[SOLVED] MS8-26sg image 1/3 the size it should be. HOT running at 24.3Khz???

[jgarrettcorbin]

Hey everyone! I'm excited to join this community. I recently found a Sega Aero City cabinet on the side of the street in my neighborhood. I've been wanting an arcade machine with a CRT for a while now, so I took it home.

I rigged up a Raspberry PI to output 15khz RGB and with some tweaking I got the monitor to look amazing. The only issues were that I lost the blues every once in a while. I light tap on the neck board with a drum stick made them come back, so I figure it was a cracked solder joint, or a tube socket issue, so no big deal there. There was also some minor blooming. I thought I'd do a cap replacement to hopefully improve this.

So I did a cap kit. I want to preface this by mentioning that I've been soldering for the better part of 20 years now. I used to work on a lot of tube amplifiers, but now most of what I solder is under a microscope. This isn't to say that I didn't make a mistake here, because thats entirely possible, but I didn't do this with a $10 radio shack soldering iron.

After installing everything back into the chassis, I fired it up and the picture was way too small and kind of dim. Adjustments didn't help. Something was wrong. If I leave it on, eventually the picture would stretch out horizontally, and it made the worst squeal i've ever heard in my life. This seems temperature related because it usually takes 10 minutes or so for it to happen. If I unplug for a few seconds, it'll be fine for a few seconds and then happen again. If I wait until it cools down, it takes 10 minutes to happen.

I have since checked over my work multiple times. I verified the new caps against their original value. I verified the polarity is correct, I checked all of my joints and places I reworked to make sure there were no dry joints or bridges. Everything seems good. I was hoping someone could point me in the right direction as to what could have happened, or at least what part of the circuit I should be looking at. I know a decent amount about vacuum tubes, but everything I know about CRTs I've learned in the last 2 weeks.

-------

Since I wrote this up in anticipation of being able to post it, I've found something else. I pulled the HOT and I have a short from base to emitter. I'm guessing this is the issue, but now I'm concerned as to why it blew. I'm going to order a couple of these, which will hopefully fix my issue. Is there any way a routine cap replacement could stress a component like this that might be on the edge of failure, so should I look at another part of the circuit before I fire it back up? -- EDIT. It came to my attention that this "short" is do to a built in resistor tying B to E, so as far as I can tell, the HOT is ok.

Thanks everyone.

[grantspain]

caps way out of spec could stressed the HOT, double check the b+ when all work done
b+ is 92v on this chassis but remember use the test point or cap ground when reading as heatshield is not ground
one real issue to watch out for on these chassis are corroded glass diodes and resistors, in particular if the chassis has been damp or any cap leakage

[jgarrettcorbin]

caps way out of spec could stressed the HOT, double check the b+ when all work done
b+ is 92v on this chassis but remember use the test point or cap ground when reading as heatshield is not ground
one real issue to watch out for on these chassis are corroded glass diodes and resistors, in particular if the chassis has been damp or any cap leakage

Thanks for the response! I replaced every electrolytic with Nichicon caps, so they should be okay. I'm assuming that you are talking about the ceramic/films caps then. I'll check over some of those.

I did check the b+ and got a stable 92v. I will check again when I hopefully get it back up and running.

[PrincessPrinPrin]

It's not a short. This type of HOT with built-in damper diode (A to emitter K to collector) also have a low ohm resistor (40-50 ohm) in parallel with the base-emitter junction.

[jgarrettcorbin]

It's not a short. This type of HOT with built-in damper diode (A to emitter K to collector) also have a low ohm resistor (40-50 ohm) in parallel with the base-emitter junction.

Wow. You are correct. B to E measures 40 ohms. I checked the data sheet to verify. Back to the drawing board I guess. Do you have any suggestions on where else I should look?

[grantspain]

power on without the video signal, if you get a raster with full deflection perhaps its something to do with whatever you have connected - a pi or something?

[jgarrettcorbin]

power on without the video signal, if you get a raster with full deflection perhaps its something to do with whatever you have connected - a pi or something?

This is actually on without a signal. The RGB+sync cable isn't even plugged into the board at the moment.

[jgarrettcorbin]

I did some more digging. I tested the voltage on the collector of the HOT and got 120v DC. This didn't seem right. I put an oscilloscope on it and here are the results. I'm concerned about the frequency. I believe it's supposed to be around 15khz, but it's 24.3khz. In my little pea brain, I imagine a frequency thats too high would draw a smaller picture. Right now i'm looking at the IC103 extension board. I did have to work around a glob of white silicon, and I might have been a little rough with it. Looking through the service manual, I see that the board can be switched from 15khz to 24khz by moving a plug. I did not do this, although the resemblance of my measured frequency to the "fast" setting isn't lost on me. I discovered this board could be switched, so I moved the harness to the "high" setting to see what would happen. Almost complete vertical collapse in that setting, but it is full width. I never tried this pre-recapping, so I have no idea if it worked or not.

[jgarrettcorbin]

I wanted to add another finding. I hooked my raspberry pi back up to the inputs which outputs a 15khz RGB signal. I get vertical collapse. The horizontal line still doesn't expand the width of the screen. It's the same width as the original raster photo that I posted.

Should I start suspecting the HA11423 chip? I've been over every circuit, component and joint in that area. I pulled and checked over the IC103 board.

[jgarrettcorbin]

The saga continues. I'm following the troubleshooting guide in the service manual. The IC103 pin # 4 is supposed to test 10v. I'm getting 4v. This would explain why the chip is misbehaving so badly. Next, troubleshoot the 12v supply...

[jgarrettcorbin]

Me again. I found the issue. Yeah, it was my fault, but I don't think I deserve all of the blame. Replacing the big 680uF capacitor requires a little bit of scraping silicon off the board. Turns out I BROKE THE STUPID JUMPER THAT SITS RIGHT UNDER IT. This broke the 12v supply. I guess for reasons I don't understand, some voltage is able to reach the HA11432 chip. At least enough for it to run kind of? Either way. That was the issue.

Now I cant get it to sync horizontally. Vertical sync is fine. I did turn some of the pots on the board in order to lubricate them, but I was careful to return them to their original position. I'm messing around with HSUB1 and HSUB2 (which I believe only influences fast mode), but I'm not having any luck.

[PrincessPrinPrin]

The HA11423 has two sections, horizontal and vertical, supplied by two different 12V lines. The 12V_H (without which the flyback wouldn't work) comes from the B+ through R913 and a 12V zener built in the IC.

[jgarrettcorbin]

WE'RE BACK!!! I solved the Horizontal sync issue. Sloppy work on my part. It happens. Anyways, thank you PrincessPrinPrin. I'd probably still be waiting for a new HOT to show up in the mail if it wasn't for your response earlier this week.

Anyways, if anyone experiences these symptoms after a cap job, check the jumper right next to C905. It probably has a wad of silicon over it. You might have broken the jumper removing the capacitor.