Colecovision Black Screen Repair Help

[dafa_123]

13 hours ago, ChildOfCv said:

TTL logic says that anything >2V is high, so those all work.

Oh you're right, I got the bit order backwards.  The bit order should be:

36 35 22   U6
---------+---
 G  G  G | 15
 G  G  5 | 14
 G  5  G | 13
 G  5  5 | 12
 5  G  G | 11
 5  G  5 | 10
 5  5  G |  9
 5  5  5 |  7

So U5 and U6 are both working.  U22 serves clock signals with 4 of its 6 gates, so you'd see a dead clock line if one of them weren't working.  You could check the two gates that help out U1 though:  Check that U22, pin 6 is the opposite of U1 pin 27 for both states.  Check that U22 pin 12 is the opposite of U1 pin 28.

 

Now for U7:  U7 pin 2 should be the opposite of U1 pin 31.  U7 pin 4 should oscillate at half of the main clock speed when U1 pin 27 is grounded, but should be high when it's +5V.  U7 pin 6 should read low.  U7 pin 8 should read high.  U7 pin 10 should read high when you're not pressing reset, or low when you are.  U7 pin 12 should read the opposite of 10 in both cases.

 

U8 works.

We assume that U1 works since it's been replaced.  

We assume for now that U2 works since it's new.

U18 and U19 are untested so far, but unless they're shorting a bus line, they shouldn't be a factor in basic functionality.  Same for U20.

 

You can test the address and data lines with U1 removed though.  The pins of interest are 1-5, 7-10, 12-15, and 30-40.  Use a resistor between 2K and 10K to pull each of these pins up to +5V and make sure they actually measure 5V.  Alligator clips for the meter probe would be useful, or you can read it from the scope since it has a clip.  Then repeat with the resistor pulling them to ground.  If any read strongly different than what is being pulled, then it will be necessary to track down the cause.

 

I rigorously tested the pins. Here are my results:

 

Check that U22, pin 6 is the opposite of U1 pin 27 for both states. --> yes it is in both states.

Check that U22 pin 12 is the opposite of U1 pin 28. --> yes it is in both states.

U7 pin 2 should be the opposite of U1 pin 31. --> yes it is in both states.

U7 pin 4 should oscillate at half of the main clock speed when U1 pin 27 is grounded, but should be high when it's +5V. --> it appears so, see attachment.

U7 pin 6 should read low. --> It's low.

U7 pin 8 should read high. --> It's high.

U7 pin 10 should read high when you're not pressing reset, or low when you are. --> That is correct.

U7 pin 12 should read the opposite of 10 in both cases. --> That is also correct.

 

I tested the address and data lines. I used a 5.1k resistor and I get 4.68v ± 0.1v on high and between 0v and 0.1v on low except for pins 31 and 37 where I get a low of ±0.8v.

 

For now everything looks Ok except maybe pins 31 and 37 on the low state. Note sure if 0.8v is significant?

 

Thanks.

 

 

[ChildOfCv]

0.8v is right on the upper edge of a TTL low (should be between 0 and 0.8), but it is interesting that they resist going to safe low levels.  However, it sounds like the Z80 is capable of driving the bus without passenger interference.

 

So the next likely suspect is U3 and U4, the two SRAM chips.  Manipulating them manually for a test would be quite time consuming though.  If you know someone who has a chip tester, have those tested.  If not, maybe just order new ones and hope for the best.

[dafa_123]

50 minutes ago, ChildOfCv said:

0.8v is right on the upper edge of a TTL low (should be between 0 and 0.8), but it is interesting that they resist going to safe low levels.  However, it sounds like the Z80 is capable of driving the bus without passenger interference.

 

So the next likely suspect is U3 and U4, the two SRAM chips.  Manipulating them manually for a test would be quite time consuming though.  If you know someone who has a chip tester, have those tested.  If not, maybe just order new ones and hope for the best.

Unfortunately, I don't know anyone with a chip tester.  Found some TMM314APL-3 on Ebay for around 5$ each. Would they be compatible? From what I gather, the only difference is that the TMM314APL are rated at 450ns and the TMM314APL-3 are rated at 300ns.

