Test Equipment Anonymous (TEA) group therapy thread

[tonyalbus]

What triggers me is that you have outlets/power chords to turn them all on at the same time 

Oh yes and i am also 'living on the edge' 

[Ice-Tea]

Two scopes, both entirely awesome in one aspect and a lot less impressive in another. Both have analog BW that would cost a cars worth if you'd buy something new/modern like that but both with very poor one-shot performance. The Lecroy is still somewhat serviceable at 100Ms/s but the HP does... wait for it: 2MHz BW so probably 10MS/s or some such.

[Martin.M]

arrived today: 883AB: Diff, from old John 

[Vince]

PHILIPS SCOPE REPAIR

Yesterday I worked on the issue of the remaining issue, the trace that does not span the entire width of the screen.
I ran out of time / enegy /motivation to report about it yesterday, so I am doing it now. If you find this delay totally unacceptable, please feel free to log on to our website, click onthe "Contact Us" tab. There you will be able to fill in a form and send it to our customer service department.

So.. what I thought should be a simple 10 second "tweak that trimmer and that's it " job, to my dismay turns out to be a full blown troubleshooting exercise, believe it or not. 

I tried to be logical and started upstream : checking the time base / sweep generator before the  X amp.

I scoped the sawtooth to make sure it looks proper, it does, and that the trimmer named on the schematic "Sweep time" did work.
That's when I realized that when they "sweep time" they don't have in mind what you and me have in mind : these two vocables  are NOT meant to refer to the trace length on the screen !
That's why I thought it "did not work" at first ! I was mislead...
Nope. It does what it says on the tin : "sweep time" refers to the actual sawtooth signal itself ! So yeah, I could see on the scope (no the DUT I mean), that playing with that trimmer did vary the duration / "time" of the ramp.... when I would have needed it to vary the amplitude instead !  Nope, amplitude is set in stone, at about 5Volts.

I read the theory of operation as well as the calibration procedures, chapters and nowhere does it tell you how to adjust the trace length !
The only time I ever saw this subject mentioned in the manual, was when they said : " Make sure the trace spans at least 10 division" .... but that's it ! They never say "if needed, adjust it by dong this or that ", no !

So at that point it all started to make sense and the conclusion was that :

1) the sweep signal is fine.
2) it normal if neither the sweep trimmer or gain trimmer in the X amplifier have zero effect on the trace length... tehy aren't meant to to begin with !
3) If the trace length is not wide/long enough, then A) the problem must lie in the X amplifier and B) the gain trimmer is not supposed to fix that, and instead it's a defective circuitry and I need to troubleshoot it.


So that's were my focus is now. Concentrating on the X amplfiier alone (schematic attached), and trying to figure out why the excursion of the plate voltages are not as wide as they should.

The amplifier is basically two identical amps, one for each plate. I drew roughly the "mirror" axis in dotted blue . Sorry looks crap I did it by hand. I don't know how to draw a dotted line with The Gimp just yet 

The lower amp is fed with the sawtooth signals (in green), via transistor V1414.
The upper amp (in blue) is fed with the DC offset voltage from the " X Position" control on the front panel, via transistor V1413.

I checked power rails, +/- 180V they are good.
I checked all 6 Zener diodes (while powered up I  mean), they all drop their expected voltage BUT for the two 5.1V ones.
They drop only 4.4V or so... both of them. Suspicious that they would be both bad, and by the same amount... so I looked at the schematic and figured it must be "normal" / acceptable : They are fed with 180V and a series resistor of 150k or so. So that gives a current of about 1+mA doesn't it. A little bit low I though, might not be enough to get past the "knee". So I checked the datasheet (they are BZX79C series) and indeed voltage is specified for a current of 5mA and even then with a 5% tolerance, they say that 4.8V is acceptable even with the proper 5mA current. It's not far fetched to think that 4.4V for only 1mA is... to be expected. So not a fault.

So that's where I am. Need to keep working on it to figure out what's wrong.

I would check all transistors but only two of them (in red) are socketed. Plus, testing them won't tel me much since they obviously "work". One of them maybe it "tired", if that's even a thing for a transistor, but how would I be able to tell anyway...
I don't have a proper desoldering gun, and they are all soldered on both sides of the PCB. It's a recipe for disaster. I don't want to damage the pad and tracks and make my life miserable, having to fix ruins.

I think I could maybe extract these transistor harmlessly for the PCB, if I sue my hot air station. I used it with success to removed DIP 16 packages (optocouplers) that were soldered on both sides. So I guess a mere TO92 3 legged trannystor should do it pretty easily. Still I run the risk of overheated the board or trannystor.
So my idea is to first do my best to try to narrow it down to a smaller subset f transistors.

My plan for that, well a starting point at least, is to look at the X Position circuitry (in yellow), which feeds the input transistor (V1413)  for the upper amplifier, because when you use that knob it clearly misbehaves : there is a massive offset of the trace to the left of the screen. You need to turn the knob at least 3/4 CW to get the trace centered on the screen.

