Tektronix 468 Repair

[willmac]

Originally Posted by RogerEvans

Just to the left of the CRT is a vertical group of five pushbuttons, the second one down is labelled 'A trig view'. This is supposed to show the signal that caused the A timebase to trigger, I have never used it (I don't have the option on any of my scopes!) but I think it is supposed to be useful if you are looking at a complicated waveform, maybe using also the delayed sweep capability, or triggering from an external trigger input. In this case none of that is relevant, simply that some of its circuitry is shared with the display of the stored digital signal. If you press the button you should see the same square wave but its amplitude might well change. It would be nice if someone who has a 465 (which is very similar) could confirm this, I am just working from the manual.

I think that the 'A trig view' is used to display the signal that is connected via the external trigger input. I don't think it is intended to be used with Ch1 or Ch2 input, but I will try it and see what I get. I could also take a look at what happens when I look at the calibration signal using the external trigger input. I would imagine that I would not see anything when Ch1 or Ch2 is selected, but perhaps something when A trig view is selected?

Also you implied that the LED display might well be showing the correct voltages for the unseen digital waveform. A nice check would be to see what happens to the cursor display of Volts when you switch from DC to AC coupling on the input signal. With DC coupling you would expect the LED to go from zero to the calibrator voltage (assuming that the calibrator does switch from zero to 1 volt or whatever, which is nearly always the case) whereas with AC coupling the LED should go from (minus half the voltage) to (plus half the voltage). That would show fairly conclusively that the processor, the input A2D circuitry and the memory were all OK.

The calibrator signal is 0.3V 1 kHz square wave. Also current loop, but I haven't anything that can easily check that.

Looking what happens with DC and AC coupling is a good idea - I will try that.

[WaveyDipole]

I have the 465 and have just tried it on mine using CH 1 and a CAL signal from two different scopes. The Trig View function does indeed show a square wave trigger signal and the amplitude varies depending on the signal being observed.

When fed into the External Trigger input and with triggering set to Ext, it showed an un-synchronized trace.

[willmac]

OK, some measurements based on the previous few messages.

Voltages on the digital storage board:

TP148, +12 = 12.21
TP315 +5 = 4.95
PP520 -12 = -11.82
TP518 -6 = -6.01

All of these are within the tolerances given in the manual.
The same voltages measured on the digital power supply show very similar voltages.

[factory]

Originally Posted by karesz*

Hello Willmac,

I could replace all these parts like for like

Be carfull please with changing all of tantalums, especially the Sprague metallic cylinders are practically military grade tantals_ I didnt find a failed from it in my praxis and their are after i.e. 30 years tight on specs...
The problematic types of Tantals are only the diverse coloured blob types! (as Chris told it too)
BTW: before exchange all blobs; you can touch that for temperature sensing, than I would change only ones there are some "warm"...
Be succesfull with these action!
Karl

Originally Posted by MrBungle

The military grade ones are the ones that blew up for me.

That will be the one in 465 thread here then;
http://www.vintage-radio.net/forum/s...7&postcount=17

Not all the axial metallic looking capacitors are tantalum.
Some are marked 30D and are ordinary aluminium capacitors, they are still made, the data-sheet describes them as general purpose aluminium capacitors.
http://www.vishay.com/docs/42041/30d.pdf

The ones marked 109D (like the exploded one in the linked thread) are definitely wet tantalum types.
http://www.vishay.com/docs/40023/109d.pdf

I've yet to find an exploded tantalum of this type yet, but I have plenty of equipment with them in, so it could happen. I've changed quite a few of the bead type though, they seem much less reliable.

David

[willmac]

I have probed the raw output from the 55V supply on the leg of the smoothing capacitor. I get a pronounced sawtooth shaped trace. I was expecting something a lot smoother than this so perhaps this is a pointer to an electrolytic smoothing capacitor that has failed, or maybe a bad bridge?

I tested the 55V at the test point and got the same results as before on my scope. It could not trigger on the signal. I switched to 'Line' trigger, and although the trace looked a little different, it still did not trigger.

