Troubleshooting an old Tek 475


 

I've got a Tek 475 that is almost 50 years old. It belonged to my father who used it in the late 70's as a service engineer for a manufacturer of laboratory instruments and data machines. My father took very good care of it, but it saw a lot of use and travel in that time, and it may be showing the effects. I had been using the scope for a few months, seeing very minor issues -- dirty contacts on channel #2 mode switch, gummy pots for channel #1 and #2 vertical position and beam intensity, some minor drift in the timebase, and some persistent high-frequency noise visible on the calibrator signal, but it may well have been present all the time -- when it suddenly experienced several failures at once last week.

When I turned it on, last week, the horizontal sweep was gone. BEAM FIND would show the beams, but they were crammed over to the left of the screen. I tried all the sweep settings and got no sweep on any of them. When I put the scope in X-Y mode to see if I had any signal from the channels I discovered that channel #2 was also out (channel #1 shows some response). On top of this the tilting bale broke lose from the righthand side of the case, just over the trigger board (like I said, this thing had been hauled all over the country for six or seven years, and it looks like the aluminum case was suffering from metal fatigue; no surprise, this is a heavy beast).

I opened it up to see if anything had released the magic smoke, even though I didn't smell smoke when it failed. I found nothing burned up, but after careful inspection I did find an electrolytic capacitor (C1059) on the upper edge of the sweep board had leaked out it's contents and corroded its positive terminal. I ordered a replacement, which just arrived today. After replacing the bad cap I now have horizontal sweep back, but now both channels #1 and #2 seem to have no signal, and the vertical position adjustments have no effect on the traces (both traces seem to be draw at the same vertical position, as I get a different intensity of line when I switch from CHOP to ALT, CH1, or CH2).

I verified that a signal is still getting to the CRT by taking trigger from the external input and pressing the TRIG VIEW button. This also verified that the trigger system is working.

I was able to trace the channel #1 signal (using the calibrator signal) from the BNC connector, through the mode switch, and to the resistor (R41) just before the input to U120. I don't seem to see any differential signal on the inputs to U140 (pins 13 & 14 and 16 & 1). The voltages I see on U140 look mostly appropriate (except for 8V on pin 7). I have not been able to investigate all the voltages on U120 because it's not very accessible. I don't, however, seem to see any kind of AC signals on any of the pins of U140, which means that I must not have AC signals coming out of U120 (pins 5, 6, 8, and 9).

I also verified that I could see the calibrator signal was present at the front of the channel #2 mode switch, but did not take off the metal case to trace it all the way through as I did with channel #1.

The voltages I am seeing on U140 are +6.5V on pins 1,13,14&16 (input from U120), +5.7V on pins 2&3, +11V on pins 5,6,8&9, and 8.8V on pin 7 (which looks wrong, from what I can tell of the schematic in the service manual).

I'm worried that U120 (and U220) is dead and will need to be replaced. I have not been able to find replacement 155-0085-00s, but I did find a source for 155-0085-01s. Can I replace one with the other?

Before I go down that road, however, I'm wondering what else I should be checking, or if I am (likely) checking things incorrectly. I have been checking for the input signals using a Tek 2213, and checking the voltages using both the 2213 and a digital multimeter. I've been checking for the signal at .2V/div on the 2213, which was working just fine between from the BNC to the back of the mode switch, but is that sensitive enough to detect the signals going from U120 to U140?

I'm a bit at a loss for what to try next. Any suggestions would be appreciated.


 

It turns out that I managed to fix the problem: While removing the case I had brushed against the channel switch circuit (U370) and partially pulled it out of its socket. I even heard the scraping sound at the time, but though it was just the springs that ground the aluminum case to the rest of the instrument. I discovered this by going back and trying to trace the signal further along the chain of amplifiers. I'm not sure why I couldn't see the input signal before, but on this third try it was readily visible on the 2213. It may help that I read through the schematics more carefully, so I had a better idea of what pins to be testing in order to find the signal. Anyhow, once I got through the 2nd and 3rd "cascode" amplifiers I rapidly traced the signal to the channel switch, and then the problem was obvious.

After reseating it (which was not the easiest thing to do, I'll tell you) everything seems to be working again.

