POLE POSITION – A RESTORATION – PART 9

Welcome back great people. Thank you for taking the time to once again return to read through the next update in this restoration.

This will be the third in a number of updates surrounding the innards of any arcade cabinet, the electronics, where, today, we will be continuing to look at the printed circuit board, following on from the previous post, which if you missed it, you can catch up HERE.

Last time, we wrapped up with myself standing on the doorstep of an amazing friend, hoping with all my hoping hope, that he was both in, and going to open his door.

I want to add a similar caveat to last time, before we get going. This area of knowledge is not my forte, this update is as much for my benefit, than anything else. I don’t pretend to know what I’m doing, any more than I propose you follow these notes on your own boards.

And finally, but of huge importance, is to overstate just how knowledgeable, helpful and patient, my friend “Lester” has been and remains as such, with advice, expertise and practical skills aplenty, but also the willingness to sit down and explain it all (a number of times) to a clown like me.

When I refer, below, to things I have done, I really mean Lester did. When I say what we did, I mean that Lester did it. And if I write something wrong here it is because I listened wrong and Lester did it the right way.

Had Lester not answered his door, this update would finish around about now….. But the door opened….

With the pleasantries partaken and the pastries proffered it was time to bench the boards and say a small prayer…

Lester had already planned ahead and very kindly ordered the required edge connectors and made a jamma adaptor, which then plugged into a very smart, compact and incredibly functional jamma to scart convertor. A device designed and developed by a very talented ukvaccer, for the community, powered by a standard switching power supply, which all worked together to display the PCB output direct onto a portable CRT unit. A very compact and versatile set up indeed. I was already taking notes..

***

But before we plugged in the board it was time to tackle the edge connectors themselves and with the two boards, each joining to the main loom, as well as each other, that was four edges to tackle.

A common issue with these is, due to a number of possible reasons, the connection is no longer perfect, which can lead to, at best, failure to boot/play properly, whilst at worst, can lead to the voltage dropping and in some cases, the power supply trying to react against this by sending more and more power to the board, eventually frying the components and causing irreparable damage.

Causes of this can be simple wear to the edge/connector, caused by years of connecting and disconnecting, wearing away the traces; or a build up of solder if, rather than make an adaptor, operators have instead wired directly onto the edge itself.

So, out with some solder braid, a large tip on the soldering iron and a little bit of time spent cleaning all of these up. It is also worth taking the time to check the loom connectors to to make sure they are all clean, in line, and going to make a good connection.

***

Okay, with that sorted, our attention returned to the board itself. And time to begin, my first lesson.

Every time the board is worked upon, before running the power through it, good practice is to use a multimeter, set to check resistance, between the ground pin, on a TTL logic chip, and the power pin on an Eprom chip. There should always be a reading returned. If not, or if it reads close to zero, then this suggests the board could have a short, possibly from the tinkering, and therefore do not switch the power on. A simple check that could save a load of work. To identify a TTL chip it usually has a prefix of 74, whilst Eproms are prefixed with 27. On a logic chip the ground pin is always the last pin on the first row. On an Eprom, the power pin is always the first pin on the second row.

With this checked, the power was put through and the familiar scrambled mess, we left our last post at, was back on the screen. Now sometimes you get lucky and sometimes you make your own luck, but with the board now physically in front of him, rather than on a video call, Lester got all hands on, poking and prodding around the 7E chip, and whilst doing so, we noticed the on screen garbage was changing. The more we rubbed and pushed, the more it changed and flickered. Clearly there was some connectivity issues at foot.

Power off, and with the chip removed from it’s socket and the old socket desoldered from the board, a load of old flux was revealed, and also some of the top green layer of the board missing. Evidence of a previous repair, perhaps?

Old flux removed, the area cleaned up, some ultraviolet curable mask applied where needed, and a new socket was fitted.

Whilst the chip was out, Lester also cleaned up all the pins and tested it on his chip reader, on which it passed on all the tests and showed as working fine. At least we can move forwards from here.

***

Refitted the chip at 7E, tested for any shorts we may have made during our handiwork and although we still had a RAM 8 error, at least it was solid and static and showed no signs of instability when pressing on the chip. We also know that this chip/socket is all good. To double check this, using the logic probe, showed all pins to be behaving as expected and not as it had read in the previous update, when we tested the low pins.