I haven't tried adding capacitors or adjusting the 55V pot yet. However the ripple looks to me like high frequency semi random noise. Is it possible for the 55V power supply to become unstable, and if so, is this what it would look like?

The A Trig view does work as expected. If I take the calibration signal and feed it into the external trigger input, then select the A trig view, I get a relatively clean square wave, correct amplitude and frequency. If I also connect Ch1 to the calibrator and display that at the same time, I get both traces superimposed. The fuzzy square wave aligns perfectly with the cleaner A trig view trace.

Looking at the 468 calibration signal on my D75 scope I can see that it is indeed 1kHz, 300 mV. With DC coupling I can see that the bottom of the square wave is offset about +160mV from ground. It does not go +- about ground, only a positive offset. On the 468 in digital mode with the same calibration signal I see a variation from about 1-2 mV up to about 290 mV as I move the cursor (which isn't visible). The values seem to be roughly the same with both AC and DC coupling, but this is a very difficult measurement to make since you can't see what you are doing. This however does support the idea that a voltage is actually being measured and a waveform has been captured.

As I wrote this I realised that I may have a silly error in my measurement technique because I can't see anything being displayed on the screen. I should probably have saved the digital waveform twice, first with DC coupling and then after looking at the voltages again with AC coupling. That would account for the voltages looking roughly the same (if they are voltage readings). I will try again when I have time, probably in a coupe of days.

[karesz*]

Dear David,

Not all the axial metallic looking capacitors are tantalum.

I agree, but what I wrote was explicit over "Sprague metallic Tantalums": what I mean; their are mostly the solid electrolytic, practically military ones, and are seldome problematic...
Regards - Karl

[RogerEvans]

The sawtooth on the unregulated side of the 55V supply is normal, as long as it doesn't drop below 55V plus a margin for the regulator to work properly, so 59-60V would be a safe lower limit.

The feedback circuit for the 55V contains a high gain op-amp so it might become unstable, which is why I suggested adding a capacitor on the output side in case C603 was failing. You would expect an unstable circuit to have a fairly periodic waveform, old op-amps had unity gain-bandwidths of no more than 1MHz so oscillation would be unlikely at any higher frequency.

It should be possible to find a connector to pull that would remove the load circuits from the 55V. If the noise disappears it is not coming from the regulator.

What amplitude is the noise on the 55V supply?

The relatively clean signal on the A trig view is interesting, it suggests that the noise on the normal trace is coming from the input attenuators or pre-amp circuitry rather than the main Y amplifier. Do channel 1 and channel 2 have the same amount of noise? What happens for alt, add and chop?

The reason for asking about the various voltages to the digital board was simply that it could have been a simple reason for the lack of a digital display. The output from the digital board back to the analogue display has lots of op-amps with +/- 12V supplies. If you lost either of those the processor and memory would probably still run but the analogue outputs would be lost.

Roger

[ColinTheAmpMan1]

Originally Posted by willmac

I have probed the raw output from the 55V supply on the leg of the smoothing capacitor. I get a pronounced sawtooth shaped trace. I was expecting something a lot smoother than this so perhaps this is a pointer to an electrolytic smoothing capacitor that has failed, or maybe a bad bridge?

I tested the 55V at the test point and got the same results as before on my scope. It could not trigger on the signal. I switched to 'Line' trigger, and although the trace looked a little different, it still did not trigger.

I am currently working on my 468, which works fine in analogue (non-storage) mode, but the first cycle or two get displayed very jittery in storage mode. I have got around to looking at the storage power supply board ("schematic 21" in the service manual; board A19 to confuse matters) and the waveform at test-point #166 (+5V) is of the right shape, but has a lot of HF noise on it. I can get the trace to trigger by using the "AC HF REJ" trigger coupling on a 7B53A time-base in a 7623A mainframe. If you have a similar coupling possibility, that might work.
I am thinking that I should replace C162 (390uF, 15VDC) and C175 (540uF, 25VDC) but this board is a pig to access. Come to that, the manual is not one of Tek's best, either.
Good Luck to both of us, Colin.

