Tektronix 2215A with short trace only when free running

[torquil]

Hello there!

I'm new to this forum, and I'm here because I recently bought a used (obviously) Tektronix 2215A which has a problem that I would like to repair. Fortunately, I have access to another 2215A which is working fine, so I can compare behaviours between the two.

In a nutshell, the problem seems to be that when the trace is "freerunning"/not triggering, the horisontal extent of the trace is compressed to about half the width of the display. When triggering on a normal input signal, the horisontal extent is fine and covers the entire width.

I went through the "baseline trace" procedure in the service manual on both scopes. On the problematic scope, the baseline trace is short. On the working scope, it covers the entire width. With a normal sinusoid input and using "PP auto" triggering, both scopes show a nice trace that covers the entire screen width.

I have a few other observations, which might be caused by the same underlying problem on the problematic scope:

1) When selecting "A trigger source: line", I get no trace. On the working 2215A with a sinusoid input from my cellphone, it shows an "unsynchronized" trace, is it should since the input is not synchronized to the line voltage.

2) Pushing "single sweep" while in "PP auto" trigger mode doesn't always result in a trace. It seems very unreliable. On the working 2215A, it always shows one trace.

3) Sometimes the trigger LED flickers when moving the "A intensity" knob.

4) Say I have a full trace covering the entire screen width with a sinusoid input and "PP auto" triggering. If I then turn the "Volts/DIV" until the trace becomes very flat, it will stop triggering and then the trace width suddenly reduces to about half the screen width. So when the triggering is not happening, the width is reduced, as mentioned. This should not happen, of course, and does not happen on the working 2215A that I'm comparing to.

5) On more thing: Say I have e.g. the baseline trace going (which occupies half the screen width) and set the "TIME/DIV" knob so that the trace moves very slowly. If I use the "single sweep" trigger button, the first sweep starts all the way from the left of the screen and travels across the entire screen, but every subsequent trace is "compressed".

So it seems to me that after a trigger event, the trace will go across the entire width, but when there has not been a trigger event, the trace is compressed in the horizontal direction...

So my conclusion is that it has to have something to do with the part of the timebase circuitry that controls the horizontal amplifier. The input to the horizontal amplifier seems to be wrong whenever there has not been a trigger event. When there has been a trigger event, everything seems fine. The vertical behaviour of the trace is working fine.

Any ideas?

The only thing I have done so far inside the scope is to check the power supply voltage test point on the main board as specified in the service manual, and the values were perfect. The 100 VDC test point measured 100.0 VDC on my DMM...

EDIT: I just learned about tekscopes at groups.io, so I posted the same question over there as well today (if it is approved by the mods).

Best regards,
Torquil Sørensen, Norway

[pbarton]

Welcome to the forum. Your English teacher did a good job, your command of the language is commendable.
Looks like your faulty 2215A has a minor problem, probably within the "A Sweep generator and Logic" which is described on page 3-2 of the 2215A manual.
The 'A Sweep Logic' is comprised of U532B, U502A, U506A, U506B and U506C on the main logic board and the schematic is shown on diagram A SWEEP GENERATOR & LOGIC <4>.
There are several test points (marked 6 through 16) on diagram <4> and two pages earlier in the service manual the waveforms for these test points are illustrated.
Why doesn't the trace go the whole way across? Because the fault is resetting the sweep logic prematurely.

[David Hess]

Is the trace shortened or compressed?

When the trace is shortened or compressed, does it start on the left side and extend toward the center or does it start in the center and extend toward the right?

Check the connectors which go between the front panel board and the main board.  A broken common signal would cause all kinds of mischief.

[torquil]

Hi! Thanks for the complements on my English (I had some help, my mother comes from Scotland).

I checked the curves shown in the service manual for the "A Sweep Generator and Logic" section. Since I don't have a normal signal generator, I had to make do with a weaker signal coming from my telephone.

1) But first, the traces shown in the manual do not indicate the horizontal scale. All the traces shown for test points (7) and upwards in the manual seem to have longer wavelengths than then one for test point (6). But in my tests the wavelengths are about the same. Is the horizontal scale on curves (7-16) in the service manual streched out just for illustration?

(2) Since I'm seeing a problem with the horizontal extent of the baseline and free running traces, should I not do this troubleshooting with input set to GND? After all, there is a problem even when the input is set to GND, so I would expect that to be simpler. But how do I know what the traces should look like for the proerly working baseline trace. Can I use the working 2215A to test on its own circuit board so that I can compare.

