Trouble with -15V Power Supply Regulator on Tektronix R7623A Scope

[nevetslab]

I've been chasing a power supply regulator problem on my back-up Tektronix R7623A scope for several days now. With it installed in the scope, having replaced the bad 14,000uF/25V buss cap that serves to filter the -23V Unregulated input to this discrete -15V regulator, it's been sitting around -4VDC in the system. I've sequentially disconnected portions of the multitude of circuit boards to see if there's one board responsible for loading it down, but thus far I haven't found it. There's still a few more board to go in that hunt, but I decided to remove the Regulator PCB assy from the instrument, and set up external supplies of -23V Input to the regulator, -50V Reference Supply (as is used for all of the circuits in the scope), and +130V for the positive reference on some of the circuits in this -15V regulator.

I haven't yet visited the copier down the street to get scans to add to my database, so I took photos of it, shown here....tall one being all of the Regulator circuits including their Rectifier and Filter circuits that I had been into and replaced the bad filter cap. Then I have one just showing the +/- 15V discrete regulator circuits. I hope this is readable. I'll get scans shortly to replace them with.



Then, here's the bench set up for this regulator, and shots showing the output voltage (on the Simpson 460D DMM) and Supply Voltage readings on the HP 3467A 4-Ch scanning DMM. That shot has the -15V regulator loaded with 200 ohms/75mA. Output under this load is NOT -15V, but -14.3V. I've found the output at 14.3V with 75mA & 150mA load, then -14.5V at a 375mA load. With No load, the output is -23.5V....more than the unregulated input. This has me puzzled. That's the third image below, showing the 130V and the unloaded -23.5V on the Simpson DMM



These are the three Power Supply readings with the regulator loaded by 200 ohms/75mA load.



The scope that has failed has this same power supply board in it, and here's the link to that Failed Tektronix 7633 scope thread:

https://music-electronics-forum.com/...lloscoe-failed

Tektronix scopes with their power supplies are complex, and they do make use of Sense Lines & Feedback from those points. Setting this regulator up separately like this, did require finding all of those circuit connections and strapping them accordingly. I had gone thru the -15V regulator thoroughly, verifying all the dual and single xstrs are measuring ok, as far as junctions go, as are the diodes and the resistors, and as the rest of the regulator circuits on the PCB assy are functioning properly, and do connect their potentials as needed to this troublesome -15V regulator, I had been getting nowhere, constantly seeing -23.5VDC on the output.

It occurred to me last night that I didn't have the -15V regulator circuit under load, so this morning, I tried different loads, not yet knowing WHAT the actual instrument current load IS for this regulator. I've tried 40 ohms/375mA, which yielded -14.5VDC, 100 ohms/150mA, yielding -14.3VDC and 200 ohms/75mA, yielding -14.3VDC as well. I haven't yet connected lower resistance loads to see if that gets it to where it should be. I have my 8 x 4 ohm/500w load bank (32 ohms total there) plus my 4 x 8 ohm/200W load bank that sits on top of that....so I can trim values in 2 ohm increments in some instances.

I'm puzzled by the no-load value of -23.5VDC, it being greater than the input of -23VDC...though that could be help from the -50V ref and +130V ref supplies.

Any thoughts on this puzzle would be greatly appreciated. I still suspect there's a board in this scope that is causing this regulator to be dragged down to -4VDC when it's installed in the R7623A scope.

All of this is in preparation to extracting and digging into my Tek 7633 scope still yet to be opened up. Same complex system, so it seemed logical to educate myself in this instrument, not knowing what problems were awaiting me here.

[dmeek]

Shorted tantalum bead capacitors are a common problem with Tektronix

Here is a great resource for anything Tektronix.

https://w140.com/tekwiki/wiki/Main_Page
https://w140.com/tekwiki/wiki/7623

[nevetslab]

Shorted tantalum bead capacitors are a common problem with Tektronix

Here is a great resource for anything Tektronix.

https://w140.com/tekwiki/wiki/Main_Page
https://w140.com/tekwiki/wiki/7623

Hey, thanks for those links. On this Regulator board, there's only one dip tantalum, and it already checked out as being ok.

I just looked at the output of the -15V Regulator as a function of load/current (current based on assumption of there being 15V).

