Tektronix 2235 prblems

[Delillusions]

Hi!

Sorry for troubling you guys, but I would like to ask you some question about a Tektronix 2235 100MHz Analog Oscilloscope. I got it little more than a year ago and it worked like a charm until now, but recently it started acting somewhat strangely.

#1 There is a high pitched noise coming from the Scope. There's "nothing wrong" with it, except, that I don't like that high pitched sound and I don't think it should give out such noise. It's most likely from the PSU and some caps are giving up.
#2 Using the Beam Find I got a small line and not a dot. According to some service manuals it should be the "Unblanking Logic". Besides the Beam Find the horizontal deflection works great. The scope it still quite accurate.
#3 This problem is maybe the most tricky. The projected wave is getting squished horizontally, so it's getting smaller. Today I used the Scope for some extensive debugging and I used it for 4-5 hours constantly. At the end the wave got squished horizontally.  It turned it off, and I noticed that the Scope was somewhat warm. It sits on the top of my table, nothing restricts the "airflow". It wasn't hot, but somewhat warm. After the Scope cooled down it worked like a charm again.

For #1 I know where to look and I have some idea about what could cause it. For #2 I somewhat know where to look, but I'm all ears for any idea. But #3 is another thing. As long the scope is nice and cool there is nothing wrong, but after it warms up (at least 4 hours at room temperature (it's around 15-17 Celsius at my "workshop") the problem appears.
I have a job to finish with a deadline that's getting closer so until I finished I don't have that much time to take it apart, but most of the work with the scope is done. Until then I'm all ears for what should I look for as the source of the problem.

Thank you for your help in advance!

[grumpydoc]

#1 common on the 22xx scopes - yes it is almost certainly coming from the power supply. Might not be a fault and probably isn't the caps themselves. If you want to "re-cap" the power supply then C906, 925, 943, 942, 961-3, 968, 970, 954 and 956 are the "usual suspects". Some are slightly odd values - 75uF can be replaced with 100uF (but watch the voltage rating on C906) and 840uF can be replaced with 1000uF.

#2 Normal

#3 Could be the anode HT drifting up (which makes the electrons accelerate more and therefore the beam is harder to deflect) or the horizontal amplifier output having a problem. Start with measuring the power supply voltage when warm - there are a set of jumpers between the B timebase board (the one which sticks up vertically just in front of the PSU screening) and the PSU where you can check all the rails.

[Delillusions]

Thank you for your help grumpydoc!
I will certainly look into those parts

Best regards!
Gábor

[markce]

Had a #1 on a 2236. Did a recap of the power supply, it was gone...

[David Hess]

If the power supply capacitors are worn out then there may be enough ripple to cause the beam finder to display as a horizontal line.  The power supply output voltages could also be low explaining the horizontal scale problem.  I would at least check the power supply before looking for other causes.

The beam finder on my 22xx oscilloscopes displays as a bright unfocused dot if there is no signal, no sweep, and no readout to be displayed.

[oldway]

A high pitched noise varying with brightness ajust and overheating is coming from a defective (internal short circuit between turns caused by isolation failure) inductor T906 (120-1439-00 or 01) in the power supply.
For this reason, this defective inductor is no more able to store energy during on state of the mosfet, as it should.
Current in the Mosfet Q9070 became then abnormal with very short on state and high peak value and dissipated power increase. (reason of overheating of the Mosfet)
That's a commun failure of 2235.
It's not occurring with 2236 because it has a fan and the inductor works at lower temperature (better cooling of the power supply)
I installed a fan in my 2235 for this reason.

[oldway]

Had a #1 on a 2236. Did a recap of the power supply, it was gone...

Only the high pitched noise was gone (coming from high peak current in the capacitors), but the true problem is still there.
It is probably a 2236 who has worked for a while with defective fan.

[grumpydoc]

The beam finder on my 22xx oscilloscopes displays as a bright unfocused dot if there is no signal, no sweep, and no readout to be displayed.

If there's no sweep then, yes it should be a dot - if there is a sweep (for instance, when the trigger mode is auto) then it will be a line maybe 3-4 divs long - probably with a brighter dot at the start. I kind-of assumed this was the situation.

If it is a short horizontal line <= 1 div then it might just be that the focus adjustment is off or perhaps the astigmatism or geometry presets need adjusting.

[markce]

It is probably a 2236 who has worked for a while with defective fan.

