Repairing and Upgrading an old PM3320A - Including adding FFT!

[idrenth]

So do you have that on a PM3320A? possible make foto and maybe video so that we can see it?

[SaabFAN]

Yes, I have the FFT on my PM3320A.

Here are two screenshots of my Scope.
The FFT-Function is on Page 2 of the Mathematics-Functions.

[idrenth]

Yes, I have the FFT on my PM3320A.

Here are two screenshots of my Scope.
The FFT-Function is on Page 2 of the Mathematics-Functions.

I might think it get unlocked from the ccd part. because that is math calculation area. There wasn't a way to make rom dump from that one? because I also have version 5.1 on it now so the only place would be the ccd part.

[idrenth]

So there is some time past by again and I went back to see what is the differences compare with my board. I know it should be in the ccd part because that is the math area. I saw 2 things different can others confirm if this is correct? @Saabfan or others?

I saw there is a bridge wire on one of the chips and there is firmware number 2 on the chip.

Can someone make a backup from the chip with a D8303 B720.2 on it for me ?

I hoop we can make this upgrade now possible for this scope after a long time.

[SaabFAN]

Funny that you're just now continuing to work on your scope. Yesterday I livestreamed the aftermath of my repair and the calibration-process of my PM3323 on Twitch. The old boat-anchor is now working and performing well, despite me having no way to generate the low frequency Square-Wave and no way to verify the high frequency square-wave for its characteristics. Once I have the Signal-Generator running, that shouldn't be a problem anymore though

You're right, there are version-numbers on those chips. My D8303-Chip Label on the PM3323 shows D8720.2.
According to the Service-Manual, these Chips are 20R8, which are 20L8 (easier to google) with Tristate-Outputs. It might be possible to replace them with GALs from Atmel or Lattice if they support Tristate-Outputs. Unfortunately, I haven't been able to read the chips with my TL866 Programmer (I get a Read Error 31 and then a Warning about wrong Serial-Number with a link to the Website of the Manufacturer). The newer programmers also don't list the chip in their Supported Chips-List. Maybe it is possible to read the chip with an Arduino. But I haven't investigated that possibility beyond one google search.
I do have a defective P2CCD Board of a PM3323 here though. If you want to, I can ship it to you. It's the board that had a transistor release the magic smoke due to a dead P2CCD-Chip on Channel B. I have held onto it to salvage it for parts or sell separately.

Pretty interesting that they have put the chip that enables FFT onto the P2CCD-Board. Either the other chips were full, or they needed to record the Samples in a different way to be able to process them for an FFT later.

Btw. I have made another interesting discovery today: When I closed my PM3323, I transfered my modified A5 and A6-Boards from the PM3320A to the PM3323, which resulted in a crash.
I have identified the problem to be the CPU-Card A6, which seems to have some hidden configuration-jumpers that select whether it is a PM3323 or PM3320A (I didn't find the Jumpers in the Service Manual). I have then transferred the 68010 and the ROMs from the PM3320A to the PM3323 A6-Board and installed it and it works fine.
Installing the A6 Card in the PM3320A makes the Scope think it's an PM3323. I can also ship the A6-Board from the PM3323 I have disassembled. Maybe you can trick it into believing it's a PM3320A, or you can find the hidden configuration-switch for FFT if it isn't on the P2CCD-Board.

[idrenth]

That would be nice to test both cards. p2ccd print and the a6. Did you also saw in the image the jumper wire? it's front from the p2ccd board. Do you also have there a wire bridge on your chip ? It's on the first image bottom with a red arrow on it.  Because I remember you had one pm3320a with fft. So we maybe can figure this out.

Also bad that it happens with magic smoke on the board.

I would like to have that fft function

[idrenth]

I am not sure you can read the 20R8 as the same as the other PLD chips. Philips have their own chips: SIGNETICS(PHILIPS)   so this one PLUS20R8-7   PLCC   dip24   PA28PL24D2P6-YA. So it could that is why you had an error. I am also not sure which one can read that chip out. I have tl866 also but not the cheap version from that.

[idrenth]

Update I found now out the chips is from MMI brand (Monolithic Memories Inc) that is bought over in years 90s by AMD. So I also found the conversion table maybe that helped from TI. Only problem my TL866A only read 1's but didn't give any error.