No load....-23.5VDC
200 ohm load/75mA.....-14.3VDC
100 ohm load/150mA...-14.3VDC
40 ohm load/375mA......-14.4VDC
32 ohm load/469mA.....-14.1VDC
50 ohm load/300mA.....-14.8VDC

It doesn't appear to be regulating. All of the discrete xstrs in this circuit have Tek's in-house numbers. Now, they use the same dual TO-5 NPN xstrs in their diff-pair for the +/- 15V circuits as well as in the +/- 50V circuits (all being 151-0232-00 Dual NPN TO-5 can parts. I haven't yet tried swapping those from, say the +15V regulator (which works), just to see if either of the dual NPN's in this -15V regulator are not working as expected. I tried swapping out the pass xstr without any change seen.

Haven't made any component swaps. I did lift CR961, as that diode was the only one in their regulators that had it in the error amplifier. That made no difference.

I'll download the 7633 manual, just so I have it where I can pull up the schematic from the computer.

[loudthud]

The tricky thing about those Tek regulators is that they depend on all those other Voltages you see in the schematic.

In the Service Manual, there is a list of parts used. Sometimes there is commercial equivalent listed. Otherwise, you'll need a spare parts scope since parts probably are no longer available from Tek.

[Jon Snell]

Looking at the photo of the Regulator Board, I would reflow the solder first. Most of the joints have the tell tale ring around the component legs.
Once the dry joints are re soldered, I would then test and check again.

[nevetslab]

I have the regulator PCB apart and going thru the solder joints of the -15V regulator circuit on the back side. As it is a double-sided TOP QUALITY PCB, with my surgical loupes and head light on for close-up view well lit, I'm not finding anything obvious in way of cold or fractured solder joints, but have still de-soldered/re-soldered the xstr sockets (all of the xstrs on this regulator board use sockets).

Meanwhile, having downloaded what dmeek provided from his tekwiki/wiki/7623 manual link, I've extracted the schematics of the Low Voltage Power Supply, then a section of that showing the +/- 15V regulators, then the Theory of Operation for the Low Voltage Power Supply in the scope.

7623A_LV Pwr_Sply_Sch.pdf

7623A_+ & - 15V Regulators Sch.pdf

7623A_LV Pwr_Sply_Theory_of_Operation.pdf

Onward with this hunt.

[Tom Phillips]

Hi Steven,

The following link is the Google search result for "7623A" +"-15V" limited to the tekscopes@groups.IO discussion site.
You may find some information in one of the discussions that will be useful for your repair.
There are some really knowledgeable members that post there so you might even want to start your own thread. Once in a while you even get a response from someone who was on the Tektronix development team for the specific model.

https://www.google.com/search?q=%22-...client=gws-wiz

Cheers,
Tom

[loudthud]

The soldering on Tek circuit boards is the best you'll find anywhere. I know, I used to work there.

The one thing to watch out for are the low profile black IC sockets made by TI (Texas Instruments). Get one of those U shaped IC pullers and pull ICs STRAIGHT out, not one end at a time.

Another weak point is the TO-3 Sockets for those transistors on the heatsink. If you tighten the screws too tight, You'll break the part that solders to the PCB.

[nevetslab]

Hi Steven,

The following link is the Google search result for "7623A" +"-15V" limited to the tekscopes@groups.IO discussion site.
You may find some information in one of the discussions that will be useful for your repair.
There are some really knowledgeable members that post there so you might even want to start your own thread. Once in a while you even get a response from someone who was on the Tektronix development team for the specific model.

https://www.google.com/search?q=%22-...client=gws-wiz

Cheers,
Tom

Hi Tom,

Many thanks for this link to tekscopes@groups.IO and the additional link. I didn't find what I was looking for there yet, but see there is a vast amount of info there to sift thru.

I've found the problem appears to be at Q943A Error Amplifier for the -15V regulator. R946 (12k 1%) & R945 (28k 1%) are connected in series between GND and the -50V Ref supply, such that there SHOULD be -15V at the junction of the two and appearing at the base of Q943A. Instead, powered up, that junction of the two resistors at Q943A Base is instead sitting at -9.88V, and there's NOT a correct B-E junction voltage there. Emitters on both xstrs are sitting at -9.92V, Base of Q943B IS sitting correctly at -9.2VDC. Both resistors measure ok, lifted up out of circuit. Q943A & B measure ok, as far as diode junctions go. I've swapped it with the error amp from the +15V supply which works fine in it's circuit, but yields the same problem here in the -15V supply. I tried returning the 12k to local ground instead of the Feedback Sense Gnd that comes from the interface PCB of the three plugins. No difference there...same problem

Re-soldered the connections both on the bottom side of the PCB's xstr sockets for the -15V regulator, along with a number of resistors. I was hoping to get success running this regulator PCB assy outside the scope using external supplies...first using external +130V, -50V and -23V for the -15V regulator source voltage. I then changed to use the on-board -50V ref regulator, supplying it with -63VDC to it's pass xstr. Kept having this same problem at the error amp Q943A and not getting the correct -15.0V at the base of that xstr.
Sigh...........

