Jump to content

Tales from Midibob's workshop Proconn


midibob
 Share

Recommended Posts

48 minutes ago, riche100 said:

i had a address issue once , someone got my postcode and house number wrong 

I bet it wasn't fixed with a chip change either!:lol:

53 minutes ago, CanonMan said:

Sounds like an issue with the address decoding...

I'm not sure how the addressing works on this tech as all the CE pins on the PIOs are driven directly from the address bus. One thing I did try was to isolate

the CE pin so that it wouldn't address the chip but it made no difference.

Link to comment
Share on other sites

3 hours ago, midibob said:

I've got a PC92 on the bench which is giving me a real case of itchy head!!

It came in as a non booter and on inspection was found to be stuck in reset. The reason for this was corroded track in the reset circuit. Once this was made good it came out of reset and .....did absolutely nothing!! These can be horrors when you get to this point as it can be any of the PIO chips, CPU, CTC etc, etc. This time I was lucky as my hand was resting on the board over IC4 (PIO) and I noticed it was running quite warm. Out with the desolder tool and in with a socket. Before putting in a replacement I thought I'd switch on in case anything else might be suspect. Well, what do know she boots, runs and does everything except the functions of IC4 which are basically the 7 seg display and the triacs.

OK, so flushed with success I thought I'd slip in the new PIO and wrap it up. Switch on no boot!! Remove chip, boots OK?? A faulty spare chip you all say, but no, tried it in another board and it's fine???? What is going on? All the other functions are testing OK so all the existing PIO's should be good. Looking back in this thread a couple of entries ago the SIO chip could be responsible although I don't see how. I'll swap that out tomorrow to see what happens.

Watch this space....

Fantastic  read Bob as always I'm sure you will over  come this issue as always 👍👍👍

Link to comment
Share on other sites

Not sure if I like the taste of humble pie or not but I did have a large slice today.:lol:

I eventually found that one of the data lines was low and looking on the solder side spotted a bridge between pins 19 and 18 on IC4. Don't know why I didn't see it before as I always check with the microscope after putting in sockets. Not sure what's on pin 18, it's not mentioned on the diagram but with the chip in it brought the whole thing down.

Of course I did start with the serial I/O chip first which was blameless but hey ho that's life!

Now it all boots with IC4 in I'm happy, well sort of as there's now two segments not being driven on the 7 seg display.:headache:Roll on tomorrow.

  • Like 2
Link to comment
Share on other sites

3 minutes ago, midibob said:

I eventually found that one of the data lines was low and looking on the solder side spotted a bridge between pins 19 and 18 on IC4. Don't know why I didn't see it before as I always check with the microscope after putting in sockets. Not sure what's on pin 18, it's not mentioned on the diagram but with the chip in it brought the whole thing down.

A short between ARDY and D0. I can see that messing things up...

 

"ARDY Register A Ready (output, active High). The meaning of this signal depends on the mode of operation selected for Port A as follows:

1. Output mode: This signal goes active to indicate that the Port A output register has been loaded and the peripheral data bus is stable and ready for transfer to the peripheral device.

2. Input mode: This signal is active when the Port input register is empty and is ready to accept data from the peripheral device.

3. Bidirectional mode: This signal is active when data is available in the Port A output register for transfer to the peripheral device. In this mode, data is slot placed on the Port A data bus unless ASTB is active.

4. Control mode: This signal is disabled and forced to a Low state."

  • Like 3
Link to comment
Share on other sites

Well the two segments missing from the 7 seg display turned out to be two of the power resistors being open circuit. They were actually damaged which had broken the wire in the resistors.

856026783_P1040412(Custom).jpeg.a1bc2ee10a9820f136d6e6b95481bfc3.jpeg

You can see the damage on the top, more than likely caused by boards piled on top of each other.

I can't find these anywhere, they are 47R 2W with 10mm lead spacing so it could be a donor board job which is a shame.

Another 'fault' showed itself as triac TR5 always on and I couldn't make out why as it was definitely being driven from IC4. I eventually found this by chance and it was due to switch 5 being in the OFF position!!!

  • Like 1
Link to comment
Share on other sites

5 hours ago, midibob said:

Well the two segments missing from the 7 seg display turned out to be two of the power resistors being open circuit. They were actually damaged which had broken the wire in the resistors.

856026783_P1040412(Custom).jpeg.a1bc2ee10a9820f136d6e6b95481bfc3.jpeg

You can see the damage on the top, more than likely caused by boards piled on top of each other.