 

Some people seem to build there own memory tester with an Arduino. I wonder if it would be worth the time to get one and try to find a program to test the SRAM?

 

 

[ChildOfCv]

3 minutes ago, dafa_123 said:

Unfortunately, I don't know anyone with a chip tester.  Found some TMM314APL-3 on Ebay for around 5$ each. Would they be compatible? From what I gather, the only difference is that the TMM314APL are rated at 450ns and the TMM314APL-3 are rated at 300ns.

 

Some people seem to build there own memory tester with an Arduino. I wonder if it would be worth the time to get one and try to find a program to test the SRAM?

 

 

If you already have an Arduino and most of the parts, or if you plan to do a lot of tinkering, then I'd say go ahead.  But the cost of the Arduino is about the same as the cost of the memory chips, so otherwise I'd just buy the chips.  Note that if you get an Arduino Uno, you may need to use an expander shield to get more I/O pins out of it.  I'm currently playing with a Mega2560, which has enough I/O on the chip for those tests.

 

The faster memory access speed should be fine.  They have the same pinout.  Apparently that part number is Toshiba's version.  I've more commonly seen 2114, which is probably Motorola's part number.  It's basically the same chip too.  So there are even more options

[ChildOfCv]

Well, slight correction.  The Arduino Uno has 19 usable I/O pins, and you only need 16 at most to test the chip.  So that should be okay.  Here's someone's sketch for doing the test.  No personal guarantees that it works though, but maybe @Danjovic can give you the cat's assurance.

 

https://github.com/Danjovic/2114-Tester

Edited December 4, 2019 by ChildOfCv

[ChildOfCv]

And speaking of Arduino, I'm back working on my all-in-one diag board for the CV.  I got tired of using shields with the Mega2560 since it likes to get flaky connections, so I started designing a board with the chip integrated on.  Here's the current board.  Even with just the processor and FTDI chip, I've already discovered 2 boneheaded mistakes!

 

 

Now to add the 3 voltage pots so that it can attempt to read the CV's voltages.  I guess that also requires soldering on the expansion port plug.

[dafa_123]

3 hours ago, ChildOfCv said:

And speaking of Arduino, I'm back working on my all-in-one diag board for the CV.  I got tired of using shields with the Mega2560 since it likes to get flaky connections, so I started designing a board with the chip integrated on.  Here's the current board.  Even with just the processor and FTDI chip, I've already discovered 2 boneheaded mistakes!

 

 

Now to add the 3 voltage pots so that it can attempt to read the CV's voltages.  I guess that also requires soldering on the expansion port plug.

Nice. Are you going to sell kits?

 

I have ordered the SRAM. Should be getting them in a few weeks. I will let you know once I have replaced them.

 

Thanks.

[ChildOfCv]

I haven't reached that point yet, but I do plan to make the PCB, schematic, and source public.  I at least have schematics and sketch for the shield version available, though the external video test doesn't yet work on them.

[dafa_123]

3 hours ago, ChildOfCv said:

I haven't reached that point yet, but I do plan to make the PCB, schematic, and source public.  I at least have schematics and sketch for the shield version available, though the external video test doesn't yet work on them.

That would be very cool.

 

BTW, I've found 2 SRAM module with the exact same P/N on Ebay. Got them from Taiwan for $3.50USD each with free shipping. It will take a few weeks for me to get them. I've already removed the old SRAM module and I also ordered some sockets. This way I can swap them easily when I test them. Hopefully, it will bring back to life this old console. If not, I'll keep digging and I will have some spare SRAM module for a future project or something.

 

At this point, it's more of a personal challenge for me to repair this console.  Plus, I really enjoy understanding at a more basic level how these old console worked.

 

I will keep you posted once I receive and install the new SRAM modules.

 

Again, thank you for your time and patience.

 

Daniel

[ChildOfCv]

Yeah I hate ordering things from China or Taiwan or Hong Kong just because of shipping times, plus that occasional lost mail thing, like when something I ordered in June didn't arrive until September despite making it to customs within a week or so.