I have looked at some of the traces, but I have to do a more detailed and systematic in order to be able to present my results in a clear manner here.

One simple fact that is clearly different from the manual is the trace for test point (10), which is a DC shifted sawtooth: When the triggering is working on the input sinusoid signal, the sawtooth I get here is exactly like in the manual, going from about 0.6 V up to 12 V. But when I ground the input so that the horizontal compression happens (or put the A trigger source to EXT so that there is no triggering), the sawtooth is suddenly from around 7V to 12V, but with the same slope. So it has around double the frequency.

But also on the traces of with lower numbering (7-9), their frequencies double when I do something which puts the scope into freerunning mode.

It would be nice to understand the "cause and effect" connections between these traces (6-16), so that I would be able to get closer to the source of the problem.

[torquil]

Is the trace shortened or compressed?

When the trace is shortened or compressed, does it start on the left side and extend toward the center or does it start in the center and extend toward the right?

Check the connectors which go between the front panel board and the main board.  A broken common signal would cause all kinds of mischief.

It is shortened, not compressed.

If I first make sure that the properly triggered trace from e.g. an input sinusoid or squarewave (from the probe adjust contact) is centered on the screen, and then ground the input so that the horizontal problem appears (or put A trigger source to EXT), then the trace ends up on the right hand side of the screen. If I turn the horizontal position knob afterwards I can get put the trace anywhere on the screen, though.

I guess this fits with the observation that the example sawtooth trace (10) for the "A Sweep Generator and Logic" in the service manual has the same slope both when it is working and not working, but that it goes from 0.6 V to 12 V when it is working and from 7 V to 12 V when the horizontal problem happens.

[pbarton]

Q1a) “But first, the traces shown in the manual do not indicate the horizontal scale.”
A1a) If you look at the service manual page that illustrates the normal waveforms 6 through 16, at the top of that page there is a ‘2215A CONTROL SETTINGS’ table.
The table states all the switch positions that you require, so that you can duplicate the exact same conditions. Within the table is the timebase setting used, ‘A SEC/DIV 0.1 ms’
If you use this setting then your display should appear identical to the illustrations.
Thus its not necessary for the illustrated waveforms to have a horizontal scale.
Q1b) “Is the horizontal scale on curves (7-16) in the service manual stretched out just for illustration?”
A1b) No, if the waveforms appear unfamiliar then you are probably viewing the fault symptoms.
Q2a) “But how do I know what the traces should look like for the properly working baseline trace.”
A2a) The ‘2215A CONTROL SETTINGS’ table informs you to use the cal waveform as Ch1 input signal. You can trust the illustrated waveforms 6 through 16 within the service manual.
Q2b) Can I use the working 2215A to test on its own circuit board so that I can compare.
A2b) If you wish, but it means opening up two scopes on your workbench.
You mention “But also on the traces with lower numbering (7-9), their frequencies double when I do something which puts the scope into freerunning mode.”
That’s a symptom of the fault you are trying to identify.
If waveform 9 is doubling in frequency, I would like to understand why. We need to follow this back to identify the problem.
So, can you tell me if waveforms 6, 11, 12 and 13 are doubling in frequency when the scope is put into freerunning mode?
You may have a fault with U502, U504, U506, U532, any stuck outputs?

[David Hess]

If I first make sure that the properly triggered trace from e.g. an input sinusoid or squarewave (from the probe adjust contact) is centered on the screen, and then ground the input so that the horizontal problem appears (or put A trigger source to EXT), then the trace ends up on the right hand side of the screen. If I turn the horizontal position knob afterwards I can get put the trace anywhere on the screen, though.

Ok, that is important.  If the shortened trace is starting on the left and terminating in the center, then the sweep is terminating early.  But what you describe is the trace starting at the center so the sweep generator is not properly being reset between sweeps and significant offset remains.

I guess this fits with the observation that the example sawtooth trace (10) for the "A Sweep Generator and Logic" in the service manual has the same slope both when it is working and not working, but that it goes from 0.6 V to 12 V when it is working and from 7 V to 12 V when the horizontal problem happens.

Exactly.  If you check, I bet you will find that the amplitude of the -a-gate signal (1) at the base of Q701 is different when the sweep is shortened.  I think that rules out a problem with the sweep generator itself.