I can't find these anywhere, they are 47R 2W with 10mm lead spacing so it could be a donor board job which is a shame.

Another 'fault' showed itself as triac TR5 always on and I couldn't make out why as it was definitely being driven from IC4. I eventually found this by chance and it was due to switch 5 being in the OFF position!!!

what can i say bob ,you nailed it again mate ,i have probably got some of these resistors on a scrap board ,i can pull them off and test them for you and put them in an envelope with a first class stamp,if this is any good to you pm your address 

Link to comment
Share on other sites

  • 2 months later...

Procon boards seem to be coming out the woodwork at a great rate of knots at present. Looks like I'm going to be doing nothing else for the next couple of weeks!

This one was a goodie. On switch on it would almost boot then make funny noises and lock up. If the board was flexed it would boot completely and then maybe lock up when pressure was released. It was all very hit and miss. Eventually I found applying pressure to IC3 (PIO) would make it go one way or the other. As this was a socketed chip suspicion was centered on perhaps a dry joint somewhere.

P1040516.thumb.JPG.7e24d0698cace9e1706b705a192c62ae.JPG

As it happened lifting the chip out showed to be the cause ......

P1040517.thumb.JPG.594b539561af134a9494705660d90c4a.JPG

Two curly pins! Once straightened and the chip re-inserted it was stable. I then noticed there was a lockout always active whereas they should all be off after boot up.  One of the gates on this chip had failed so a new one had to go in anyway.

Working with loads of PC Plus's I've seen many with the bonfire starting under or around the battery area. The reason for this has been well documented by Andrew in the past so I don't take any credit for this but here is my method to get around the issue.
The 53V rail passes under the battery and the whole area becomes conductive after years of battery rot ingress. The high voltage then finds it's way into the supply rails and if you're lucky, sets fire to the board!

'LUCKY' I hear you say, well if you're unlucky it sets fire to the board and blows all the 5v logic chips.:lol:

To see how bad the leakage can get if you connect a meter on Ohms range with the black on a ground point and the red on the 53v supply (obviously power disconnected) any slight bit of humidy around the battery area will drop the resistance down to a few K Ohms instead of being open circuit. I usually test this with a straw and just direct a small puff of breath. It's quite frightening to see how the resistance plummets!

As Andrew previously documented you can't really cure this as the board has become contaminated. I must admit I haven't tried the Scorp 1 method but it's just too time consuming and I don't know if it would work anyway. The workaround I'm using is to completely disconnect the 53v rail both sides of the battery area by cutting the PCB track and then making good with a patch wire. Andrew may have done this anyway in which case I doth my cap but here we go...

The two cuts are made here.....

 

P1040519.thumb.JPG.d8402bace3696c65bd2fcd9ffc2ddc1c.JPG

The patch wire is inserted into the via just below the left hand cut. Sometimes I've found the right hand cut isn't far enough back and can still be susceptable to stray voltage leaks, in which case you need to make a cut nearer to the power socket......

P1040518.thumb.JPG.008240cdc2a53e2ed79c7f6a6c447b19.JPG

Once done just tack a wire on the underside.

P1040520.thumb.JPG.feabd669476708926519f83d7ac28e6b.JPG

Testing has shown no stray voltage to be apparent even under high humidy ( a good puff :lol:) round the battery area. I would still recommend cleaning with decent IPA as it can get under the 5v regulator heatsink too.

  • Like 2
Link to comment
Share on other sites

Really great thread, I really want to learn about z80's and what the PIO's are doing (been doing some reading). I'm not so keen on the sound of contaminated +55v rail and battery area 😂

Link to comment
Share on other sites

Another bizarre fault today. The board would boot sometimes, and other times only partially ie. the reels would index and then it would click to initialise but then just make a high pitch noise? Eventually I discovered that it was due to IC1 (Z80 PIO) by substitution but couldn't understand why. After much head scratching I found that if it was stuck with the funny noise just touching the chip would make it fire up correctly??? If that wasn't strange enough, only when it was touched in the middle. Further proding and poking led me to pin 16 and touching the meter on this pin would also fire it up.

Now the really odd bit, this pin isn't used and therefore not connected to anything. I can only think that internally it should have damping resistor to 5v or 0v which was damaged. Anyway a replacement sorted it.

