More MDFS Repair

for bbc micro/electron hardware, peripherals & programming issues (NOT emulators!)
Post Reply
User avatar
BeebMaster
Posts: 2664
Joined: Sun Aug 02, 2009 4:59 pm
Location: Lost in the BeebVault!
Contact:

More MDFS Repair

Post by BeebMaster » Mon Feb 11, 2019 10:22 am

I'm back on MDFS surgery. This is the "green" one, also known round here as MDFS 103 (because of the main ROM version number), or 2337 being its serial number. It's completely unmodified, so it would be nice to get it going again. It did used to work till the battery leaked.

Now I just get power light and (if no clock), no clock. Nothing else ever happens, it doesn't even get to all lights on.

Although it's pretty green, it isn't actually as manky looking as MDFS 2300 which does now work again.

I've been over the board again with vinegar and washed it all off. Nothing fizzed during the vinegar application and it still looks green.

I've tried its ROM, Z80 and 6818 in MDFS2300 and that one still worked with each of those chips in it. Similarly, MDFS2300's parts didn't bring 2337 back to life.

I tried to measure the Z80 on the 2300 and 2337 with my new Oscilloscope but I didn't exactly get the readings shown here. I was getting waveform type patterns on MDFS2300 but it was all over the place with 2337.

That was before swapping the Z80s so I have ruled out a broken Z80 anyway.

I'm suspecting the area around the battery/clock/charging circuit which was the main culprit on MDFS2300, so I am going to replace all the components there. I managed to break a leg on the X2 crystal when trying to measure it, so that hasn't helped.

Just checking - it's a 32.768kHz crystal, same as in the RISC PC next to the clock chip? For some reason I bought a handful from China (more than likely via E-Bay) a while back but they seem to be readily available from Farnell etc.

I've got everything I need to do the replacement except CV1, the variable capacitor. Can anybody point me to a part number please? I can't see a component value on the MDFS schematic or parts list.

When I replace the diodes D1 and D2, how do I make sure I put them in the right way round? It's hard to the tell the markings on either MDF2300 or 2337. On my new parts, D1 is red and appears to have a stripe but it's more or less in the middle. D2 is the black chunkier type with a grey stripe at one end.

While I'm at it, it may be worth getting a few other crystals like the X1 and X3 in the MDFS, and any other sort commonly found in Acorn kit. What values/packages might those be?
Image

User avatar
danielj
Posts: 6778
Joined: Thu Oct 02, 2008 4:51 pm
Location: Manchester
Contact:

Re: More MDFS Repair

Post by danielj » Mon Feb 11, 2019 12:44 pm

Parts are here. viewtopic.php?f=3&t=10542&s&start=90#p129253
Values of the crystals are on Ottley's excellent schematic:
viewtopic.php?f=3&t=9407

Clock crystal is indeed one of the 32khz ones.

There should be markings showing which way 'round the diodes go on the motherboard? If you've got a picture I expect I can tell you.

The RTC won't stop the boot process. First thing that happens is all the lights get switched on. If that's not happening it can't read the ROM properly, or something's stopping it from reading the ROM...

Check if the processor starts attempting to read from addresses and whether you see anything on the data lines? (this is off the top of my head)

d.

User avatar
BeebMaster
Posts: 2664
Joined: Sun Aug 02, 2009 4:59 pm
Location: Lost in the BeebVault!
Contact:

Re: More MDFS Repair

Post by BeebMaster » Mon Feb 11, 2019 1:27 pm

Thanks, I couldn't see the X values on the Schematic or parts list (or CV1).

I was going to replace the whole of the clock/battery circuit, (ie. D1, D2, Q1, C12, C13, R53, R54, R55, R57, R58, R59, X2, CV1) as it was repairing around this area which fixed MDFS2300 previously, and I think I've already had a go this area on MDFS2337 which may have made it worse on a previous occasion. However if this part of the board doesn't stop it getting to all lights on I will have to investigate further first.

I presume though that X2 being broken will stop it booting?
Image

User avatar
danielj
Posts: 6778
Joined: Thu Oct 02, 2008 4:51 pm
Location: Manchester
Contact:

Re: More MDFS Repair

Post by danielj » Mon Feb 11, 2019 1:30 pm

I don't think so. It'll stop the clock running, but I don't think that will stop anything happening - the hardware doesn't touch the RTC until much further on in the boot sequence... If the lights aren't coming on it's not reading the ROM. That's what I'd want to get to the bottom of first.

d.

User avatar
BeebMaster
Posts: 2664
Joined: Sun Aug 02, 2009 4:59 pm
Location: Lost in the BeebVault!
Contact:

Re: More MDFS Repair

Post by BeebMaster » Mon Feb 11, 2019 1:45 pm

OK.