 

[dafa_123]

On 12/5/2019 at 5:52 PM, ChildOfCv said:

Yeah I hate ordering things from China or Taiwan or Hong Kong just because of shipping times, plus that occasional lost mail thing, like when something I ordered in June didn't arrive until September despite making it to customs within a week or so.

 

I got the RAM and replaced it. No luck. Same symptoms. I double checked the lines between U1, U3, U4 and they are all good. Not sure what to look for next.

 

 

[ChildOfCv]

Well, chips that share connection with U1 pin 31 are:  U19 (input), U2 (ROM), U3 and U4 (RAM), and U7.  Chips that share connection with U1 pin 37 are:  U2, U3, U4, and U6.

 

Did you socket the RAM?  If so, you can pull U1 and pull them as well as the ROM to see if any of them make pins 31 and 37 behave.

[dafa_123]

On 12/26/2019 at 7:05 PM, ChildOfCv said:

Well, chips that share connection with U1 pin 31 are:  U19 (input), U2 (ROM), U3 and U4 (RAM), and U7.  Chips that share connection with U1 pin 37 are:  U2, U3, U4, and U6.

 

Did you socket the RAM?  If so, you can pull U1 and pull them as well as the ROM to see if any of them make pins 31 and 37 behave.

Yes, the RAM chips are in sockets. Basically, I put everything I replace in socket.

 

Not sure what you want me to do? Removed U1, U2, U3 and U4 and check the lines with my meter?

 

I checked the all the lines with my meter between U1:31 to U19 (input), U2 (ROM), U3 and U4 (RAM), and U7. It's all good. Same for U1 pin 37.

 

Thanks.

 

 

[ChildOfCv]

2 hours ago, dafa_123 said:

Yes, the RAM chips are in sockets. Basically, I put everything I replace in socket.

 

Not sure what you want me to do? Removed U1, U2, U3 and U4 and check the lines with my meter?

 

I checked the all the lines with my meter between U1:31 to U19 (input), U2 (ROM), U3 and U4 (RAM), and U7. It's all good. Same for U1 pin 37.

 

Thanks.

 

 

Well, I meant that you should repeat the test where you use a resistor to pull to ground, then measure the voltage.  You had two address lines that were cranky, and it may be useful to see if one of the chips caused it.

[dafa_123]

Sorry for the delay in my response. I've been quite busy.

 

I managed to redo the IC test. Here are the results:

 

U5:7 --> 6.15v
U5:9 --> 4.8v
U5:10 --> 4.7v
U5:11 --> 4.65v
U5:12 --> 4.8v
U5:13 --> 4.8v
U5:14 --> 4.7v
U5:15 --> 3.45v

 

U6:7 --> 4.7v
U6:9 --> 4.95v
U6:10 --> 4.95v
U6:11 --> 5v
U6:12 --> 4.6v
U6:13 --> 4.6v
U6:14 --> 4.6v
U6:15 --> 5.8v

5  4  3   U5
--------+---
G  G  G | 15 --> 0.1v
G  G  5 | 14 --> 0.1v
G  5  G | 13 --> 0.1v
G  5  5 | 12 --> 0.1v
5  G  G | 11 --> 0.1v
5  G  5 | 10 --> 0.1v
5  5  G |  9 --> 0.1v
5  5  5 |  7 --> 0.1v

The other pins will vary between 4.85v and 6.2v

 

22 35 36   U6
---------+---
 G  G  G | 15 --> 0.12v
 G  G  5 | 11 --> 0.13v
 G  5  G | 13 --> 0.11v
 G  5  5 |  9 --> 0.13v
 5  G  G | 14 --> 0.13v
 5  G  5 | 10 --> 0.13v
 5  5  G | 12 --> 0.13v
 5  5  5 |  7 --> 0.13v

The other pins will vary between 3.45v and 4.75v

 

I've noticed that the HIGH value on U6 is a lot lower then on U5. Could there be a power regulation issue?

 

Thanks.