Hi oldway, the original fan was there, and it worked. High pitch noise was gone after a complete recap, also on primary of smps. Also modified the fan because it was an early one with asymetric loading on the pws.
Which windings are likely to have an insulation  short?

[oldway]

I already had three 2235 oscilloscopes with this problem.
I have not disassembled and uncoiled the defective T906 inductors to find which winding presented the default because I did not see the usefulness of this information.
The inductance is not repairable anyway and it seems quite difficult to make one yourself because it is very compact.
NB; The ferrite core must have a gap.

The best way to verify if T906 is defective or not is to feed the scope through an isolating transformer and to look with another scope if the waveform of current on R907 (0.2R) is correct or not. (see waveform 39 - 150mV peak)

[Delillusions]

Hi!

I'm back again. I've finished with my job, so now I and my 2235 have some "free" time for each other. I disassembled her and she's currently sitting on my workbench while completely exposed. Now I'm waiting for her to warm up. While she's cool I mad some measurements and all of her power rails are spot on except one. The -8.6V rail is actually -8.55V The Service manual show me, that it should be between -8.56V and -8.64V, so it missed the target by 10mV.

I looked at the caps too and they looks OK, but they are quite old, so I'm going to change them. When I'm finished with the warm power measurements I'm going to measure the sizes of the caps. They are no good, if they can't fit in.

Just in case I want to ask something. Do you think, that the +8.6V or the +5V rail can handle the power requirement of an extra small (~8cm) PC fan? I'm pretty sure they can, but just in case I want to know your opinion.

Well I let her run for almost 7 hours and she still didn't showed me that squished screen. She showed me that squished screen only once, after that I made sure, that she's not going to warm up that much. I suspect, that now she couldn't warm up because she was completely exposed. Tomorrow

During the test I regularly measured the power rails and all of them was spot on, and after 20 minutes or so the -8.6V rail( the one that missed the lower limits by a hair) increased and stayed within the limits, around -8.62V.

Thank you in advance!
Gábor

[David Hess]

A high pitched noise varying with brightness ajust and overheating is coming from a defective (internal short circuit between turns caused by isolation failure) inductor T906 (120-1439-00 or 01) in the power supply.
For this reason, this defective inductor is no more able to store energy during on state of the mosfet, as it should.
Current in the Mosfet Q9070 became then abnormal with very short on state and high peak value and dissipated power increase. (reason of overheating of the Mosfet)
That's a commun failure of 2235.
It's not occurring with 2236 because it has a fan and the inductor works at lower temperature (better cooling of the power supply)
I installed a fan in my 2235 for this reason.

I already had three 2235 oscilloscopes with this problem.
I have not disassembled and uncoiled the defective T906 inductors to find which winding presented the default because I did not see the usefulness of this information.
The inductance is not repairable anyway and it seems quite difficult to make one yourself because it is very compact.
NB; The ferrite core must have a gap.

The best way to verify if T906 is defective or not is to feed the scope through an isolating transformer and to look with another scope if the waveform of current on R907 (0.2R) is correct or not. (see waveform 39 - 150mV peak)

I have often wondered if this is a common failure in other 22xx oscilloscopes, even the ones that include fans, where the power supply becomes "finicky" and suffers mysterious unexplained failures over an extended period of time.  When troubleshooting them, suspicion falls onto the buck transformer but naturally the power supplies behave when watched and most people do not have the facility to test the transformer anyway.  Sometimes the current waveform is distorted and sometimes not but when remotely troubleshooting, it is difficult to distinguish measurement mistakes from reality and having someone work on the primary side of an off-line circuit is challenging enough.

Fortunately or unfortunately depending on how one sees it, none of my 22xx oscilloscopes have displayed this behavior which is just Murphy playing his game since I would be particularly able to get to the bottom of the issue.

[David Hess]

I'm back again. I've finished with my job, so now I and my 2235 have some "free" time for each other. I disassembled her and she's currently sitting on my workbench while completely exposed. Now I'm waiting for her to warm up. While she's cool I mad some measurements and all of her power rails are spot on except one. The -8.6V rail is actually -8.55V The Service manual show me, that it should be between -8.56V and -8.64V, so it missed the target by 10mV.

10 millivolts of difference is not going to cause the problems you are having.  The specification is tight because it affects calibration and at that level I would question the calibration of the multimeter if it is a cheap one.

I looked at the caps too and they looks OK, but they are quite old, so I'm going to change them. When I'm finished with the warm power measurements I'm going to measure the sizes of the caps. They are no good, if they can't fit in.