Actually I am now beginning to suspect the area next to the battery (D3, C3, R49, R6). C3 has already been replaced on a previous occasion but I am not getting steady readings from R49 and R6. In particular R49 seems to be settling around 1.9K when it should be 10K.
Image

User avatar
danielj
Posts: 6778
Joined: Thu Oct 02, 2008 4:51 pm
Location: Manchester
Contact:

Re: More MDFS Repair

Post by danielj » Mon Feb 11, 2019 2:20 pm

What does it read when it's not in circuit?

User avatar
BeebMaster
Posts: 2664
Joined: Sun Aug 02, 2009 4:59 pm
Location: Lost in the BeebVault!
Contact:

Re: More MDFS Repair

Post by BeebMaster » Mon Feb 11, 2019 2:21 pm

Just replaced it, reads 10.0K after being removed! The new one is reading 2.4K fitted, no change to the operation.
Image

User avatar
danielj
Posts: 6778
Joined: Thu Oct 02, 2008 4:51 pm
Location: Manchester
Contact:

Re: More MDFS Repair

Post by danielj » Mon Feb 11, 2019 2:28 pm

Which would suggest some sort of short somewhere... Look back at the schematic and check the connections on both sides of R49 go where they're meant to go... (+5V and back to an IC?) then check the resistance of the +5V side to ground, and the IC side to +5V and ground...

d.

User avatar
1024MAK
Posts: 8277
Joined: Mon Apr 18, 2011 4:46 pm
Location: Looking forward to summer in Somerset, UK...
Contact:

Re: More MDFS Repair

Post by 1024MAK » Mon Feb 11, 2019 2:59 pm

I’ve not got any time to help much today. But I did want to say that in circuit testing can be affected by parallel current paths. This where the test current from the test instrument flows via other components and via the supply and return/common/0V/GND wiring and back to the component under test.

If you have another unit with the same design of board (or very similar), it’s often easier to compare any unexpected results with a known good unit.

With regards to the fault as described. The first thing to check with Z80 based systems are:
  • Does the Z80 have power to it’s +5V and GND pins?
  • Is /RESET going low then high during power up?
  • Is there a good clock signal on pin 6?
If you are not used to using an oscilloscope, it’s best to practice on a known good working system, so that you can get an idea of what real life waveforms look like. As a lot of microprocessor systems don’t have many simple repetitive waveforms, sometimes it is easier to use the “single” / “once-only” / “one shot” triggermode on modern storage/ digital oscilloscopes.

Mark
Last edited by 1024MAK on Mon Feb 11, 2019 3:02 pm, edited 1 time in total.

User avatar
BeebMaster
Posts: 2664
Joined: Sun Aug 02, 2009 4:59 pm
Location: Lost in the BeebVault!
Contact:

Re: More MDFS Repair

Post by BeebMaster » Mon Feb 11, 2019 3:04 pm

R49 goes to +5V and LK3, an unjumpered link. I get 0 resistance from +5V side to another +5V point on the board (pin 1 of the keyboard) and 0 resistance between the link pin and the other side of R49. But still only reading 2.4K.

Also just changed D3, no help.
Image

User avatar
BeebMaster
Posts: 2664
Joined: Sun Aug 02, 2009 4:59 pm
Location: Lost in the BeebVault!
Contact:

Re: More MDFS Repair

Post by BeebMaster » Mon Feb 11, 2019 3:15 pm

The Z80 is OK, it's been tested in another working MDFS. Here's a picture of pin 6:
IMG_3757.JPG
Image

User avatar
BeebMaster
Posts: 2664
Joined: Sun Aug 02, 2009 4:59 pm
Location: Lost in the BeebVault!
Contact:

Re: More MDFS Repair

Post by BeebMaster » Mon Feb 11, 2019 5:18 pm

It's difficult to get a picture of the working MDFS before all the lights come on, but this is what happens at A15 on switch on:
IMG_3783.JPG

Then it's all over the place as the boot sequence starts:
IMG_3786.JPG

The broken MDFS only gives one image:
IMG_3788.JPG
I used the same setting so I don't know why the image appears more stretched.
Image

User avatar
BeebMaster
Posts: 2664
Joined: Sun Aug 02, 2009 4:59 pm
Location: Lost in the BeebVault!
Contact:

Re: More MDFS Repair

Post by BeebMaster » Mon Feb 11, 2019 5:21 pm

Ahhh....the broken MDFS's Z80 was in a ZIF socket. I've put it back into the motherboard socket and it now gives the same size waveform as the working one.
Image

User avatar
BeebMaster
Posts: 2664
Joined: Sun Aug 02, 2009 4:59 pm
Location: Lost in the BeebVault!
Contact:

Re: More MDFS Repair

Post by BeebMaster » Mon Feb 11, 2019 5:42 pm

R49 in the working MDFS is measuring 10K, compared to 2.4 in the broken one...so definitely something suspect in that area.
Image

Post Reply