Phantom drive?
in [Commodore]
|
Prev: Starcraft for C64
Next: Breadbox vs C64c - reliability?
From: Peter Schepers on 7 Nov 2009 16:43 In article <7lklk9F3e5fghU1(a)mid.individual.net>, Paul F�rster <paul.foerster(a)gmx.net> wrote: >Hi Bill, > >On 2009-11-06 19:33:16 +0100, "Garberstreet Electronics" ><willy46pa(a)comcast.net> said: >> I'm not a big Commodore guy, but a chip is a chip, and I'd suspect >> that there is a short preventing the 7406 from functioning properly, >> or it is blowing the chip as soon as it is powered up. At least he >> has narrowed this down to a single location. Now to find what it is >> that's causing it. > >... well, I think, with Ruud's tip about the CIA -> serial port pin >test, I can narrow the problem even to the pins 3, 4, 8, 9, 12 or 13 of >U22 because these pins would be the ones affected when bending the CIA >pins away. Funny enough, the DEVICE NOT PRESENT message appears >correctly if I bend *any* of the pins 5, 6, 7 or 9 of the CIA away. I vaguely remember dealing with one C64 that had a serial port issue. I found that one or more of the traces between the serial port and the 7406 were bad, either shorted together or to something else. I believe I ended up x-acto'ing the existing traces out and running jumpers between the port and the 7406. The only way I found out what was wrong was by making sure each pin on the 7406 had the correct voltage, and was pinned directly to the serial port and the 6522. All the data lines are being pulled up by 1K pull-up resistors, and one connects directly to the 6522. >Still, the remark about blowing the 7406 when powering up is very >important. Let's just hope that this is not the case. TTL chips (74xx series) are very hardy. Their pins can take shorts to both GND and Vcc. It's only Vcc over-voltages that can hurt, but that would be affecting most of the chips in the machine. PS.
From: Paul Förster on 8 Nov 2009 04:15 Hi Peter, On 2009-11-07 22:43:39 +0100, schepers(a)ist.uwaterloo.ca (Peter Schepers) said: > I vaguely remember dealing with one C64 that had a serial port issue. I > found that one or more of the traces between the serial port and the 7406 > were bad, either shorted together or to something else. I believe I ended > up x-acto'ing the existing traces out and running jumpers between the port > and the 7406. > > The only way I found out what was wrong was by making sure each pin on the > 7406 had the correct voltage, and was pinned directly to the serial port > and the 6522. All the data lines are being pulled up by 1K pull-up > resistors, and one connects directly to the 6522. .... I was thinking about something along that line. I had the idea tonight when I realized how simple it should be according to the schematics. I was thinking about putting a 7406 on a small board along with 1K resistors (RP6) and the diodes CR9..17, wire that to CIA2, while disconnecting its pins 5-9 from the board and wiring them to the new extra chip and try again. If that works, then the fault is beyond the CIA. If it does not work, then the CIA itself might be an issue, tho I've alreay tried 3 different CIA chips. I could try that without having to bend CIA pins. The CIA sits in a socket so I can make an adaptor for it and not have to manipulate it physically -- and possibly damage it by breaking pins or something like this. What do you mean by correct voltages? It should be max. 5V in that whole area of the board. I have no oscilloscope so I can't analyze it in detail. I have a standard multimeter that can do resistence, voltage, capacity, amps, transistors, tubes and some other stuff. Would that be enough to check the voltages on the individual pins? I thought, since signals change so fast there is no way to measure the voltages other than Vcc and GND with a standard multimeter... hey, my electronics knowledge is still *BASIC* but it's growing slowly. :-) > TTL chips (74xx series) are very hardy. Their pins can take shorts to both > GND and Vcc. It's only Vcc over-voltages that can hurt, but that would be > affecting most of the chips in the machine. .... that's good news. So maybe the chips are not fried after all. They definitely shouldn't have to stand over-voltages because where in the whole system should be anywhere over the max. 12V (a little more to be precise) that comes from the 12V DC input power supply...? Yesterday, I also had the strange effect again that taking out the 7406 for checking purposes would not give me a picture *once* when powering on again. So I suspect a slack joint at the 4066 (U21) because they are next to each other and I touched it when removing the 7406. Strange enough, the picture immediately came back when powering it on the next time. Maybe U21 has a slack joint (too)? I'll check that... In any event, a slack joint on U21 shouldn't affect operation of U22 because they have nothing to do with each other according to schematics. Please allow me a dumb question: In the schematics the small boxes representing the inverter gates of U22 have a "1" printed inside them. What does that mean? -- cul8er Paul paul.foerster(a)gmx.net
From: Paul Förster on 8 Nov 2009 13:19 Hi, On 2009-11-08 15:52:33 +0100, Paul F�rster <paul.foerster(a)gmx.net> said: > ... I was just curious and checked RP6, the 1k Ohm SIL resistor array. .... hmm, it's not RP6. I've exchanged that with a definitely good one (tested before soldering in) from another board. It now shows the same odd behavior on pins 4 and 6. And the old one which I have soldered out shows proper 1k Ohm on *all* pins. It's interesting that a resistor changes its value when soldered in. Or that'd mean there is some parallel resistance which is lower (~670 Ohm)... -- cul8er Paul paul.foerster(a)gmx.net
