uPB8228C replacement

Ebay?

Datasheet:

Intel also made the 8228. Very likely that Intel designed it, and NEC copied.

Oops, must have made a typo when I searched Ebay yesterday. There's one on Ebay UK too which I've just bought.

Cheers,

Gareth.

Damn, replacement didn't work.

Reason for suspecting this device was that the systems _WR line was held at around 1 volt, other Address/Data lines look normal. Isolating this device from that line produced a solid zero, so worth a punt, I thought.

On boot up there doesn't appear to be any activity on the _WR line at all.

Sigh. Could be anything.

Gareth.

Do you have a schematic? The _WR line comes from the CPU, but perhaps it goes somewhere else. Could the CPU itself be bad?

No I don't have a schematic, but it is a small PCB. The _WR line also goes to 2 DRAM chips, but nowhere else as far as I can tell.

Yes, the CPU could be faulty, the RAM could be faulty, the EPROM could be faulty, and a few other things too. This is a very old board. (Apart from the EPROM, and now the chip I replaced, none of these chips are socketed)

I will go back tomorrow and see if I can see anything else on the scope that might look a little suspect, and what, if anything, is actually happening - I didn't get much further than noticing the _WR was definitely screwy.

Cheers,

Gareth.

Just a thought... If you can isolate it again, try putting a 2.2K resistor pullup on the signal to see if the CPU can drive that lower than .8V

Normally I'd suspect the DRAM chips as well, but not so much if isolating the _WR signal to the 8228 seemed to correct the voltage level.

Also, the 8080A had three power supplies (+12V, +5V and -5V), but you probably already checked/knew that.

Oops, sory about that double post, finger slipped.

Its a public holiday in the UK today but just popped into my workshop cos I'm curious, and this actually belongs to a friend of mine.

I did try a pullup resistor on the _WR line which gave a good zero, as did simply removing the 8228. Only when the 8228 was connected did the line hang at around 2v. A bit weird, but that's what lead me to replace it.

But I think I may be closer now. I'd checked the 3 power supplies, but scoping all the 8080A lines just now reveals one other dodgy looking one, D6. The pulse train is only 3v pk to pk instead of 5v. This line connects only to D6 on the 8228, and on removing the 8228 it remains at 3v. So it looks to be the 8080A processor then, with a faulty D6 (and possibly _WR).

Thanks,

Gareth.

Not supposed to be a pull-up on that /WR line maybe ? Little sm resistor gone open ? Worth hanging a 1k up to rail in case some internal pull-up on one of the chips has gone open ?

Arfa

I tried a pullup on _WR when first investigating. Although it produced a nice zero, there was no activity.

I have since found another duff line, which was only producing 3v - I quickly tried a pullup on that with the pin off the board and in the air, and it suddenly shot up from its 3v and stayed at full value swings, but with an obviously faulty waveform (large RC curves).

So yes, I suspect it it the output tri-state interfaces/pullups that are failing on this device.

There are no sm devices on board, this is from the 1970s/1980s!

Cheers,

Gareth.

Well, managed to get an NEC 8080AFC off the Interweb, socketed it onto the board, and the AMS is now working again.

Hurrah!

Thanks JW and Arfa,

Gareth.