willbritton wrote

To clarify, when you say you have shorts, you are measuring resistance with a digital voltmeter across the pads where the capacitor was, and are reading 0 ohms? Are you measuring with the battery connected or disconnected?

You haven't circled all the removed capacitors, does that mean you are not reading 0 ohms across the removed components on the non-circled ones? What are you reading across those?

How did you decide which capacitors to remove, are you following some kind of guide or using a schematic?

willbritton wrote

Okay, so before you took the caps off, you were reading a short across some of them, and so you took them all off but even with them all removed you still get the short? Are you still definitely reading the short across every single one of the caps you removed?

It might be a tough one to track down, since all but one of those caps are across the power supply (according to a schematic I found elsewhere) and that is indeed a low resistance. It's a little curious that you managed to get anything out of the system at all with what looks like a hard short across the power supply. Again - can you just confirm that you are making your measurements without a battery or external power supply attached? Also without a cartridge inserted?

That you also read a short across C49 is a little confusing since that's not across the supply - do you still read close to 0 ohm there? (C49 is centre left of the unit with volume knob to the left)

One possibility that's worth considering is that you might have a bit of metal or something inside the cartridge slot, which could explain if the fault is intermittent and that you were seeing something useful on screen when you (presumably) had a cartridge inserted. Can't remember how accessible things are on the GG, but you could shine a torch in the slot and maybe have a go blowing it out with compressed air or something to see if there's something in there shorting across the supply.

willbritton wrote

So what state is the board in now, is it populated with all new caps, or did you remove all the caps you installed? Did you take continuity readings at the point when all the caps were removed (i.e. when you took the photograph)?

GG1.jpg (3.29 MB)

GG2.jpg (4.14 MB)

willbritton wrote

Thanks for the extra info, and sorry for all the questions!

Man, I'm a bit baffled by this. I still think that if you really had a hard short across your power supply when you first powered it up you'd be in a bad place and couldn't possibly have seen anything on the screen as you did.

This leads me to believe one of three options:

1. Your measurement of the short is erroneous - maybe check your multimeter with some known resistances in case there's a fault with it?
2. The short condition has arisen since you removed the capacitors, possibly because of the process of installing and removing the capacitors - check for any solder strands or bridges you may have left behind; even an almost microscopically thin solder strand can cause a short. You might check not only the empty pads but other components near them to see if you accidentally dropped a blob of solder across some other component.
3. It isn't really a short circuit at all but something that looks like one, although tbh I'm not sure what it could be here. If you still had charged capacitors in circuit then I'm almost certain that the act of probing for resistance would have discharged them. I can't really think of what else could be causing something that looks like a constant short.

My money's on the second option, so if it were me I'd painstakingly go over the board looking for some debris or stray solder. If that's not it, then of course there's a possibility that any one of the remaining components has shorted out, which would be a very tough thing to troubleshoot.

Sorry I can't offer much more, I'll keep thinking about it and get back if I have a brainwave. Otherwise, good luck, hope you can get it working!

EDIT: if there really is a short you might be able to track it down by systematically measuring different points in the circuit in relation to a single known reference point, e.g. some ground point near to where the power supply enters the board. You'll find that although you get continuity, the measured resistance will vary slightly at different points around the board. The lower the resistance, the closer to the other side of your short you'll be.

Quote

5.5 DIODE & CONTINUITY TEST
1. Set the function range switch to the position.
2. Connect the black test lead to "COM" terminal and red test lead to V/OHM input terminal; (Note: the polarity of
the red test lead is "+").
3. This range has an "AUDIBLE CONTINUITY TEST" function. Built-in buzzer sounds if the resistance between
the two probes is less than 30±10Ω.
4. Connect the test leads across the diode and read the display value.
Note:
a. When the input is not connected, i.e. at open circuit, the figure "1" will be displayed
b. Test condition: Forward DC current approx 1mA. Reversed DC voltage approx. 2.8V
c. The meter displays the forward voltage drop and displays figure "1" for overload when the diode is
reversed.

willbritton wrote

Thanks for that video - useful!

