Wells Gardner D9200 - Horizontal Collapse?

[Rocketeer2001]

Hi everyone! First time poster here.

I've recently acquired a Big Buck Hunter II: Sportsman's Paradise and it has seen some action. The guy I got it from said the machine would turn on, but the monitor wouldn't work and crackles, pops, and sparks would shoot out of the back. After a lot of research I learned I have a dreaded 27" Wells Gardner D9200 monitor. This is where the rabbit hole of woe and sorrow started. Apologies in advance for the long post.

I began by fixing some broken ground wire connections and fixing some solder joints. After that, the monitor would work, but still made crackling sounds every now and then, and the picture jittery at times. The common fix for these issues is to replace caps, so I replaced all 76 of the electrolytic capacitors (and made a list of every single cap if anyone wants it). That smartened up most of the issues. I could then see the monitor has some burn-in (not much I can do about that) and the image was rotated slightly counterclockwise. I adjusted VR101 to rotate the image, but it was already cranked as far as it would go so nothing could be gained there. I foolishly thought "I've manually fixed a rotation problem like this on a TV before, I'll just do that here". I didn't know at the time that the yoke is adhesive bonded to the tube's neck.

I could see there was hot glue all over the yoke, so I used a soldering iron to melt it away. I managed to get it to the point where I could wiggle the yoke, and I could twist it about 1 or 2 degrees, but it still offered resistance. This was enough to get the image straight though, so I just taped it in place temporarily while I adjusted other settings like color and gain, etc. I really wanted to get the colors better than they were, so I was attempted to hook up a laptop using a VGA cable so that I could run a CRT test pattern, but that darn chassis board would not play nice with my laptop. It might have been a resolution setting in the laptop, or it's something on the chassis board. I eventually gave up and put all the connections back to stock.

Shortly after this I wanted to see if I could copy the contents of the hard drive onto a CF card to use that for a boot drive because the original hard drive did not sound happy and I anticipated that any day now it would pack it in. That process is a whole nother story, but during the attempts to use the card is when the D9200 suddenly started making a horrible high screeching frequency noise and the image started experiencing what I think is horizontal collapse. When the machine boots up the text is all wobbly / jittery / zig-zaggy (see photo), but when the game is finally loaded, it actually looks like it's displaying "fine" but is just squished into 2" wide (see photo).

Here's the things I've checked so far:

Q412 - checks out fine
Q413 - checks out fine
Q414 - checks out fine
Q415 - checks out fine
Q425 - checks out fine
Q426 - checks out fine
Q427 - matches Q426
Q428 - had to remove it from the board, but checks out fine

D417 - checks out fine
D423 - checks out fine
D424 - checks out fine
D426 - checks out fine

VR402 - is good
VR101 - not good! Supposed to be 10k ohms but only got up to 4k. Likely the cause of my crooked image in the first place. I'll be replacing this one.

R312 - good
R314A, R314B - each read 0.4 ohms, but I didn't remove them from the board, but if I did I think they would read the expected 1.3 ohms
R318 - good
R415 - good
R418 - good
R424 - good...I think. It takes a while, but it eventually climbs up to 15k ohms.
R462 - good
R480, R479, R78 - each read 0.3 ohms, but I didn't remove them from the board, but if I did I think they would read the expected 1.5 ohms
R853 - good
R854 - good
R855 - good

Things I know I should check but haven't:
IC 404
IC 403
I don't have a way to test them powered up. The monitor is still in the cabinet and I have the chassis board on the workbench.

Some of those parts I de-soldered from the board to measure them individually, but some of them I tested still in the board. The only ones I removed were ones that were throwing out suspicious numbers. If it was close to the stated resistance value found on the schematics, I left it on the board to check.


I've done a lot of internet browsing for possible answers and spent a couple weekends attempting to fix this, but I'm still at a loss, so any insights you have would be greatly appreciated. Or should I just yeet this thing into the sun and find a 27" TV with S-video input and use that instead?

[grantspain]

do you get full deflection with video signal disconnected?

[Rocketeer2001]

Good question! I have not attempted to turn it on with no signal connected. If it displays the boot text normally with no signal, that would mean there's something wrong with the IC that does the resolution switching, yes?