Now that I've lost my 'scope virginity, I would like to address some of the less pressing problems with this scope. In the service manual is clearly says that the pots are sealed and don't need to be lubricated, but I'm not sure that they were thinking about the 50 year service cycle when they wrote that. Three of the pots are very stiff, and one of those is quite "scratchy." I would like to lubricate them, but I'm not sure what that would involve. I can see that there are 4 screws on one end of the pots, so I guess I could open them up (after desoldering them?) and re-lubricate them, but what would I use? Silicone grease? Lithium grease? WD-40? Or do I even need to go to that length; can I just spray them with WD-40 and be done with it?

I also need to clean the contacts on the channel #2 attenuator switch, but I've seen plenty of posts about how to do that (100% cotton bond paper and anhydrous isopropyl alcohol).

I also found out that, in fact, my scope, despite having a fairly early serial number, actually has 155-0085-01s for the 1st "cascode" amplifier chip, so that answers any question I had about replacement with that same part number.

Also also, I'd never heard the term "cascode amplifier" before this. You learn something new every day.


Simon
 

The leaf contacts in the attenuators can be cleaned with isopropyl alcohol. You might get away with brushing them lightly with isopropyl alcohol (IPA) using a fine paint brush, but the recommended way is to soak a small strip of paper in IPA and slide it under the open contact, close the switch and slide the paper back and forth. You need to do this on the AC-GND-DC contacts as well. The attenuator blocks (6 pins) can be pulled carefully out of the sockets on the circuit board and it is worth cleaning the pins with a cotton bud before re-inserting them, and drop a little IPA on the sockets. Do them one by one so that you don’t get them mixed up !. You have to pull them out to gain access to all the leaf contacts.
You might find that pulling transistors out of their sockets and cleaning them in the same way helps stability. I had some problems with the horizontal amplifier, which was cured by doing this. After 50 years corrosion can build up in the transistor sockets.
Simon


John
 

I would start by checking all of the power supply rails including +5V, +15V (reg and ureg), +50V, +110V and +105V/+160 unreg, as well as -8V and -15V. Sometimes the Tantalum caps fail with a short. Occasionally you get problems with the large cans. If you have problems with these supplies than voltages at other parts of the circuit might be incorrect. The vertical pre-aps, for example, require a -8V feed. A failure in any of these power rails might affect several circuits.


 

tenareze32 wrote:
You might find that pulling transistors out of their sockets and cleaning them in the same way helps stability.
I had some problems with the horizontal amplifier, which was cured by doing this. After 50 years corrosion
can build up in the transistor sockets.
That would explain why (accidentally) re-seating the channel switch seems to have cleared one of the minor problems I observed. I was having this odd cross-talk issue where, when you had one channel reading a square wave and set to one of the highest sensitivity range, you would get "noise" on the other channel that was synchronized to the rising and falling edges of the square wave. The noise was unaffected by the sensitivity setting of the other channel which led me to believe that it must be happening after the both channels are "mixed" going to the display.

After I re-seated the channel switch IC (because I accidentally knocked it out of its socket while disassembling the scope) that problem seems to have almost entirely gone away. I will consider giving the rest of the scope the same treatment, but the biggest fix is going to be cleaning the channel #2 attenuator switch contacts, as that is the most debilitating issue remaining with the scope (basically none of the attenuator positions below .1 V/div work reliably, which I can live with, if I'm careful, but it's easy to stumble over when you're concentrating on other things).


 

John wrote:
I would start by checking all of the power supply rails including +5V, +15V (reg and ureg), +50V, +110V and +105V/+160 unreg, as well as -8V and -15V.
While the scope seems to be working just fine, this IS the biggest worry in the back of my mind, so I am planning on doing this while I'm in there. I did have a careful look at all the caps while I was in there this time, as I was fully expecting to find other bad caps than just the small one on the sweep board, but everything else looked okay. Certainly none of the tantalum caps had burned up, and none of the other electrolytics had leaked and corroded anything.

I've got a list of the all the other electrolytics in the scope and will be ordering replacement in the next few weeks. I haven't made a list of the tantalums yet (it will be a hell of a lot more work than the electrolytics were), but I also don't know what to replace the tantalums with once I have such a list. I got burned a couple times by the capacitor plague of the late-90s/early-00s and I'm as wary of modern tantalums as many people seem to be of the old ones.

Is there specific advice on what to use to replace the old tantalums?


