Spectrum 48k Boot pattern

A place to discuss the ZX80's and ZX81's younger brother - the Sinclair ZX Spectrum
User avatar
Erm
Posts: 49
Joined: Thu Jun 20, 2019 7:07 pm

Re: Spectrum 48k Boot pattern

Post by Erm » Wed Aug 21, 2019 2:54 pm

On the 4116 chips the +5v and +12v was fine, but the -5v showed a small (0.4+) voltage on all of them.

My original Sinclair transformer puts out 14v (without load) so I'll also try a modern one that is actually 9v. Is it therefore likely that the TR4 and TR5 are dead? I just have a multimeter.
I'm Chris, a Brit in Germany (Chr$ on the QL forum).

User avatar
1024MAK
Posts: 2531
Joined: Mon Sep 26, 2011 10:56 am
Location: Looking forward to summer in Somerset, UK...

Re: Spectrum 48k Boot pattern

Post by 1024MAK » Wed Aug 21, 2019 5:55 pm

Erm wrote:
Wed Aug 21, 2019 2:54 pm
On the 4116 chips the +5v and +12v was fine, but the -5v showed a small (0.4+) voltage on all of them.

My original Sinclair transformer puts out 14v (without load) so I'll also try a modern one that is actually 9v. Is it therefore likely that the TR4 and TR5 are dead? I just have a multimeter.
If the issue 3B board has the updated DC/DC circuitry, then there can be no feed through of the nominal +9V, as Sinclair added an in line capacitor. In this case, using the Sinclair ZX PSU should not be a problem. Note that the Sinclair ZX PSUs are unregulated and typically do output between 13V and 16V with no load (not connected to a Spectrum).

Can you please post up a photo of the right hand side of your board, that is the side where the speaker is. Then we can see if the DC/DC converter circuitry is the older design, or the later design.

Mark

User avatar
Erm
Posts: 49
Joined: Thu Jun 20, 2019 7:07 pm

Re: Spectrum 48k Boot pattern

Post by Erm » Wed Aug 21, 2019 7:57 pm

Certainly, here it is.
Spectrum48kPCBRight.jpg
I'm Chris, a Brit in Germany (Chr$ on the QL forum).

User avatar
1024MAK
Posts: 2531
Joined: Mon Sep 26, 2011 10:56 am
Location: Looking forward to summer in Somerset, UK...

Re: Spectrum 48k Boot pattern

Post by 1024MAK » Wed Aug 21, 2019 11:31 pm

Thanks for the photo.

Your issue 3B board does not have the updated/later DC/DC converter circuitry.
So it is possible that the unregulated +9V may be feeding through the coil to the +12V rail.
If the voltage on the +9V input is slightly higher than the voltage on the +12V rail, then that is likely the reason.

Meanwhile can you test the voltage across diode D11 (1N4148).
Also test the voltage across capacitor C74.

There is a board layout here

Mark

User avatar
Erm
Posts: 49
Joined: Thu Jun 20, 2019 7:07 pm

Re: Spectrum 48k Boot pattern

Post by Erm » Thu Aug 22, 2019 9:26 pm

Thanks so much for your time and expertise here, it's not the first time you've helped me with words of wisdom.

I think I've done this right:

On the input side of the regulator it measures 11.3v.
The 12v leg on the RAMs are about 12.3v.
Voltage measured on the anode side of D11 is -1.6v.
Voltage measured on the + side of C74 is 5.07v.

I also removed a leg from D11 and did a diode test with my multimeter. Black on cathode showed a voltage loss of 0.6v. Round the other way was a loss of about 1.6! Same story for D12, although sometimes they tested ok.
I'm Chris, a Brit in Germany (Chr$ on the QL forum).

User avatar
1024MAK
Posts: 2531
Joined: Mon Sep 26, 2011 10:56 am
Location: Looking forward to summer in Somerset, UK...

Re: Spectrum 48k Boot pattern

Post by 1024MAK » Fri Aug 23, 2019 11:06 am

Erm wrote:
Thu Aug 22, 2019 9:26 pm
Thanks so much for your time and expertise here, it's not the first time you've helped me with words of wisdom.

I think I've done this right:

On the input side of the regulator it measures 11.3v.
The 12v leg on the RAMs are about 12.3v.
Voltage measured on the anode side of D11 is -1.6v.
Voltage measured on the + side of C74 is 5.07v.

