C128 black screen / blank screen

[gerha7]

Hello,
I recently acquired a flat Commodore 128 but with no power supply (PSU). I have a working Commodore 64 and 1702 monitor (and IBM 5153 CGA/RGB monitor) to test with. I know the monitors and monitor cables work. I have a C64 PSU which I purchased a little over a year ago from C64PSU.COM. I also ordered from them a C64 to C128 PSU adapter.

C128 Motherboard: Artwork No. 310381 REV 7 Mfg date: 11/85

When I turn on the C128 the red power light does come on but I get a black screen. I've tested with the 1702 with a Radar Rat Race cartridge (that works on my C64), but still a black screen. I've tried testing with the 5153 monitor, with the 80 column key pressed, still nothing. I've tried all the combinations of the 40/80 column key and the Commodore key, a cartridge plugged in and not plugged in. Tried the RF connector to my TV. Still a black screen.

I've disassembled the machine, inspected and didn't see any bulging or cracked capacitors, broken/severely degraded traces--but some traces don't look ideal. I did spot some corrosion by capacitor C99 and what looks like an exposed trace (see pic).

I removed all socketed IC's, cleaned the pins and sprayed with DeoxIT, still black screen.

I do notice that the MOS 8502 CPU, MOS 8564 VIC composite video and voltage regulator (U59) get hot in about 5 minutes.

I've looked thru the C128 Diagnostic Instruction and Troubleshooting Manual (1986)(Commodore)[PN-314060-01] and other forum posts regarding the same black screen issue. I don't have an oscilloscope, just a manual digital multimeter. Below are the readings with the keyboard disconnected and no other peripherals connected.

The components that seem suspicious are:
CR13 rectifier
C97 capacitor
CR11 diode (possibly)
U59 voltage regulator (possibly)…not sure if testing it in-circuit is recommended/conclusive

Any thoughts as to what components I should replace or what else to test? Thanks.

Item to test (tested w/ keyboard disconnected)	Expected result	Actual result
Male 5-pin DIN power plug	5 VDC, 9VAC	5.1 VDC, 10.7 VAC
6581 SID (U5) pin 25	+5 VDC	4.74
6581 SID (U5) pin 28	+12 VDC	11.89
8502 MPU (U6) pin 40 0 VDC on power-up, w/+5 VDC in 1 second	+5 VDC	4.72
Datasette port, pin 2	+5 VDC	4.74
Capacitor (+ leg) C107	+5 VDC	4.75
Capacitor (+ leg) C99	+5 VDC	3.45
Connector CN11 - pin 1	+5 VDC	4.80
Capacitor (+ leg) C111	+12 VDC	11.91
Anode of diode CR10	+18 VDC	17.48
AC inputs of rectifier CR13	9 VAC on both	11.5, 11.3 VAC
Capacitor (+ leg) C97	9 VAC	5.28 (not sure which is + leg)
Connector CN11 - pin 3 & 5	9 VAC	10.4
Connector CN11 - pin 1	+5 VDC	4.8 (amps fluctuate 2.7 to 0.1)
Anode of diode CR11	9 VAC	13.94 VDC, 29.8 VAC Not sure if measured correctly
Anode of diode CR15	5 VDC	4.31
Voltage regulator IC (U59) pin 3 When testing continuity in-circuit, the ground and output pins give 193 Ohms. I assume all pins tested should test open/no reading?	11.9 VDC	11.93 (1.45amps)
Z80 CPU pin 11	+ 5VDC	4.74

 

[dmac]

>Male 5-pin DIN power plug 5.1v/10.7v

Is this perhaps tested without the computer attached?

I notice every other voltage reading is considerably lower than it should be, and suspect that you've adjusted the calibration knob on the PSU to a perfect 5.1v without the computer attached.

Please ensure PSU is configured so that you get 5.1v or almost in the computer (as measured through e.g. any expansion port that provides 5v power).

