Star Rider, Discan system (TM) from Williams-The game that does laserdisc as only Williams could....
Program crashed mid game.
Startup tests pass and goes to high score screen
On trying to display disc image for demo mode or game start screen goes black then various unhelpful disc error numbers appear (505/507/520/521).
Test mode disc test shows reset and walking bit test pass. Search to and step disc fail. Which suggest the PIF board is working and the manchester decoder is not.
One could observe when it issued a seek to frame that it was in fact showing the right image on the screen but frame number is reported as '??????' insted of a decimal number.
I was pretty sure the fault was on the VGG pcb but I swapped it out with another that had other issues but a working decoder to confirm.
I rather cheekily asked Matt Ownby if he had any thoughts since he has done so much work on this in the last few years and using a Dexter pcb and the very detailed information he has provided on 'mycoolprojects' and 'star rider notes' on the wiki at daphne-emu helped a great deal although I've yet to get round to compiling and burning the helpful looking test routines to an Eprom.
So, as suggested I started by looking at U85 which converts the NTSC signal to a digital value for the decoder. I didn't fancy trying 'slice' on this device as it had an analogue input so I stuck a logic probe on pin 6 and pin 2 to confirm it was doing something although it didn't allow me to check for the correct timing delay but I could see it was active a prod further down the line at a logic gate suggested the values were different sometimes, but not all the time, which is good.
I thought it was time I got familiar with the wonders of 'slice' and since the board was largely running/active I thought it might be a good option so I catalogued pretty much all of the logic on page 5 of the schematics, which is the manchester docoder and then tested them, there were a few suspect devices (probably I could have got them to pass with a little more practice) but essentially I ended up replacing the 161 timers at U67-69 inclusive (my HP logic comparator didn't like them either) and an LS02 and an LS164. All of this work and I fired it up and no change... it didn't suprise me since all those devices tested fine out of circuit in a 'minipro'
I then replaced U85 on a whim and since this did nada I re-read some of Matts wiki and went in to page 4 of the VGG schematics, specifically the timing section on the top half of the page. I must have been getting more proficient at using slice by then because I got most things to pass or at least look very likely to work expect U93 (LS32). It kept failing on pin 11 so I disabled it from the test and left it sampling an average for several 100 cycles at the default sample period of 10mS and a few samples after at much longer sample periods. I then discovered the delights of the raw data view on slice and the ability to press 'E' to see the next error, etc and it all looked pretty black and white. I then tried the logic probe on pin 12 and 13 and they were pretty busy whilst pin 11 was stuck. The chip was pulled and tested bad, I replaced a tested new chip fired up the game and the seek to test passed. Then on step disc and the subsequent test that allows you to manually play or select a frame it failed (grrr). That said it was sporadically reading the frame numbers and was some sort of improvement but since this was a 'worker' it seemd less likely there were multiple faulty devices unless there'd been a mains spike or something.
Now, it had occured to me that from Matts wiki notes the timing on U85 is imporant as it's basically a hack and since the supply of the original chip was not that cheap and they all look old and nasty I'd replaced the MC14049U(BCP) with a CD4049UBE, I'd also relocated the aparently undocumented capacitor that goes between pin 3 and 8 to the underside of the board so I could fit a socket and not cook any new chips soldering directly to the top of them.
Anyhooo, I left the capacitor under the board but changed back to a tired looking MC14049U that I got off ebay ages ago (so tired I had to sand the corrosion off the legs) and fitted that and bobs your fathers brother we now have gameplay again. Hurrah.
If there is a lesson to be learned I should have paid more specific attention to the signals coming into the manchester decoder from the rest of the pcb. Incidentally, slice did a rather excellent job of calculating the timing frequencies of the numerous clocks in the timing cicruit and was close enough to book values (which probably aren't spot on) for me to carry on. On a related note If anyone has anywhere to share slice projects in addition to those which gadgetfreak has already helpfully supplied I have a big chunk of the VGG pcb done and that's basically the most complicated board in the stack. My next repair will probably be a star wars so I think the sample projects help there but after that maybe Space Harrier, so if anyone has catalogued that it'd be greeeat
karlcdoe2022-07-29 18:12:37
Program crashed mid game.
