Cheese Chase: Fixed with pics
- Thread starter guddler
- Start date
it's not the DAC on my one
Just found a Slick Shot, which has a Mosel MS176-50, so working on Muppz assumption that they were all pin compatible back then, put it in - and no different (slower part, but I guess it should do something!)
put the WM70C178 into the slick shot, and it's happily displaying a screen (ok, it says 'cue ball missing', but apart from that .... ) - aha, stick it into test mode and I now have a pool table displayed, so I would say it works
Just found a Slick Shot, which has a Mosel MS176-50, so working on Muppz assumption that they were all pin compatible back then, put it in - and no different (slower part, but I guess it should do something!)
put the WM70C178 into the slick shot, and it's happily displaying a screen (ok, it says 'cue ball missing', but apart from that .... ) - aha, stick it into test mode and I now have a pool table displayed, so I would say it works
OK then - just to tie up this and turn it from a bit of a waffle fest into a proper fix log, I'm now going to stick up a few posts showing the reasoning behind the diagnosis. It's not a bad example of how a scope can be helpful at the same time!
So to summarise, the original fault was that the screen was mostly a washed out white but the game could just about be made out playing away in the background and indeed you could start a game and so on. There was another fault in that at the start of the game instead of speech you got white noise but I decided to tackle the screen issues first.
There are no schematics for this game but I managed to get hold of one of the chaps that worked on it and he gave some good tips suggesting that if there was no sync or video then it was likely the GFX chip, a TMS34010.
So we start there. I checked the video signals coming out of the GFX chip:
HSYNC - a little out, but nothing you wouldn't be able to adjust out on the monitor, or could even be my cheap digital scope
VSYNC - the same. Note 50Hz (PAL). Unusual for an arcade game, but fine.
And video blank - impossible to know if it's right or not but theres certainly plenty of activity:
I confirmed that the BLANK signal seemed to be doing it's job as you'd expect because grounding it made the screen go black (permanently blanked). OK, I think that made sense anyway!
guddler2013-10-01 22:53:51
So to summarise, the original fault was that the screen was mostly a washed out white but the game could just about be made out playing away in the background and indeed you could start a game and so on. There was another fault in that at the start of the game instead of speech you got white noise but I decided to tackle the screen issues first.
There are no schematics for this game but I managed to get hold of one of the chaps that worked on it and he gave some good tips suggesting that if there was no sync or video then it was likely the GFX chip, a TMS34010.
So we start there. I checked the video signals coming out of the GFX chip:
HSYNC - a little out, but nothing you wouldn't be able to adjust out on the monitor, or could even be my cheap digital scope
VSYNC - the same. Note 50Hz (PAL). Unusual for an arcade game, but fine.
And video blank - impossible to know if it's right or not but theres certainly plenty of activity:
I confirmed that the BLANK signal seemed to be doing it's job as you'd expect because grounding it made the screen go black (permanently blanked). OK, I think that made sense anyway!
guddler2013-10-01 22:53:51
So from that then I assumed the GFX was OK. So i spent a while buzzing out where signals went - which was a bit of a nightmare on this board as it's 4 layer and has vias all over the shop. Eventually I worked out that BLANK was going to a 28 pin IC down near the edge connector. See earlier for how that panned out. At this point I was suspecting something was up with BLANK.
Eventually then we got the pinout for what we knew now was a DAC and I checked the signals:
Red (Green / Blue just the same):
Now RGB signals aren't usually that high of a voltage but that doesn't look right to me - I'd slowed it down quite a lot on the scope to see that it was fairly uniform. Given the game was running and all three colour outputs were doing the same there was no way that's right.
So I checked /RD and /WR on the DAC too and look what we have. A whole lot of nothing much...
/RD:
/WR:
At this point I wanted to try and work out if the DAC was holding the /WR & /RD signals high or if they just weren't doing anything. So I removed the IC. Problem is, without the DAC the game very quickly gives up and locks up. I guess that's because it has a colour pallette in it as well so the CPU knows it's not working.
