Issue 4a docs needed

[eccles]

Hi folks,

I'm trying to recover an Issue 4a spectrum that I put aside years ago for when I wanted to start repairing/enjoying my childhood computer collection once again.

This particular board I got in the early days of ebay for spare parts, and judging by it's state it looked like someone had tried to modify it (and failed) - I later found reference to a power supply mod in the zx spectrum service manual along with warnings not to use it as it was incorrect.

So I thought a sensible approach would be to remove the mod and return the board to factory - expecting to need to replace blown out transistors and regulator and such.

But here's where the problems start - having found schematics for the Issue 4a, and consulting with circuit diagrams, photographs, I find the parts list and schematics don't match the board layout and photographs.

Could anyone with knowledge of such things speak to this - is there 'known good' set of diagrams and parts lists someone could link to for the Issue 4a?

[1024MAK]

Hello and welcome. You may want to introduce yourself in the Welcome area

Beware, the circuit / schematic diagrams for most of the different issue boards have various errors. For the issue 4a and 4b I have not been able to confirm what went wrong, but it maybe that the circuit labelled 4a is actually for a 4b and vice a versa... But I don’t have examples of both boards, so this is speculation.

Issue 4b are not supposed to need this modification, as they are supposed to have it included as standard. But I’m not sure how true this is.

For the DC/DC converter circuitry, you may want to visit my blog page here. I recommend you compare the actual circuit on your board to the diagram and the notes on this page. If possible print the diagram on paper so you can mark it as you compare it to the actual circuit on your board. Mark it as you go .

Depending on what design of DC/DC converter circuitry you have, there are good reasons to upgrade to the improved version of the circuit. Protection against the nominal 9V going above 11.9V and causing the DC/DC converter to no longer generate the -5V supply which can lead to failures of the 4116 DRAM chips. Better regulation of the -5V rail (as in closer to the specified -5V). The overall result is better reliability.

Note that some components are changed from one type to another. How many component positions this applies to depends on which issue board you have. The issue 4a may need less changes.

If you have any questions, please do ask. Photos of the relevant section of your board would be helpful .

Mark

[eccles]

Thanks for the resources - the diagram on your blog helped me track down the issue on a 4S board I had and got it working with the help of a new TR4 and a little resoldering - that spectrum now awaits a keyboard membrane before I can make a working unit to play with.

I've managed to trace out the components with the help of the diagram on your blog post. As I had initially returned it to what I thought was factory spec, it currently has the following differences:

R60 is currently 68ohms
C77 is currently a ceramic 100nF
C78/80 is missing
D17 is missing
It has R55, not D18
It has D16, not C79
It has D12, not R79
It has R54, not D19

I'll look to reinstate these changes and end up with something identifiable as the DC/DC converter circuit in the diagram, then i'll test again to see if have the missing 12/-5v rails

[eccles]

The modification instructions on your blog state, for two steps:
Replace R54 with D19 (anode to ground) 5.1V Zener
Replace R55 with D18 (anode to C46) 1N4148

But the diagram you provide for modifications, it suggests D19 should have the cathode to ground and D18 the cathode towards C46

Could you confirm for me before I make a mistake?

[1024MAK]

The modification instructions on your blog state, for two steps:
Replace R54 with D19 (anode to ground) 5.1V Zener
Replace R55 with D18 (anode to C46) 1N4148

But the diagram you provide for modifications, it suggests D19 should have the cathode to ground and D18 the cathode towards C46

Could you confirm for me before I make a mistake?

Ahh! I glad that someone is awake and paying attention! Yes, some errors have crept in
Well spotted
The schematic diagram is correct.
Ignore “Replace R54 with D19 (anode to ground) 5.1V Zener”, it should say “Replace R54 with D19 (cathode to ground) 5.1V Zener.
Also ignore “Replace R55 with D18 (anode to C46) 1N4148”, it should say “Replace R55 with D18 (cathode to C46) 1N4148”.

I wrote the text separately off line to doing the circuit diagram and obviously dropped a clanger!

I’ll correct the blog entry tomorrow...