Okay, I was just reading the instruction manual for your multimeter and something caught my eye:

Quote

5.5 DIODE & CONTINUITY TEST
1. Set the function range switch to the position.
2. Connect the black test lead to "COM" terminal and red test lead to V/OHM input terminal; (Note: the polarity of
the red test lead is "+").
3. This range has an "AUDIBLE CONTINUITY TEST" function. Built-in buzzer sounds if the resistance between
the two probes is less than 30±10Ω.
4. Connect the test leads across the diode and read the display value.
Note:
a. When the input is not connected, i.e. at open circuit, the figure "1" will be displayed
b. Test condition: Forward DC current approx 1mA. Reversed DC voltage approx. 2.8V
c. The meter displays the forward voltage drop and displays figure "1" for overload when the diode is
reversed.

So I'm actually wondering whether the value you're seeing on the display really is a resistance measurement. Perhaps you could switch to non-continuity resistance mode (start with 200 ohm) and make the measurements again - see whether it's reading the same values?

Otherwise, makes sense that C42 and C45 are showing an open circuit.

Madsmaten wrote

In 200ohms mode I get absolutely nothing on any of the pads.

willbritton wrote

Madsmaten wrote

In 200ohms mode I get absolutely nothing on any of the pads.

Ah, now that is interesting! Try changing it to 2k and measuring across C49 - it should read 470 ohm. If it does, and if the same resistance triggers your continuity reading, then I'm starting to think that your short isn't a short at all...

Madsmaten wrote

Why did it act like there was a continuity read?

willbritton wrote

Madsmaten wrote

Why did it act like there was a continuity read?

Bad meter!!!

But, maybe good news for your Game Gear...

As suspected, the original problem almost certainly wasn't a short circuit. Since the meter does at least seem to measure resistance properly on the 2k setting, you should be able to verify that there isn't a short circuit across any of the capacitor pads. Assuming you can indeed confirm that, my suggestion would be to carefully re-install the new electrolytics and power up again. If there is still a problem, then at least you know it has a different cause and can start a new investigation.

willbritton wrote

Probably not realistic to try and remotely diagnose your board just watching a video of you probing with a scope like that, but might be useful at some point if we find a reasonable culprit, so good to have to hand!

Right, let's start with the basic symptoms:

1. I assume you're inserting a cartridge - to my recollection the GG doesn't do anything at all if there's no cartridge inserted. Now, do you know for sure that the cartridge works; or have you tried multiple cartridges in the same unit and are seeing the same behaviour?

2. Give a little more info on the exact behaviour - specifically I'm wondering if you're seeing the unit continuously reboot, or does it do the splash screen just once and then random stuff after that? If it's not rebooting continuously, what happens if you just wait - do you ever see glimpses of something that looks like the game? Similarly with the sound: do you ever hear snippets of game sound?

3. What is happening with the screen and the backlight while all this is going on? Does the screen look like it's constantly showing something, albeit garbage, or does it go completely blank at all? Is the fluorescent backlight on the whole time, does it flicker or go out?

4. Do any of the LED indicators on the board (is there just one of them?) flicker at all or do they stay steady?

5. Lastly, what is your power supply situation? Do you have the battery connected, or are you just using the power adapter? I'm not sure if the unit will run without the battery connected, but if so you might want to try the 3 combinations of just battery, just power adapter and both to compare whether there's any difference in behaviour.

Having said what I said about the scope, one thing I would check straight out is the master clock, which you should be able to probe here on one side of R24 as indicated in the picture below.

Not sure on the capabilities of your scope there, but ideally it will be able to tell you what the frequency is, also you can adjust the vertical and horizontal scale to get two or three cycles on the window so you can look at the shape of the trace. If you can't get a stable trace, you may need to set the trigger value lower. Since we're looking at the crystal here the voltages won't be full swing so you'll need to play around to get something decent looking.

In general when you're probing you want to take a ground reference as close as possible to your measurement point, rather than using a long lead to connect to a common ground somewhere else on the board; in this case the other side of R24 should be perfect. If you happen to have the type of probe with a little spring clip you can attach for the ground point, that would be ideal, but I don't know what you've got there.

Madsmaten wrote

2: It's random what I see. Somtimes white / black screen, sometimes artifacts, sometimes random sprites.
Pretty often see the blue screen "This game is produced from license with Sega" screen.

Madsmaten wrote

Interestingly whenever I try to scope that pad you specified the Game Gear shuts down?
Been using the big ground pad on the right to the R24.

willbritton wrote

Madsmaten wrote

2: It's random what I see. Somtimes white / black screen, sometimes artifacts, sometimes random sprites.
Pretty often see the blue screen "This game is produced from license with Sega" screen.