 

[gerha7]

Yes, the Male 5-pin DIN power plug 5.1 VDC, 10.7 VAC was measured directly on the pins in the 5-pin male connector--not attached to the computer.

There is no calibration knob/adjustment mechanism on the PSU, it just has an on/off rocker switch. Here's a link to the type of C64 PSU I'm using. Its input is 115VAC 0.5A, output1 5VDC 4.0A, output2 9VAC 1.1A.

The 8502 CPU, 8564 VIC and voltage regulator (U59) get, what seems to me, abnormally hot. Almost to the point where I can't keep my finger on them. I know a hot chip can mean a bad chip, but all three? I guess it's possible the old/original PSU from the previous owner fried some components/IC's. As it seems most likely a power/voltage issue, and the PSU seems to be fine, what component(s) would be most suspect? I have a working C64 but am hesitant to put any of its chips in the C128.

 

[dmac]

Should have got that option, which seems to be an addon (from the linked page).

POKE LEVEL UP WITH "ADJUST & PROTECT MODULE" ADD-ON OPTION
(...)
As the current PSU is universal and 4A capable it's possible to power the C128 computer with use of C64-tio-C128 cable adapter. At lease the basic voltage adjustment is the great option when switching from C64 to C128 and opposite as C128 draws much more current and causes larger voltage drop on the cable, connectors and computer's power switch.

While the issue can be anything, insufficient voltage should be your working theory at this point.

 

[gerha7]

I've tested all the electrolytic caps with and ESR meter and some are out of spec (C80, C91, C105).

While I'm waiting for some new caps, I've been looking over my C128 board again for any issues. I did find what looks like a broken trace by the decoder IC (U3), see the attached. It looks like there was a dab of solder that may have come off. The red arrows point to the exposed/disconnected traces. I assume I should re-apply a pad of solder to connect these two traces?

Seems odd that there is a pad of solder that doesn't go through the PCB (connect to the other side of the PCB). So, why even have a pad of solder and instead just use a trace?

 

[dmac]

Seems odd that there is a pad of solder that doesn't go through the PCB (connect to the other side of the PCB). So, why even have a pad of solder and instead just use a trace?

https://www.zimmers.net/anonftp/pub/cbm/schematics/computers/c128/mainboardC128_rev7.png

It seems to go to another layer (notice trace in a different color), which might not actually be the back.

Solder will likely refuse to stick where it should. I would put a small trace of copper between the exposed track and the pad. Worst case, wire between the pad at the start and the pad at the end, bypassing the entire track.

 

[gerha7]

Thanks dmac for the PCB image. Yes, I was trying to find where the trace led to...since it went under some IC's I couldn't tell for sure where it terminated/connected. Looking at the back of the motherboard, I see a bodge wire with a resistor. The bodge wire looks like it terminates/connects to the same spot as the [broken] trace would. When I did a continuity check on the two ends of the broken trace (i.e. put my probes by the two red arrows in the previous post's photo) I got a ohms reading of 68, which is the spec for the resistor in the bodge wire.

So now I'm not sure if the missing pad is really an issue...was it fixed by the bodge wire with resistor, or do I still need to connect that broken trace?

Attached is a pic of the bodge wire. For orientation, the top of the photo is the front of the machine, the bottom is the back.

 

[dmac]

Thanks dmac for the PCB image. Yes, I was trying to find where the trace led to...since it went under some IC's I couldn't tell for sure where it terminated/connected. Looking at the back of the motherboard, I see a bodge wire with a resistor. The bodge wire looks like it terminates/connects to the same spot as the [broken] trace would. When I did a continuity check on the two ends of the broken trace (i.e. put my probes by the two red arrows in the previous post's photo) I got a ohms reading of 68, which is the spec for the resistor in the bodge wire.

So now I'm not sure if the missing pad is really an issue...was it fixed by the bodge wire with resistor, or do I still need to connect that broken trace?