So first image is from the chip D8303 DB720.1 version without the fft version. PAL20R8AJCS chip.

in the second image was the closes version I could find from AMD
PAL20xxACNS  --->  TIBPAL20xx-25CNT

I check the conversion what tl866a programmer Maybe have listed. 3rd image
GAL20V8B-25LP    <-----  TIBPAL20xx-25CNT

TL866A knows lattice GAL20V8B IC and I selected that one.

It readed the chip but I didn't get any thing back. only 1's

[Chris56000]

Hi!

Sorry if I'm interrupting the party, a couple of Members have mentioned that the PM3320 and PM3320A Service Manuals differ – could someone tell me where I can get a PM3320A Service Manual please?

I'd like to have a read of them both to see if I can spot how the F.F.T. option comes on/is enabled, but so far I've only found the original '3320 S.M. for download in k04BB!

Chris Williams

PS!

I've just realised I had it – it's actually included in the 3323 Service Manual rather than being separately produced!

[Chris56000]

Hi!

Reading back all "1s" with GAL devices usually means the device has had the security bit set – although occasionally, depending on the particular device type, you may read a row of "zeros" on the last or last but one row of the fuse–map even if the device is secured!

There has been a circuit published that can determine the Boolean expression for each o/p of a GAL but unfortunately it will NOT work if the device is the "registered" type, denoted by "R" in the middle of the type number!

My experience with GAL devices in industrial fault–finding is that they're not all that unreliable, and when they do fail it is usually an obviously defective input or output pin that can be found with a simple "V–I" or component–tester type device, nearly always a very low resistance compared to the good pins, or a fairly obvious internal o/c!

Obviously the usual caveat applies, make sure the +5V digital logic supplies powering the PCB under investigation are in perfect condition and all internal decoupling is in good order – unfortunately it does look like Philips do secure their programmable devices, so your only option is a parts mule or replacement board of the same type as the suspect unit!

Chris Williams

[idrenth]

I am not sure yet. maybe the programmer can't read it because the gal needs a clock osc. without the clock it can't get anything out of the gal because it will give 1's. I remember to program a GAL chip with abel in the past and you can get a state table to check if everthing is correct. Problem I don't have that old program and the gallep 4 programmer. Still more research is needed to find this out.

[idrenth]

Had talk today with @SaabFAN and share info and we talk about the options. Till we talked about diffrences on the CPU board and there is a diode parts on logic chip what I had on the 3320a and I didn't. Maybe this unlock the FFT option and says which CPU it is. In the bottom, the board is the pm3320A without the FFT and top on with PM3323 CPU board it looks like the left one is for saying the pm3323 and right one is for the FFT option. We didn't confirm this yet. But would help to figure this out maybe to the next step.

[SaabFAN]

I can confirm now that the FFT and most likely all the other Options are being configured by connecting Diodes from the Select-Lines that originate at D1746 on the CPU-Board to the Databus-Lines connected to D1723.

So far I have identified two config-diodes.
- D1723 PIN 2 ->|- D1746 PIN 18: SELECT PM3323
- D1723 PIN 9 ->|- D1746 PIN 11: ENABLE FFT


In the Service Manual there's also a table that lists the Address-ranges for some of the Options (See Attachments).

So if you own a PM3320A, PM3323, or PM3340, you can apparently enable the FFT-Function by simply connecting a Schottky-Diode from Pin 9 of D1723 (74LS245) to Pin 11 of D1746 (74LS541) on the CPU-Board.
For the PM3320 (early Version without the "A"), there might be additional modifications necessary. For example upgrading the Firmware from 2 ROMs to the 4 ROM-Version and upgrading the RAM.

[idrenth]

Nice!! we finally found it. I saw it was a Bat85 same as V1707 diode. So i will use that one. Now I will also test it if it doesn't have any problems with the 3320A pld chip has the firmware D8303 DB720.1 or D8303 DB720.2 is needed because I have PM3320A with D8303 DB720.1.

[idrenth]

I can also confirm now if you have the D8303 DB720.1 or D8303 DB720.2 firmware on your P2CCD board the FFT get unlocked from the PM3320A with signal Schottky diode Bat85 between D1723 PIN 9 to D1746 PIN 11. I place temporary a 1N5819 on the back from the board because I rather don't place soldering on the side from the ic if there is space left on the back. I will soon replace them with the right components what philips use BAT85.