Staying around the 7E, we know it is doing what it should be doing, but if the data is corrupted before it gets there, then the familiar garbage in, garbage out adage is true. So, looking back at the schematics, this whole sound processor circuit has a number of other chips, including the 2 Eproms at 7F and 7H which are linked to the PAL chip at 7C and the Z80 CPU chip at 7D.

Also on this diagram, on the right hand side are chips 6D, 6E and 6F which have outputs labelled AB1 through to AB14. Moving to another schematic diagram for the area of concern on the video board, you can see, below, that these AB feeds appear at the top left corner going into the custom chip at 7E and through a number of logic OR gates, into the bank of chips at row 7 and 8 at F through to K.

So these two areas of each board are definitely linked, but where to start. Well, with no time to waste, Lester reached to a set of drawers to his left and produced a very nifty tool, with it’s name, SLICE, emblazoned across the top.

I don’t pretend to understand this fully, but managed to nod in the right places often enough to allude enough of an air of competence, for Lester to think I was keeping up. In short this piece of kit was able to sit on top of the chips, in situ and measure in real time, their inputs and outputs. But it was also able to mimic the intended chip and give the required outputs as expected.

And as we continued to probe and interrogate each chip along the schematic path, more and more pieces of kit, some having a definite Soviet era appearance, began to fill the available bench space.

From the readings obtained and after some head scratching, we (he) replaced 6F. Desoldered old socket, installed new socket, cleaned the pins, checked for shorts, power on and…. RAM 1 error…

Tracing everything back to the Z80 CPU chip at 7D on the CPU board, we wanted to implicate/eliminate this. We were now using a Fluke 9010 machine, which bore more than a passing resemblance to a cash till from the 80s, maybe one from the iconic clothing store, C&A, remember how the money pods were sucked up into the ceiling… Anyway, this machine allowed us to send specific commands to the CPU chip and see how it read them and what it wrote as it’s response, (I think).

After sending a number of different commands, we were starting to note a common error code being returned, showing a decode error at databus 3, with the schematics pointing us to 8F on the video PCB.

***

8F removed, tested, passed, cleaned and returned. But, remembering from the last post/update, 8F is linked to 7F, so the same procedure was applied to 7F, which showed consistent failing under testing. Had we found the fault?

Having sourced a working 7F from another board, which passed all the tests, this was replaced, the board checked for shorts and powered up. RAM 1 error was no more.

But RAM 8 error had made an unwelcome return…. So it appears that we may have both created and fixed the RAM 1 error, only to return, full circle, to where we were when I originally powered the board up.

No one said this was going to be fun..

It was time to make a smart decision. It was time to make the right decision. It was time, for lunch.

But this was not to be just any old lunch, no, Lester drove me through valleys and over fells to a tiny hamlet, where someone’s front room had become, a pie shop!!

Oh yes, and not your average run o’ the mill, pie shop. This was the gold standard! I opted for the minted lamb with chilli peas, oh yes, with a rocky road chaser. And by Gods, they were good.

But this frivolous camaraderie was not, a poorly PCB, going to fix and so, to the bench we returned.

If you recall the manual suggested RAM 8 to be aligned to chip 7E on the CPU board.

Pin 18 is listed as the Chip Select (CS) pin. Using the logic probe on this, showed it was not receiving any instruction. The schematics show this instruction to be labelled CMOSRAMCS.

Searching through the schematics found this signal being generated within the RAM Battery Back-Up Power circuit.

You will also remember, that the battery had previously, and rightly, been removed, but as it was only there to allow scores to be saved, and coupled with the knowledge that the board had worked previously, with this removed, I couldn’t see why this would affect the boot process.

I have also mentioned the option to replace the chip at 7E with a NVRAM chip, which replaces the battery circuit and allows for your score to be saved, but we did not have one of these at hand and were not completely convinced that it would resolve the error, but definitely something that I will be looking to do at some point and document in a future entry.

For now though, we decided to try to bypass the circuit, shorting pin 18 with the CMOSRAMCS signal from the battery circuit, and………. The game attempted to boot!!!! The boot procedure ran to the first test screen, before crashing to garbage.