From: Peter Schepers on 8 Nov 2009 21:54 In article <7lnghuF39k6rmU1(a)mid.individual.net>, Paul F�rster <paul.foerster(a)gmx.net> wrote: >Hi Peter, > >On 2009-11-07 22:43:39 +0100, schepers(a)ist.uwaterloo.ca (Peter Schepers) said: > >> I vaguely remember dealing with one C64 that had a serial port issue. I >> found that one or more of the traces between the serial port and the 7406 >> were bad, either shorted together or to something else. I believe I ended >> up x-acto'ing the existing traces out and running jumpers between the port >> and the 7406. >> >> The only way I found out what was wrong was by making sure each pin on the >> 7406 had the correct voltage, and was pinned directly to the serial port >> and the 6522. All the data lines are being pulled up by 1K pull-up >> resistors, and one connects directly to the 6522. > >... I was thinking about something along that line. I had the idea >tonight when I realized how simple it should be according to the >schematics. I was thinking about putting a 7406 on a small board along >with 1K resistors (RP6) and the diodes CR9..17, wire that to CIA2, >while disconnecting its pins 5-9 from the board and wiring them to the >new extra chip and try again. If that works, then the fault is beyond >the CIA. If it does not work, then the CIA itself might be an issue, >tho I've alreay tried 3 different CIA chips. It's a bit of work what you're proposing. You will need to wire the outputs from the 7406 to a serial port as well. You can eliminate the diodes from your test, though. Thinking about the diodes... have you checked all of them? They are there for over/undervoltage spikes. It's easy to verify they are OK with the diode checker on your multimeter. >I could try that without having to bend CIA pins. The CIA sits in a >socket so I can make an adaptor for it and not have to manipulate it >physically -- and possibly damage it by breaking pins or something like >this. With enough bending out, the pins will break off. Be careful. >What do you mean by correct voltages? It should be max. 5V in that >whole area of the board. I have no oscilloscope so I can't analyze it >in detail. I have a standard multimeter that can do resistence, >voltage, capacity, amps, transistors, tubes and some other stuff. Would >that be enough to check the voltages on the individual pins? What I mean is to make sure that the voltages on each pin seem right. 5V and GND are easy. The 7406 is an open collector inverter so it needs the 1k pull-ups for proper operation. If pin 1 (input) is low, pin 2 (output) will be high and vice-versa and for each inverter set (3 in/4 out, 5 in/6 out, 9 in/8 out, 11 in/10 out, 13 in/12 out). >I thought, >since signals change so fast there is no way to measure the voltages >other than Vcc and GND with a standard multimeter... hey, my >electronics knowledge is still *BASIC* but it's growing slowly. :-) If you're not using the port, the voltages on the pins will not be fluctuating. >> TTL chips (74xx series) are very hardy. Their pins can take shorts to both >> GND and Vcc. It's only Vcc over-voltages that can hurt, but that would be >> affecting most of the chips in the machine. > >... that's good news. So maybe the chips are not fried after all. They >definitely shouldn't have to stand over-voltages because where in the >whole system should be anywhere over the max. 12V (a little more to be >precise) that comes from the 12V DC input power supply...? There's no 12v here. The diodes are in this circuit to prevent voltage spikes (+ or -) to be absorbed. >Yesterday, I also had the strange effect again that taking out the 7406 >for checking purposes would not give me a picture *once* when powering >on again. So I suspect a slack joint at the 4066 (U21) because they are >next to each other and I touched it when removing the 7406. Strange >enough, the picture immediately came back when powering it on the next >time. Maybe U21 has a slack joint (too)? I'll check that... In any >event, a slack joint on U21 shouldn't affect operation of U22 because >they have nothing to do with each other according to schematics. No, the 4066 is used as a switch for the POT lines. No relation. By slack joint, do you mean a bad solder joint? If you suspect one, you can simply re-solder the joints. >Please allow me a dumb question: In the schematics the small boxes >representing the inverter gates of U22 have a "1" printed inside them. >What does that mean? I have no idea. I'm looking at schematic #252312 as it seems to match what you have, and I see the "1" as well. Doesn't make sense to be. Even the symbol for the inverter is non-standard. Regarding your question about RP6 and the resistance difference on pins 4 and 6... they are the only ones that connect back directly to the 6522. I suspect that the port resistance is in parallel and changing the values. Also, I assume you are checking things with no drive connected to the serial port. PS.
From: Andreas Beermann on 9 Nov 2009 03:26
On 8 Nov., 19:19, Paul Förster <paul.foers...(a)gmx.net> wrote: > Hi, > > On 2009-11-08 15:52:33 +0100, Paul Förster <paul.foers...(a)gmx.net> said: > > > ... I was just curious and checked RP6, the 1k Ohm SIL resistor array. > > ... hmm, it's not RP6. I've exchanged that with a definitely good one > (tested before soldering in) from another board. It now shows the same > odd behavior on pins 4 and 6. And the old one which I have soldered out > shows proper 1k Ohm on *all* pins. It's interesting that a resistor > changes its value when soldered in. Or that'd mean there is some > parallel resistance which is lower (~670 Ohm)... > -- > cul8er > > Paul > paul.foers...(a)gmx.net Hallo Paul, to make the search more systematic please try to do the following: 1. check if you've connected something to the user port that's fiddling arround with pin 9 which is shared with the ATN pin! if no: 2. please check if it's OK after removing CR9,CR11-17. If yes, re- insert one diode after the other (polarity must be correct!) or Simply replace all with new diodes (could also be a 1N4148) if no (do NOT re-mount the diodes so far!): 3. remove capacitors C88 and C94 and check if it's OK thereafter. If yes, replace with fresh ones if no (do NOT re-mount the capacitors so far!): 4. remove the CIA, U22 and RP6. Now all signals of the serial port (except RESET) should not have contact to *anything* (with the diodes and capacitors removed as well!) In case you measure any Ohm-value < 100kOhms from any of the serial signals to ground you have detected a shortage on the board. see if that leads you any further... Andreas |