Another one had me going for a little bit. It wouldn't completely boot and during the reel indexing there was no activity on reel 3. Checking on IC1 confirmed that no drive pulses were present which they were on reels 1&2. When the reel indexing sequence finished they did change state which made me think it possibly wasn't IC1?

Then a thought occured to me, if the reel sensor was stuck it would think it had already indexed and would never get round to spinning the reel. Checking the reel sensor input on IC3 showed reel 3 to be stuck low and with a new chip all was restored to normal.:D

  • Like 5
Link to comment
Share on other sites

.... and some more.

On switch on the board would clock up 18 credits, in fact every time it was switched on another 18 credits? Sometimes just touching the coin inputs with the scope probe or the meter would clock up credits. This little head scratcher turned out to be the 12v supply feeding the coin mech. Transistor T95 had a broken leg and hence no 12v was delivered to the coin mech so the inputs were floating. This was caused by bad storage of the board and it's a real bug bear as I see so many faults similar to this with broken/crushed components, broken tracks due to a deep scratch etc etc.

Still, all in a days work.:lol:

Had a couple that were dead but looked like they should be running? Previous repairs had taken place with regards to the transistors in the reset circuit but what hadn't been checked was the resistors.

image.png.46e1cf49884a154f34c9dddd5f2be99b.png

It's always good to change all 8 resistors in the middle and the three, or even four, at the top. I always use the fingernail test, applying slight pressure to the side of the resistor. If they are really bad the wire comes away from the body. They usually have green legs but not always so don't be fooled! In this case one of the 10K's failed the fingernail test and another one fell apart on removal.

You can do a manual reset to prove the rest of the board is OK just by momentarily shorting CPU (26) to ground. If it then boots you know the reset circuit is misbehaving.

Another real head scratcher was the reels not spinning correctly (or at all). I use my Arduino to fake a reel set and I have LED's to indicate the motor drives. In this case there was no drive to the odd reel motor coils although the evens were OK? This was a PC 92 and the reels are driven, odds from IC1 and evens from IC2. There were no outputs on IC1 whereas IC2 looked fine? You'd obvioulsy go straight to IC1 and swap it out but then what do you do when the spare is exactly the same!:headache:

After not too long (I surprised myself here) I thought I'd just check the signal paths and eventually got round to the DATA bus. All were showing a signal on the scope but D7 didn't look right. It looked like there was lots of information missing due to big gaps in the waveform. Checking D7 on the CPU showed what I expected to see so I'd sussed it there was a break somewhere. That was good news as it's faults like this that cause me to jump in the Thames!

If you've ever tried to trace a track on these boards it's horrendous, and for the life of me I couldn't find where the connection was between all the PIO's and the main DATA bus back to all the other chips. To make some headway I just put a tack wire back to the CTC chip and BINGO all the motor drives were now working.

Get that wine bottle open!!

  • Like 7
Link to comment
Share on other sites

Down to the final two in this present batch and of course I always leave the worst till last!!

This one wouldn't boot and I couldn't see why not? There was activty on all the DATA and ADDRESS buses but nothing happened. Someone had had a good go this already and 4 of the PIO chips had been socketed and swapped out. In the head scratching that followed I socketed the other two so that I could test all the chips in a good board. So all PIO's and SIO tested OK. The CTC was then swapped but still the same. The RAM was already socketed but a replacement also made no difference. Attention then turned to the Synth chip but that proved to be innocent as was the CPU.

Next stop the Thames I was thinking....... but one last thing was to check all the pull up resistors to the buses which should all measure 10K Ohms. As I was checking I noticed two of the DATA lines only measured 5K!! Blow me there was a short between D1 and D6!! As I've mentioned before trying to trace the PCB tracks on these boards is horrendous and if you have a short it could be anywhere. With the panic hat now removed and logic hat back on I thought it had to be caused by the previous repairer and that indeed was the case.

The only place I could see where the two DATA lines were anywhere near each other was under the RAM socket so a quick removal was in order.

P1040525.thumb.JPG.931f1ed49c438c5f0583d4e358e43869.JPG

Once I'd seen this I was cock-a-hoop as I new I'd sussed it. Can you see it?
Once cleaned up and the socket re-fitted and chip put back in I was thinking it'll be time to wrap up. Shriek, it still doesn't do anything! The short had definitely gone but not a dickie bird.:eek:

Quickly slipped in the test RAM chip again and BINGO she lives!! Phew. This is why repairers always hate it when a board has already been worked on as any faults unknowingly applied can be real pigs to find.