I usually change all of the capacitors if I find or suspect a bad one on the theory that the rest are approaching the end of their usable life anyway and select replacements based on form factor first and capacitance second.  Higher voltage capacitors can always be used and will tend to last longer anyway.  Of course I use low impedance high temperature capacitors intended for switching applications like the Nichicon PW series but there are lots of options there.

Just in case I want to ask something. Do you think, that the +8.6V or the +5V rail can handle the power requirement of an extra small (~8cm) PC fan? I'm pretty sure they can, but just in case I want to know your opinion.

You may want to check the 2236 schematic which shows that the power is definitely available as the fan was effectively driven from the -8.6 volt and +5 volt supplies making 13.6 volts.  I would not use the -8.6 volt supply directly though to prevent noise from being introduced.

[oldway]

it is difficult to distinguish measurement mistakes from reality and having someone work on the primary side of an off-line circuit is challenging enough

You are kidding ?
To work on the primary side of a SMPS using an isolating transformer is challenging ?
I do this everyday....
Measurement mistakes ?
If someone is not able to use a scope and compare with pictures of the service manual, better not to try repair a scope.

[David Hess]

it is difficult to distinguish measurement mistakes from reality and having someone work on the primary side of an off-line circuit is challenging enough

You are kidding ?
To work on the primary side of a SMPS using an isolating transformer is challenging ?
I do this everyday....
Measurement mistakes ?
If someone is not able to use a scope and compare with pictures of the service manual, better not to try repair a scope.

It is challenging when you are not the one doing it personally.  Besides the safety issue of trying to walk someone through off-line side measurements, getting uncorrupted measurements when a bad (or good) inductor is saturating and/or spewing flux all over the place can be challenging even if you know what you are doing.  Oddly enough modern DSOs tend to make this worse for the unwary when they emphasize the noise over the signal.

[oldway]

When inductor T906 (120-1439-00 or 01) is faulty, waveform of current on R907 (0.2R) is very different from the correct value.
It is not challenging at all, there is no way to make a mistake.

[Delillusions]

When inductor T906 (120-1439-00 or 01) is faulty, waveform of current on R907 (0.2R) is very different from the correct value.
It is not challenging at all, there is no way to make a mistake.

I agree, that it's quite hard to make a mistake, but it's also quite hard to measure it, since this is my first and still my only oscilloscope. Not to mention, that I don't have access to another one.

[David Hess]

When inductor T906 (120-1439-00 or 01) is faulty, waveform of current on R907 (0.2R) is very different from the correct value.
It is not challenging at all, there is no way to make a mistake.

I have seen waveforms from good ones that worked fine and I have seen waveforms from bad ones that failed either immediately or eventually but I have also seen waveforms from good ones that failed regularly at least when not being measured.  My hypothesis is that the transformer develops an intermittent short but I have no direct evidence of this yet and I find it difficult to believe that it could temporarily heal itself.

None of my 22xx oscilloscopes have displayed this problem and I have never gotten my hands on one that did to investigate it personally.

[markce]

You may want to check the 2236 schematic

Please use 2236A (improvement in this respect over 2236) schematic as an example and use a common low power fan.

[David Hess]

You may want to check the 2236 schematic

Please use 2236A (improvement in this respect over 2236) schematic as an example and use a common low power fan.

I would have suggested this except that I am unaware of any free electronic 2236A service manuals online.  If there is one, a link to it would be convenient.

[Rutger]

Here is a link to a website with service manuals.

http://www.ko4bb.com/manuals/index.php?dir=Tektronix/Tektronix_-_2235

I am thinking of get a 2235 with horizontal sweeping issues, anyone experience with this issue on the 2235?

[David Hess]

The 2235 is the basic model for at least the 2221, 2221A, 2224, 2230, 2232, and 2236 so horizontal diagnostics are shared between them all and if necessary custom parts may be found in the other oscilloscopes of that series.

As far as what is wrong, it could be a sweep logic problem, triggering problem, sweep generator problem, or horizontal output problem.  Of those, the last is the most likely and the easiest to repair.

[AheadOfTheTimes]

My 2235 also makes the high pitched noise when warm.  Actually it's the Opt 01 version, aka AN-USN-488.  Can I surmise from this thread that adding a fan has cured it in some scopes?  Wondering whether adding a heat sink on top of that inductor might do the trick, or would changing a single cap (which one?) possibly eliminate the noise?