For now, I would take a very close look at the amplitudes of the signals at 7 and 8 since they are documented in the service manual.

Also check the levels of 6, 9, and 15.  I wonder if U506A is being driven such that its output is oscillating when the automatic sweep is occurring.

The trigger signal at 6 drives the automatic mode timeout at U502 and the -a-gate generator at U506A.  Conceivably that is how the state of the automatic mode timeout could get back into the -a-gate generator.  Of course the automatic timeout also triggers the -a-gate generator through U532C which is why 15 should be checked for amplitude.

I suspect either the level of 6 or 15 is bad or U506A is damaged.

If waveform 9 is doubling in frequency, I would like to understand why. We need to follow this back to identify the problem.

It doubles in frequency because the sweep is starting half way across the CRT so takes half as long to complete.

(1) The dash in front of "a-gate" means that the signal is active low.  On the schematic, this is shown by placing a line above "A GATE" just like the -Q and -CE markings on the flip-flop signals.

[torquil]

Q1a) “But first, the traces shown in the manual do not indicate the horizontal scale.”
A1a) If you look at the service manual page that illustrates the normal waveforms 6 through 16, at the top of that page there is a ‘2215A CONTROL SETTINGS’ table.
The table states all the switch positions that you require, so that you can duplicate the exact same conditions. Within the table is the timebase setting used, ‘A SEC/DIV 0.1 ms’
If you use this setting then your display should appear identical to the illustrations.
Thus its not necessary for the illustrated waveforms to have a horizontal scale.

I have selected those settings from the service manual on the faulty scope. The difference is that I'm using the 1 kHz sinusoid of somewhat a lower amplitude, because it comes from my telefone. On the working scope that I'm using for the measurements, I have to select 1 ms / DIV in order for the it to show something that looks like trace (6).

When the faulty scope is triggering and apparently working correctly without a short trace on its screen, trace (6) on the working scope looks just like in the manual, with the correct amplitude (when it is set to 1 ms/DIV, as mentioned). The LO and HI voltages are perfect, 3.4V and 4.4V.

In the service manual, trace (7) is a factor of 10 slower than (6). In my case, the difference is around a factor of 2, This is when it is triggering and apparently working when just viewing the display. The waveform is the same, though, a squarewave. The voltages I'm getting on trace (7) are 3.4V and 4.3V.

My measured trace ( 8 ) looks like (7), but with different voltages. The LO voltage here is 0V, while the high voltage depends on the intensity knob. At low intensity the HI voltage is 3.4V, and at maximum intensity it is 2.9V. This seems to be the opposite behaviour compared to what is indicated in the service manual, where it looks like the highest voltage should occur when the intensity is lowest.

Trace (9) looks very wrong. Instead of a square wave it is a series of short pulses. The wavelength is like trace (7) and ( 8 ). But the HI state only lasts around 5 microseconds. The LO voltage is 3.3V and the HI voltage is around 4.3V.

Trace (10) looks just as expected, but again its period is the same as for traces (7), ( 8 ) and (9), so twice the period of trace (6), not 10 times as indicated in the service manual.

Trace (11) is spending most of the time in a HI state with 4.8V, with only short pulses of LO voltage at almost 0 V for a few microseconds. Same period as the others.

Trace (12) is similar, Spends most of the time at 4.8V but with short pulses where it falls down by around 0.4V instead of down to 0V. Same period as the others.

Trace (13) is a flat line at 3.6VDC...

All these measurements were taken while a 1 kHz sinusoid were inputted to the fault scope and its display shows around one period, since the timebase is set to 0.1 ms/DIV.

If waveform 9 is doubling in frequency, I would like to understand why. We need to follow this back to identify the problem.
So, can you tell me if waveforms 6, 11, 12 and 13 are doubling in frequency when the scope is put into freerunning mode?

I will flip the A trigger source between INT and EXT to provoke the horizontal problem, and write "triggered" and "freerunning" for those cases:

Waveform 6:

Triggered: As above.
Freerunning: Flat line at the HI state, 4.4VDC

Waveform 11:

Triggered: As above
Freerunning: Similar as in the triggered case, except the frequency is four times as fast.

Waveform 12:

Triggered: As above.
Freerunning: Similar, but four times as fast.

