exxos's Atari, Amiga & retro forum

rubber_jonnie wrote: 21 Sep 2023 16:25 Also interesting, but neither the original + board nor the spares board work with the DivMMC cart, so the plot thickens, even though the original board will load from tape.

And yes, the edge connectors have been cleaned, but I need to have another look and see if I can clean some more.

My Spectrum 48k, the original subject of this thread works just fine.

RE the Divmmc, apparently, this is often the result of a weak M1 from the CPU

Edit:

http://blog.tynemouthsoftware.co.uk/201 ... gnals.html

rubber_jonnie wrote: 21 Sep 2023 21:40 Ok, so despite the 30V on the 12v line, the -5v line works as advertised...

Looks like a 5V zener on it so would explain why.

Is that "9 to 11V" actually about right ?

I'm wondering if its a capacitor charge pump circuit now. If that was the case, you could only double the voltage. Its hard to tell because the "dots" for joining points on the schematic seem to be missing half the time and some look to be not joined. Its very misleading when stuff isn't drawn properly to start with :roll:

The -5V looks like a capacitor charge pump.

RED is the 5V flow when TR4 is OFF. C46 gets charged to 9V minus about 1V drop across D11.

When TR4 is ON, its shorting the left side of the cap (charged now to about 8V) to 0V which means the other side of the cap is now -8V.


Assuming C78 is connected to the top of TR4. That could be a charge pump as well. But I am slightly confused about what's connected to what to figure it out. There doesn't seem to be any current path for C78 to charge it. So Id assume its using the pulses from the coil to get the step up voltage. But you wouldn't really put a cap in series like that :roll:

EDIT:

I suppose what happens is when TR4 is ON, C78 is connected to 0V. The coil charges. TR4 turns off. Then its got the voltage from the coil AND the 9V source giving C78 a higher voltage which is pushed onto the 12V rail via D15. I assume C49 is also taking a bit of punishment there as well. There doesn't really seem to be any proper regulation on it.

I assume 12V at the top is the feedback to control the voltage (regulation) via TR5 as it controls the coil primary. C43 is probably there to slow down the feedback loop to stop it running-a-muck.

chronicthehedgehog wrote: 21 Sep 2023 21:44

rubber_jonnie wrote: 21 Sep 2023 16:25 Also interesting, but neither the original + board nor the spares board work with the DivMMC cart, so the plot thickens, even though the original board will load from tape.

And yes, the edge connectors have been cleaned, but I need to have another look and see if I can clean some more.

My Spectrum 48k, the original subject of this thread works just fine.

RE the Divmmc, apparently, this is often the result of a weak M1 from the CPU

Edit:

http://blog.tynemouthsoftware.co.uk/201 ... gnals.html

Cool, thanks :)

exxos wrote: 21 Sep 2023 22:09 The -5V looks like a capacitor charge pump.

RED is the 5V flow when TR4 is OFF. C46 gets charged to 9V minus about 1V drop across D11.

When TR4 is ON, its shorting the left side of the cap (charged now to about 8V) to 0V which means the other side of the cap is now -8V.


1.jpg
2.jpg


Assuming C78 is connected to the top of TR4. That could be a charge pump as well. But I am slightly confused about what's connected to what to figure it out. There doesn't seem to be any current path for C78 to charge it. So Id assume its using the pulses from the coil to get the step up voltage. But you wouldn't really put a cap in series like that :roll:

EDIT:

I suppose what happens is when TR4 is ON, C78 is connected to 0V. The coil charges. TR4 turns off. Then its got the voltage from the coil AND the 9V source giving C78 a higher voltage which is pushed onto the 12V rail via D15. I assume C49 is also taking a bit of punishment there as well. There doesn't really seem to be any proper regulation on it.

I assume 12V at the top is the feedback to control the voltage (regulation) via TR5 as it controls the coil primary. C43 is probably there to slow down the feedback loop to stop it running-a-muck.

Gonna have another look in the morning, thanks.

Your issue seems quite weird...

HigashiJun wrote: 22 Sep 2023 07:48 Your issue seems quite weird...

It certainly does seem weird, that's for sure...

Ok, so going through the schematic for the PSU section again, I have tested all of the resistors and they are all the correct values.

D11 and D18 have both been replaced. They are standard 1N4148 diodes and D11 was the one that had originally gone pop.

In terms of the physical connections between components, I've buzzed them out and all seems to be fine too.

I have some replacements coming for D17 and D15, which are rectifier diodes, just on the off chance they are not playing ball, and in addition I'm going to do a recap as the circuit looks to be a charge pump type based on what somebody mentioned earlier and a little Googling.

Since it's highly reliant on the caps, most of which are electrolytic's, I think they may be bad and replacement is inexpensive, so why not?

Clearly the circuit is operating, since it is generating a high voltage compared to the input from the PSU, but clearly whatever the feedback mechanism is, it's falling short of stopping at 12v.

It certainly makes me wonder if the previous TR4 failure was related to caps, and that then killed some of the 4116's and also some of the other caps, leading to what I'm seeing now.

So, caps came today from Retroleum and I replaced all the electrolytics on the + original board. No joy, still 30v. I also swapped out D15 with the one in the spares board, but no change.

Measuring the voltage at the left hand side of D15, I get about 8.5v, but the right hand side is 30 on the + board and 12v on the spares board, but the only other connection is D19, the Schottky diode between -5v and +12v, so I'll see if swapping that helps at all.

I did also find this on the spares board:


It was difficult to see as it originally had solder on it, but I confirmed this when I removed it:


Yep, it had been cut. I honestly thought that would solve my problem of games failing to load, clearly with one pin cut the RAM IC wouldn't function.

It seemed to get further than before, but still fails to load games. I think that since there is still an issue, there is a further problem with the upper 32K, I could do with a diagnostic ROM to test and and tell me which IC is bad, certainly there isn't one that's hotter than the others using the thermal camera. I have a spare 32Kb IC so may try piggybacking to see if that helps.

If swapping D19 make no difference, then I'll remove TR4 and see if it all runs nicely with one of the boost converters that arrived today.

rubber_jonnie wrote: 23 Sep 2023 15:21 Measuring the voltage at the left hand side of D15, I get about 8.5v, but the right hand side is 30....

So the 30v must be coming from elsewhere. 8.5v left side means no current flow if the right side is 30v. 30v can't feedback as the diode blocks it. So sounds like the PSU is mostly working then.