Attached is a pic of the bodge wire. For orientation, the top of the photo is the front of the machine, the bottom is the back.

Very interesting. It probably is, but I wonder what the history behind it is.

As for the c128 not working, I still think power supply can't handle the current c128 demands, and voltage drops below operational. There must be reason the psu maker recommends the adjustable voltage option for c128.

 

[cjs]

4.72 V isn't terribly low, just barely outside the ±5% spec typical for '80s ICs. I wouldn't dismiss that as a problem (especially given the suggestion from the PSU manufacturer), but probably my next step would be to grab a 'scope and verify the obvious signals on the board, starting with clock generation and working along from there. If you don't have a 'scope (and even one of those cheap $30 jobs will find truly broken signals), a logic probe at least will ascertain that a signal is present.

 

[gerha7]

Regarding the bodge wire with 68 ohm resistor underneath the motherboard, as far as I can tell it connects to (among other things):
…pin 6 of the SID (U5) - 1 MHz, TTL-level input for master clock
…pin 3 of the FlipFlop (U56) - 1 MHzA
…pin 11 of the Latch (U12) - 1 MHzA
I'm using the SAMS Computerfacts manual, I'm not sure why it denotes "MHz" for the SID (U5) but "MHzA" for U56 and U12. Also, I noticed pins 25 of both the CIA chips (U1, U4) are connected to the 1 MHz line and it shows the 68 ohm resistor with MHz and MHzA (page 12). Looking at the attached schematic from that page, does this mean I should connect the traces as indicated in post #5?

FYI, still waiting for capacitors, and I've ordered an oscilloscope (OWON HDS242)...been wanting to get one for a while.

 

[gerha7]

The replacement capacitors arrived. I replaced the C106 capacitor (1000 uF 25V) since its ESR was ranging from 0.125 to 0.130 ohms (in-circuit) and 0.132 (out of circuit), upper limit spec is 0.08. The replacement capacitor (1000 uF 25V) tests at 0.051 (out) and 0.058 in-circuit. In post #5, I mistakenly typed C105, should have been C106.

Good news! Now I get something on the screen, even though it's garbage. Other times when I turn it on, it displays a solid color or sometimes blank. I plugged in my Radar Rat Race cartridge to force it into C64 mode--still garbage. See attachments.

The voltages haven't improved on the few components I tested: Datasette port pin 2, 6581 SID (U5) pin 25, 6581 SID (U5) pin 28 are all pretty much the same as in my initial post. I may just wait until I get my oscilloscope to test the IC's before I replace any more caps. The IC's are running much cooler, except the voltage regulator (U59) still gets hot.

 

[dmac]

Awesome.

I would try and get my hands on whatever modern test cartridges exist for the c128.

 

[dabone]

There's actually a diag rom for the 128 that can plug straight into the motherboard at u36 if you have a spare eprom and a way to program it.

C-128 Diagnostic Rev. 789010

World Of Jani (Click here)

 

[dabone]

But grabbing a deadtest / destest (Brand new diag rom) cartridge is a good idea.

 

[gerha7]

I may have to get a diagnostic cartridge/harness kit if it comes to that. Unfortunately, I don't have an EPROM programmer.

I assume I should patch what looks like an unintended break in the trace in post #5, as indicated in the schematic in post #10.

 

[gerha7]

Using my scope and the troubleshooting steps in the C128 Diagnostic Instruction and Troubleshooting Manual (1986), here's what I found starting on page 2-1. Steps 4, 5 and 6 tested OK but step 7 (pin 23 on the VIC U21) should have tested at 2 MHz but I measured 1.023 MHz. Step 8 also failed, pin 25 on the VIC U21 should have tested 4MHz but it measured 2.28 MHz.

Per the manual I jumped to Section 2.4 - System Clocks.
All the steps passed (2.4.1 thru 2.4.5) except on step 3 of section 2.4.5 (4 MHz clock) which says to measure signals on pins 10 and 12 on IC 7407 (U60). It should test at 4MHz but I get around 2.04 to 2.27 MHz for both pins. Therefore, it says that U60 and Q6 may be defective.