Waveform 13:

Triggered: As above, just a flat line.
Freerunning: No difference.

[torquil]

Exactly.  If you check, I bet you will find that the amplitude of the -a-gate signal (1) at the base of Q701 is different when the sweep is shortened.  I think that rules out a problem with the sweep generator itself.

When the scope is triggering, the signal at the base of Q701 is a squarewave with LO at 0V and HI at 1.3V. When freerunning, the signal here is a periodic "Dirac" pulse signal of four times the frequency. It stays mostly at 0V, but has brief pulses at 1.3V.

For now, I would take a very close look at the amplitudes of the signals at 7 and 8 since they are documented in the service manual.

Also check the levels of 6, 9, and 15.  I wonder if U506A is being driven such that its output is oscillating when the automatic sweep is occurring.

I hope my description of (6) and (9) in my other reply was adequate.

Waveform 15:

When triggering and with no horizontal problem: Flat trace at 3.3V
When freerunning (a trigger source at EXT, and shortened trace): Mostly 4.3V. Short "Dirac"-pulses down to 3.3V. Twice the frequency of trace (6) when triggering, or four times the frequency of (7) when triggering.

[pbarton]

When the faulty scope is in its correct working mode, does the VAR HOLDOFF knob (front panel top right had corner) behave correctly. Like your working 2215A?
Are the waveforms 14 and 16 changing state, in the working and failing modes?

[torquil]

When the faulty scope is in its correct working mode, does the VAR HOLDOFF knob (front panel top right had corner) behave correctly. Like your working 2215A?

I created a toneburst signal using an Android app called "Keuwlsoft Function Generator" to provoke triggering problems on the working scope. It was few cycles of 2 kHz sinusoid, and some silence between each pulse train. On the working scope, the trace flickered and I could stabilize it by increasing the VAR HOLDOFF, as expected.

On the faulty scope when it is triggering, the VAR HOLDOFF doesn't seem to do anything to the flickering trace. I don't see any effect on the flickering trace when turning the VAR HOLDOFF knob on the faulty scope. I used the same settings on both scopes.

Are the waveforms 14 and 16 changing state, in the working and failing modes?

I will have to postpone checking this until this evening, when I will hopefully have time to open the scope again :-)

[torquil]

OK, I have some more information, as well as a bunch of photos of traces 6-16 measured in both the working and faulty scopes.

1) I exchanged the timing boards A4 between the two scopes, but it did not cure the faulty scope, so the problem is not there.

2) I have measured traces 6-16 on each scope by using the other as the measurement device. The faulty scope can still be used as a measurement device as long as one is aware of the problem. Only flat traces will be limited in horizontal extent because then it is not triggering. Whenever there is a nontrivial waveform it will trigger and behave as normally. Note that the timebase calibration might be a bit different between the two, especially since there is something wrong with the "CAL" knob for the timebase on the working scope. But I believe it is minimal. The CAL knob on the working scope does not click into place, but I believe it is in the "calibrated state" even if the knob looks a bit randomly rotated in the photos (it usually has no effect on the trace).

I have grounded the input on the scope that is analysed, because I have a problem even with no signal on the faulty scope. The scope that is measured is always set to 0.1ms on the timebase. The measurement scope is set so that there is 1V/DIV vertically except in the photo of trace 10 because that is a big signal. The timebase is 0.2 ms/DIV on the measurement scope unless otherwise stated. It can also be seen in the photos. Also note that I have adjusted the vertical position so that 0V is two divisions below the horizontal centreline (because then I get most of the traces in DC mode without having to adjust vertically or change the Volts/DIV)

First, here is the signal when not probing anything, so you can see what is the 0V reference:

Working scope:

Faulty scope:


Comment: When using the faulty scope as a measurement device here, the trace is short because the waveform is flat and then there is no triggering. This is the problem I'm trying to solve.

Trace 6:
Working scope:

Faulty scope:


Ignoring the horizontal problem when using the faulty scope to measure, these are the same. Around the same postive DC voltage, around 3.3 VDC.

Trace 7:
Working scope:

Faulty scope:


Comment: Here there is a clear difference: the low and high voltages are about the same (3.4V and 4.3V), but the faulty scope is doing something strange. The HI state on the working scope lasts 0.2ms, but on the faulty scope it is much shorter, and occurs more often. The latter is something that happens in several of the traces below. In this photo, I wonder if I forgot to get it correctly triggered, and perhaps this is freerunning. I will check again next time I open the scope.