[grantspain]

the max res on these is svga, some of these chassis revisions only go vga as the components are not on the chassis - i wonder if you went past the max res and damaged something, i doubt it would be ic, maybe a fet or transistor but looks like you checked most of them. q418 checks ok?
see what happens with no signal

[Rocketeer2001]

It seems like it should be a transistor to me too. I did take out Q418 and checked for resistance between pins, and it did give results. I have read other posts where these components were fine out of the board, but when they were on the board and powered up, they would start to overheat and mess up the values, thus not work properly. Maybe that's what's happening?

I should probably explain how I'm testing the parts. I'm not electronics expert, so I've just been googling these things.

For Mosfets I was putting the multimeter in diode mode, then measuring the center terminal to each outer terminal. I should see around 0.5V on one leg and nothing on the other.
I could also measure their resistance between outside terminals and get around 1.2 ohms.
I could also measure for continuity between the outside terminal. There should be none.

For Transistors, I'd use diode mode on the meter, measuring between the center leg and either outside leg. They should read around 0.6V

And because I'm not sure if any of that is correct, if I had the component off the board I tossed it in my transistor tester. It's a common gizmo that tests much more than just transistors (see pic).

[grantspain]

yes diode test is what i use, in general a bad fet or transistor will read a dead short if bad. I have had a transistor fail and not read dead short but will have an odd reading

[lilshawn]

i suspect a cap has gone out of spec in the horizontal tuning (search 400 area)... or a FET has gone bad (one of the ones the chassis uses to switch to tune the Hfreq)

also suspect anything attached to CS1/CS2/CS3... find your Hf logic in the table for your frequency you are running and check for shorts in the caps/opens in the resistors/transistors/fets attached to them (IRF640's etc) C455 C456 etc. so on.

[Rocketeer2001]

Sorry for the tardiness, it's been a busy week. Finally got some time to work on this today.

Before I put the board back in the cabinet and turn it on without an input signal, I wanted to check over a few more components since the board is already on the workbench.

I read that the deflection yoke could cause horizontal collapse if it's busted, so I tested the resistance of the coils and got 0.2 ohms for the horizontal and 2.3 ohms for the vertical. These numbers seem low, right? Then again, I don't know what the resistance should be on a D9200. If it's noted on the schematics, I couldn't find it. Anyone have a D9200 they could measure to give me a ballpark? I attached a pic so you can see the stellar condition it's in.

I then removed transistor Q418 to check it again. I use my LCR tester to check all sorts of parts, and normally it will identify what the part is and give you a little diagram with the resulting values (see pic of Q428 and Q430 to see what I mean). Q418 was identified as a "resistor", which seems odd to me (see pic). I did a diode test afterwards with my multimeter and I got 0.262V between center and one leg, and 0.334V between center and the other leg, which is what I think are normal, so Idunno.

Following lilshawn's advice, I checked a few other transistors and didn't see any issues. Then checked the big C456 polypropylene capacitor and got 0.304uf as a result (see pic). Schematics say it should be 0.24uf, so that seems close enough. Decided to check the other 4 big polypropylene capacitors but they were all stuck together with some kind of glue, so they had to be removed as a clump (see pic). The results I got are below [figures in these brackets are what the schematics say the values are]:

C430 [1.2uf, 440V] = 1.18uf, ESR 0.23 ohms
C438 [0.39uf, 440V] = 0.303uf, ESR 0.09 ohms
C455 [0.15uf, 250V] = 0.319uf, ESR 0.10 ohms   This one looks to be double the value it should be. Too far out of spec?
C475 [394pf, 440V] = 390.4nf, ESR 0.09 ohms  This one is rated in picofarads, but my result was in nanofarads. Right numbers, but wrong scale. I should have got 0.3904pf.


What do you guys think is the biggest issue here? The yoke, Q418, or these last couple capacitors? Or all the above?

[Rocketeer2001]

Apologies, I'm trying to upload all the photos but I keep getting errors. I'll try again in a bit, I guess.