Startup tests pass and goes to high score screen
On trying to display disc image for demo mode or game start screen goes black then various unhelpful disc error numbers appear (505/507/520/521).
Test mode disc test shows reset and walking bit test pass. Search to and step disc fail. Which suggest the PIF board is working and the manchester decoder is not.
One could observe when it issued a seek to frame that it was in fact showing the right image on the screen but frame number is reported as '??????' insted of a decimal number.
I was pretty sure the fault was on the VGG pcb but I swapped it out with another that had other issues but a working decoder to confirm.
I rather cheekily asked Matt Ownby if he had any thoughts since he has done so much work on this in the last few years and using a Dexter pcb and the very detailed information he has provided on 'mycoolprojects' and 'star rider notes' on the wiki at daphne-emu helped a great deal although I've yet to get round to compiling and burning the helpful looking test routines to an Eprom.
So, as suggested I started by looking at U85 which converts the NTSC signal to a digital value for the decoder. I didn't fancy trying 'slice' on this device as it had an analogue input so I stuck a logic probe on pin 6 and pin 2 to confirm it was doing something although it didn't allow me to check for the correct timing delay but I could see it was active a prod further down the line at a logic gate suggested the values were different sometimes, but not all the time, which is good.
I thought it was time I got familiar with the wonders of 'slice' and since the board was largely running/active I thought it might be a good option so I catalogued pretty much all of the logic on page 5 of the schematics, which is the manchester docoder and then tested them, there were a few suspect devices (probably I could have got them to pass with a little more practice) but essentially I ended up replacing the 161 timers at U67-69 inclusive (my HP logic comparator didn't like them either) and an LS02 and an LS164. All of this work and I fired it up and no change... it didn't suprise me since all those devices tested fine out of circuit in a 'minipro'
I then replaced U85 on a whim and since this did nada I re-read some of Matts wiki and went in to page 4 of the VGG schematics, specifically the timing section on the top half of the page. I must have been getting more proficient at using slice by then because I got most things to pass or at least look very likely to work expect U93 (LS32). It kept failing on pin 11 so I disabled it from the test and left it sampling an average for several 100 cycles at the default sample period of 10mS and a few samples after at much longer sample periods. I then discovered the delights of the raw data view on slice and the ability to press 'E' to see the next error, etc and it all looked pretty black and white. I then tried the logic probe on pin 12 and 13 and they were pretty busy whilst pin 11 was stuck. The chip was pulled and tested bad, I replaced a tested new chip fired up the game and the seek to test passed. Then on step disc and the subsequent test that allows you to manually play or select a frame it failed (grrr). That said it was sporadically reading the frame numbers and was some sort of improvement but since this was a 'worker' it seemd less likely there were multiple faulty devices unless there'd been a mains spike or something.
Now, it had occured to me that from Matts wiki notes the timing on U85 is imporant as it's basically a hack and since the supply of the original chip was not that cheap and they all look old and nasty I'd replaced the MC14049U(BCP) with a CD4049UBE, I'd also relocated the aparently undocumented capacitor that goes between pin 3 and 8 to the underside of the board so I could fit a socket and not cook any new chips soldering directly to the top of them.
Anyhooo, I left the capacitor under the board but changed back to a tired looking MC14049U that I got off ebay ages ago (so tired I had to sand the corrosion off the legs) and fitted that and bobs your fathers brother we now have gameplay again. Hurrah.
If there is a lesson to be learned I should have paid more specific attention to the signals coming into the manchester decoder from the rest of the pcb. Incidentally, slice did a rather excellent job of calculating the timing frequencies of the numerous clocks in the timing cicruit and was close enough to book values (which probably aren't spot on) for me to carry on. On a related note If anyone has anywhere to share slice projects in addition to those which gadgetfreak has already helpfully supplied I have a big chunk of the VGG pcb done and that's basically the most complicated board in the stack. My next repair will probably be a star wars so I think the sample projects help there but after that maybe Space Harrier, so if anyone has catalogued that it'd be greeeat
karlcdoe2022-07-29 18:12:37