[RogerEvans]

You could add another capacitor (in parallel with C162) somewhere where you can get at +5VS and ground, as a temporary measure. I don't know just how inaccessible is the area around C162, but if as appears from the manual it is a wire ended component, you can snip the leads of C162 close to the body of the capacitor and solder in a new, and usually much smaller, replacement from the top side of the board. Some of the high current power supplies have such large ground planes that they are a right pain to get hot enough to remove the old components anyway.

Roger

[willmac]

Originally Posted by ColinTheAmpMan1

I can get the trace to trigger by using the "AC HF REJ" trigger coupling on a 7B53A time-base in a 7623A mainframe. If you have a similar coupling possibility, that might work.
I am thinking that I should replace C162 (390uF, 15VDC) and C175 (540uF, 25VDC) but this board is a pig to access.

I will check using the 'AC HF Rej' triggering on my other scope. I think I have already tried it, but I can't be sure.

The main board that runs underneath the body of the scope carries the electrolytic caps for the main power supply and a lot of tantalum caps as well. I don't think I can change these parts with my level of skill without removing the board first, and that means a total strip down of more than half of the scope. Definitely an access problem.

It may be we have one smiliar problem with noise. Will be interesting to see how that works out. I think my problem with no storage display is a bit different to yours.

[willmac]

Picking up some loose ends on the evaluation of the 468 scope.

Using the AC HF Reject triggering did not help to trigger on the noisy 55V at the test point. Still just very fuzzy. I'm not sure what the filter spec is on my D75 for the low pass filter, I will have to look it up.

The sawtooth on the unregulated side of the 55V supply is not quite as bad as I thought. The peak to peak voltage of the sawtooth (top of it to bottom of it) is about 1.2V. The DC voltage of the supply at this point is about 74.8V which is higher than I expected, but my DVM agrees with my scope more or less.

Ch1 signals are worse than Ch2 signals for noise and both are worse than the A trig View. If I look at the noise using the calibrator signal I seem to see several components. There is a ripple visible at the top and bottom of the square wave. Each horizonal bar of the square wave signal (top or bottom) has about 20+ ripples, and the calibrator square wave frequency is approximately 1kHz. There is a much lower frequency imposed on this ripple, at maybe 2kHz. There is also what may be higher freqency noise that makes the whole trace indistinct and fuzzy.

Looking at the 468 calibrator signal with my other scope, I see some ripple on that, but nowhere near the noise I see through the 468 Ch1 or Ch2.

My rather uninformed conclusion is that I probably have some power supply based noise as well as some problems in the Ch1 and Ch2 attenuators/amplifiers etc. This is as well as the yet unexplained lack of any digital signal display.

I still haven't tried added capacitance to try and reduce power supply noise yet, but I will do that.

Thoughts and comments would be very useful.

[RogerEvans]

The 468 has a switching buck regulator to generate the +5VS on the storage power supply board from an 18V input. Test point 165 should give you a good measure of its switching frequency. Is that the same frequency as the ripple on the Ch1, Ch2 traces? Also put the scope on test point 166 which is the +5VS after the regulator and see if there is much noise and ripple at that point.

If you change the vertical deflection factor does the ripple stay constant in divisions or constant in mV?

What happens if you ground the input selector? You might need to increase the sensitivity to get the scope to trigger or even trigger externally from TP165.

The 468 has a Ch1 out connector. Can you connect that to your second scope while the Ch1 input on the 468 is on the calibrator. Is the result the same as on the screen of the 468?

Is the Trig View signal on the 468 similar to or noisier than, looking at the calibrator output with the second scope?

A lot of questions but the problem is identifying the sources of the ripple and other noise and where they might be getting in to the signal chain.

Finally a trivial one - wiggle the vatiable Volts/Div knobs several times in case the tracks have become noisy. I can't see in the manual that they are switched out of circuit when in the Cal position.