 

tanareze32 wrote:
The leaf contacts in the attenuators can be cleaned with isopropyl alcohol. You might get away
with brushing them lightly with isopropyl alcohol (IPA) using a fine paint brush, but the recommended
way is to soak a small strip of paper in IPA and slide it under the open contact, close the switch and
slide the paper back and forth. You need to do this on the AC-GND-DC contacts as well. The attenuator
blocks (6 pins) can be pulled carefully out of the sockets on the circuit board and it is worth cleaning the
pins with a cotton bud before re-inserting them, and drop a little IPA on the sockets. Do them one by
one so that you don’t get them mixed up !. You have to pull them out to gain access to all the leaf contacts.
So I followed this exact procedure for the attenuator switch for channel #2, but this did not fix the problem, in fact it might have gotten worse. What I have now is an attenuator system that makes good contact on only half of the lower range settings (the other half show, instead of a nice square wave, up and down spikes that rapidly decay back to the ground voltage at the square wave's rise and fall points), and the other half register an order of magnitude lower sensitivity than they are marked for (e.g. the 50mV range registers as if it were .5V).

Obviously this means that one or more of the attenuator blocks are not being switched in or out properly (or that one of the attenuator blocks is faulty). I will need to review the schematic to figure out which one (or more) is implicated by this failure mode.

I also reseated about 3/4 of the transistors on the vertical amplifier board. Some resisted being removed from their sockets more than I was willing to force them, and others were blocked by other parts or cables (which I was also unwilling to force). The leads on everything that I managed to remove were nice and shiny, so I'm not going to sweat it until I see a specific failure pattern that implicates one or more of these transistors. I also checked that the logic chips had not crept out of their sockets, but they all seemed very securely seated.


John Gord
 

Jeff,
A known source of problems in the attenuator blocks is failed pin-to-hybrid solder joints. You can unplug the attenuators and then carefully remove the plastic lids to inspect and, if necessary, reflow the joints. I don't remember if it requires a special flux or solder type.
--John Gord


 

John Gord wrote:
A known source of problems in the attenuator blocks is failed pin-to-hybrid solder joints. You can
unplug the attenuators and then carefully remove the plastic lids to inspect and, if necessary,
reflow the joints. I don't remember if it requires a special flux or solder type.
Oh dear. That sounds like a whole new level of difficulty.

I did unplug the 100x block (because it looked that might be the culprit from the readings I was seeing, and from what I saw in the schematic) and was checking the resistance between the different pins when I noticed that one of my probes was in 1x mode while the other was in 10x mode (DOH!). I replaced the 100x attenuator block, powered up the unit, set both probes to 1x, did my tests again, and the differences in readings between channel #1 and channel #2 were gone.

There's still some problem with the switch contacts in some positions on channel #2, but it's definitely improved. The resistances I read on the block pins seemed to make sense (I didn't write them down, sadly, before I noticed that I had the probes set up wrong), so I don't think that any of the attenuator blocks have failed. I may need to go in and clean the contacts again, though.


Ed Breya
 

I don't know about the 475 front ends, but in the 7K ones they use what sounds like the same kind of contacts, operated by the cam. If the 475 is the same, then there are likely contacts on both sides of the attenuator board. The ones between the attenuator tubs are usually the easy ones, for access - they select the tub through-connections. The ones on the other side of the board are the bypass contacts that pass the signal straight through if all are closed. Each attenuator section has four contacts associated, to either select it or bypass it. The contacts are arranged opposing in pairs, on opposite sides of the board, and have integral plungers pins molded into the plastic carriers. When the cam lobe pushes the back side contact to close, the opposing one is pushed open. Even on the 7K plug-ins, which are fairly open in construction, the back-side contacts are difficult to access, and some disassembly is needed to do a proper cleaning. I would imagine this would be even harder to do on a 475. I have two 475s, but haven't seen their insides in many years.

Ed.


John
 

I wouldn't worry too much about all the Tantalum caps in the scope. There are rather a lot of them. You are unlikely to see them burned up. They just short internally. There are a handful in the power supply circuit between the various power rails and ground which is why I mentioned them and these can drag a line down if they are shorted. They are easily checked them with a multi-meter and can be replaced equivalent or nearest value electrolytic.

It is quick and easy easy to check the power supply outputs with a DMM. In approximately the centre of the main board, in between the bridge rectifiers and the cluster of 4 large metal can transistors, there is a group of test points with the voltages marked on the PCB. Page 171 of the pdf shows the board layout and page 172 the circuit diagram.