I haven't yet tried replacing R945 and R946. Both measure ok lifted. The common emitter resistor R942, being 150k is fine. CR941, which blocks the DC voltage coming in from the -15V Sense Voltage to appear at the base of Q943A is ok. Regulator output voltage sits at around -4V. Outside the scope, the circuit output voltage just isn't regulating. It keeps leading me back to Q943A and it's reference voltage via R945/R946 being wrong.

[nevetslab]

The soldering on Tek circuit boards is the best you'll find anywhere. I know, I used to work there.

The one thing to watch out for are the low profile black IC sockets made by TI (Texas Instruments). Get one of those U shaped IC pullers and pull ICs STRAIGHT out, not one end at a time.

Another weak point is the TO-3 Sockets for those transistors on the heatsink. If you tighten the screws too tight, You'll break the part that solders to the PCB.

I agree....their construction standards, along with soldering has always been excellent. I bought into the 7000 series scopes when they hit the used market in mass as the industry shifted to 11000 series. Grew up with a Tek 547 & 1A4 4-trace plug in, while my first Tek scope was the non-plugin 7000 series called the 5031 Dual Beam Storage...build factor being that of the 7000 series scopes. This is the first hard failure I've had with the 7000 series. I still haven't extracted the 7633 from the bench. Started my learning curve with this 7623A, thinking this would be a good place to start. Having seen the -15V regulator sitting at -12.6V when I began, finally finding and replacing C811 (14,000uF/25V) with a 15,000uF/50V part, when I powered back up, I no longer had a trace and the -15V supply output was 0V. It's been a battle ever since. Had a trace when that original buss cap (which measured 68nF instead of 14,000uF), the trace I had showed serious sweep problems, no storage, though the time base would control the sweep speed of the ugly waveform on the screen. That's the last time I saw a trace thus far.

[nevetslab]

This morning, I swapped out Dual Xstrs Q943 (where the problem lives with Q943A having improper Vbe) with Q919 same error amp used in the +15V discrete regulator. Problem remains, no proper Vbe at Q943A. I then replaced C943 1000pF/500V ceramic with a fresh 1000pF/200V X7R Ceramic, and CR941 (1N4152) with a 1N4148. Same problem. Voltage at Q943A base still sits at -9.88V, Emitters are at -9.92V, instead of it being at -15.0V and having proper Vbe. Collector of Q943A is definitely at GND.

Lifting both resistors R946 (12.0k 1%) and R945 (28k 1%) out from their GND and -50V pads on the PCB, and externally connecting -50V to R945 and GND to R946, with Q943 removed from it's socket, I get -15.0V as I should get. With Q943 back in it's socket, though not energized, that junction at the 12k and 28k resistor stays at -15.0V. If I connect GND to the circuit board, as well as to R946 (12k) to Gnd, then this junction voltage drops to -14.7V. Nothing powered up on the PCB....the 28k resistor is still lifted off of the PCB, and has -50V connected to it (with the GND from -50V connected to the circuit GND of the board. This has me stumped.

7623A_+ & - 15V Regulators Sch.pdf

[nevetslab]

Hi Steven,

The following link is the Google search result for "7623A" +"-15V" limited to the tekscopes@groups.IO discussion site.
You may find some information in one of the discussions that will be useful for your repair.
There are some really knowledgeable members that post there so you might even want to start your own thread. Once in a while you even get a response from someone who was on the Tektronix development team for the specific model.

https://www.google.com/search?q=%22-...client=gws-wiz

Cheers,
Tom

I just joined the TekScopes@groupes.io last night, and posted this issue I'm stuck at on the 7623A -15V Regulator circuit. I read all of the related 7623A and 7623B posts on their power supply issues. I did find one that revealed the problem I found on C811 14,000uF/25V buss cap for the -15V regulator circuit. Internally, that connection that's a contact-pressure joint stops conducting and the result is little to no capacitance (where I ended up with around 68nF instead of 14,000uF. Figures that one part is the hardest item to remove in the assembly, next to the power xfmr itself.

So, with new members on this TekScopes group, I have to wait for the moderators to review my first post. I didnt' see what the limits were in word count. I kept it as short as I could while detailing the issue. All other supply regulator outputs are correct...just this one, and seems to be at the ref voltage circuit that feeds Q943A base...and there being no proper Vbe there.

