Tektronix 465B, 110V rail is 102V RMS with 40volt dips.

[kmc_resistor]

Hello EEVBlog Forum,

New member, getting more into EE and purchased a Tektronix 465B. Everything seems to function normally, but there is a noticeable amount of ripple on the display and it isn't quite as sharp as others I've seen on the forum and on YouTube. Thinking of Dave, I decided to Take It Apart and test the power lines and run through the calibration routine in the service manual.

The good news is the +55V, +15V, +5V, and -8V are all perfect and clean. The 110V line, however, reads with my Fluke meter as 101.6V, well out of tolerance (manual says 106.7 to 113.3). I borrowed the Rigol DS1054Z from work this weekend and took some readings and discovered that there are some huge dips in the 110V rail. See attached screenshot taken at test point TP4340, where you see 40V drops, so the 109V max becomes 102V RMS (at least that's how I read the Rigol output).

Can anyone tell me if the unregulated 110V supply is supposed to be fully rectified before it reaches the Q4425 Regulator ? Mine definitely is not.
Looking at the Block diagram on page 3-26 of the service manual I checked the incoming voltage at the diode CR4431, and that is 129V RMS with a huge ripple, then at the Base, Collector, and Emitter of Q4425 everything reads 102V RMS (the same as at the test point).

The block diagram doesn't show the 110V line going through a Rectifier--so I wasn't sure what to expect at this point in the circuit.

 I'd love to clean this bit up since everything else seems to be functioning well.

thanks for your help,

-kai

[tautech]

Welcome to the forum.

Notice the frequency of the waveform and one might suspect a open diode in the rectifier bridge, check that.
Dave found a crook bridge on this HP:
https://www.eevblog.com/2015/10/03/eevblog-803-hp1740a-analog-oscilloscope/
But it might be in Pt. 2
https://www.eevblog.com/2015/10/06/eevblog-804-hp1740a-oscilloscope-repair-part-2/

Another handy Tek repair doc:
http://www.sphere.bc.ca/test/tek-parts/troubleshooting-scopes.pdf

[tggzzz]

The block diagrams are only intended to illustrate the key "interesting" parts of the circuit's operation. For detailed fault finding, you need to look at the full circuit diagrams nearer the end of the manual.

The service manual will specify the allowable ripple on each voltage line, at the beginning of the "calibration" section. For a 465 (I'm not going to download a 200MB file to check the 465B), the allowable p-p ripple on the 110V line is 20mV. Therefore I strongly suspect you do have a problem.

Since the ripple appears to be at twice line frequency, I would guess that the rectifiers are OK and that the relevant smoothing capacitor is faulty. A  quick and dirty test is to take another capacitor with roughly the same capacitance and an adequate voltage rating, and solder it across the suspect capacitor. If it reduces the ripple proportionately, then the existing capacitor is faulty.

(Before soldering, ensure scope is disconnected from the supply and the 110V has discharged to a safe level)

If you have to replace a large capacitor, then be careful not to overheat the PCB tracks, since they will lift. A previous owner has managed to do that with one of my 465s, and I'm going to have to repair it when I replace the bodged replacement capacitor.

[RJFreeman]

You need to locate C4439 (550uF at 100V) - this is one of the cluster of large capacitors mounted on the back of main bottom PCB at the bottom of the CRO - this is the most likely culprit, as C4439 is the main filter cap for the 110V rail.

From memory these capacitors are a pig to get to, and out, so I would suggest replacing all these capacitors while you are at it, as they are probably all near end of life
C4419   5000uF 25V
C4429   1200uF 100V
C4521   5500uF 30V
C4531   5000uF 25V

[kmc_resistor]

Thank you for the quick replies everyone.

Those caps do look like they will be a bear to get to, but as you suggested, definitely all overdue ! This unit was built in 1979.
I'll go ahead and order those replacements and perform the test that tggzzz suggested.

After poking around inside this scope for awhile on my own, nice to be able to reach out to folks who can make suggestions.

thanks again and I'll report back,

-kai

[RJFreeman]

of course if you want to just check, you should be able to tack say, a 680uF, 110V capacitor across C4439, just to prove the point, before you pull it all apart....