 

John wrote:
It is quick and easy easy to check the power supply outputs with a DMM. In approximately the centre
of the main board, in between the bridge rectifiers and the cluster of 4 large metal can transistors, there
is a group of test points with the voltages marked on the PCB. Page 171 of the pdf shows the board
layout and page 172 the circuit diagram.
Yes, I saw those on my first pass. They are quite well marked with what voltages should be present. There are also five unmarked test points, one next to the INVERT switch, and the other four behind the A TRIG HOLDOFF pot.

I measured all the test points against the test points marked GND (which I had previously tested for continuity). The only standout was the UREG +50V point which read +68.5V, but I figure that's what "UNREG" means. Everything else was within 1% of it's marked value (the regulated +50V was well within 0.5%).

The five unmarked test points read as follows:

next to INVERT switch: +34.06V
immediately behind A TRIG HOLDOFF: +14.64V
2nd back from A TRIG HOLDOFF: -2.411V
3rd back from A TRIG HOLDOFF: -3.776V
4th back, next to UNREG +50V: +137.9V

Until I find these in the service manual I have no way to know what they should read.

I did NOT check the -2450V point. I do not have a high voltage probe handy (though I probably have one somewhere). I figure that the CRT appears to be working correctly, so I can just assume that the high voltage power supply is working correctly (until I have a high voltage probe).

I have not checked ripple, but I plan to do that with my 2213 later today.


 

I checked with the 2213 and there is basically no discernible ripple on any of the labeled test points (<< 0.1V on all but +50V and +110V, < 0.5V on +50V, and < 1.0V on +110V) except for UNREG +50V which has ~2V trianglar ripple at 60Hz.

The unlabeled test points are a little bit more interesting: +14.64V and -3.776V have no discernible ripple, but -2.411V is actually an asymmetrical square wave with a period of 8ms and a negative pulse of 1.8ms, +34.06V has a 0.2V ripple at ~1.2KHz, and +137.9V has an 8V sawtooth ripple at ~120Hz.

My father took meticulous care of his instruments, and this one spent the majority of the past 40 years in a cool (~65°F) dehumidified basement. I did try to "bake" it dry about 15 years ago in an un-air-conditioned room with a dehumidifier, but I doubt that made any appreciable difference. I'm still surprised that it's has as few problems as it evidently has. I had fully expected all of the electrolytic caps to have gone dry, if not (as with the cap on the sweep board) leaked their guts out.


 

Ed Breya wrote:
I don't know about the 475 front ends, but in the 7K ones they use what sounds like the same kind of
contacts, operated by the cam. If the 475 is the same, then there are likely contacts on both sides of
the attenuator board. The ones between the attenuator tubs are usually the easy ones, for access -
they select the tub through-connections. The ones on the other side of the board are the bypass
contacts that pass the signal straight through if all are closed.
There are indeed contacts on the underside of the vertical attenuator boards, but it looks like you actually have to unsolder both the BNC connectors and the connection to the main vertical amplifier board in order to remove the attenuator boards, so I'm not going to do that. Instead I was able to slip a narrow strip of bond paper in through the side and clean most of the contacts that way. There were a couple that I just couldn't get to because they were blocked by the socket housings for the attenuator blocks, but I got to more than half of them, and that seems to be have been good enough. The dirty contact problems on channel #2 seem to be gone, so I'll take that as a win.

I probably need to adjust the horizontal timebase, since I replaced a capacitor on the sweep generator board, and because I have suspected that the timebase may have drifted all along (of course, maybe that was due to the bad capacitor, so that might have been fixed), but I don't have any of the equipment needed to do a real calibration, and the timebase isn't off by enough to make gross adjustments worthwhile.

I also wanted to trim up the probe calibration signal, because it seems to run just slightly fast (or the timebase is slightly slow), but it looks like there is no way to adjust the calibrator frequency, only the amplitude; the square wave period seems to be set by soldered in component values, so I'll let that be as it is (until and unless it becomes a real problem).

So I guess that this is a success. I have regained full function on the instrument, and comparing its readings to those from my multimeters and from the 2213 at least give me good bounds on how out-of-cal the 475 actually is.

Does anybody know how much it costs to have a scope this old professionally calibrated?


