Issue 6 VM12 & resistors confusion

[Pr0f]

I use to have all sorts of issues desoldering - until I hit on a method and a tool that seems to work for me.

1) Old solder - often the solder joints on the board are not keen to 'free up and flow', but a little extra new solder can help no end to promote good repair work - and that's before you remove any!

2) A good desoldering tool - I now use a pistol grip iron with a vacuum pump - was about £85 on Amazon - but seems to have gone up a little, but it's quite well made and works a treat on clearing out the old joints - removal of IC's with all legs intact - removal of a 40 pin or larger socket or connector - possible - it makes replacement of chips for socket and chip an easy chore, and as it's temperature controlled - you can keep the heat regulated to avoid lifting those tiny tracks.

https://www.amazon.co.uk/DESOLDERING-ST ... 510&sr=8-7

[stephen_usher]

That's why I invested in a Hakko FR-301. It's expensive but not as expensive as ruining old machines in the long term. (It's also compact and fits in a carry case so it's easy to put up in the loft after use.)

[tofro]

Yep. Using lots of (good) flux and new (lead) solder really is the key to success. Too much doesn't hurt - You'll need to clean up afterwards anyhow.

And a de-soldering iron and patience help a lot. Don't force it.

Don't expect a solder joint that's well over thirty years old to come apart easily with no flux added.

Tobias

[Ruptor]

Ruptor has shot himself due to the stress of trying to fix a QL. Well I could have but haven't yet. Guess what, changing the LS245 didn't fix the problem but I think you lot knew it wouldn't. I have lifted both the 8301 and 8302 pins so it isn't them causing the problem and that leaves the two RAM chips. I stuck a 20K pull up and then pull down resistor on the data line to see what happens. Pulling up made little difference but pulling down caused the memory test to put FFFFFFFF out and read back FEFEFEFE every time which I thought was strange. Why is the output word fixed if it is supposed to be taking a value from ROM? I only had one LS245 lying around that I checked in my programmer so I know it is good and I don't have RAM or turned pin sockets either I used socket strip for the 20 way. If it is the lower bit in the lower byte is that IC1?
You guys must be rich if you can spend £260 on a station and stick it in the attic. Even the £100 is a bit much given that I don't intend to make a job out of fixing old boards. Fixing my QL is a one off if I can do it.

[martyn_hill]

Hi Ruptor

That's disappointing, but as you say, only one of the potential causes of what you have observed.

Pulling up made little difference but pulling down caused the memory test to put FFFFFFFF out and read back FEFEFEFE every time which I thought was strange. Why is the output word fixed if it is supposed to be taking a value from ROM?

My earlier post, since corrected by Marcel - was misleading here; we have yet to get to the stage of testing with the 'random tweed' pattern (that comes from the ROM image) as the very first test of writing all '1's /0xFF to each location is failing to be read-back correctly (hence, 0xFE.)

Incidentally, when DB0 was pulled HIGH, just how much further through the RAM test did it get before failing? What value was written that then failed to be read back in this case?

As for the cost and time of fixing the board - it's a perfectly valid point. Some of us (well, maybe just me) clearly have too much time on our hands that we spend on the relative luxory of just 'fixing stuff' for the hell of it...

[Ruptor]

Hi Martyn

we have yet to get to the stage of testing with the 'random tweed' pattern (that comes from the ROM image) as the very first test of writing all '1's /0xFF to each location is failing to be read-back correctly (hence, 0xFE.)Incidentally, when DB0 was pulled HIGH, just how much further through the RAM test did it get before failing? What value was written that then failed to be read back in this case?

It fails straight away at $20000 with ones and with the random pattern as well but sometimes manages $20004, $20008 or $2000C more often with the resistor high.

As for the cost and time of fixing the board - it's a perfectly valid point. Some of us (well, maybe just me) clearly have too much time on our hands that we spend on the relative luxory of just 'fixing stuff' for the hell of it...

I have time while caring for mum and having a break from my main work on Electrotherapy devices that can cure all diseases but nobody is interested in even in the current climate. I never throw away things I like, treasure or think useful I believe could be fixed if I had the time like my MHC6600 Sony system and QL, both half fixed so far. I am reluctant to throw things out because every time without fail I have a use for what has gone. For instance I remember having some empty double Euro sized boards that were unstuffed dynamic RAM cards with circuitry. They had four banks of 16 sockets on each half of the board that must have been 1 Mb cards that could be useful now but do I still have them? If I had a proper solder sucker I could get the sockets out at the very least and a reason to get one. Thinking.

[Ruptor]

I gone & dunit. I saw it on EBay for £50 and the bloke said it had only been turned on for 4 hours. I like a bargain and pressed the buy button convincing myself I could take the RAM chips out and flog it again given that new ones are £100 but I won't. It looks like being turned on is all that has been done with it because the only sign of use is the discoloured tip. Why would someone buy a solder sucking machine for 4 hours use? The sponge is brand new and there is a tiny fleck of solder on the spring as if it has been used on one hole. Perhaps he ripped off the tracks and gave up.
While waiting for the sucker to arrive I searched the attic high and low looking to see if I had any 4164 chips thinking I might have a tube because I had those RAM cards but I must have chucked the chips with the cards when I moved house if I had any. Look what I did find instead. A bunch of 68010 a 68000 and some virtual memory management chips I think but what use are they? Ideas on a postcard. Gone to figure out how to use the sucker and rip the RAM tracks off my QL.

[Pr0f]

oooh - purple ceramic goodness

[Ruptor]

Looking for 4164 chips I don't want to pay pounds locally for old tech although I can buy them from China for 70p and wait a month. Now I have the sucker wouldn't it be better to remove all the RAM and replace it with an SRAM like the kit offered for the Spectrum internal lower 16K?

[Pr0f]

I looked at doing that - you could put SRAM in one of the ROM sockets and wire the missing address lines to give you static ram in the address map - but you need to make RAM available for the ZX8301 which is producing the video display - and that's engineering to work with Dynamic RAM - fast page mode capable Dynamic RAM. You could use 4 bit wide DRAMS of larger capacity - but that would involve a lot of rework!

If you get the DRAMS out cleanily - there's a good chance most of them would be ok - I have a few spares if you only need a couple.