Trace 8:
Working scope:

Faulty scope:


Comment: Around the same low and high voltages (0V and a bit more than 3V), but longer lasting high voltages on the working scope. Higher frequencies of short high voltages on the faulty scope.

Here is a time zoom with 1 uS/DV of the short pulse happening in the faulty scope:


Trace 9:
Working scope:

Faulty scope:


Comment: As in trace 8, more frequent, and much shorter pulses on the faulty scope. Voltages are about the same.

Here is a time zoom to 5 us/DIV to show the structure of the pulse happening in the faulty scope:


Trace 10:
Working scope:

Faulty scope:


Comment: 5V/DIV vertically since it is a big signal. The top voltage is the same, around 12V. The faulty scope doesn't start from the bottom of the voltage ramp.

Trace 11:
Working scope:

Faulty scope:


Comment: The same low and high voltages, but the faulty scope jumps down to often, and doesn't stay down there at the low voltage.

Time zoom at 50 us/DIV on the short pulse happening in the faulty scope:


Trace 12:
Working scope:

Faulty scope:


Comment: The same here. Too frequent and the low voltage period is absent. In the service manual, the low voltage period here is the HOLDOFF period, so it is completely missing on the faulty scope.

Time zoom to 50 uS/DIV showing the short pulse on the faulty scope:


It looks just like in the service manual, except it is missing the HOLDOFF period.

Trace 13:
Working scope:

Faulty scope:


Comment: The faulty scope has just a flat line here...

Trace 14:
Working scope:

Faulty scope:


Comment: Both show a flat line. The horizontal problem appears when measureing with the faulty scope because it is freerunning due to the flat signal.

Trace 15:
Working scope:

Faulty scope:


Comment: As other traces above. The lov voltage periode is shortened on the faulty scope, and occurs more often.

(Attempt at) time zoom on the short pulse happening on the faulty scope:


Trace 16:
Working scope:

Faulty scope:


Comment: Both are flat at around the same voltage.

My main observations are as follows:

1) Waveform voltages seem fine.
2) Trace 8 shows that the faulty scope doesn't have a correct "A retrace" period. Or, it is very short and happening too frequently.
3) Trace 10 seems very much correlated to chat is happening on the screen. The voltage ramp is shortened, as well as the trace width.
4) Trace 11 shows that the faulty scope has no "A end of sweep" period of low voltage.
5) Trace 12 shows that the faulty scope has no "HOLDOFF" period, but the waveform up until that is supposed to happen looks fine (apart from duration, of course).
6) The voltage ramp-up that should happen in trace 13 is missing. This missing voltage ramp-up is called "HOLDOFF TIMING" in the service manual.
7) Trace 15 should have a period of low voltage, but it is drastically shorteden and happens too often, similarly to several of the other traces.

So, does this info narrow down the list of suspected components? As mentioned above, trying the known healthy timing board A4 from the working scope did not cure the problem, so the problem is likely not here.

Best regards,
Torquil Sørensen

[pbarton]

The problem is on the main logic board and the schematic is shown on diagram A SWEEP GENERATOR & LOGIC <4>.
There is something wrong with HOLDOFF, which is U504B (96LS02) as shown by waveform 12 on the faulty scope.
Waveform 13 is also wrong and you also mentioned that the VAR HOLDOFF knob is not behaving like your working scope.
Without a working HOLDOFF circuit the ‘A Sweep Generator Logic’ is running much too fast and the timebase sweep does not have sufficient time to complete a scan from beginning to end.
It's locked in a loop without a HOLDOFF delay, hence the short trace.
You may like to check that the wiper/mid contact of the VAR HOLDOFF potentiometer is galvanically connected to pin 14 on U504B, just to verify that a couple of interboard connectors are in place.
If that’s OK, you will need to replace U504B 96LS02 Dual Retriggerable Resettable Monostable Multivibrator.

