Need some help with Elevator Action

[patzik]

Graphics go through some logic stuf from the RD GD and BD bus and end up as SN11/SN21/SN31 for RD, SN12/SN22/SN32 for GD and SN13/SN23/SN33 for BD. Using the video probe I can see these graphics clearly, except the signals from the GD, which show garbage resembling the corruption.

So I decided to run the board, without the graphics board attached, mean all graphics should be blank right? True for RD, and the datalines on the 2 2016's are continuously high, same goes for BD. Both show nothing anymore on the video probe and no data.

But GD, probing this show me a pattern on the screen, the same as the corruption. And looking at the datalines of the 2 associated 2016's there is activity. So I am definitely close, now to find out how and why this sram is filled.

patzik2021-01-05 20:12:01

 

[patzik]

The screen without the probe is red, now here on this picture you see the distortion in green. The black part of the screen on the right is normally empty, it only gets filled with stuff from the video probe and is a 'copy' of the right part of the screen. A bit of an odd thing, but in fact is really handy because it's always empty except for stuff from the probe.

 

[patzik]

So I decided to look more at this, and replaced (socketed) ic79 and ic 65, 2 74ls157's that move the data from the GD BUS. And no change again. So I see the /G on these ic's and start wondering what would happen if I disconnect both.

Graphics are not good, but the corruption is gone. When I pull the 2 legs low, I get the corruption again.

The signal is low when the game is running, only with short activity when changing from title screen to gameplay.

Ok, now I am really confused, any suggestions???

patzik2021-01-05 22:16:52

 

[patzik]

So I was thinking, why is there activity on the GD bus with the graphics board removed. And why was disconnection the /G pin in ic79 and ic65 removing part of the graphics and also the corruption, but this returned when pulling the signal low. And then it started to make sense. There are 2 sram ic's ic107 and ic106 and my thought is that switches the data lines between these 2 ram ic's. That would mean that removing the problematic ram would stop the activity, it's not the ram itself, it's the stuff that is loaded into it but this would narrow the search down. Removing the rams one at a time confirmed this, removing ic107 made the signals disappear. Hmm, This bring me back to the beginning, this was one of the first things I looked at

but now I understand the board much better and have ruled many things out.

The only thing different between the 2 rams is the chip enable, so this is probably it. This comes from CDR2RQ via ic91 an 74ls08 with one input tied high. And is generated by ic73 and ic59 with some inputs.

So I started looking at the generation first and replaced ic59, this made no difference. Then I went back to that ic91 the 74ls08 and replaced this.

Now I'm not there yet!!! There are more issues with this board, and we'll get to that later....

 

[Lurch666]

Very nice work.

You have worked out the problem by slowly eliminating some things and figuring out the rest.

It's great to finally crack a difficult problem and get a good result-It's what I enjoy about fixing these boards.

 

[patzik]

Yeah, very satisfying if you can finally find the fault after many hours, but it can be frustrating if you don't. Now the board is not done yet, here's the todo list.

When I started the sound wasn't working, and there was this strange thing when coining up, I hear a small pop and the screen got a little bit faint whenever I coined it up. Didn't look into it, wanted to fix the graphics issue first. Then at one point I started having problems where it didn't want to coin up, not thinking I held the coin button pressed and then saw some magic smoke. I had no idea why, so I just detached the 12v line and removed all roms, the z80 and the 8910's from the board. From this moment on, it's not coining up anymore, but that might be the z80 removed, not sure but it looks like it generates a coin lock signal needed for coining up. So I need to fix this part.

Also need to fix the sound

And there is still a small issue with the graphics, the are 2 columns that are shifted one pixel line up. Really strange, and I don't really know where to start looking for this, so if anyone has a suggestion, let me know. Here's 2 pictures.

patzik2021-01-07 01:32:24

 

[patzik]

I looked into the schematic relating to the coining up, it's quite a different way that I am used to. It has 13v into it, I am guessing for the coin meter, but it also generates logic signals. For player one there is a jumper to configure, and I think this was configured wrong on my board, it had the jumper for 'use counter pc board' but the other trace wasn't cut.
I don't really understand how this part works, maybe someone can explain it to me. The coin signal comes from ic28, I get a high input on pin5 when coining up, but not at pin2. As I see it, both are needed.

I was also not getting the 'coin lock' signal, so I put back the z80 and roms, but this made no difference. So I started looking at that, and quickly found there is no clock. Now I don't know if something happened or if it was like that when I started working on it.

Here's the related part of the schematic I also looked at the p2 coin part and this is also not a high on pin 12 of ic28. How does this part work?? 13v some diodes some resistors and a transistor???

patzik2021-01-07 01:51:41

 

[Lurch666]

Have you checked the dip switches?

Dip switch C have settings to turn 'coin display' off and coinage to one or two way.

Don't know if that makes any difference but worth a look.

Is pin 4 of IC28 low?

 

[patzik]

Don't really know what happened, but as said I removed the z80 ay8910's and roms. This might have caused it because I noticed it said 9 coins for 1 credit. Maybe it was coining up I don't know.
That said, I put back in these ic's and cut the jumper. The board was strangely configured, it had the jumper from C to 2, but the trace from C to 1 was not cut. I think this is what caused the smoke, I think that big 6ohm resistor got hot. What is very strange is that with the board configured this way pressing coin you see the picture get just a little bit dimmer, like the 5v is also drawn down or something, but what I can't explain is that those 2 columns that are 1 pixel line high also go to the correct spot! Really weird!

With the board configured as default I can't coin it up, there is no high signal coming into pin2 of ic28. With the board configured with the jumper and the trace cut (use counter drive pcb) it works fine.

