Are you testing the voltages on the power supply test points? Or also on the mpu. I ask as i’ve seen instances where crappy connections rob you of good voltages on the PS leading to borderline at the mpu.

Replacing the caps would off forced you to unplug the connectors and replug them in again after. Maybe this cleaned up the connections temporarily but now they are oxidizing again. Just a thought.

You can pull the 12v connector pin off the MPU and see where the voltage reads but I doubt the reset circuit of the MPU is causing the 12v to drop out. Its probably the large 5v load of the MPU making the 12v dip.

How are the power supply board connectors and fuse clips? Can look for excessive voltage drop happening across fuse clips, connectors, bridge, etc. Ground input plug 3J2 on PS has been known to burn up.

If you have any mods that are tapped into the 12VDC check those too.

I've personally never had this problem on the Sys 11 games I've owned but this bit from http://techniek.flipperwinkel.nl/wms11/index2.htm#reset may be applicable:

Low +12 volts from the Power Supply board.
If your +12 volts is low (below 10.5 volts), check/replace the large power transistor 2N6057 on the power supply board. Sometimes the socket used to connect this transistor is corroded or simply not making good contact.

It's a bootleg copy of the cfh guides that he took down several years ago.

BR1 on power supply board tests good and neither F5 or F6 are blown?

Check voltages at all the listed voltages / test points in this schematic. Does the new Q5 test good out of circuit, use continuity checks to be sure you didn't pull any traces or through-holes, check that you installed the replacement in the proper orientation.

pasted_image (resized).png

3 and 4 are good.
Are C7 and C8 originals?
Measure voltages at pins 11 and 5 of IC1.

I've not had to troubleshoot that circuit much except for a weak transistor but in your case I believe it has to be a bad Q5 or the regulator (comparator) or an associated component is bad, trace broken etc. Since it was working, albeit poorly, before I question Q5 first and foremost. Check the voltages at all the connections (EBC) to Q5. As you can see C should = TP3 so 12.2V in your case (a little low), B should = ~6.5V. Also check voltages at pins 5, 6, 12/11 on the IC.

That's a darlington pair transistor, NPN so put the red leg on the base and the black on the emitter and you should have about 1.2 - 1.4 voltage drop, maybe a little less. Other direction should test very high or open. Since you have your original voltages back I agree you have a bad transistor or you installed it wrong and may have blown it up. Again it's a been a while since I worked on that supply in a Sys 11 but I'm trying to help, maybe GRUMPY or one of the other Sys 11 gurus will chime in. Ed from G-P-E (apologies if it's not actually Ed) has a good question about the two caps but I'm assuming you replaced C7 and C8 when you say you re-capped the supply?

Edit: Testing these types of transistors is a little different and can confuse people not used to them because, as you can see in the schematic it is one package containing two transistors, a diode and two resistors (to bias the second transistor). The pin out should match the attached drawing.

pasted_image (resized).png

I’ve got a couple games suffering from this issue. Very intermittent. New caps on power supply were no help. Replaced the 2n6057 on Pinbot and it seems better but not perfect. On both I only see voltage drop while sound board is powered. New BR on Cyclone seemed to help but not enough. It can go a couple of days without trouble, then it can’t make it through five minutes.

I pulled the 6057 on Cyclone, sanded the contacts, applied deoxit to the socket, and mated the connection several times. I did observe corrosion on the 6057 contacts. The old 6057 measures .4 and .6V, same as a new one. Putting it back together the reset issue was much worse.

Theory is the 2n6057 socket is causing the trouble. Maybe all that effort turned a questionable socket into a bad one. Thinking of just soldering the 6057 in directly. Is there a better appproach?

IMG_1507 (resized).jpeg

Are you certain the replacement 2N6057 are legit? It's super common to run into fakes/knock-offs apparently. I would more readily assume that to be the case than a bad socket.

Post a couple close-ups of the replacements, it may help someone spot whether they're legit or not...

Edit: it's possible I'm thinking of a different one, the LM323K which looks similar to this in shape/package but is a different type of component. I'd still try to verify that what you have is a legit part though...

Thanks for the answer. The 2n6057 I put in Pinbot came from Marco and has the right markings. The one I pulled out of Pinbot tested good, making me wonder if I replaced it for no reason. The one I pulled out of Cyclone also tested good so I put it back, and the reset issues were a million percent worse after futzing with the socket. That’s why I’m thinking maybe socket

Marco doesn't have a monopoly on non-counterfeit parts. They're subject to the market just like everyone else is.... someone just got remarked 2732's from them, and others have gotten bad parts as well.

Well here’s where I am. TP3 used to wander around between 8.5V and 10.5V, which is how we get intermittent resets. This was known back in the 80s—there’s even a note about it in the schematic.