Thanks for your help!!

[Tom Phillips]

I just joined the TekScopes@groupes.io last night, and posted this issue I'm stuck at on the 7623A -15V Regulator circuit...

Sounds good. I'll watch for your thread.
Many of the members at TekScopes use the software in the old style newsgroup mode which sends them an email for every single post. I set my profile so that I receive one email each morning with the list of topics that are new and that have been updated within the last 24 hours. Then I review the topics that interest me using a browsing mode that shows all the posts in a given discussion in chronological order. Those that like the old newsgroup email mode don't seem interested in changing their ways so I no longer try to convince anyone to try other methods. The groups IO software is not as feature rich as the V Bulletin SW used here and they require photos to be posted separately in a photos folder. Anyway...I've learned a lot of good information from reading posts there and I hope you receive some useful feedback.
Cheers,
Tom

[nevetslab]

Yesterday, having finished up with the existing shop work, while I still had a GK 800RB in pieces on the Check-out bench, waiting for a 2N21193G/2N21194G power xstrs to come in to complete that surgery, I moved those aside and put my Tek R7623A scope back up on the bench, pulled the top cover, rear cover and dropped the regulator PCB assy down onto the bench, to resume trouble-shooting the -15V regulator. With it connected to the system, instead of sourcing about 1 amp to the -15V circuits, it was sourcing around 2.2A, with the supply sitting low at -4.3VDC. And still smelling something burning, while yet to find the culprit.

This morning, I resumed the search, and began disconnecting power supply cables to different boards in the scope while monitoring the output current from that -15V regulator circuit. I had never seen the supply at the correct -15.0V, even isolated on the bench powered externally. I had again disconnected all of the load to the scope from the supplies of this regulator board, but this time, I now saw -15.05V on the output, as well as at the base of Q943A, which I had yet to also see. That let me off the hook with removing/replacing the 6-pin TO78 body's socket, not being sure it was ok. Powered down, re-connected P902 cable to the scope, and powered back up, I now had the nominal 1A load current flowing, and still had -15.05VDC output, as well as at the base of Q943A. Powered down, pushed the 7B53A time base & the 7A18 and 7A22 Vertical plugins back in and powered back up. Now load current flowing was 1.55A, still -15.05VDC output, and the smell of something burning in the vicinity of the logic board or back side of the plug-in interface board, not real accessible....was no longer drifting up to my nose.

I turned my attention to the CRT controls, not yet seeing a trace. With a few moves on the 7A18 dual trace plugin and the CRT controls, I found the trace, got stable trace in Line Trigger and then trimmed up the display, now it appearing like a proper scope (so far). I let it run for a while, cover still off, power supply not yet mounted back into place. Powered down, mounted the power supply assy, then powered back up, monitoring the -15V load current. Still the same, so I put the rest of the cover panels back on. Turned it around so I could see it from my desk, powered back up, and have test signal running on it (B & K 1027 Sine-Random signal...noise-modulated sine), and it's been running now for an hour and a half.

Still makes me nervous NOT having found WHERE the problem was, nor WHAT was yielding that burning smell. I had again done unplugging and re-connecting power supply ribbon cables this morning...having been down that road many times with this scope over this past month, but pleased to see what might be again a working scope. I know I'd feel better having found the cause of all this.....though I did find the bad filter cap and extracted that with all the grief that took. NICE to see the $20 investment in this mainframe scope working again!

However, this scope being here was an exercise in preparing me to tackle the 7633 Storage Scope that's still embedded on the bench. If this R7623A was NOT the Rack Mount version I'd have already swapped it out with the still-dead scope on the bench. Trying to get the wide-body rack mount scope into the instrument stack would take some effort in reorganizing the bench. As it is, that's gotta happen just to pull it out from there to dig into it, so that moment seems like it's coming soon. Gotta psych myself up for that move.

[Tom Phillips]

Nice that the scope is working but I share your uneasiness that the root problem was not found. Perhaps whatever was getting hot and causing that burning odor eventually went open circuit. That brings up the question of what could burn up but allow the scope to fully function. One thought is that it was a decoupling circuit component. Time will tell if you later notice a feature that's not working properly.

I have heard of people who used the brute force method of connecting a non-current limited supply to the errant power rail and then watching for the magic smoke. That would be a last ditch approach and not one that I'd like to try unless had already thrown in the towel and decided to put the device on the e-waste pile.

I'll keep my fingers crossed for you.

Cheers,
Tom