[kmc_resistor]

Thanks, I'll try and put something else across there first.
Also, seeing that those 550uF in that packaging can be
hard to come by.

[tggzzz]

Thanks, I'll try and put something else across there first.
Also, seeing that those 550uF in that packaging can be
hard to come by.

Yes.

I'm currently awaiting arrival of a "PCB adaptor" to convert modern snap-in and radial capacitors to that footprint. Unfortunately I've made a cockup and the separation between the + and -ve will be dodgy for 100V; I may decide to re-spin the order.

If it works, then I'll make it available for other people to buy.

[PaulAm]

There are gerbers for an adapter in a 7000 series repair thread, I think.  If I can find that I'll edit it in.  They're not tough to make up.  I've used a drawing program, PNP blue and home etching to make some up when in a pinch.

I usually put rather long leads on the adapter board so I can thread the new cap in from above without having to take anything apart.

Use 105C caps if you can find them.

It can be challenging to remove the old cap without damaging the PCB  There are THREE ground pins from the can along with the positive terminal.  If you don't use an adapter board you need to jumper the ground pins together since Tek used the can as a jumper.

This is a well known failure mode and one of the prime reasons you can get these scopes very cheap.   Yes, they are not a DSO, but they are still very useful.

[tggzzz]

There are gerbers for an adapter in a 7000 series repair thread, I think.  If I can find that I'll edit it in.  They're not tough to make up.  I've used a drawing program, PNP blue and home etching to make some up when in a pinch.

Just so. I think it took me a couple of hours, and most of that was spent figuring out how to create a non-rectangular outline, and double checking dimensions.

I usually put rather long leads on the adapter board so I can thread the new cap in from above without having to take anything apart.

I've considered that, and have neither required nor prevented it. I'll probably start with standard  0.1" headers, and if that doesn't work I'll use wires and the bit of the PCB that I've designed to be a spacer.

[kmc_resistor]

Thanks for the suggestions. I'm still hunting for the parts, but would definitely consider switching to modern equivalents.
Especially as some people seem to be selling new-old stock, or used. Seems like there is plenty of room in here for
another board.


 

[tggzzz]

Thanks for the suggestions. I'm still hunting for the parts, but would definitely consider switching to modern equivalents.
Especially as some people seem to be selling new-old stock, or used. Seems like there is plenty of room in here for
another board.

Definitely don't buy NOS electrolytic capacitors without a very good reason - and electrical compatibility is not a good reason, if you think about why you are replacing them in the first place!

There are many documents around suggesting replacement strategies and tactics; judge then for yourself. Do make sure the replacements have a sufficient ripple current spec at  a suitable temperature - fortunately modern caps with equivalent specifications are much smaller, so it is easy to fit a higher capacitance and/or higher voltage replacement.

For LV PSU capacitors on 100Hz rails, my attitude is don't replace unless there is a problem, where "is a problem" can be either with a specific capacitor or with a similar capacitor in a similar location with a similarly dubious rating. 15V tants on a 13V rail definitely fall into the latter category!

The principal high current caps in switching PSUs are ambiguous, but need replacements designed for switching PSUs, e.g. Nichicon PWs.

HV caps are also ambiguous, especially because it can be problematic to identify the specific source of a problem in an HV supply.

[SingedFingers]

You can desolder the big electrolytics easily enough in these. Use desolder braid and a 50W iron with a 3.2mm+ tip (I use a Weller TCP). It'll take a lot of wick to clean them up however, probably half a roll.

I put BC/Vishay (ex Sprague) long life electrolytics in the unit I fixed. They are much smaller than the originals and don't fit the holes so I stuck cable tie bases to the chassis and strapped them to them then soldered insulated link wires to the A4 board.

I only replaced two dead ones rather than a full replacement.

[kmc_resistor]

Thanks, SingedFingers. My first step is to test a replacement for C4439 (550uF at 100V) and
your workaround might be the way I proceed. Solder wick is cheap enough !

[tggzzz]

For those large capacitors, I've found it beneficial to have a decent hand-held solder sucker. Mine has silicone nozzle. Why? Because the holes in the PCB are large, with one in particular being very large. Hence there's a lot of solder to remove.