[grantspain]

the transistor reading seems odd to me, i would expect collector/base 0.450 and emitter/base 0.500 or around that sort of range

your yoke looks pretty bad, the yoke in my experience are one of the most common failures on d9200 - they just corrode due to the crt suface paint flaking off and rotting the windings or actually burning out. One of the horizontal pins on the connector has been burning for some time, it could be poor connection

c455 is well out of spec but are you going from the component marking or the schematic, c475 looks to be completely wrong but if its marked 394 then the cap is 390nf

I think first thing to do is check all the codes on the caps and if they all match then replace the yoke connector pin

I am leaning towards corroded yoke but that normally causes dead short to HOT and blows, the fault you have on a chassis with a good yoke would be down to an out of spec high voltage poly cap - it could be any of the poly caps rated from 250v to 1.6kv . An example of this is a polo 25 chassis i had this week whereby the 3.3nf 1.6v cap was bad, chassis ran but eht was high and the width control had no proper range

[grantspain]

i take back the readings on q428  should be about 0.640 on those junctions

[princess prin prin]

I tested the resistance of the coils and got 0.2 ohms for the horizontal and 2.3 ohms for the vertical. These numbers seem low, right? Then again, I don't know what the resistance should be on a D9200.

Low resistance (and inductance) is what you find on the horizontal coil of monitors that can do 31KHz so 0.2 ohm is fine to me. If you have an LCR meter measure the inductance, it should be in the 0.2-0.3 mH range.

I then removed transistor Q418 to check it again. Q418 was identified as a "resistor", which seems odd to me (see pic). I did a diode test afterwards with my multimeter and I got 0.262V between center and one leg, and 0.334V between center and the other leg, which is what I think are normal, so Idunno.

Q418 is an NPN transistor so it tests like any other NPN, with Base-Collector being one P-N junction and Base-Emitter being the other one. This means that you'll have voltage drop readings only on those (with red probe on Base). You should have about 0.45V on B-C and 0.50V on B-E (all 2SC5144, all test similarly):

[Rocketeer2001]

The one wire on the yoke connector does look like it's been getting hot, but the actual pin inside is still clean and makes a good connection.

I can probably use the LCR meter to test inductance of the yoke. I've never tried, but it's a feature-packed little device, so I'll give it a go.

I did notice that C456 on the schematic is 0.24uf, but the actual part has a '304' on it (see pics), so you're probably right that the components used don't match the schematic and I should just check the labels on the parts instead. It just means I have to split that glue to get them apart. I hope the labels don't get ruined by this and become illegible.

If Q418 is just a big NPN transistor, then I would expect the LCR meter to identify it as so. It identifies all functioning NPN transistors, so I'm leaning towards this being the culprit. I kind of tested it backwards in diode mode with the negative probe on the base.

Lets see if the pics will upload this time...

...nope, still get this error that says "your attachment has failed security checks and cannot be uploaded". Not sure what security check it's failing. It's under the size limit and it's just a jpeg, like the other photos I posted. Is there a limit to how many photos can be posted in 24hrs or something?

[grantspain]

i thought it was q428 that was reading bad, its a 2sc945c and is part of the frequency circuit and would certainly explain your issue as the chassis would not be able to find the right frequency
q418 is the horizontal output transistor, 2sc5144 and should read exactly as princess prin prin has shown

[PL1]

Lets see if the pics will upload this time...

...nope, still get this error that says "your attachment has failed security checks and cannot be uploaded". Not sure what security check it's failing. It's under the size limit and it's just a jpeg, like the other photos I posted. Is there a limit to how many photos can be posted in 24hrs or something?

Unfortunately, the forum software sometimes throws false positives during the security checks.   
https://forum.arcadecontrols.com/index.php/topic,164181.0.html

The best known workaround is to either crop or resize the image by about 3-5% and try uploading it again.


Scott

[Rocketeer2001]

I hooked the LCR meter up to the yoke and got:
2.8 ohms, 0.23mH for the Horizontal and
9.3 ohms, 5.71mH for the Vertical
Those readings look much better than what my multimeter spat out, and is close to that 0.2mH range you mentioned, so it looks like the yoke is good! (whew).

I hadn't actually looked up the spec sheet for Q428 because it didn't look vastly abnormal at a glance. Is the hFE or Uf no good? Is that the 0.64 you were talking about?
Makes me second guess if any of the transistors I tested were actually good. If the LCR meter didin't show a fault or incorrectly identified it, I assumed it was good.

