An upgrade too far?

[Trevor_B]

I've just "broken" my much modified ZX81..
It's a series 1 with 16/32k RAM internal (see later in thread), Composite output, VLA81, ZX81x2 ROM - and a modern 7805 just because I had one..

Todays upgrades were the ZX81x2 ROM (i have both options - a programmed RWAP direct replacement and a 128 that I had programmed). Today's was the simple swap out of the direct replacement into the original slot with no modification.
I swapped the ROM, powered up- all was fine. I ran a few simple programs to see the improved speed and yep - it was working.

While the machine was apart I decided to rewire the previously discussed 16k internal to take advantage of the VLA81's ability to see 32k without needing additional circuitry ("A13 to Pin35" from previous threads).

Again - I soldered the wire, powered up and all was fine; a quick program to test the additional memory showed that the range c000-ffff wasn't being seen (I got an error 1) so I powered off, re-read the VLA81 instructions and switched to Mode 1 (Mode 1 provides an upper 32K memory map and support for HIRES programs) by switching switch 3 on the VLA81.

Powered up again - and I got a nice solid white square and no cursor...
So - I switched VLA81 back to what it was before, and got a nice solid white square and no cursor.

A recheck of wires I'd soldered proved they were all still nicely connected, so I'm back to where should I look next. As much as I could reinstate the original ROM, I don't think this could be the cause. I thought about re-rewiring the A13 chip RAM chip back to where it was before, but ran out of time. This would mean that I'm still limited to 16k internal and no hires - which was one of the things I was looking to achieve.

I'll point out that I was running this through a composite -> HDMI convertor, if that has any bearing.

Any pointers where I should be looking first?

Trev

[Moggy]

Whilst not being the smartest here and have to rely on others for all things technical I have found from bitter experience that whilst these new ULA chips are a blessing they are far more fragile electronically speaking than their Ferranti cousins and once fitted should be left alone without too much over handling.

I have blown three of these in situations the older ULAs would (by my own cack handed experience) have survived.

Put an EPROM chip back in its socket with one of the address pins accidentally left out ? blown ULA.

Wobble a memory module(Memotech especially) ever so slightly? blown ULA and so on.

Having said all this I should add that my zeddies are probably powered up and in use and abuse more than most around here.

A quick check I've found is to power up and feel if the ULAs chip gets hot very quickly, if so it's most certainly blown and yes all the usual static discharge caveats apply/have applied before any one mentions it.

May not be the cause but worth mentioning I think.

EDIT.

Also does anyone know if these new ULAs are buffered on their video output Pins? I only ask as they seem to connect to the outside world directly, with no transistor or what ever in between and I thought that was a bit of a no-no with ULAs?

[Trevor_B]

tbh I was thinking it could be ULA related, but hoped it might be a bit more robust, especially as I only flipped one of the built in switches.

I've got another on order - hopefully this one won't take 8 weeks to arrive like the last one.

I still have the original ULA which I could refit in the meantime to see if that's the cause. That would also mean rewiring the memory upgrade again to make it a 16k version again - and possibly back to square one.

[Moggy]

Does the chip get hot when powered up like I suggested? If it doesn't you may have got a lucky break in that respect and the fault lies else where.

[1024MAK]

Definitely check to see if the chip on the vLA is running hot.

Unfortunately a white screen without a curser only tells us that it has power and that the ULA or equivalent is producing the video signal level for err, white.

So this could be a RAM fault, a ROM fault, a Z80 problem or a ULA problem...

I don’t think the video converter is the cause.

If you have a logic probe or a ‘scope, testing the usual Z80 pins will tell us if it’s trying to do anything.

Mark

[Trevor_B]

It looks like it was the ROM;

I switched back to the old rom (one of those "rare" 622's) and it's working fine. That's all I changed so hopefully I'll now be able to test the c000-ffff memory range too.

Thanks for the pointers; I ran out of time earlier and probably would have started on the RAM connection first.

It does mean that I'll probably have a spare VLA81 for sale very shortly... one careful(ish) owner, only used for church on Sundays etc...

Trev

Update:
It works but only seems to be seeing 8k RAM, even after switching the VLA back to "normal" mode. I've it away for now - in short it's slightly worse than before I started but hopefully nothing I can't back out.

[Moggy]

The ULA ok ,excellent news, keep persevering Trevor it will be worth it in the end.

[1024MAK]

Next step, 4K of RAM

Yes, keep going...

Mark

[Trevor_B]

I'm now not 100% convinced I had 16k before I attempted the 32k upgrade - I'm going to switch back and test again. More soldering... aargh.

While the (soldering) iron is hot, is there something I need to do with LK1 / LK2? From the text and photos on the "reversible upgrade" page (https://web.archive.org/web/20200929054 ... rsion.html), it mentions making a link to LK2, but but further info (pic 4 - no link, pic 6 - LK1, pic 8 LK1 to LK2,pic 9 just LK1, pic 10, just LK1). It doesn't help that my board is a green board - like pic 2, with two points marked LK1 and LK2 - do I need to connect them?

Sorry about the raft of questions - I thought I had this in the bag but I'm now doubting what I had to start with..

Trev

[1024MAK]

LK1 / LK2 is irrelevant if pin 21 is bent out and wired elsewhere.

The bent out pins should be 1, 21, 23 and 26. And all these should be wired to address lines.

Did you need to fit a socket, or was a socket already fitted?

Mark