And if you see the "...license with Sega" screen, say, does it stay there or does it disappear? If it disappears, what happens afterwards, does it ever reappear?

Madsmaten wrote

Interestingly whenever I try to scope that pad you specified the Game Gear shuts down?
Been using the big ground pad on the right to the R24.

When you say it shuts down, what do you mean? Does the screen suspend, or does everything (including LED and backlight etc.) just turn off? What happens when you remove the probe from the pad, does it turn back on again? Do you get a reading on the scope at all?

willbritton wrote

Thanks again for the videos, very useful.

I'm not quite sure why probing that point should be causing the console to lose power entirely - that's pretty dramatic - and looks like it stays off, I guess until you restart it...huh.

But the interesting thing is you do get a legible and very steady looking BIOS splash screen (which to my understanding only shows if the TMR SEGA cart header is detected), as well as an audio Sega jingle. This suggests to me that all the major components of the system are broadly functional and I'm at a bit of a loss why it seems to crap out specifically at the point that it does.

I see you probing D0 which is pin 13 of CN1 at around 1:20 in this video and there's a lot of activity on it, which suggests to me that the CPU is busy doing something - i.e. not just stalled or halted or powered down.

I think if I were in your shoes right now I'd probably just double-check I'd installed all the capacitors with the right polarity and with the correct values, and then I don't know, I'd probably think about flashing some basic test cartridge ROMs to try and test out functionality one thing at a time, but I don't suppose you're in a position to flash your own carts are you?

One other thing I might play around with, if I were feeling bold, is point M9 on the back of the board is connected to the ~WAIT signal for the CPU. If you were to solder a little dangling jumper wire to M9 you should be able to freeze the CPU by connecting it to GND. That will enable you to read all the bus pins individually in a frozen state and start to piece together what the system is up to. That's assuming you don't have access to a full logic analyser which would allow you to record the system running in real time.

Otherwise, I'm running a little low on ideas for the time being. Will have a think and get back to you if I think of anything else to try.

Madsmaten wrote

One thing in the installation guide and the caps I recieved was that the voltage on some of them didn't match the same value of the orignals.
But as far as I understand the uF is what matters, not the actual voltage? It can be over, but never under right?
Should I try another cap set? I've notice you can get small ceramic ones, which looks really clean ngl.

willbritton wrote

Madsmaten wrote

One thing in the installation guide and the caps I recieved was that the voltage on some of them didn't match the same value of the orignals.
But as far as I understand the uF is what matters, not the actual voltage? It can be over, but never under right?
Should I try another cap set? I've notice you can get small ceramic ones, which looks really clean ngl.

Yes, you're right that the capacitance is what's important here. The max. voltage rating of all your capacitors will be well in excess of the +5V supply here. The only possible problem you might find is if there are any attached to the 34V LCD driver circuitry or the high voltage fluorescent backlight driver circuitry. I can see you've used some 50V rated caps in your photo, which should be enough for those applications.

In any case, electrolytics just tend to just blow up rather violently when you put too much voltage across them; I don't think you'd see the kind of problems you're seeing.

You can indeed get small ceramic capacitors these days in much higher capacitances than back in the day, and depending on the pad spacing you may find you can even use some lovely little surface mount devices which would be much neater than soldering radial legs to pads, but you do need to be careful about the lower voltage ratings with SMD capacitors.

Big electrolytics tend to have significantly higher equivalent series resistances and inductances than small ceramics, which is normally a bad thing, although with an old design like this I would be slightly cautious about replacing components in circuits which may have been tuned based on the original equivalent values. That said, we're not exactly talking rocket science here, and in all likelihood you could replace them with equivalent ceramic caps and not see any difference.

Madsmaten wrote

So like all sane persons would do, I went out a purchased 2 more broken Game Gears.

willbritton wrote

Madsmaten wrote

So like all sane persons would do, I went out a purchased 2 more broken Game Gears.

Haha, this made me chuckle!

I’ve never replaced a screen on a GG before so will gladly defer to wasup’s experience, but I would add that I’ve had plenty of old brittle flexible ribbons tear on me (and from devices considerably newer than those Game Gears) so be very careful when physically handling the ribbon on the screen you want to keep. If you can, try to avoid twisting it and keep the plane of movement squarely across the ribbon so it’s being loaded evenly.

Good luck!