ZX Spectrum Issue 3 repair

[1024MAK]

I don't know if the ZX81 heatsink will be suitable. It may be okay. The L200 has protection against overheating, so the only way to know would be to try it.

I have not built either of the two circuits as shown. But have built PSU designs using both the L200 and the LM317T before. My previous designs used a built in mains transformer, bridge rectifier and various other parts.

As long as my calculations for the resistor values are correct, they should work okay.

Edit: a quick calculation shows that the ZX81 heatsink may be too small if using with either a 12V DC input or when running at the maximum current, or when the output voltage is lower than 4V. However, if you use TWO ZX81 heatsinks arranged to give maximum flat area, this may be okay. Use a thin smear of heatsink compound to aid the heat transfer between regulater chip and heatsink, and between heatsinks.

Edit2: here are pictures of two of the PSUs built by me:

IMG_4399.JPG

L200 based PSU front

(366.63 KiB) Downloaded 280 times

IMG_4400.JPG

L200 based PSU rear

(464.71 KiB) Downloaded 280 times

IMG_4401.JPG

LM317T based PSU front

(287.59 KiB) Downloaded 280 times

IMG_4402.JPG

LM317T based PSU side

(283.75 KiB) Downloaded 280 times

IMG_4403.JPG

LM317T based PSU bottom

(790.28 KiB) Downloaded 280 times

Mark

[1024MAK]

And here are pictures of my smallest ever general PSU (this one does not have a mains transformer!):

IMG_4408.JPG

LM317T small PSU top

(691.48 KiB) Downloaded 279 times

IMG_4409.JPG

LM317T small PSU connections on side

(471.46 KiB) Downloaded 279 times

IMG_4410.JPG

LM317T small PSU top/side view

(476.3 KiB) Downloaded 279 times

IMG_4412.JPG

LM317T small PSU heatsink on bottom

(485.07 KiB) Downloaded 279 times

Mark

[PokeMon]

I have something similar directly in my desktop including AC power.

20170423_184616k.jpg (115.75 KiB) Viewed 4091 times

20170423_184733k.jpg (139.78 KiB) Viewed 4091 times

[Tezz]

The L200 looks very professional Mark, that would come in handy right now. With the issue 3 board still troubling me, I thought I'd look at the 3B board this morning with it's lower RAM removed, I reinstalled the heatsink and ULA and did the initial tests which didn't indicate anything too concerning so I thought I'd risk trying the DC voltages with the power supply connected for the first time. Happily +5v +12 -5v 0v all tested correctly so I moved into the other room and hooked it up to the TV RF hoping to see the expected output with no lower RAM but it outputs a noisy black screen instead. I tried swapping with the issue 3 ULA but no change. As with the issue 3 board nothing else is socketed on this 3b board. Does the output by any chance indicate a typical fault symptom? If not I have some more learning to do with trial and error diagnostics. I have several lower RAM replacement modules now so I can hopefully eventually install one once the fault/s have been sorted The only info I've found online this morning is someone repairing a Spectrum +2 with a black screen output which turned out to be a faulty transistor ZTX650 which is what's commonly installed on the 48k Spectrum TR4 so perhaps it will just be the common fault with that transistor being taken out by the bad RAM on this board. I have some transistors due to arrive tomorrow so hopefully I can see some success with at least one of these repairs.

I just read on world of spectrum repair guide that the black screen output is common with an upper RAM fault. With 12v showing present, does that indicate TR4 is working? It suggests cutting legs on each RAM IC but I will probably remove them and install sockets to see if I can discover which are faulty.

One circuit I'd love to build would be a RAM tester for 4116 and 4532/4162. I suppose a microcontroller would be needed also to run a full RAM test like a pic or pi zero.

[1024MAK]

Funny you should mention a stand alone DRAM tester, guess what I'm working on...! (Wanted link).

If the -5V is present, then both TR4 and TR5 are okay.

However, check the +12V on pins 14, 15 and 16 of IC14 (LM1889 colour encoder) and for +5V on the modulators 5V power connection.

It is possible to disable the upper 32k bytes of RAM without removing it. This assumes that any fault with the upper RAM is not a bus interface fault. I'll see if I can find my post on WoSf...

Do you have a logic probe?

Mark

[1024MAK]

Given that so far I cannot find my post about disabling the upper RAM, I'll give you this one instead

Your black screen, what colour is the border? Is the border the same colour every time you try this (switch off for at least 20 seconds between tries)?

Mark

[Tezz]

Ah oops I mixed up -5v and 12v for the transistors purpose. I don't yet have a logic probe. I keep meaning to add that time my list so will make sure to order one tonight, luckily they are not too expensive as I've spent quite a lot so far with everything. I'll check Ic14 this evening as you suggest, thanks again.

[Tezz]

Given that so far I cannot find my post about disabling the upper RAM, I'll give you this one instead

Great, thanks for that, much quicker to see if the upper RAM is the problem

Your black screen, what colour is the border? Is the border the same colour every time you try this (switch off for at least 20 seconds between tries)?

Mark

From memory the whole screen was black but I'll retest a few times to see if there's any change.

Will I get an output of flashing blocks with no lower RAM and upper RAM?

[Tezz]

Ok, now tested IC14 pins 14 15 16 with the meter set to 20 on voltage test. Shows 11.16 on the meter. The modulator 5v connector shows 4.64. Hopefully within expected range. I will try disabling the upper RAM with the wire jumper on IC23 in the morning.

Edit: just tried the jumper on ic23, not soldered but held in place with blutack. Output was the same, all black noisy output. No border showing. Just totally black with noise.

[1024MAK]

Ah oops I mixed up -5v and 12v for the transistors purpose.

TR4, TR5 and the associated circuitry do produce both the -5V AND the +12V. I ask people to check the -5V, as with issue 3 and earlier boards (that have not had the official modification), the nominal +9V can pass through the transformer ("the coil") to the +12V rail, so if the switching circuit of TR4 and TR5 is not working, but the mains adaptor output voltage is around 11 to 12V, people can be mislead into thinking the +12V rail is okay...

Mark