Page 1 of 4

Another repair case, plase help if you can

Posted: 14 May 2014 22:25
by laios
Hi to everyone
I recentely I bought one of theese "untested" from ebay which apear from time to time.
It was more than obvious but I said to give a try, perhaps it was only a regulator if I was lucky,
well I wasn't , a bad psu at first , I made a replacement following the very good schematic with the 2 transformers.
Now I had power , and I got black screen continouus tone an a lot of snow, the one big cap 4700uf was almost loose
I changed both the 4700 caps with new ones , now I had black screen without snow, I changed both regulators and TIP2955
but was stupid enough to forget one of the insulations so I burned the TIP.I replaced all again for sure.Measurings both voltage and passive resitstance were good as stated somewhere early in this thread.I noticed a OKI chip of the main ram of 4164 was hot,
I carefully unsoldered all of them ,put sockets and replaced them all.I also unsoldered all of the video ram 4116 put sockets and replace them too.Now I have a nice black screen no continous sound only RF output.Somehow I think the TMS video processor is warm but I think
this is a common issue with this processor.I look for some replacement.Just want to mention this has 3 roms one has a label say "SPAIN"
, I bought it from Spain, and has some wires from this to some point of motherboard.Also it looks that some repair was done as I someone
replaced the 74S04 just up the XTAL1. I anyone with more experience has any suggestion to guide me will much much helpfull

I try to post some photos in next days

thank you very much

Re: Another repair case, plase help if you can

Posted: 15 May 2014 18:55
by 1024MAK
Hi :D

Well you have already done most of the normal repair steps.

If you have another Z80 CPU, try changing it.

The PCB is designed for one type of ROM chip. If there are wires (normally yellow in colour) wired on the bottom track side, this is likely to be to enable different ROM chips to be used.

Apart from a multimeter, what other test equipment do you have.
A 'scope (best) or a logic probe (with pulse detector) are often very helpful with tracing some types of fault.

Mark

Re: Another repair case, plase help if you can

Posted: 15 May 2014 20:33
by Martin A
Usually Dave pops up with a pointer to his site, so I thought I'd nip in first

http://primrosebank.net/computers/mtx/d ... Manual.pdf

Black screen and tone is a sort of catch all for anything that stops the MTX booting. As the VDP initialisation code and the Kill sound routine haven't been called. And they're both early in the start-up sequence (after the ram test and auto-start rom test)

As Mark says it's worth checking a known good Z80 if you have one. It doesn't have to be the 80's vintage chip. A modern CMOS version should work just as well.

If you have no picture, but also no tone, it could be the MTX has got to ready, but the video in missing. You could try Ctrl G to check whether the bell sounds (as on page 28 of the service manual)

Re: Another repair case, plase help if you can

Posted: 16 May 2014 20:16
by laios
Thanks for your replies
I have already had a look at the site you advised me and also the service manual ,
it is realy a great help
but my electronics knowledge is little primitive even coming to an 80's machine

I socketed the keyborad and tried Ctrl-G ,nothing happens
So I continue with Z80
I do have a z80a new but that's for my Harlequin zx clone , I'd prefer to use another one

is this ok ?
http://www.ebay.com/itm/Toshiba-TMPZ84C ... 338aec5d01

at the moment got no camera ,as soon as I solve this I'll post some pics

Re: Another repair case, plase help if you can

Posted: 16 May 2014 20:55
by Martin A
That looks OK, I found Toshiba listed as a 2nd source. I'd guess the -6 on the one in the photo means that one's a 6mhz part.

I've had a modern 10mhz CMOS Z80 in one of my MTXs for a while, so being both CMOS and capable of faster speeds shouldn't matter.

Mouser is listing new Z80's at 3.31 euro's+ VAT (making them around a euro cheaper?), but the postage is a lot more unless you're ordering other stuff to get free delivery

Re: Another repair case, plase help if you can

Posted: 17 May 2014 00:53
by AndyKey
More likely duff DRAM than duff Z80.
Changing and/or socketing the chips probably didn't help, and may have made it worse.
I know of 3 attempts to change DRAM that didn't work afterwards, even though visual and electrical inspection showed nothing wrong.

I'd try removing them and using external RAM card.
Or replacing ROM with Testcard ROM that produces video and sound output without needing any DRAM to do it.
I am lucky in that I have such tools on my shelf, but I am out of the country as of tomorrow and am therefore unable to loan them for 3 weeks.

Re: Another repair case, plase help if you can

Posted: 17 May 2014 22:36
by laios
Mr. Andy you realy give me hope , many thanks.
Although to replace something ,pending everything has been done corectly and make it worst , Is something I realy can't understand
the fact that it realy happened 3 times confuses me quite a bit, anyway at the moment I don't have any other option.
I also have ordered the z80 so I will change it anyway.
Thanks again very much for your help

Re: Another repair case, plase help if you can

Posted: 18 May 2014 00:16
by 1024MAK
The reason for the MTXs failing to work after the existing DRAM chips have been removed, sockets fitted and different DRAM chips fitted is not known. I would not worry about that as most of these were attempted upgrades (32k to 64k). Instead follow a logical process and keep notes about what you find (or don't find).

One possible problem is track or pad damage (including loss of plated through connections). Also there may be almost invisible solder splashes that maybe causing short circuits. So remove the DRAM chips from their sockets and go through the pain of resistance checking. Also remove the Z80 CPU and the other large socketed chips. Be careful to lever them out gently, working from each end in turn, so as to avoid the chip flying out. Otherwise you may bend or break some pins :evil:

First test for short circuits. Pick one socket out of main DRAM bank. Your meter needs to be on a resistance range, for example with a manual type, use the 20k or 200k range. Test each "pin" to each other pin, so test pin 1 to pin 2, pin 3, pin 4 etc. Then pin 2 to pin 3, pin 4, pin 5 etc. Record any that are lower than 1k.

Next, do the same on the video DRAM bank.

Now test on all the CPU pins using the same method.

If you get this far, that's good progress.

Next, we are looking for good continuity. So switch to the 200 ohm range or continuity range. Now on the sockets of the main DRAM (4164's) test that each of the following pins is connected to the same pin on each of the other sockets: 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 15, 16.

On the video DRAM (4116) sockets, do the same for pins 1, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 15, 16.

Now the next stage gets a bit more tricky. So that can wait for now. So get on with the procedure I listed above, report back and good luck ;)

Mark

Re: Another repair case, plase help if you can

Posted: 21 May 2014 22:04
by laios
I had some free time and invested it to the mtx
I unsocketed and checked the main 4164 DRAMS
the resistance check with multimeter passed ok
then I checked the sockets with condinuity check as was the instructions , was also corect
then I socketed the rams and just of curiosity I pluged it on
On my surprise I see that screen instead of the black screen, it was the same character repeated something between pound symbol and dollar , that was not static but changing
sorry for the quality was taken with cellphone
Image
however at the next power on I had again the black screen ,and this time less stable , I have to tune manual to find it between 35 to 36
channel, any ideas ?
I will continue with video ram tomorow

Re: Another repair case, plase help if you can

Posted: 24 May 2014 00:10
by Dave
That display is typical of a VRAM problem.

Look at the service manual (available on my manuals page) for instructions on disabling each VRAM chip in turn by grounding one of the pins (I can't call which) through a 10 ohm resistor to see if the display improves when one of the chips is disabled.

That may give you a pointer as to which chip(s) is/are faulty

regards
Dave