Hardware problem with 464. AY chip fault?

[Novabug]

Hi guys


I recently picked up 2 464's, one model E which works fine and a model D which has a problem. See the first picture. It does this everytime I switch on, with the delete/error key sound repeating. I have checked and cleaned the keyboard and membrane and tested with them connected. Same result.


[size=78%]


I have made the judgement it is the AY chip at fault. Hope that is right. I have acquired an AY-3-8912A from a Spectrum plus 2, Will this chip be a suitable replacement? Does it have a different pin config being an 'A' variant? On the speccy, it is also socketed, do I have to solder a socket in to the cpc board or just put the chip straight on?


Any help would be very much appriacted as I really would like to get this CPC up and running again.


Cheers


Novabug


[/size]

[gerald]

See the first picture. It does this everytime I switch on, with the delete/error key sound repeating. I have checked and cleaned the keyboard and membrane and tested with them connected. Same result.

Do you get the same result without the keyboard membrane connected ? If so, the AY is likely damaged. (TBW, I can't see any picture  )

I have acquired an AY-3-8912A from a Spectrum plus 2, Will this chip be a suitable replacement?

Yes

do I have to solder a socket in to the cpc board or just put the chip straight on?

A socket is a welcome addition if you have one. Amstrad did not put one just to save some money.

[TFM]

Oh yes, to socket things is always a good idea. 

[Novabug]

Do you get the same result without the keyboard membrane connected ? If so, the AY is likely damaged. (TBW, I can't see any picture  ) Yes
A socket is a welcome addition if you have one. Amstrad did not put one just to save some money.

I don't have a socket yet, but could get one. Could I solder it straight in anyway?
I assume the 'A' variant on the Speccys chip is irrelevant and has the same configuration as the normal 8912?




Yes, it does the same without the membrane connected. Sorry, picture hasn't displayed the first time. Cheers to the reply.

[gerald]

I don't have a socket yet, but could get one. Could I solder it straight in anyway?

You can solder the replacement AY, but a socket is still useful if you still have problems after the replacement.

I assume the 'A' variant on the Speccys chip is irrelevant and has the same configuration as the normal 8912?

The pinout and functionality are identical. I guess A variant uses a different technology node.

[Spirantho]

Not quite.... the AY-3-8912A is very similar to the non-A version but not identical.
I think in this case you'll be ok, but from memory if you put the wrong type in an Oric-1, some keys don't work. Tested with several chips. I always thought they were identical too... that took quite some loss of hair before I realised it (when trying to fix an Oric)!

[gerald]

Not quite.... the AY-3-8912A is very similar to the non-A version but not identical.
I think in this case you'll be ok, but from memory if you put the wrong type in an Oric-1, some keys don't work. Tested with several chips. I always thought they were identical too... that took quite some loss of hair before I realised it (when trying to fix an Oric)!

Interresting !
From datasheet point of view, the differences are on the timing side (setup/hold), H/L voltage level and power consumption.
The voltage level differences could be the reason for your ORIC problems as the matrix is scanned using ~2.5V (from a polarisation resistor) rather than a proper low voltage like on the CPC.

[Spirantho]

It surprised me too.
It was just certain keys, always the same ones. Might be a timing issue too.
Suffice to say I think in an Amstrad you'll be fine with either, but I have spares of both to be safe.

[Bryce]

I had the same problem with the AY when I fixed my Oric (which turned out to be a dodgy processor in the end anyway  ).

Definitely get and use a socket on the AY.

Bryce.

[Novabug]

So even using the Speccy plus 2 chip will be ok, even with the plus 2's different key layout?


Could anyone give me the dimensions to the right socket required. Dip 28, 36mm x 15.24mm? Socket pitch for example? 2.54

[gerald]

So even using the Speccy plus 2 chip will be ok, even with the plus 2's different key layout?

The layout is managed by the SW. The AY only provide the IO port which is used for reading the keyboard matrix.

Could anyone give me the dimensions to the right socket required. Dip 28, 36mm x 15.24mm? Socket pitch for example? 2.54

That's the right dimension.
Now I am wondering : do you feel confident unsolder old AY from the PCB ?

[Bryce]

He doesn't need to unsolder the working AY from the Spectrum, that's socketed already. He only needs to remove the broken one, so assuming he doesn't destroy any tracks on the PCB he should be fine.

Bryce.

[Novabug]

I have worked on PCB's in welding machinery, so I feel fine about desoldering the chip, using a solder pump.


This socket ok? 1 PC. 28 PIN (2x14) DIL / DIP IC Socket Bargain | eBay

[Bryce]

Yep, that's the socket you need.

Bryce.

[Novabug]

Right, the AY has been replaced with the one from the Speccy plus2, (AY-3-8912A) It seems the A variant does has an effect on the keyboard input. Most of the keys when typed display 2 characters at the same time, which makes any programming impossible. But, I can load games and the joystick is mapped correctly. Thoughts chaps?

[Bryce]

Yes. Who manufactured the AY you've used? Parts from GI and Microchip will work properly, other makes (which you seem to have) will cause this problem. You can solve the problem by getting yourself a GI or Microchip part or by adding pull-up resistors.

Bryce.

[Novabug]

It's a GI.

[Novabug]

Fitting and results -
Retro Special - AY chip replacement on Amstrad CPC464. - YouTube

[Neil79]

Oh Hai NovaBug, that bloke from Dizzy thingy  

[Bryce]

It's a GI.

Ok, then it's most likely that the 74LS145 has failed.

Bryce.

[Novabug]

Great.... gonna end up replacing the entire chipset soon. You positive the 'A' variant is not the cause?

[Bryce]

You could try adding pull-up resistors to see if that solves the problem first.

Bryce.

[Novabug]

You could try adding pull-up resistors to see if that solves the problem first.

Bryce.

I'm not certain how to do that, and since then, others have suggested it the SPECCY chip that;s causing it. This is becoming a minefield. HA

[asapreta]

You could try adding pull-up resistors to see if that solves the problem first.

Bryce.

Hi, sorry to revive this very old thread but I got a problem that may be solved with some the information on this page.
It seems the 464 board has the place for these pull-up resistors near the keyboard connector.


[attachimg=1]


But I dunno the value. I'm guessing they are about 1Kohm, connected to ports 7 to 14 on the AY8912(A) in this case. As I couldn't find a schematic with this information, I am here asking for help. Can anyone confirm this?


Thanks!

[gerald]

Hi, sorry to revive this very old thread but I got a problem that may be solved with some the information on this page.
It seems the 464 board has the place for these pull-up resistors near the keyboard connector.


[attachimg=1]


But I dunno the value. I'm guessing they are about 1Kohm, connected to ports 7 to 14 on the AY8912(A) in this case. As I couldn't find a schematic with this information, I am here asking for help. Can anyone confirm this?


Thanks!

It is only documented on the Amendment service manual, on the 6128 cost down schematic (page 21) as NR102 and it's a 2.2kx8 resistor network.