OK so now I have a functioning board I can actually do some testing. First thing that became apparent was some of the 7 segment displays were showing extra segments. I've seen this quite a lot recently and it mostly due to the 100 Ohm standoff dropper resistors (bottom left hand corner) which get bashed about due to poor storage. They then break their solder joints and touch each other.

  • Like 3
Link to comment
Share on other sites

Good work Bob.you have good patience to do this.do you ever feel like throwing a PCB across the room.but I guess you are used to issues and just get on.good to see proper repair and analysis.

Link to comment
Share on other sites

I can't say I ever get angry enough to do that. Normally just a shrug of the shoulders and go and make a cup of tea!:lol:

And so on to the last one. After yesterdays success the first thing I thought I'd do was to check all the DATA lines but this time all were in order. They do say lightning never strikes the same place twice! So where do we go from here? ADDRESS and DATA lines are all running but there's no IRQ on the CPU (pin 20). This called for a quick check in the Midibob book of horrible faults. Page 1, don't jump in the Thames! (or 'DON'T PANIC' if you're a Douglas Adams fan)

If the board doesn't boot and you've tried everything else, before you start lifting all the PIO's start with the SIO. Blow me, hole in one! The SIO chip was bad. In with a replacement and BINGO she lives. These boards will actually fire up without all the PIO's but not the SIO, which is good to know.

This set of Procon boards has been a real challenge but I've managed to succeed and keep my sanity. Time to move on to the next box of faults as I've got quite a few folks waiting and still unopened boxes in my hallway. Repairing boards is one thing but keeping the missus happy is something else!!!!!

  • Like 5
Link to comment
Share on other sites

  • 3 weeks later...
  • 3 months later...

Haven't posted here for a while although plenty of boards have gone through the workshop, no time as per usual.

Anyway one today had me head scratching for a few minutes.

The board came in with 'no boot' as the fault which is usually most of the transistors, resistors and track around the battery area. In this case someone had been here before and had done quite a few repairs. On first inspection I noted the volume pot was missing and on boot up there was nothing although the RESET was good and there was activity on the ADDRESS lines. The DATA lines however were very low, under 2v. First suspect was the CPU which was innocent. Fitting the volume control just gave a continuous farting noise?

So where does one go from here?

The DATA lines go to all the support chips, RAM, CTC and the Synth chip!! A bit of luck intervened here before I started socketing everything. Checking the waveform of what there was on the DATA bus by the SIO chip showed about 2v p to p. Travelling back towards the CPU I noted it got slightly less per each chip ending up at the Synth chip where it was the lowest. Probably only talking about 0.2v per chip but that made it about 1v difference. Now the synth chip looked like it had been removed and replaced  but probably not swapped and it hadn't been socketed (grrrr!) Anyway, once removed, socket fitted and replaced it then booted first time and ran as it should.:clap:

The chip in question is the AY-3-8910 and these are now becoming difficult to source at a reasonable price. prieces on Ebay seem to vary between approx £18-40 which is ridiculous. I don't even look at the Chinese rubbish as the phrases fingers and burnt spring to mind! Even the Yamaha equivalent YM2149 is getting scarce although I have ordered a couple.

  • Like 3
Link to comment
Share on other sites

  • 3 months later...

Hi folks, this will be the last workshop thread this year and it's a real humdinger. Are you sitting comfortably? .....

Must be a good 6 months or more ago now I was given a PC 92 board that was in a bad state and probably more fit as a donator board. As the owner had done me a few good turns I thought I'd have a go at trying to repair it.
There was a fair bit of corrosion around the battery area (as per normal) and some of the large multiplex power transistors had been ripped out along with their vias and associated track then put back with tack wires! The reg transistor and associated heatsink were missing and the smoothing cap had been ripped out along with one of it's vias. These are larger than I stock, to take the cap terminals, and the only place I could get them from was Digikey in the US!
 
Unfortunately I didn't take a pic of the board pre repair but this is how it ended up...

Finished-1.thumb.JPG.db1314bb40368610dbd87385621ba4ac.JPG

Finished-2.thumb.JPG.2f1e3fa6259ce57e3ef5e42a587f64fd.JPG

So, first thing was to remove all the bad corroded components and give the board a good clean around the battery area. Next step was to try and work out how many tracks were corroded around the battery area. Anyone that's worked on these knows what a pain this is as there's so many connections going all over the board passing through this area. It's a nightmare! MPU4 is childs play in comparison.