Click HERE to see the progress for yourself.

So, does this mean the problem IS in the battery circuit? We checked the transistor at Q2, where both CMOSRAM signals are linked to, and this tested okay. But shorting out Q2, had the same effect, with the board booting to the test screen, before crashing to garbage.

We still were not convinced that there was anything in this circuit that would prevent booting, but what we did have was a board that was trying to boot for long enough, to allow us to try some other avenues.

Starting with the custom chips on the CPU board, Lester began to remove one at a time, powering up the board with the custom removed, and seeing if the board booted further, not as far, or no change.

The working theory behind this was little more than an expectation that most/all the customs would be checked during the boot process, and we expected to see ‘something’ with each removal/power up. And we did, until we got to the custom chip at 9FA.

As you can see, this chip sits, again, within the sound (generator) circuit, but what was interesting is that when we removed it, there was no change to the boot process, whether it was in or out of the board….

We swapped it out for the same chip from the other board and the game managed to boot all the way through the self tests and into attract mode. Result!!

But, before we had chance to break out the champagne and cigars, after about 10 seconds, the graphics started to glitch and random characters began to appear, with this worsening until the screen was, once again, filled with garbage and it crashed.

However, this was definite progress.

Click HERE to see just how close we were..

Putting the board and ourselves to bed for the night, the next day saw a new sunrise, a new raison d’etre and….. a new error to appear on the board…

Continuing on with the above process of removing and testing, one custom at a time, focusing around the sound generation circuitry, to try and understand how this interacts within the boot-up process, we ended up with another error and the board no longer booting, again..

If you recall, when I spoke to the previous owner, they told how it had worked reliably, up until he had swapped out some chips to help another PP owner in the community, try to pinpoint a fault on their board.

So, essentially, we had both been doing the same chip swapping, and had caused an error. Add to this the issues we found with the original socket that had been on chip 7E, and couple this with a quick web search, exposing the unreliability of these sockets themselves and the decision was reluctantly made to remove all, yes, ALL the sockets from the board and replace with modern equivalents.

I know this seems extreme, but we are chasing ghosts across the PCB where chips are testing good, when removed, but causing errors on the board, and until we change the sockets themselves, we cannot confidently rule out, these being the reason.

Thankfully Lester had a very good desoldering station which at least made this somewhat easier, but it was a slow and painfully boring process. With each chip removed, he also took the opportunity to clean up all the legs on the custom chips.

These things are sooooooooo fragile. A lot of them are also unobtainable, so this whole process was something of a trial by fire… and, maybe inevitably, despite the utmost care being taken, a couple of these rare beasts did succumb to this process.

At this point, Lester (who will argue he had been from the start) was on his own. With a dremel in one hand, the custom chip in the other, the illuminated magnifying ring in place, he gently ground back the package of the chip, to expose enough of the metal connection underneath to allow him to solder a new leg in place.

From someone, me, who had never been able to win a game of the children’s board game Operation, I have never been more impressed with this technique. Custom chip saved.

After much rinsing and repeating of the above process, we eventually got to a place, where, with all the sockets on both boards removed and replaced; with all the custom chips cleaned and refitted; with all the edge connectors cleaned and ready; it was time to (check for shorts, and) power up the board.

And we got this…

Click HERE to see for yourself, could it really be…

Yay!

There is no sound on the video as we hadn’t rigged this up on the bench, but using an audio probe provided some promise that the speech and sound would hopefully be correct. We hadn’t rigged up a credit button or any controls either, as I can sort all that out when I get the working PCB back home.

And this I where we will wrap this entry up, quickly, before it breaks again…

There is still some electronics stuff to cover, I have been looking into the reliability of this game and what can be done to improve this, so I’ll go over what I found and what I have done in another update.

Until then, I’ve been busy with other areas of the cabinet, so I will document this next, to break things up a bit and bring you up to speed with where we currently are with the whole restoration.

If you like what I’m up to, please subscribe, read through my earlier posts, hit like and share on the social media of your choice, via the buttons below, etc.

I genuinely would love to hear what you think, good or bad, and will always reply to any comment, kind enough to be left.

Thank you for taking the time to read through, see you all soon for the next instalment.