Roger

[MrBungle]

Just a note. Some of the older 400 series scopes have pretty fuzzy traces. My 475 is terrible.

[PJL]

I am not too clear what the faults are from reading the thread.

First step is to check all the power rail DC voltages which I believe you have already done and all are in spec. The tants are generally used for rail decoupling and go dead short and either explode or stop the scope from working at all. I would not get too concerned about general rail noise until you have evidence that this is the problem.

I suggest we focus on the non-storage functionality at the moment. Can you summarise each fault with the analogue functionality and provide example pictures of the traces you see on screen. Once that is working you can move on to the much more complex storage section with the knowledge that the vertical amplifier, timebase and triggering are all working.

[willmac]

Major progress with the 468 today. You should know that I am a rank amateur at electronics and I have just had this very firmly emphasised to me.

I had been thinking about why I was seeing such a lot of noise on this scope, both on the screen and whenever I probed in it. I was really puzzled since although there was clearly something badly wrong, it wasn't obvious what was causing it. The nature of the noise was also very difficult to analyse since it seemed to consist of a whole mixture which I couldn't readily seperate on my other scope.

I was doing this work in my main workshop which is used for machining with lathes, milling machines etc. Most of these are inverter based and there are numerous other bits of industrial electronics around. The workshop has a very high level of lighting provided by a lot of multiple double tube florescents fitted with digital ballasts.

I think you can probably guess where this is going...

On a hunch I switched off the lights and some other equipment. Suddenly I had clear stable traces on the screen of the 468 which responded as I think they should to the controls. In order to probe voltages etc, it is obviously necessary to have the main case removed. I didn't think about the possible consequences of this, indeed it was not a problem when I repaired my D75. I experimented with some sheet steel to find out where the RFI was getting in and was not too surprised that the vertical amplifier board is incredibly sensitive to RFI. A few other areas as well. I don't know if it should be like this, but now that I know, I can make some temporary shielding while I fix the other problems.

Is this a common issue?

I must apologise for wasting peoples time analysing the noise problem - I should have thought about this much earlier. Anyway, I will repeat all the measurements with some better shielding and let you know the new status. Hopefully I will be ready to move on to the digital storage display issue soon.

[MrBungle]

This always gets me and I forget every time

You'll find the noise better with the cover on.

My first Tek 465B I spent a good two hours hunting for a noise issue which went away when I put the cover on.

[PJL]

Test the analogue functionality is fully working first.

[willmac]

Yes, I will do this. Even if I can't get the digital side working properly, the scope still has value to me as a purely analogue instrument. For that I need to have confidence in the analogue functionality.

[RogerEvans]

I certainly don't regard any time spent trying to help as time wasted. I hope you enjoy your hobby as much as I do and get your 468 working to your satisfaction.

Roger

[Craig Sawyers]

Originally Posted by willmac

On a hunch I switched off the lights and some other equipment. Suddenly I had clear stable traces on the screen of the 468 which responded as I think they should to the controls.

Is this a common issue?

Yes, indeed it is. It is a problem with compact flourescent lamps, and LED lighting too. Both are supplied by switched mode supplies and radiate like a radio antenna. I have no idea how on earth these get though regulatory EMC approval.

When I need to make high sensitivity measurements, I have to turn them all off, and also the central heating - the Honeywell controller uses a switching supply too.

And these are just the obvious sources. If you put WIFI; digital cordless phones; laptops, tablets and desktop computers; mobile phones; broadband over power line, and HDTV over power line, are all sources that you can pick up when trying to use a spectrum analyzer (or for that matter a sensitive radio receiver).

How pervasive this is was brought home when I was supervising a major EMC test programme at a test house. The product was in the massive sealed EMC chamber. The test engineer piped up "Er - who's using a mobile phone?". Turned out one of the other of my engineers was using her phone in the corridor outside. I generated a clear signal on the scanning frequency analyzer.

Craig