The 9 coins for 1 credit was due to the 8910 being removed so it works ok now, and coins up.

Now for the sound, and that small graphics glitch. Working on the sound first. I get no sound at all, what I have done is put the arduino ict in the z80 and verified the 2 roms, which are ok, so I assume the sound data bus and sound address bus are also ok. I couldn't find the memory range, so I tested both 2114 rams out of circuit and they are ok too. I took out the 4 8910's, and tested these on another board. 3 worked fine, 1 had no sound so I replaced this one. The clock is fixed and the z80 is also tested.

Yet no sound at all. The only think I noticed is that there is a 8MHz crystal on the board, and the schematic shows a 6MHz crystal. Otherwise I see no stuck signal on data or address lines.

I find this really hard to troubleshoot, the sound is like a separate computer, but without a picture it is a bit like trying to fix it in the dark.

Oh and the amp is working find, I get hum when touching it. I probed the ABC lines of the 8910's and they look to be doing nothing.

 

[patzik]

Just noticed ic14 is missing! Anyone know what kind of roms this is, and if this could be the cause?

Never mind this is the rom that is jumpered according the wild western to elevator action conversion.

But I did realize that I did not check the other rom. What kind of rom is this?

patzik2021-01-07 18:41:52

 

[qjuk]

Have you checked the 3x LM3900 at IC1, IC8 and IC17? I had to replace one of these on one of my boards to fix a partial missing sound problem.

IC14 is normally fitted with a Fujitsu MB14257. It’s present on both of my EA conversion PCB’s, but as you previously said, I don’t think EA uses it.

 

[patzik]

Found this on another forum... So I think it is used, my board is jumper hence the missing rom.

From the conversion manual of WW to EA.

9. a. Install the Black Jumper Module "26-00016-001" in socket number IC14 on the GAME board.

Most Taito SJ series boards don't have a socket or a chip here. There are hard-wired jumpers that connect pins in the following manner:

J1: connects pin 2 to pin 19
J2: connects pin 3 to pin 17
J3: connects pin 10 to pin 18
J4: connects pin 10 to pin 14

This socket is marked WW16 ROM3 in the WW manual.

 

[patzik]

I did look at those LM3900, but didn't see anything on the inputs. I looked at the ABC (which is the sound) of the 8910's and there is nothing there. But I did get a bit further. replacing that one 8910 that was broken did give me some sound back, I just didn't hear it because of a bad pot. Don't you just hate multiple faults, in the same circuit. This 8910 was ic63 and is responsible for the background music during the game, so this is back now. I could not find anywhere what sounds the 4 8910's make so this will be a nice reference for future repairers.
I did dive into the workings of the 8910, simply said there is a shared data/address bus, bc1 and bdir tell the 8910 what gets written on the bus or read from the bus. When I look at ic63 I see no activity on bc1 and bdir until the music starts, so this makes sense. The thing different with this ic63 is that it get these 2 signals from a different place then the other 3, so my guess is now this is where to problem lies.

I looked at these 2 signal on the other 3 8910's and they don't make sense, even with sound there is activity, but it looks like pulses of pulses, just strange. It might be my scope so I will connect a logic analyser to see what's actually on it.

The bc1 and bdir of these 3 8910's come from ic47, a 74ls153 and I don't have this in my spares so I can't swap it right now. Inputs are AB0, /CS5, /SRD, /SWR the last 2 are inverted by ic58.

All 4 signals come from the sound address bus and AB0 /SRD /SWR are also used to address the sound ram so I guess these are fine (as I have sound from that 1 8910 and the z80 is not crashing as far as I can see). /CS5 comes from ic60 (74ls138) which has inputs AB11, AB12, AB13, AB15, AB16 and MREQ from the sound address bus. But I think there might be a mistake here cause as far as I can see there is no AB16.

edit, I checked out AB16, and it is indeed a typo, it's AB14.

patzik2021-01-09 16:07:22

 

[patzik]

Put in a new ic60 just to try and see if this is where the problem was and no difference. The while doing some troubleshooting, I removed the z80, and I still had music... Looked at the schematic again, and saw my mistake. ic63 (the 'working' 8910) is not on the sound data bus, but on the data bus. This is the bus that is running the game, so that explains why it is working.
So back to the start I guess...

What I find strange is that there is activity on bc1 and bdir of the 3 8910's that produce no sound, even when there should be no sound, this should not be like this. What also is strange that with this activity, there is no sound, I would expect noise or other sounds. Anyhow, code is not running properly that is for sure. As said before I put the arduino ict on the z80 of this board and was able to verify the 2 sound roms, so some part of it is ok..

patzik2021-01-08 15:19:59

 

[patzik]

I did find the memory range for the ram on the sound board, it’s 0x4000-0x43FF
I tested it with the arduino ict and it tests ok.

patzik2021-01-08 16:30:46

 

[patzik]

Can't find what's wrong with the audio, and today powered up the board and get this...
Guess it's one step back again

 

[patzik]

Well. several hours later tracing signals (mostly on the video pcb) and even replacing an ic I finally got the colours back and correct.
It was looking like green was missing, but there was white, so that needs green too. Anyhow after I was probing ic67 on the game pcb all came back to normal. Thinking back, this should have been the first place to look. This memory ic is 9 bits and outputs 3 bits per color, so kinda makes sense now, I think it was just a bad contact in the socket...

Back to the sound, hoping someone can help me out. I'm at a point that I just don't know where to look anymore.

 

[Georgian2]

Did you find a solution to this sound problem? I converted an jungle king to time tunnel. All the sounds from sound CPU are missing...