[ChildOfCv]

The output voltage of a logic chip is affected by the number of inputs that it drives, so small voltage variations are of little consequence.  As long as a high is well over 2V, and low is under 0.8V, you're fine.

 

I should have been more explicit about what to test though.

On 12/4/2019 at 2:43 PM, dafa_123 said:

You can test the address and data lines with U1 removed though.  The pins of interest are 1-5, 7-10, 12-15, and 30-40.  Use a resistor between 2K and 10K to pull each of these pins up to +5V and make sure they actually measure 5V.  Alligator clips for the meter probe would be useful, or you can read it from the scope since it has a clip.  Then repeat with the resistor pulling them to ground.  If any read strongly different than what is being pulled, then it will be necessary to track down the cause.

To be even more explicit, clip one side of a resistor to your measuring probe and the other side to +5, then measure all the above pins with the probe.  Re-clip the other side of the resistor to ground, and repeat the test.  If any pins refuse to reach their target voltage, try pulling ICs U2, U3, or U4.  If the affected pin(s) are 3, 4, or 5 and symptoms continue, try removing U5.  If the affected pin is 35, 36, or 37, try removing U6.  If it's 31, try removing U7 or U19.  If it's pin 30, try removing U9.

[dafa_123]

Sorry again for the long delay in my response. I was swamped at work and home. I have managed to find some free time to work on the console.

 

When all IC are present, High values are OK but Low values I get:

 

U1:31 --> ±0.82v
U1:37 --> 0.77v

 

I removed U2, U3, U4 and U19 and I get the same results.

 

If I rermove U6 I still get 0.76v on U1:31 when pulled low but U1:37 is now 0v.

 

When I replace U6 with a new 74LS138, I get:

U1:31 --> 0.6v
U1:36 --> 0.6v
U1:35 --> 0.6v
U1:31 --> ±0.76v

 

I have not found a replacement for U19 at my electronic store. but I am note sure that it's bad. I will have to order one from the Internet if I want to replace it.

 

To recap, I now have U2, U3, U4, U6 and U19 removed and I still get ±0.76v on U1:31.

 

When I follow the diagram, I see U1:31 going to U19:19 and U7:1. The IC U7:1 is an inverter, so a low signal on U7:1 becomes high on U7:2 and vice versa. If I check the diagram, I see U7:2 connected to U18:19 and to a 5v source through a 3.3k resistor. When I do a high/low test, I get the proper voltage on U7:1, U7:2 and U18:19. I do find the low value of U1:31 quite high. I an note sure where this can be coming from and if it might be an issue?

 

From the values that I am getting, should I replace U6 with a new 74LS138?

 

Thanks. 

[ChildOfCv]

Hmmm.  What's the voltage between U6 pin 8 and ground?  Between U6 pin 16 and ground?  It's starting to seem like the problem is the ground reference, and perhaps we've just been chasing ghosts for a bit.

 

But if that's the case, then the CPU, ROM, RAM, and all the clock and address logic should be working.  And assuming that the cartridge contacts work, that ought to be enough to run a basic system.  You should be able to plug a DK cartridge in, wait about 10 seconds, and then see waveforms similar to what is in the technical manual.

[dafa_123]

14 hours ago, ChildOfCv said:

Hmmm.  What's the voltage between U6 pin 8 and ground?  Between U6 pin 16 and ground?  It's starting to seem like the problem is the ground reference, and perhaps we've just been chasing ghosts for a bit.

 

But if that's the case, then the CPU, ROM, RAM, and all the clock and address logic should be working.  And assuming that the cartridge contacts work, that ought to be enough to run a basic system.  You should be able to plug a DK cartridge in, wait about 10 seconds, and then see waveforms similar to what is in the technical manual.

There is something weird with the power. I have repopulated all the sockets and connected the Coleco to 110v with the power switch off. I was preparing to redo the checks with the oscilloscope and I noticed that, even with the power switch off, I get some voltage on some of the IC.

 

For example:

 

U3:8, U4:8 are getting 2.8v and some other pins are getting ±0.1v when the power switch is off.