I replaced nearby caps and the 12V bridge rectifier. None of the parts I pulled tested bad. I saw a momentary improvement that was probably wishful thinking. I saw another forum post about diodes in this regulator circuit. Despite the fact that D1, D5 and D6 seemed ok in circuit I replaced them. Out of circuit they test good.

Now TP3 is a volt or so higher than it was and seems stable at around 10.5V. My only working theory is that the bridge rectifier and/or those diodes were on their way to failure, even though they tested OK. It is known that a component might behave differently under load than when examined by a multimeter but I just have no idea.

Here’s a few pics with captions—

replacement part from Marco. Seems to behave the same as the part I pulled.

TP3 over 15 minutes. The two blue lines are where it was before replacing D1, D5, D6.

I couldn’t find a diagram that legibly identifies the components in the board. Here’s what I traced out.

Schematics show this as a simple series pass transistor type regulator. Too low of input voltage and the regulator circuit will 'drop out' resulting in glitches or drops in regulated output voltage. Too high of input voltage and the regulator circuit has to dissipate the excess input voltage as heat. Sweet spot for this circuit is near the 12V value but varies depending on your specific 2N6057.

2N6057 is LONG obsolete. Those in the photo would be pretty old stock. For awhile, 2N6059 was the manufacturer recommended replacement part.
As of about 5 or 6 years go, 2N6059 also went obsolete.

Recommended replacement is now 2N6284. You can find these at Mouser -- search for 2N6284G.

Sockets for TO-3 transistors are horrible and have had them cause all sorts of problems in the past.
Not a bad idea to replace socket. Get AMP (TE) part number 8080-1G1-LF which is also available at Mouser.

Thank you! And thanks also for existing, I got a bunch of your stuff in my games over the years

That 2N6057 looks like a remark. The top of the cap looks like it's been painted silver with the part number printed on it. Genuine parts have the part number printed on metal.

Well what do you know…

IMG_1897 (resized).jpeg

spblat
I have taken the liberty to add your photographs to the PinWiki, here:
https://www.pinwiki.com/wiki/index.php/General#Counterfeit_Integrated_Circuits_.26_Blacktopped_Parts

If you would prefer that I not, please LMK, and I will remove the images.
--
Chris Hibler - CARGPB #31
http://www.ChrisHiblerPinball.com/Contact
https://www.youtube.com/c/ChrisHiblerPinball
http://www.PinWiki.com - The Place to go for Pinball Repair Info

Chris --
On your webpage, the two 6821s also have another indication on them that shows them to be fakes.
Right side photo of 6821, date code is "0438" (38th week of 2004). Left side photo of 6821 - Jerry's photo, date code is "0640" (40th week of 2006).
Motorola issued a last time buy for 6800 and 6821s with a date of December 1999. Last time deliveries were in scheduled for 2000 but I was told they stretched into 2001. Absolute latest date code for 6800s and 6821s would be 2001.
Jerry's part may have worked but the date says it was made 6 years after it was discontinued. Definitely a remark. And since it misrepresents the age of the part, it qualifies as a counterfeit.

I still have a bit more info on my website - if you want to, copy what you want from that page because I'm not sure how much longer that will be up.
https://www.greatplainselectronics.com/Information.asp?region=133

My favorite is the photo of a 6810 memory at the bottom of the page with the 40 pin package.

Thanks G-P-E Ed.
Would it be OK to copy that linked page in its entirety into the PinWiki?

I’m going to miss your site.
I still use it to find what I need.
I suppose I’ll always have the wayback machine.

I hope all is well with you and the family and that being unburdened by running GPE is working out for you.

I’m going to send you a PM about another topic you may be able to help with.
—
Chris Hibler - CARGPB #31
Http://chrishiblerpinball.com/contact
Thank you for checking out the PinWiki - http://www.PinWiki.com/

Chris -
Sure, copy what you want from that page.
I have the website paid for another year, I think.
After that, who knows what will happen.

-- Ed

Thanks Ed.

I’m honored, thanks Chris. In my particular case there were several red herrings and an obvious solution.

1. Not the bridge rectifier
   
2. Not the regulator—original, counterfeit or otherwise
   
3. Not the caps
   
4. Not the diodes

Replacing these (yellow circles) did no harm but it would have been better for me to have noticed the heat coming off the AC input connector. It would have been better for me to have noticed the yellow discoloration on that plug. It would have been better for me to have realized that producing all that heat upstream of the 12V regulation circuit makes a perfect hypothesis as to why the PSU could sometimes go a few hours or days between resets.

Next time I’ll see those clues earlier

IMG_1954 (resized).jpeg

Toasty