Fit the new rivet for the main cap and re-fit a heatsink. Fit a total of 14 rivets to renew all the multiplex transistor broken vias and battery terminals and replace all the broken tracks. Then replace 11 of the transistors, 24 resistors and 7 diodes.:headache:

With these boards I always fire them up first time without the 53v connected as contaminated boards leak through to the 12v and 5v lines and anywhere else that's in the vicinity. Once powered up I check to see what voltage if any I see on the 53v line. If it's any more than a couple of volts it needs investigating and the 53v supply should not be connected until resolved. In this case it was about 12v and the 53v track which runs under the battery was leaking across to other tracks.
I usually cut the 53v track at both ends either side of the battery area and run a patch wire as even deep cleaning doesn't appear to effect a cure.


Andrew mentioned this a long time ago in one of his repair threads about the boards getting contaminated and conductive so all credit to him.


So, with the 53v still not connected I switch on and .....nothing! Well it comes out of RESET so that's a start I suppose.
Everything appears to be running ie activity on the ADD and DATA lines. Hmm hold on though there's nothing on A11 on the CPU?
From here things start to get a bit hazy and as we're looking back quite a few months it's difficult to remember exactly, but I digress.
Removing the ROM showed all ADD lines pulsing OK so it was initialising to some extent then crashing.


Nothing was making any sense until one day I switched on, it bleeped and the 7 seg came on. The only thing was it didn't go any further as there were no motor drives? I was coming around to the point of having to remove all the 40 pin chips and test them. So 7x 40 way sockets later, all chips tested OK and I'm back to square one.
The 7 seg wouldn't always come on and sometimes would just flash for a second and then nothing. Once the board had been on for while (warmed up) it would then start but only to the same point of no motor drives. It was then I discovered that connecting the 53v supply it would boot first time and every time hot, cold but still only as far as the motor drives?
The months are passing and I'm really getting nowhere with this. One day I switched on and was connecting and disconnecting the 53v supply and all of a sudden it booted completely!! This made absolutely no sense whatsoever but it did prove that the board was a goer and all I had to do was to find why the 53v supply was affecting it.
This went on and on for a while getting nowhere as usual until one day I had it running and out of interest operated one of the DIL switches (3) which crashed the board?? Out came the DIL switch which was contaminated and leaking voltages all over the place. Whilst on this track I then checked the board under all the diodes in the top left hand corner and there was bad tracking all over the PCB so all the diodes had to be lifted and the board deep cleaned.
This cleared all the leakage but I still had the original fault that it would only boot half way with the 53v connected sometimes not at all with it not connected and sometimes boot completely?? This was driving me insane!


OK, we're nearly at the end of this monumental saga. This morning some  6 months later I was trying a different fault technique as I was trying to find out why the 53v supply was getting in to places it shouldn't be. Using an oscillator connected to the 53v supply with the PSU unplugged I searched round with the scope looking to see if I could pick up any leakage. Nothing was really showing until I got to the SIO chip then touching pins 13+14 showed almost the same voltage I was applying! Further investigation showed that this connection goes to IC U4 (74LS92) which is part of the main clock circuit, it also runs right across the battery area and under the heatsink. It had a break in it somewhere so it was out with the patch wire. By this point I was getting a bit excited and checking on the diagram showed that this feeds a 75kHz clock to the SIO chip. On switch on it all sprung in to life first time, with or without the 53v connected.

Bingo I'd sussed it after a good 6 months. This must rate as the worst board I've ever done but it wasn't going to beat me. I do hope Santa reads this.:lol:
Funny thing is I have another board that won't get passed the motor drives and it could be the same fault of the 75kHz pulse not getting to the SIO chip. That would really make my Xmas if it was.

Merry Xmas to all.

  • Like 5
  • Haha 1
Link to comment
Share on other sites

Well the other board had the little main clock daughterboard on it and although I'd thought I'd strapped out all the corroded tracks (the main clock was working) I'd not checked the secondary clock (75kHz). Lo and behold a track was broken on the 74LS92 chip. I could have sworn I tap'd them all out but must have missed this one. Once patched out the 75kHz pulse was received at the SIO chip and  it passed the motor sequence and fully booted.:D

  • Like 3
Link to comment
Share on other sites

 Share

×
×
  • Create New...