Being this is a DSO however I was able to set "Acquire" to single shot and trigger to 50%. Then I was able to capture what happens when the game first boots (you could also do this with a logic probe too)
Looky looky:
It's not very scientific, but to me I believe it does at least show that the signal can go fully low without the IC so there's a good chance the IC is shorted internally and holding it high.
Remember no working PCB to compare to and no schematics so this is all guess work to a certain extent. I suppose educated guess work from experience though.
Eventually then we got the pinout for what we knew now was a DAC and I checked the signals:
Red (Green / Blue just the same):
Now RGB signals aren't usually that high of a voltage but that doesn't look right to me - I'd slowed it down quite a lot on the scope to see that it was fairly uniform. Given the game was running and all three colour outputs were doing the same there was no way that's right.
So I checked /RD and /WR on the DAC too and look what we have. A whole lot of nothing much...
/RD:
/WR:
At this point I wanted to try and work out if the DAC was holding the /WR & /RD signals high or if they just weren't doing anything. So I removed the IC. Problem is, without the DAC the game very quickly gives up and locks up. I guess that's because it has a colour pallette in it as well so the CPU knows it's not working.
Being this is a DSO however I was able to set "Acquire" to single shot and trigger to 50%. Then I was able to capture what happens when the game first boots (you could also do this with a logic probe too)
Looky looky:
It's not very scientific, but to me I believe it does at least show that the signal can go fully low without the IC so there's a good chance the IC is shorted internally and holding it high.
Remember no working PCB to compare to and no schematics so this is all guess work to a certain extent. I suppose educated guess work from experience though.
So finally we arrive at tonight and a replacement DAC. Obviously at this point it's a simple case of popping the new chip in and bingo the video is now properly working. Oddly enough the speech issue is now sorted too. I don't really get that but at the end of the day, *SHRUG* !!
Just to confirm what these signals SHOULD look like:
Red (Green / Blue similar):
This was in the test screens with a plain red screen which I think is why it also happens to be around the same frequency as HBLANK. If you think about it the red beam will be on for all of the time apart from the blanking period. Either that or I've mixed my screenshots up
Don't think so though as the voltage is only 2.76v
Finally, the /WR signal:
Much more like it!
Note that I did work out that on this board there never seems to be a huge amount of activity on /RD, if any and that there are long periods with no activity on /WR. Basically /WR is active whenever the colour palette is changed which appears to be done completely rather than in small bits. You get a load of activity at the start of attract mode and then again at the start of a game level. That's when the entire colour scheme changes on the screen. And it then follows that the CPU never really needs to read the colour palette theres not really any need for it to be anything other than a write.
There we are. My first proper fix log write up for the VAC site - hope someone enjoys it
Just to confirm what these signals SHOULD look like:
Red (Green / Blue similar):
This was in the test screens with a plain red screen which I think is why it also happens to be around the same frequency as HBLANK. If you think about it the red beam will be on for all of the time apart from the blanking period. Either that or I've mixed my screenshots up
Finally, the /WR signal:
Much more like it!
Note that I did work out that on this board there never seems to be a huge amount of activity on /RD, if any and that there are long periods with no activity on /WR. Basically /WR is active whenever the colour palette is changed which appears to be done completely rather than in small bits. You get a load of activity at the start of attract mode and then again at the start of a game level. That's when the entire colour scheme changes on the screen. And it then follows that the CPU never really needs to read the colour palette theres not really any need for it to be anything other than a write.
There we are. My first proper fix log write up for the VAC site - hope someone enjoys it
It's a DS1000E. Nice write up on it here: http://geoffg.net/DS1000E.html
If anyone is that interested then I can dig out the invoice from the place I bought it. Of particular note is the fact that it's still possible to first downgrade the firmware, then hack it so that the 50Mhz version becomes the 100Mhz version as they're both identical from a hardware point of view.