The Q418 transistor definetly doesn't read 0.5v, so maybe it is kaput.

I broke apart the group of poly caps so I could read their labels and only got a 75% success rate. That white paint or epoxy or whatever it is absorbed the ink from the label, so when I scraped that off the ink went with it. I have no idea what C455 is supposed to say, which is of course the only one I really wanted to know, but all the other ones matched the schematics. Well, mostly; the other 3 said they were 440V MMP on paper, but the actual parts are 400V MPA. Not sure if that's a big deal.
So do I trust the schematics that C455 is supposed to be 0.15uf? Mine reads 0.319uf, which would be wrong.

C456 reads 0.30uf when it should be 0.24uf. Big enough of a variance to warrant replacing?

Do Q418 (and maybe Q428) seem sketchy? Should I just replace all those parts and see what happens?

[princess prin prin]

The HOT is definitely at fault (leaky) as you should have OL with the probes connected like that (or whatever your meter displays when the probes are disconnected).

Your 2.8 ohm reading isn't correct for the horizontal yoke (looks like your meter was doing a better job here).

[Rocketeer2001]

Prin prin, I took out the Q418 again and hooked it up to my multimeter in the same configuration you did. It said 0.32v and 0.08v. I think we can confidently say that its broken. I'll get a new one of those.
What would cause that to fail? Or is being 21 years old reason enough?

Should I replace C455 as well? I mean, it must be a cheap part. I might as well just get one and hope the schematics are accurate and it's a 0.15uf, 250V.

Anything else I should be ordering just to be safe?


Scott, thanks for the suggestion about cropping the photos! It's a bit inconvenient, but it does work.

[grantspain]

don't get a 2as5144 to replace q418, get one of these
https://www.mouser.com/ProductDetail/onsemi-Fairchild/FJL6920TU?qs=CR8BwU6haLdFGBuY0BHhSA%3D%3D

i would also replace q428, 2sc945c and c455

[Rocketeer2001]

Hey Grant, I did discover before your reply that the stock Q418 part number of Toshiba 2SC5144 was hard to find, and the ones I did find were questionable. I did find the alternative part number for the Fuji FJL6920TU, so I'll go with that.

I'm thinking I'll remove these transistors and check them again with the LCR meter, since I'm clearly a novice at determining which are good and bad. Some of them are similar to Q428 so might be failing too:

Q412 - C945C-Y
Q413 - IRF644
Q414 - 2N7000
Q415 - IRFS640A
Q423 - C945C-Y
Q424 - A733C-Y
Q425 - C4770
Q426 - IRF630A
Q427 - IRF630A
Q428 - C945C-Y (this is the CS2)
Q430 - C102M (this is the CS1)

I don't have a CS3 so there is no Q429 or Q416 on my board.

See anything else worth checking? Just want to make sure I order all the parts I need at the same time.

[grantspain]

i think the 3 components, q412, q428 and c455 will do the job.
but if you have time it can't hurt to check all the compoents you list
i reckon q412 is causing the frequency problem, c455 is causing the horizontal size problem. q428 is the odd one here as its reading strange but not short, definitely needs replacing though

[Rocketeer2001]

Grant, I hope it's ok if I lean on you for your expertise here. I used my LCR meter to measure all these parts. See anything amiss? In my little label, I have the Q number, then the schematic number, and then the actual number that's printed on the part:

[Rocketeer2001]

and the last 4

[grantspain]

as far i can see all those read correct

[Rocketeer2001]

Ok!

Interestingly, that Q428 that you didn't like the looks of before, is now ok. I have similar conflicting results with Q430. I think the first time I removed them to test them I slapped them in the meter perhaps too quickly, and they were still too warm from the heat transfer of the soldering iron, which would give me screwy results. See the pics of the before (the first test) and after (yesterdays test). I just realized after looking at those numbers more carefully, that maybe I mislabeled my original photos from my first test, and they're actually the reverse...I'll go test those two transistors again just to be sure, haha.

Guess I'll just be ordering the Q418 and the C455. Dangit, I just noticed that in the schematics it says this cap is a 250V, but in the parts list it says its a 440V. If I recall, it's ok to get caps with a higher voltage rating than what they came with, so should I just get it as a 440V?