I also removed a leg from D11 and did a diode test with my multimeter. Black on cathode showed a voltage loss of 0.6v. Round the other way was a loss of about 1.6! Same story for D12, although sometimes they tested ok.
Thanks for those results.

So it looks like the oscillator part of the DC/DC converter is working okay, as the +12V rail is at a higher voltage than the +9V input from the PSU, so that’s good news.
1024MAK wrote:
Wed Aug 21, 2019 11:31 pm
Meanwhile can you test the voltage across diode D11 (1N4148).
Also test the voltage across capacitor C74.
By across I meant connect one meter lead/probe to one leg and the other meter lead/probe to the other leg...! But no matter :lol: There is no need to redo these by the way.

Your result for D11 shows that either C46 (1uF) is faulty or there is a problem downstream of C46. Is the resistance of R55 (56Ω) okay? If yes, then I suggest renewing C46, then if that does not help let me know.

This should be similar to the circuit on your board:
ZX Spectrum Issue 3 DC/DC converter circuit schematic
ZX Spectrum Issue 3 DC/DC converter circuit schematic
Mark

User avatar
Erm
Posts: 49
Joined: Thu Jun 20, 2019 7:07 pm

Re: Spectrum 48k Boot pattern

Post by Erm » Fri Aug 23, 2019 3:07 pm

Right, well a little progress, although the screen output is unchanged.

I changed C46, R55 is spot on.

I changed C74 as a lead fell off. I also changed C61 and D11 (I currently don't have new diodes but stole one from a parts ZX81 - one of the keyboard connector diodes which have the same value written on them). I suppose I could do the same for D12. I also disconnected R54 and checked that (was fine).

Anyway, I now have a negative voltage! It's -4.3v on every lower ram chip and they stay fairly cool instead of 2 or 3 getting boiling hot as they were before. That's not enough negative volts is it! I'm just waiting for a better desoldering solution. I suppose I should now wait for that, remove the lower ram, put sockets in and then check the voltages at the board with chips removed?
I'm Chris, a Brit in Germany (Chr$ on the QL forum).

User avatar
1024MAK
Posts: 2531
Joined: Mon Sep 26, 2011 10:56 am
Location: Looking forward to summer in Somerset, UK...

Re: Spectrum 48k Boot pattern

Post by 1024MAK » Fri Aug 23, 2019 4:17 pm

-4.3V is correct for that version of the DC/DC converter circuitry. If you want to get it closer to -5V, you have to perform the recommended Sinclair upgrade. I’ll post full details later. But here is a list of parts required:

DC-DC CONVERTER MODIFICATION

The following new materials are required:

Diodes:
• D15 - BA157 (it may be possible to reuse the existing D15)
• D17 - BA157
• D18 - 1N4148
• D19 - BZY88C 5V1 500mW 5.1V Zener

Capacitors:
• C46 - 1uF 50V WKG 105°C Electrolytic or ceramic axial (only required on issue 1 and issue 2 boards)
• C47 - 1uF 16V WKG Electrolytic axial or ceramic axial
• C49 - 560pF ceramic axial (if not already this value)
• C74 - 4.7uF 16V WKG Electrolytic radial (if not already fitted)
• C78/C80 - 22uF 25V WKG Electrolytic axial
• C79 - 1uF 16V WKG Electrolytic axial or ceramic axial

Resistors:
• R60 - 220Ω 5%, 0.25W
• R79 - 2.2kΩ (2k2) 5%, 0.25W

Note that if you cannot get electrolytic capacitors at the stated voltage rating, it’s okay to use a capacitor with a higher voltage rating.

Mark

User avatar
Erm
Posts: 49
Joined: Thu Jun 20, 2019 7:07 pm

Re: Spectrum 48k Boot pattern

Post by Erm » Fri Aug 23, 2019 4:53 pm

Well that's even more encouraging then. I may well do the DC-DC mod, but first I'll wait for the new RAM and see how/if it works with that. Thanks for now.
I'm Chris, a Brit in Germany (Chr$ on the QL forum).

User avatar
1024MAK
Posts: 2531
Joined: Mon Sep 26, 2011 10:56 am
Location: Looking forward to summer in Somerset, UK...

Re: Spectrum 48k Boot pattern

Post by 1024MAK » Fri Aug 23, 2019 5:12 pm

I've put the full details of the DC/DC converter modification on my blog here ;)

Mark

Post Reply