It is necessary to heat up the solder sufficiently fast that the heat doesn't have time to spread elsewhere. That implies a large, heavy tip, and a higher temperature than is necessary for soldering. Bias your operations to apply heat to the lead, not to the PCB track since that is only glued to the fibreglass. It will take several attempts to remove all the solder; when the first attempt doesn't succeed, move to another lead and try that before returning.

Eventually you will reach the state where you have removed as much as possible, but each lead will still be attached to the hole by a small amount of solder. At this point heat the lead alone, and waggle it with the iron.

As ever, you have to be very careful with solder "splashes" not being cleared up, and finding their way into switches, or just shorting.

[oldway]

Here are all the informations you need:

http://www.auroragrp.com/agi/tek/Tek465-PS-Cap-Replacment.pdf

[MarkL]

There are gerbers for an adapter in a 7000 series repair thread, I think.  If I can find that I'll edit it in.
...

That would be this thread:

  https://www.eevblog.com/forum/testgear/tektronix-76037613-repairs/msg743296/#msg743296

[PaulAm]

I just found that too.  Nice pics and gerbers are about halfway down the first page.

[SingedFingers]

Those are really nice but I'm not sure I could be arsed with it

[kmc_resistor]

Oldway, that PDF is invaluable ! Many thanks.

Although this seems like a lot of work for an old scope, part of my interest in spending more time with electronics is to be able to troubleshoot and repair equipment that needs just a bit of love. I maintain the darkroom and photography facilities at work and I'm often having to repair enlarger timers and other camera gear. So makes sense to invest in this 465B for my personal use.

Amazing resources here on the EEVBlog Forum !

[SoundTech-LG]

ditto...  nice photography.

Many ways to accomplish that repair. One suggestion is to use silicone (RTV) for securing the caps, not double stick tape... it tends to dry up and crumble.

[RJFreeman]

Although this seems like a lot of work for an old scope,

An old Scope?
Man the 465B, would have been the rolls Royce of CRO's when I was a lad, ok maybe the 466 would have been better, but I dreamed of having a 465B....
The 466 is the Storage Version - although I think strictly speaking the 466 might be a 465 (non 'B) with storage.
 

[kmc_resistor]

Yes, old, but not diminished in my mind at all. Certainly couldn't afford the equivalent scope today.

New caps arrived from Mouser today. I pressed the replacement for C4439 (this one is 560uF at 100V)
against the board just to see if it would make a difference and sure enough, the 110V line immediately
jumps from 101 to 107.8V (well within spec) and the trace cleans up quite a bit too. There was some jittery
lines being displayed right as the trace was about to return and those disappear.

So, now on to the surgery to swap all of these out !

[james_s]

Scopes are not like computers, an old high end scope will still outperform a lot of low end new stuff on the market today. Same is true of a lot of test equipment, a friend of mine has a Fluke DMM old enough that it has a nixie tube display. It has something like 6 digits and despite being 40+ years old it still compares dead on with a modern meter.

[Armadillo]

Like these "SPRAGUE" caps for example - excellent past its design life time and still going strong..... I don't think you can get them anymore though..

[kmc_resistor]

(I hope this isn't a double-post, but I had an error when I last submitted this)

Stage 1 is complete. All the old caps have been removed and the board cleaned up.
This might help someone else doing this job, I found that I could use a Side Cutter tool
to clip two of the ground the leads off of each capacitor, making it only necessary to
desolder 2 leads to remove them. Did all the work with my Chem-Wik and 40W soldering iron.

Now to add in the ground lead connections and new caps...

[kmc_resistor]

Thank you everyone for your pointers and suggestions.
Happy to report that the new caps are installed and everything is working. All of the power rails are now within spec and the 465B seems happy with the new hardware.

Shout out to Ron Childress for making that detailed PDF on how to replace these power supply caps back in 2008.

What a great first experience on the EEVBlog Forum.

all the best !

[james_s]

Good to hear it's back up and running. I'll make a mental note of this thread for when my own 465 eventually needs caps.