[lilshawn]

higher voltage is okay. you typically spec a cap for about 2x the voltage... then the next closest working value up from that.

so for a 150v power rail, i'd spec out like a 350v capacitor (150+150=300) i'd probably only have a 300v, a 400v, or maybe even a 1000v in my bin and it would probably be completely fine to use any one of them... and for a 12v power rail a 25v cap. (12+12=24) and although it would be perfectly okay to use 16v capacitor...i'd probably spec a 35v cap anyway anyway since it's more common than 25v caps and the size difference isn't too much and it adhers to the "rule"

the rating is often fairly overspec'd for it's labeled working voltage anyways. it's basically the rating the dielectric insulator breaks down and allows current to pass through it unimpeded. chances are, a good cap will (albeit unhappily) take 2x it's rated voltage before ACTUALLY breaking down... it would probably change it's value much sooner running it constantly over voltage like that... but even factoring in your back-of-the-napkin estimates of doubling the working voltage of the circuit, you are really in the 3x neighborhood of being overspec'd anywhos.

i think that rail maxes out at like 195 or 212v? something like that? depends on the resolution... but i think a 250v is probably perfectly fine... it doesn't adhere to the 2x "rule" per se, but is still within the working voltage it's rated at... and probably why someone changed the part list to say 450v

[Rocketeer2001]

Ok, good explanation! I picked out a 400V one. They had a 250v, but it was oddly tiny in physical size, and they had a 450v but was also tiny. The 400V one was the closest in physical size to the old one, and had all other specs matching what I needed.

When the parts show up, I'll get them installed and report back.

Thanks for the help so far everyone!

[Rocketeer2001]

The new parts arrived in a day, so I installed them today in the board and put everything back in the arcade cabinet. Turned it on and for a split second it seemed like it was working! But then I saw that new Q418 horizontal transistor sizzle around it's edges and the heatsink with electricity, then the whole board let out a horrible screeching noise and we're back to collapse.

I'm thinking whatever part broke the original HOT has now cooked my brand new HOT. Or my new HOT was faulty, or the new C455 capacitor was faulty, or something else?

[grantspain]

what heatsink compound did you use and insulation sheet?

[lilshawn]

check D416 and D417.

and yes, you should be using a mica sheet and silicone heatsink compound.

the hot must be isolated from the heatsink.

[Rocketeer2001]

I used thermal paste for a computer's CPU heatsink. I put that on both sides of the original mica sheet and placed that between the HOT and the heatsink. I don't know if the paste has silicone in it; is that specific ingredient a requirement?

I had checked D417 before and it was ok, but I'll check it again. I didn't check D416 yet, but I will. I'll check D422 too.

I tried disconnecting the input signal to the board and turning it on just in case it was the frequency switching transistor running amuck, and I don't get a constant screeching. More like 1 second intervals of a clicking sound (a relay?) with arcing occurring around the HOT.

[lilshawn]

i say "silicone" because lots of thermal pastes for computers contain metals (silver etc.) or carbon and cannot be used in this application due to their electrical conductivity.

you need something like MG chemicals 860 ... basically plain white heatsink grease.

if it's gray, chances are pretty good it has conductive stuff in it.

normally it wouldn't matter... but at voltages and currents like this... it does 100%

[Rocketeer2001]

The thermal paste I use is grey, so I put my multimeter in continuity mode and shoved the probes into a glob to see if it conducts. It did not. Then again, it is just measuring using the voltage the multimeter can provide, which is only 1v or so. I'm guessing at the higher voltages the HOT puts out, this stuff could conceivably conduct. I'm uncertain, but I'll get some of the MG 860 stuff just to be safe.

The mica sheet was in there, but upon closer inspection it looks like it's de-laminating? Also now has some char marks. I should probably just get a replacement.

