48k power supply inverter circuit

[gmc]

Any electronic gurus out there that can explain how the inverter circuit works.

Like a idiot I plugged in an expansion cartridge while the spectrum was still on. Plugged in skew so it blew TR4. I've replaced TR4/5 and powered on a TR4 blew again.

Then tried to isolate the -5V section and removed R55, replaced TR4 again (TR5 still tests fine) and it blew again. Agghh!
Already checked -5V to ground and there is no short. Checked all resistors, diodes, coil, no issues.

+12V is working, reading 11.9V
+5V is also stable.

So there something screwed with the oscillator section. Checked Tr4/Tr5 and nothing is oscillating.

If I remove R62 and R55 should that isolate the inverter section so I can test?

Would also like to connect up my current limiting power supply to stop TR4 from blowing but it seems the 5V components draw over 300mA so how would I limit this?

[1024MAK]

I have explained it somewhere, but can’t for the life of me find it again. I have also come across another person that explained it well, and yes, guess what, I can’t find that web page either!

Often either the shorting out of edge-connector signals kills one or more of the 4116 (or equivalent) ‘lower’ DRAM chips, or the damage to TR4 and/or TR5 (which results in the loss of the -5V rail) kills one or more of the 4116 (or equivalent) ‘lower’ DRAM chips...

A dead 4116 (or equivalent) DRAM chip then overloads the +12V and -5V supply rails, killing replacement transistors in the TR4 position, which in turn may kill replacement transistors in the TR5 position. Which in turn, if not checked/tested and if needed, replaced, will then kill any new replacement transistors in the TR4 position... Welcome to hard core fault finding

Worse still, sometimes the ULA, Z80 or ROM chip(s) also get damaged

Which issue board is this?

Mark

[gmc]

Thanks for the reply. Looks like I've really stuffed things up.

I'm wondering if there is any way to connect up a current limiting power supply to stop TR4 from blowing (But all the chips seem to draw a lot of power so you can't really limit it unless you can easily isolate the -12/12/-5/5 supplies)

Or do I just blindly replace all the 4116's. But as you said it could be any of the chips

It's issue 3.

BTW: Just found a great writeup on the DC-DC circuit.

https://www.tablix.org/~avian/blog/arch ... converter/

[1024MAK]

Spectrum DC to DC converter

How and why did the ZX Spectrum use so many voltages

In terms of where to go next, have a read of this post.

Also you should consider updating the DC/DC converter to the latest modification to improve reliability, see here.

Mark

[gmc]

Thanks, great links.

I will let you know when its fixed and what chip w blown

[gmc]

Ok, so after many sleepless nights and trying to troubleshoot this I found the issue.

First I isolated the 5V, -5V and 12V lines so only the DC-DC circuit was being used. Hooked up a current limiting power supply and still too much current was being drawn. Set it to limit at 100mA.

Replaced almost all the components in the DC-DC circuit, problem persisted.

I then removed all the components and rebuild the circuit on a breadboard. Problem still there. Agghhh.

Only then did I see there were some wires on the coil that were melted. Not completely melted but a darker shade of purple. Strange thing is the resistance of the primary and secondary coils was fine. Also no shorts between primary and secondary. Probably melted enough to mess with the oscillation.

The new coils online at way overpriced so I ordered some 28 AWG wire and rewound the coil.

For anyone interested:

0.32mm wire. 28 AWG
Secondary: 74cm, 39 turns.
Primary: 21cm, 15 turns.

After rewinding the coil the DC-DC circuit is working like a dream. -5/12V all stable. I put the components back on the PCB with the DC-DC mod. Everything working 100%. Luckily no blown ram chips.

[Firefox]

a darker shade of purple.

Very well done! That's some serious restoration, that is.

[1024MAK]

If a fault causes TR4 to go short circuit, if the Spectrum is left on, the coil will overheat. This may cause the thin varnish insulation to fail, causing short circuits. It’s not easy to detect with a meter on the resistance range, as the resistance of the coil is low anyway...

Glad you got it sorted

Mark