Also, during one of my testing sessions, pretty much every IC I tested, fluctuated wildly. The scope would display '?', then around 2MHz then 4MHz, when I was testing a 4MHz pin, for example. I turned the C128 off then back on, then the MHz measurements were stable….but again, some were testing way lower than they should.

 

[gerha7]

Quick update. I replaced IC U60, and replaced the power switch (SW1) on the motherboard, but voltages haven't improved...still hovers around 4.72 - 4.74, and I still get a blank screen. On the VIC chip (U21), the specs/schematics state that pin 23 is the 'changing' system clock which can be either 1MHz or 2MHz...it's always testing around 1.023 MHz....so it doesn't seem totally abnormal/out of spec. But according to the documentation pin 25 should always be 4MHz, but it always tests around 2.28 MHz.

I also discovered something about the IC MMU (U7). The service manual (p 29) states pin 43 (Z80EN) "This output is used to enable the Z80 processor and disable the normal operation of the 8502 processor. It goes low to indicate Z80 mode, high for all other modes." Without a game cartridge inserted, I tested pin 43 and it shows 0.05 VDC, testing with a game cart inserted it tests at 0.07 VDC....both low voltages. So for some reason it's signaling Z80 mode when it seems it should be in C128 or C64 mode. In all my testing so far the keyboard hasn't been connected. I'll have to do some more digging to see why the MMU thinks it should be in Z80 mode.

 

[gerha7]

Still chugging along.
The PSU I have from c64psu.com, has an adjustment at the bottom of the unit. I've turned it up to the point where it tests 5.0 VDC on pin 2 of the cassette port on my C128. Still no video.

Similar to my last post, I tested pins 10 and 12 on IC 7407 (U60), should be 4 MHz but still tests 2.27 MHz. Replaced VIC chip (U21) still no video. I can't guarantee the VIC chip I purchased from eBay is operational, although I get the same measurements on all the pins (voltages, signals) as the one that came with the machine. I replaced U60, no video. Replaced transistor Q6, no video. Unfortunately the service manual doesn't go any further after replacing U21, U60 and Q6. Testing U60 (7407 Hex buffer/driver [open collector]), and using the IC diagram and truth table, I see a few discrepancies between what I measured and the logic diagram and truth table. Attached is an image of the IC specs from the service manual and SAMS. My U60 measurements are:
Input --> Output
Pin1: high (3.61v) --> Pin2: high (4.98v)
Pin3: high (4.98v) --> Pin4: high (4.98v)
Pin5: high (4.98v) --> Pin6: high (4.98v)
Pin9: high (4.98v) --> Pin8: low (0.12v)
Pin11: pulse (2.27Mhz) --> Pin10: pulse (2.27MHz)
Pin13: pulse (2.27MHz) --> Pin12: pulse (2.05MHz)
Pin14: high (4.99v)

First, according to the truth table (Service Manual p.61), if the input is high, the output should also be high, but pin 9 and 8 are high --> low. I tried a different IC and the outcome is the same.
Second, according to the logic chart (from SAMS p.39), pin 1 should be pulse and pin 2 low…not sure if this is an error.

What's also odd is that pin 43 on the MMU-Memory Management Unit (U7) is low 0.05v which indicates Z80 mode. High indicates 'other modes', (C128 or C64). Even if I insert a game cartridge which should force the C128 into C64 mode, pin 43 is still low.

Any thoughts on what is causing pins 10 and 12 on U60 to read 2.27 MHz when it should be 4MHz? Why would it be about half what it should be? Or is it just a symptom of the MMU holding the C128 in Z80 mode….because <insert SWAG here> the Z80 can't properly complete its chores before turning control over to the 8502 CPU?

New year, same old problem…no video. : (