Phantom drive?
in [Commodore]
|
Prev: Starcraft for C64
Next: Breadbox vs C64c - reliability?
From: Paul Förster on 13 Nov 2009 11:34 Hi Peter, On 2009-11-13 04:09:29 +0100, schepers(a)ist.uwaterloo.ca (Peter Schepers) said: > Forget what I wrote before, don't know what I was thinking. > > Try measuring your serial port readings without and with a device > connected as follows: [...] .... I've spent some time measuring all sorts of connections but no new findings so far. :-( I will have to solder everything in again to try out what you proposed. I'll do so tomorrow or Sunday. I'm not even sure I'm looking for shorts of any kind anymore as *all* I check is the same on Ruud's dead board, except for the 300 Ohm resistence difference between anode and cathode of CR14. Could it possibly be a bad PLA (U8)? If that would be the case, then I would have a replacement PLA which I desoldered from Ruud's dead board. I don't know though, if that one is good either. :-/ But today, I bought those 64 pin sockets. So in case I'd really have to desolder the PLA, I can put a socket under it and try the other one out as well. Also, when I finally get the board *fully* working again, including all ports and stuff, I'll reveal what I was up to, what I modified and why I modified it -- promised! Then you all will understand why I go to such great lengths to get this specific board fully working again. :-) I even bought such an expensive tool as a Weller WDD-161V plus an oil-free compressor to get things straight. Knowing that and knowing how much money that took already should show how much I am interested to get it fully working. And once I reveal what has happened if everything works out then you definitely *will* call me crazy, no matter if that is a precondition in this group or not. :-P I already "hear" people say "oh my god! I would never have done that!" So far, thanks to all of you who helped me in the past and will do so in the future to get this working. Your help has already given me a lot of valuable insights. -- cul8er Paul paul.foerster(a)gmx.net
From: Peter Schepers on 13 Nov 2009 12:54 In article <7m5g4vF3gh6h1U1(a)mid.individual.net>, Paul F�rster <paul.foerster(a)gmx.net> wrote: >Hi Peter, > >On 2009-11-13 04:09:29 +0100, schepers(a)ist.uwaterloo.ca (Peter Schepers) said: >> Forget what I wrote before, don't know what I was thinking. >> >> Try measuring your serial port readings without and with a device >> connected as follows: >[...] > >... I've spent some time measuring all sorts of connections but no new >findings so far. :-( I will have to solder everything in again to try >out what you proposed. I'll do so tomorrow or Sunday. I'm not even sure >I'm looking for shorts of any kind anymore as *all* I check is the same >on Ruud's dead board, except for the 300 Ohm resistence difference >between anode and cathode of CR14. I've been wondering if were in the right area. When the C64 goes to load from a device, first it has to check that the device exists by sending a command on the bus and waiting for a response. I don't think the signal level of one serial bus comm line can cause this problem. >Could it possibly be a bad PLA (U8)? If that would be the case, then I >would have a replacement PLA which I desoldered from Ruud's dead board. >I don't know though, if that one is good either. :-/ But today, I >bought those 64 pin sockets. So in case I'd really have to desolder the >PLA, I can put a socket under it and try the other one out as well. I don't know. I thought the PLA basically controlled the memory map of the computer. Maybe it's not enabling U2 when needed (see pin 23 of U2) so the kernal can't talk to the CIA and it's waiting for a response. That check is part of the testing sequence I sent. PS
From: Peter Schepers on 13 Nov 2009 18:49 In article <7m5g4vF3gh6h1U1(a)mid.individual.net>, Paul F�rster <paul.foerster(a)gmx.net> wrote: >Hi Peter, > >On 2009-11-13 04:09:29 +0100, schepers(a)ist.uwaterloo.ca (Peter Schepers) said: >> Forget what I wrote before, don't know what I was thinking. >> >> Try measuring your serial port readings without and with a device >> connected as follows: >[...] > >Could it possibly be a bad PLA (U8)? If that would be the case, then I >would have a replacement PLA which I desoldered from Ruud's dead board. >I don't know though, if that one is good either. :-/ But today, I >bought those 64 pin sockets. So in case I'd really have to desolder the >PLA, I can put a socket under it and try the other one out as well. I didn't realize you were working with one of the newer boards, with the large 64-pin decoder chip. I've not seen one of those go bad, but the CIA2 line comes directly from it. It's a bear to desolder because of the fine pins. I just tried using the C64 without U2 to see if that would work, but no I still get the "device not found" error. PS.
