1. I'm making some real ridiculous error, overlooking something simple

2. The Willem I received is defective

3. 27C400 isn't working on the PCB 5.0 revision
     
4. Current drop on the wall-wart PSUs? (tried 3 so far)

5. Another older software version might work better?

Any ideas from the masters? Anyone want to sell me an earlier PCB revision which works with 27C400, or recommend one that does?

Again, any help is much appreciated!

edit

Ran the hw test, VCC is 12.9v, VPP 6.4. Most of the other pins had around 5.75v. Verified that VCC and VPP is getting to the 16-bit adapter. (Hope this info helps a little.)

Re: PCB 5.0 and 27C400

By: Damion - 2009-01-20 06:17:01

Just a small update: Tried again on another computer (#4), this time an old iPaq running Win2000, also followed the instructions for Win2000 software setup on the Willem site. Same deal here as well - reading 27C400 chips is no problem, blank check also no problem. As soon as I attempt writing, it craps out after b0rking the first several addresses of the EPROM... (ugh!)

Does anyone think it's worth trying to install Win98 on something, and having another go? Again, reading is OK, it's just writing that fails.

Thanks again!

Re: PCB 5.0 and 27C400

By: DL - 2009-01-23 19:49:06

Hi,
Damion wrote on 01/19/09 at 06:17:53:
VCC = 12.5v
VPP = 6.2v
A nit, but you have VCC and VPP swapped.
Quote:
The first few addresses or so may be written OK, then it will crap out with something like "Error at 000003, Chip = FF, Buffer = 6B" or equivalent. Sometimes, it will erroneously write "00" instead of dying with the usual blank (FF). There was one single occasion where it miraculously did write, however, I had to set tWP to max in order for it to burn without randomly failing through the process. After resetting that particular computer, it never worked again.
The datasheet for M27C400 says Ipp and Icc can both be 50mA.   Since a few addresses are programming correctly, it might be a VPP droop problem.

Remove your adapter, enter Test H/W tab, set programmer as indicated, install a 240 ohm resistor between pins 1 and 16 of the socket, enable VPP.   Measure voltage right after enabling, leave enabled for 5-10 seconds and measure again.   VPP should not drop below 12.25V.   VCC should also be able to provide 50mA without dropping below 6.0V for reliable programming.   Note - programming and verifying at the same VCC does not insure reliable programming.

Can you post a close-up of your adapter?   I have the adapter from Willem's shop, so don't know about Sivava's.  A schematic would be helpful, but AFAIK, Sivava doesn't provide them.  Even though they probably copied ATH's design...

HTH,
Dennis

Re: PCB 5.0 and 27C400

By: Damion - 2009-01-24 22:53:55

Thanks DL! I will run the tests, take some photos of the adapter and get back here within the next few days.





Re: PCB 5.0 and 27C400

By: Damion - 2009-02-09 21:21:20

OK - ran the tests, and voltage is holding rock-solid. I'm beginning to think the burner is defective, as now it's not reading correctly (on any computer), either. I may try another manufacturer for the programmer (got a tip from another user on one that seems to work) and go from there.

DL your help was much appreciated, I'll make sure to get a pic of that adapter up shortly.

Re: PCB 5.0 and 27C400

By: DL - 2009-02-11 15:58:14

Hi,
Damion wrote on 02/09/09 at 21:21:20:
OK - ran the tests, and voltage is holding rock-solid.
Could you describe the test conditions?  I find it difficult to believe that any Willem type programmer (except the Willem 4.1) would provide good voltage levels at 50mA.  All other designs use a transistor to switch VCC and VPP and they don't take the saturation voltage of the switch into consideration.  The 4.1 has an optional relay which would limit VCC voltage drop to less than 50mV.  It would still have problems with VPP voltage drop, which is bad for programming EPROMs and older FLASH.
Quote:
I'm beginning to think the burner is defective, as now it's not reading correctly (on any computer), either. I may try another manufacturer for the programmer (got a tip from another user on one that seems to work) and go from there.
Have you tried testing the programmer?  You can check all signals in the Test H/W tab.

HTH,
Dennis