U6 is getting some voltage on most of it's pins between -0.21v up to 0.42v

U1 is also getting some power between 0.1v to 0.3v on some pins.

 

I would expect with the power switch off that everything should be at 0v.

Could this be a power switch issue?

[dafa_123]

Can I have a bad capacitor that is leaking power or something? Is there a way for me to test the capacitors without removing them from the board?

[nick3092]

Keep in mind -5v is not switched, and goes directly to the board. If the wall wort is plugged in, any circuits on the board using -5 will have power. Only +5 and 12 get switched. I don't think that's causing your readings though.

 

Check your power switch. I had a Coleco that the factory did a sloppy solder job on and +5 was always on because they bridged it. Only 12v was getting switched. The only way I noticed was when I got an atarimax cart and the led would light up when the unit was switched off.

[dafa_123]

1 hour ago, nick3092 said:

Keep in mind -5v is not switched, and goes directly to the board. If the wall wort is plugged in, any circuits on the board using -5 will have power. Only +5 and 12 get switched. I don't think that's causing your readings though.

 

Check your power switch. I had a Coleco that the factory did a sloppy solder job on and +5 was always on because they bridged it. Only 12v was getting switched. The only way I noticed was when I got an atarimax cart and the led would light up when the unit was switched off.

I have an Ultimate SD Wafer Drive with a LED on the front. When I plug it into a working Coleco, it blinks several time when loading the menu. With the faulty Coleco I only get a quick blink at power up and then nothing.

 

I will redo the oscilloscope test tomorrow and post my findings.

 

Thanks.

[ChildOfCv]

I have to wonder if the meter sometimes has a mind of its own too.  I have a Meterman meter that decided to go stupid when it's on the continuity/ohms reading and just begins beeping.  Maybe you should try a different meter and see if these ghost voltages go away.

[dafa_123]

On 1/12/2020 at 6:39 PM, dafa_123 said:

I have an Ultimate SD Wafer Drive with a LED on the front. When I plug it into a working Coleco, it blinks several time when loading the menu. With the faulty Coleco I only get a quick blink at power up and then nothing.

 

I will redo the oscilloscope test tomorrow and post my findings.

 

Thanks.

Here are the new Oscilloscope results with Donkey Kong inserted. I did not do all the test since many of them look wrong from the start. Let me know if you want me to test something more specific.

 

U1:17 NMI Signal looks wrong. Sometimes it stays near 0v when I switch the Coleco ON.
U1:18 stays neay 0v.
U1:22 does not have the same wave.
U1:24 does not look right.
U1:27 appears dead.
U1:19 seems wrong. Sometimes it stays near 0v when I switch the Coleco ON and I have to cycle ON/OFF a few time for it to go up.
U1:20 seems dead.
U1:28 is wrong.
U1:35 is wrong. Sometimes it stays near 0v when I switch the Coleco ON and I have to cycle ON/OFF a few time for it to go up.
U6:5 looks Ok.
U7:3 Not sure, lokks Ok
U8:1 looks wrong.
U8:3 and U8:8 look wrong
U8:9 looks wrong

[ChildOfCv]

Just wondering:  Where are you clipping your ground lead?  We've already seen that the clock signal is good, but your measurement above (pin 6) shows it weak and with a 0.5V offset.  If the clock is really that weak, then no wonder nothing works.  But as I say, I'm sure we've seen better.

 

To see anything on pin 16, you need a spinner hooked up -- Super Action Controller, trackball, or steering wheel.  And you have to be rolling it around.

 

Pin 17 (NMI) is holding at 3.5V which also seems weird.  It ought to be closer to 5V when not pulled low.

 

If pin 18 can be trusted, then the processor is in a halt state.  But then if IORQ can be trusted, then it's also asking for I/O.  I don't think it can be both.

 

Pin 24 (wait) also looks weak.  Pin 27 (M1) is being held low.

 

If all of this can be trusted, it would seem that the CPU is not being driven by the clock.  Well, many of the signals are consistent with that.  Get a scope on U8 pin 14.  And also verify U8 pins 8 and 9.  Finally, see if the signal improves after removing U1.