Couple of negatives to be aware of: Quite a bit of noise on the signals. This may be my Jamma rig though. You can see it in the screen shots. The shots of the WR and RD signals in particular. For what we're using them for though I can't see it as anything more than a simple annoyance. If you were an electrical designer engineer or something it would be a problem. But then you'd be spending £1000+, not £300.
Other thing is when I tried to copy the screenshots from internal memory to USB last night it got to the 7th and corrupted the FAT32 partition, thus losing me several 1000 photos that were on the stick. Oops. My bad for not using a clean memory stick. I ended up connecting the scope to the PC by USB and using windows to screen grab from the remote control software after first recalling each shot from the internal flash memory. Bit tedious but it worked. I'll be looking into this as there must be a better way!
If anyone is that interested then I can dig out the invoice from the place I bought it. Of particular note is the fact that it's still possible to first downgrade the firmware, then hack it so that the 50Mhz version becomes the 100Mhz version as they're both identical from a hardware point of view.
Couple of negatives to be aware of: Quite a bit of noise on the signals. This may be my Jamma rig though. You can see it in the screen shots. The shots of the WR and RD signals in particular. For what we're using them for though I can't see it as anything more than a simple annoyance. If you were an electrical designer engineer or something it would be a problem. But then you'd be spending £1000+, not £300.
Other thing is when I tried to copy the screenshots from internal memory to USB last night it got to the 7th and corrupted the FAT32 partition, thus losing me several 1000 photos that were on the stick. Oops. My bad for not using a clean memory stick. I ended up connecting the scope to the PC by USB and using windows to screen grab from the remote control software after first recalling each shot from the internal flash memory. Bit tedious but it worked. I'll be looking into this as there must be a better way!
Hmm, £270 on ebay (and about the same from their web site)
think I may need to do some fund raising
LOHT or Pulstar MVS anyone (should cover it ...)
oh, and I have a program that should recover the pics from the memory stick if you need them!
Macro2013-10-02 11:23:15
think I may need to do some fund raising
LOHT or Pulstar MVS anyone (should cover it ...)
oh, and I have a program that should recover the pics from the memory stick if you need them!
Macro2013-10-02 11:23:15
yeah, would be nice please. It was a load of pics of me flying my heli where a mate of mine had the camera on fast shoot for the whole flight! So would be nice not to lose - I didn't realise I had them in the first place
You can use my email if needed which you should still have. If not, it's martin@ <mydomain>
You can use my email if needed which you should still have. If not, it's martin@ <mydomain>
well, sort of a postscript to this thread, I decided I liked Guddlers
'scope, but on having a look at them discovered they also do one with a
16 channel logic analyser built in ...
so, 2 channel 'scope and 16 channel LA in one (small) box - got to be a winner!
'scope, but on having a look at them discovered they also do one with a
16 channel logic analyser built in ...
so, 2 channel 'scope and 16 channel LA in one (small) box - got to be a winner!
NICE!!! You will have to let us know how you get on with it.
I deliberated over that for quite some time as I was buying a logic analyser as well at the time. In the end I bought a separate USB based one (ZeroPlus LAPC) and I think for me that has been the right choice. I've used the logic analyser mostly down in the house while working on Arduino projects and stuff like that, hooked up to the PC, whereas the scope lives on the test bench which is up in the shed.
Oh, the other thing that swayed it for me was all the bus protocol analysers that the USB one came with, but as on now I've not used any of them!!
I deliberated over that for quite some time as I was buying a logic analyser as well at the time. In the end I bought a separate USB based one (ZeroPlus LAPC) and I think for me that has been the right choice. I've used the logic analyser mostly down in the house while working on Arduino projects and stuff like that, hooked up to the PC, whereas the scope lives on the test bench which is up in the shed.
Oh, the other thing that swayed it for me was all the bus protocol analysers that the USB one came with, but as on now I've not used any of them!!
guddler said:
I almost contrived a situation a week ago where I'd be able to use my RS232 proto decoder
The bit I really want (68000 disassembler) won't work with the software I have - no option to add my own decoder so it's save the trace and post-process, instead of being able to hover over a signal group capture and get the opcode! That would be sweet