Mark

[eccles]

So where i'm at:

All modifications to DC-DC converter are done
All electrolytics have been recapped
Have done some checking with a multimeter and are happy the DC circuit now looks as per the diagam
Have tested resistors in-circuit
Have replaced TR4/TR5 and tested them again outside of circuit

Only two items in the DC circuit not confirmed (and not failed open circuit) are the ceramic caps C49 and C43

Testing one of the 4116's I find I have 5.6V at the +5v line (OK), 5.8V at the +12V (not ok), and 0.6V at the -5V (not ok)

Additionally, using my cheap rate-limiting DC bench supply, it wants to start off consuming about 1A, and then creeps up slowly from there - I only powered it on long enough to get these meter readings. My booting Issue 4S never consumes more than 700mA, so I know this is a clue.

Any suggestions where to go next? I'd considered removing all the 4116's (IC6-13) as these appear to be the only consumers of +12V and -5V in the board - at least I could rule them out as dragging those rails down, and could fit sockets.

Thoughts?

[1024MAK]

Any suggestions where to go next? I'd considered removing all the 4116's (IC6-13) as these appear to be the only consumers of +12V and -5V in the board - at least I could rule them out as dragging those rails down, and could fit sockets.

The +5V supply (pin 9 on a 4116 chip) comes from the 7805 voltage regulator. 5.6V is too high. It should be between 4.75V and 5.25V (although most 7805 chips manage an output much closer to 5V than this). You can test the output voltage between the right hand pin/leg and the heatsink tab. Be careful not to short between any of the legs/pins (the middle leg is also 0V/GND).

The current consumption for a 48K ZX Spectrum with no expansions fitted should never be more than 800mA.

So I recommend the following:

• Renew the 7805 voltage regulator.
• Isolate the +12V supply to the 4116 DRAM chips by carefully cutting the appropriate PCB track, OR by cutting pin 8 on all the 4116 DRAM chips (in such a way that it can be joined again with a bit of solder), OR remove all of them.
• Don’t fit any sockets yet. Leave the positions free for now.
• Using your DC bench supply, power up again and see what the current consumption is. And test the +5V, +12V, and -5V supply voltages again.
• Post up your results.

Mark

[eccles]

Latest update.

4116's now all removed

Voltage at the regulator, 4.96V

Voltages at 4116:
+5v - 4.96v
+12v - 10.6v
-5v - -4.74v

Familiar high pitch whine now, and max consumption of ~600mA, so progress is made

I'll have to order up some 16 pin sockets - presumably I can put 1 in at a time until I find the duff one.

Any further suggestions?
Thanks for the handholding!

[1024MAK]

The +12V is a little lower than I would have expected.

What voltage is the (9V nominal) DC input at?

What voltage do you have at TR5 base?

Before soldering in new sockets, do very carefully check (visual inspection and electrical continuity test) the tracks. As once the sockets are soldered in, you don’t want to find a problem later that is under a socket...

Mark

[eccles]

Hope you had a great new year.

Back with updates on these two (4a, and 4s) repairs.

The Voltage at the base of TR5 is 4.26V
I've checked connectivity across the address lines and voltage/grounds of the 4116 sockets and they all seem ok

So I added a socket and tested the 4116 one at a time to hunt for chips that are pulling voltages down - a full 4 of them are pulling -5V and 12V down - instantly obvious from the current draw being reported by my desk supply.

So I think i'll order a new bank of 8 4116's for this board before going further. With no 4116's in at all, the 12V line is a tad low (as you pointed out) at around 11V - if you can conceive of some checks as to why this might be, let me know

The 4s i'm looking at is a different article - i'd replaced the TR4 and TR5 in it, and it had been booting to the Sinclair Research message intermittently, but I hadn't at that point got a working keyboard membrane, so put it aside.

Armed with a new membrane, I booted her up - keyboard worked OK, but the unit wouldn't respond to incoming audio signal on the earphone port (so I couldn't load any software)

Took it apart and found the 9v regulator had a broken leg. Replaced it, now it won't boot at all getting a flickering black/white screen, or a pure black screen.

This 4s board had X1 fallen off the board, and the RF modulator unsoldered when I got it, so i'm going to switch it to composite in order to rule out the modulator, but if you've got any onward suggestions for this board, that'd be useful - +5v, +12V and -5V all check out fine on this one.

Edit: Came up with the radical idea of looking at other posts on this forum (yes, I know, i'm embarassed by this lack of forsight) and will do some of the testing shown elsewhere and will report back