[torquil]

The problem is on the main logic board and the schematic is shown on diagram A SWEEP GENERATOR & LOGIC <4>.
There is something wrong with HOLDOFF, which is U504B (96LS02) as shown by waveform 12 on the faulty scope.
Waveform 13 is also wrong and you also mentioned that the VAR HOLDOFF knob is not behaving like your working scope.
Without a working HOLDOFF circuit the ‘A Sweep Generator Logic’ is running much too fast and the timebase sweep does not have sufficient time to complete a scan from beginning to end.
It's locked in a loop without a HOLDOFF delay, hence the short trace.
You may like to check that the wiper/mid contact of the VAR HOLDOFF potentiometer is galvanically connected to pin 14 on U504B, just to verify that a couple of interboard connectors are in place.
If that’s OK, you will need to replace U504B 96LS02 Dual Retriggerable Resettable Monostable Multivibrator.

Hi!

The connection between the holdoff pot wiper and pin 14 is fine. So are the capacitances of C519 and C520. I will follow you suggestion and replace U504.

Btw, I did some DMM resistance measurements with the scopes unplugged, and noticed a big difference on the resistance between U504 pin 9 and ground, when using the red DMM probe on pin 9 and the black on ground. On the working scope, the resistance is around 2.6 MOhm, but on the faulty scope it is only around 190 kOhm. Measuring in with the opposide probe arrangement, I get 4.5 kOhm and 4.1 kOhm, respectively.

The resistance to ground on the other side of the diode CR503 (not towards pin 9) is the same on both scopes and seem independent of measurement polarity.

I will report back what happens when I replace U504, but it will take some time to get hold of it.

I really appreciate all the help from you guys!

[torquil]

Well, I got hold of a replacement for U504 along with a suitable socket. But I'm having a hard time removing the old IC (not done yet). I'm having problems with removing the old solder using my solder sucker.

And during my attempts, I'm pretty sure I broke the nearby germanium diode CR503. I ended up applying some force to one of its legs in an attempt to check if U504 was loose after removing solder, because I did not understand how fragile the germanium diode was. I could hear the sound of breaking glass as its leg sort of gave in to the force that I applied (I havent tested it yet).

Removing ICs from circuit boards is apparently not my field of expertise.

The following info is in the service manual for CR503:

A1CR503   152-0075-00   Semicond Device;SW,GE,22V,40MA   14433   G866 (where 14433 is ITT Semiconductors)

At https://www.sphere.bc.ca/test/tekequiv.html it says:

152-0075-00    Ge    Diode    Tek spec

Would you guys know of a "standard part" that is a suitable replacement for this, e.g. from Digikey? I don't like the term "Tek Spec".. :-) Is it absolutely necessary to order a "proper" Tektronix part?

[pbarton]

There are some 152-0075-00 on eBay.com but expensive. You could use any small signal Germanium diode, as its not a critical component. eBay is a good source, search for 'Germanium diode' and select a supplier that's not too far away and look for something of a similar physical size. Please note the orientation before you take the old one out.

[David Hess]

Any low barrier small signal schottky diode like a BAT42 or BAT43 should work.

A 2N3904 could also probably be used.  Connect the emitter and collector in place of the diode and wire the base to +5.2 volts through a 22 kilohm resistor.  Now it operates as a very low forward voltage drop diode.

[torquil]

I replaced U504 with  new one in a socket, as well as a new germanium diode cr503. After turning it on, I got a nice baseline trace across the entire screen, which was better than before. But it was too good to be true, becuse the trace disappeared after about 10 sec, when the screen went blank.

Beam find still works, and in xy mode I can move the dot all over the screen. I guess I have to go back and check the service manual traces again.

But perhaps I have some questionqble solder joints somewhere in there as well, because the trigger LED flickers when I move or twist the scope, on the auto trigger setting.

[torquil]

I went through waveforms 6-16 in the service manual, after having replaced U504 and CR503. The result is that all of the traces are now only flat lines.

Then I checked traces 1 and 2. These are illustrated in the service manual as quite nice square waves. When I measure inside the faulty scope, I see lots of ringing. The period is quite right, but the should there be this much oscillation?:

Trace 1:



Trace 2:

[pbarton]

No, there should not be that much ringing.
It looks like something is not terminated correctly.
However, the waveform is generated and shaped, buffered driven from the one chip U537.
So, as long as the chips supply +5.2V is OK and there is nothing obviously wrong with R544, R545, R547, R548, R549 or C545, C547, then I would suggest that you replace U537.

EDIT: Or you could try lifting one end of C547 and re-check the waveform 1 (on TP357) to see if the ringing disappears. If it disappears, then U537 is probably OK.