From: Peter Schepers on 14 Nov 2009 07:27 In article <7m5g4vF3gh6h1U1(a)mid.individual.net>, Paul F�rster <paul.foerster(a)gmx.net> wrote: > >... I've spent some time measuring all sorts of connections but no new >findings so far. :-( I will have to solder everything in again to try >out what you proposed. I'll do so tomorrow or Sunday. I'm not even sure >I'm looking for shorts of any kind anymore as *all* I check is the same >on Ruud's dead board, except for the 300 Ohm resistence difference >between anode and cathode of CR14. This is on the SRQIN line which is not used by the serial port. I don't think this has anything to do with it. >Could it possibly be a bad PLA (U8)? I tried simulating a bad PLA with the CIA2 line missing (or not working) to CIA2 (U2) by bending out pin 23 on U2, but it didn't make the error either. However, I found a detailed doc online that explained the IEC protocol very well. Reading it for a while steered me in my original direction. In order to get the "Searching for..." error there had to be an IEC protocol failure. The only way that happens is if one of the incoming communication lines (CLK or DATA) is already in the wrong state (active low, so not high). It was a comment from your original post that I re-read last night that tipped me off: > ... I was just curious and checked RP6, the 1k Ohm SIL resistor array. > I found that pins 4 and 6 show a resistive value of about 670 Ohm while > pins 2, 3 and 5 showed the proper value of 1k Ohm. I compared that to a > good board where all 5 pins show a value of 1k Ohm. Those two pull-ups for the DATA and CLK lines should not measure lower than 1K unless something is paralleling them. I think this is the problem. So I simulated it last night by shorting pins on the serial port. I shorted GND (pin 2) and CLK (pin 4), but only got a "Device not found" error. Then I shorted GND and DATA (pin 5) and now the C64 will hang at the "Searching for...". Note at this point that I cannot RUN/STOP out of this, and I also had to run with the standard kernal, not JiffyDOS, or this simulation didn't work. As I mentioned previously, I've seen this before. It's been a while since I've seen a C64C board, but I believe they are more that 2 layer boards, and the traces for DATA and possibly CLK have become somewhat shorted with something else like a GND or Vcc plane, and likely just between the 7406 and the port. My simple resolution years ago was to cut the traces for both lines from the CIA2/7406/RP6 to the serial port and run new jumpers. Your idea to run lines from bent-out pins from CIA2 to an external board with a 7406 and pull-ups would likely work fine. Did you do this? You also mentioned that the edge connector for the user port were cut off. Note that the ATN line goes there (to pin 9), but CLK and DATA do not. I hope it didn't get affected (shorted) buy the cutting. PS.
From: Paul Förster on 15 Nov 2009 12:28
Hi Peter, On 2009-11-13 04:09:29 +0100, schepers(a)ist.uwaterloo.ca (Peter Schepers) said: > Try measuring your serial port readings without and with a device > connected as follows: > > Pin Upluggd Pluggd/off Pluggd/on > --------- ------- ---------- --------- > 1 (SRQIN) High Float High > 2 (GND) Low Low Low > 3 (ATN) High Float High > 4 (CLK) Low Low Low > 5 (DATA) High High High > 6 (RESET) High Float High .... first of all the good news first: *Success*! Yippie! It's working again. :-) What has happened: I printed your three posts and started checking. To start checking, I had to solder everything in again, as I have mentioned before. After I did that, I found that I didn't receive any sensible information at all at the serial port pins. So I decided to recheck my putting a socket under the chips, starting with CIA2 and then going forward to U21, U22 and U3. When I decided to put sockets under all chips (except U18 and U23 for space reasons), I had to learn. I started with the character ROM (ok, bad decision...). Remember when I asked about the ideal temp to desolder chips? There was a background to that question of course... Anyway, I had difficulties getting Kernel/Basic and the character ROM out successfully. I accidentally made some pins a little too hot and so a few traces came out with the chip. :-( I measured everything through according to schematics and put wires where I accidentally ripped off the traces. I measured everything through and all was good -- so far... Since the "power supply" is mounted on the corner where CIA2 is located, I decided to leave that chip in until I have more experience getting chips off boards. The reason is that that corner is not easy to work on because of an ugly cap and a voltage regulator sitting there halfway mounted on top of CIA2 and the CHAR ROM. Some time later, after practicing desoldering chips I decided to try to get CIA2 socketed anyway. The funny thing is, I didn't rip anything off but I seem to have damaged the connections D4-D7 (CIA2 pins 26-29) somehow. Is that a multi-layer board and I hit some invisible connections in another layer? Was multi-layer possible back when Commodore printed these boards? So I decided to re-check *every* pin of CIA2 and the neighboring CHAR ROM thoroughly again. As it turned out, D4-D7 of the CHAR ROM was fine but had no connection to CIA2, which according to plan, it should. So I put four small wires in and soldered the rest in again (C88, C94, the diodes, RP3 & 6 and sockets for U21, U22 and U3). Btw., ceramic caps don't have + or -, only electrolytic capacitors do, right? Since I didn't know + or - of the ceramic lenses I put them in with the inscription in the direction as they are on the other board. I checked everything and tried LOAD without a drive connected. A DEVICE NOT PRESENT error occured as should have. That made me confident and I checked the first column of your table with the drive unplugged. Signals appeared as you listed them in column 1. Ok, not all. Pin 1 showed strange values. But you recall that I mentioned the board has been cut. The connection between pin 1 and the diode has been cut. So I verified it with Ruud's board and put a wire there too, just for the sake of completeness, even though SRQIN (what does SRQ stand for anyway?) is not used. That gave me a high value on pin 1 too. I connected a drive. Checked it and everything works fine. Oh, before I forget, Peter, you mention in one of your other posts that you couldn't RUN/STOP out of the SEARCHING FOR... message. This was true here too. I could only RUN/STOP+RESTORE out of that message. I don't know if that's important but I thought I'd mention it. That's the story. Peter, a very big thank-you to you! :-) Your tips led me to the right place to check. Also special thanks for taking the time and undertaking the effort to try and simulate what would or could be wrong. Many thanks to Andreas Beermann too by whose advice and tips I learned a lot of things. And last but not least, many thanks to Ruud for providing a dead board which did supply a few replacement parts and which I could use to compare with my board. Also thanks very much for the CIA pin-bending-check tip. :-) That was a real tough one for an electronics newbie like me. But I've learned a lot. Thanks again. :-) I hope such a desaster never happens again because it already costed me 10 days which I would have liked to use for something more constructive. At least I have a very fine and big printout of the schematics, a new (de-)soldering station and quite invaluable knowledge now. :-P -- cul8er Paul paul.foerster(a)gmx.net |