I removed the transistor to test it to see if it's toast, and surprisingly it's still good! (see pic)  However, I had a suspicion that the part was arcing/grounding too easily, so using my multi-meter I probed within the mounting hole and the back of the transistor and I get continuity. Not everywhere in the hole, but in a few locations. It must have had a thin layer of insulation or something because it was shorting out to the mounting screw and thus into the heatsink. Some of the part has crumbled away around the hole from where the arcing was occurring (see pic), but there are other parts of the hole that look perfectly fine, yet the meter still shows continuity when probed. I'm thinking maybe I got a faulty part (in conjunction with my possibly incorrect thermal paste and de-laminating mica).

I then proceeded to test D416, D417, and D422. The first two were good, but D422 is suspicious. It shows 0.28v in diode mode when the other two show 0.5V. Could this be a contributor?

Also tested my new capacitor and thankfully it's fine!

[grantspain]

you can compare d422 to d423, if they read the same then all good
the insulaton sheet you need is for TO-264  transistor package type

[Rocketeer2001]

Grant, Good suggestion on comparing to D423 since it's the same part number. After checking, it's got the same voltage; Looks like my D422 is good!

Thanks for providing the package size of TO-264. I wouldn't have known what to look for otherwise. Even so, I'm having trouble finding that size that won't cost an arm and a leg. If I buy 300 of them, it's a bargain!

How do we feel about using silicone sheets rather than mica with compound? If I use the sheets, I can cut it to whatever size I need and I wouldn't need the compound:
https://www.amazon.ca/100x100mm-Efficient-Conductivity-Resistant-Silicone/dp/B096ZNHY8F/ref=pd_sbs_d_sccl_3_2/144-4490169-7451254?pd_rd_w=heB6B&content-id=amzn1.sym.4b791088-03f0-49af-8af1-ae2f52d6208b&pf_rd_p=4b791088-03f0-49af-8af1-ae2f52d6208b&pf_rd_r=6FAWM4WQCSZ53FYBVCVG&pd_rd_wg=4316Q&pd_rd_r=4a0bdac6-e605-491f-bdf8-ecf523b85029&pd_rd_i=B096ZNHY8F&psc=1

Arcade Parts & Repair sells both kinds, but because I'm in Canada it's over $20 for a $2 part...
https://www.arcadepartsandrepair.com/store/integrated-circuits/transistors/mica-transparent-insulator-to-264-package-5-pieces/
https://www.arcadepartsandrepair.com/store/pinball-kits-parts/pinball-parts/mica-silicon-insulator-to-264-package-5-pieces/#reviews

Or I get the MG compound and just reuse the original mica.

[grantspain]

not sure, it would probably be ok if you got the thinnest one. i would still use thermal past with it.

your original mica sheet is knackered, thats what is causing the arcing and it most likely to have been the fault all along

[princess prin prin]

You can also use a TO-3PL (22*29*0.12mm with hole). This is cheap if you don't mind waiting:

https://www.ebay.com/itm/324673667376

[lilshawn]

typically the silicone pads used for transistors has a fiberglass fabric inner layer to keep the silicone from squeezing out and puncturing when the transistor is screwed down.

they are also only 0.3mm thick.

[Rocketeer2001]

Grant, thanks for confirming the old mica should be tossed.

I did some reading and although the thermal pads are easier to use, I couldn't find consistent claims that it would insulate against the higher voltages, or that it was substantially better at thermal conductivity than the mica. Looks like it's a good substitute for computer CPU heatsinks though. Shawn, I did read about the various silicone pads and how some have that fiberglass layer, but I wasn't finding good sources for it. A lot of items just claimed it was a silicone thermal pad, but didn't really mention enough details regarding its properties. We know the mica works and lasts a long time, so I'll just go with that and get the accompanying compound.

Prin, I did stumble across the size compatibility of the TO-3PL, and that same ebay link! I ended up ordering some of those and payed a few bucks more for faster shipping. It's still going to take about a month, so our little adventure here will have to pause for a bit. TO BE CONTINUED...

[lilshawn]

typically the cut to fit silicone pads like what you had linked are for filling space between a device and a heatskinking material like the back cover case of a device... where you want to remove excess heat, but not real concerned about how much.

you quite often see them in TV's where the main video processor has a pad that makes it so there is contact with the metal backing of the LCD panel. it sucks out SOME heat...but not a real lot.

you also see them on VRMs and ram  chips for video cards to get some heat out to the heatsink.

more of a space filler than anything else.