 

On Fri, Nov 6, 2020 at 04:57 AM, Jeff Dutky wrote:


I probably need to adjust the horizontal timebase, since I replaced a
capacitor on the sweep generator board, and because I have suspected that the
timebase may have drifted all along (of course, maybe that was due to the bad
capacitor, so that might have been fixed), but I don't have any of the
equipment needed to do a real calibration, and the timebase isn't off by
enough to make gross adjustments worthwhile.
My memory urges me to write the following:
If you replaced one of the physically larger timing capacitors (metal cylinders, the smallest about 2cm length), be aware that these are matched sets (I think 3 pcs, marked by a single letter, like "J", "K" or the like, the same for each set). At calibration time, you can't adjust their timings individually.

Raymond


 

Raymond wrote:
My memory urges me to write the following:
If you replaced one of the physically larger timing capacitors (metal cylinders, the smallest about 2cm length),
be aware that these are matched sets (I think 3 pcs, marked by a single letter, like "J", "K" or the like, the same
for each set). At calibration time, you can't adjust their timings individually.
The cap that I replaced is labeled C1059 and is the smallest capacitor on the Timing Circuit Board. In the parts list it is indicated simply as "3.6 uF, Elect., 125 V, 10%" so I don't think it's part of the matched set. There are two sets of capacitors on that same board (C1071,72,73 and C1082,83,85 that are specifically marked in the parts list as matched capacitors of 10uF/0.1uF/0.001uF) but this is not one of them.

Thank you for the warning, though; it is well taken.


 

Okay, I'm not nearly qualified to understand what C1059 was doing in the sweep generation circuit, I mean I can trace out what it's connected to, but my EE fluency ends with "which way to Kirchoff Cafe" and "My hovercraft is full of eels." The failed part still appears to act like a capacitor (in the 200nF range), at least from what my multimeter can tell (I don't have an ECR meter or a working comparator bridge to check it any real way). Whatever the case, it was clearly the culprit in disabling the horizontal sweep. I suppose it also may have been distorting the timebase, but even before this failure the timebase wasn't off by more than a few percent (if that, as I was judging the timebase by comparison to the built-in calibrator signal, which appears to be off by 5% or more), and if C1059 played any significant part in the timebase itself then I would expect a change of an order of magnitude in C1059's value to have more than a 1% effect on the timebase.

As it is, I checked the value of the replacement capacitor to make sure that it was within 10% of the specified value. I even ordered several replacement parts so I could select the one that was closest to 3.6uF, and both the candidates were right about 3.5uF, so I feel pretty good about the transplant not screwing up the host circuit.


Göran Krusell
 

Hi, the R1059 and C1059 low pass filter is filtering the +110V voltage, the values are not critical.
Göran


Tom Lee
 

Minor nit (sorry to be picky): It's double-h, double-f in Kirchhoff.

(can't help it; it's an occupational hazard)

-- Cheers,
Tom

--
Prof. Thomas H. Lee
Allen Ctr., Rm. 205
350 Jane Stanford Way
Stanford University
Stanford, CA 94305-4070
http://www-smirc.stanford.edu

On 11/7/2020 01:57, Jeff Dutky wrote:
Okay, I'm not nearly qualified to understand what C1059 was doing in the sweep generation circuit, I mean I can trace out what it's connected to, but my EE fluency ends with "which way to Kirchoff Cafe" and "My hovercraft is full of eels." The failed part still appears to act like a capacitor (in the 200nF range), at least from what my multimeter can tell (I don't have an ECR meter or a working comparator bridge to check it any real way). Whatever the case, it was clearly the culprit in disabling the horizontal sweep. I suppose it also may have been distorting the timebase, but even before this failure the timebase wasn't off by more than a few percent (if that, as I was judging the timebase by comparison to the built-in calibrator signal, which appears to be off by 5% or more), and if C1059 played any significant part in the timebase itself then I would expect a change of an order of magnitude in C1059's value to have more than a 1% effect on the timebase.

As it is, I checked the value of the replacement capacitor to make sure that it was within 10% of the specified value. I even ordered several replacement parts so I could select the one that was closest to 3.6uF, and both the candidates were right about 3.5uF, so I feel pretty good about the transplant not screwing up the host circuit.




 

(Dr.) Tom Lee wrote:
Minor nit (sorry to be picky): It's double-h, double-f in Kirchhoff.

(can't help it; it's an occupational hazard)
but you let "ECR meter" slide?

The worst part is that I looked it up before I posted that, and I STILL spelled it wrong.