The schematic as you can see, is nice and tells you what voltages the circuitry is supposed to produce at the base of the input transistor : it should vary between -1.5V and +4.1V. I do get that ! So that's good. It means the X- position problem/offset is not due to the circuitry, but due to the amplifier itself. So maybe the input transistor itself, or associated transistors : V1406 on its left (don't know what that one does, manual doesn't explain), or the stuff on its right, i.e. the amplifier per se.

So I will concentrate on these transistors first, and leave the lower amplifier alone for now.

I guess I could check all resistances first, you never know I might get lucky and find a bad one ! Worth a shot.

Wish me luck... 

[Robert763]

Vince,
Have you checked R1416 and preset R1417 ? If the supplies are correct, sweep is linear and the times ten works the combined value of these must be too high.

Robert.

[Zoli]

PHILIPS SCOPE REPAIR
...

The signals on V1413 and V1414 collectors should be symmetric to 0, and the same amplitude(AC scoping); this should help to separate the input differential from the output. It should be safe to probe, since is a relatively low impedance point. If you intend to probe the plates, I recommend 100:1 probes, and low(<~1kHz) deflection frequency.
And a little teaser:

[tautech]

Those pots marked Gain refer to horizontal gain. They are the adjustment to fill the CRT.

Edit to add
Vince those circuit voltages on the schematic are only valid when the scope is set to the manual default settings and you probably already know that.

When you have the trace correctly filling the screen you'll need to go back and check/adjust the sweep length.

[Vince]

Thanks chaps for your suggestions. Will try to answer all in one.

@Robert : nope didn't check these gain resistors. I think I must have sounded too enthusiastic yesterday when I said that X10 MAG "worked". I mean, it does speed the sweep x10, so that what I call it "working", however it does not change anything at all regarding my trace length problem. To be clear : nothing does ! No matter what control I play with, the length of the trace, where it starts, where it stops... it's set in stone... which definitely points to a problem in the X amplifier I guess, te only thing that's common to everything. The last thing before the deflection plates...

@Zoli : Oh you are scaring me now... no it's not that dangerous I think ! Supply is only 180Volts, and probes are rated for 600V IIRC ? 180V with a x10 probe is only 18V presented to the scope...
Plus the waveforms show that the plat voltage doesn't even go anywhere close to 180V anyway.
The manual also clearly say that in order to probe the amp to observe the waveforms given, one must use a standard x10 probe.

@Tautech : yes of course I set the scope up as required by the manual so the waveforms they give would actually make sense !     Time base needs to be set to 0.2ms/DIV.
Gain trimmer : nope according to the manual as I said, it is not used to adjust the length / excursion of the trace. There is no adjustment for that. Don't believe me ? Download the service manual and have a read for yourself !    No you don't have time for that, so you will have to believe me instead ! 
What that gain trimmer does, is modify the slope of the sawtooth applied to the deflection plates, but it does not modify the amplitude, hence does not modify the length of the trace.

Anyway, thanks all to your suggestions... I did scope the amplifier output / deflection plates. I do get a nice linear ramp, all is good there.
What I observed though, is that the amplitude of the ramp is a bit lower than it should. Which makes sense I guess... 
As you can see the ramp I scoped looks like 3.4 DIV high, but the waveform in the manual is more like 3.8 DIV eh ? That is, about 10% higher.
The length of the trace on the CRT starts about half a DIV from the left edge of the graticule, and ends also half a DIV before reaching the right edge of the graticule. It's symmetrical (this alone is a clue I guess...). So, in total I missing a full DIV of excursion.... one DIV out of 10.... so I missing 10%. ... and I am also missing 10% amplitude on the ramp.... so it all ads up eh.... good.

Then.... with the trace centered on the screen, the resulting ramp as you can see is clipping a bit top and bottom.
Then if I turn the X position control knob on the front panel, fully CW or CCW, I can get a lot more clipping either top or bottom.

So, looks like the problem is clipping. Now I need to probe some more to figure out if the clipping is present also before the final drive transistors, i.e. garbage in, garbage out, or if the drivers are given a good ramp with no clipping, and it's the drivers that clip, somehow. With a supply rail at 180V you would think their is ample headroom not to clip a mere 75V signal or so...

Anyway I will do some more probing.... tomorrow. 22H40 here already, eyes are telling me it's time to go to sleep, yet I have still some more stuff to  do other than electronics. Uninteresting stuff, but it needs to be done regardless 

See you tomorrow for some more !!! 

[Zoli]

Thanks chaps for your suggestions. Will try to answer all in one.
...
@Zoli : Oh you are scaring me now... no it's not that dangerous I think ! Supply is only 180Volts, and probes are rated for 600V IIRC ? 180V with a x10 probe is only 18V presented to the scope...
Plus the waveforms show that the plat voltage doesn't even go anywhere close to 180V anyway.
The manual also clearly say that in order to probe the amp to observe the waveforms given, one must use a standard x10 probe.

The idea of 100MΩ||4pF(typ.) sounds much better to me then 10MΩ||13pF(typ.) when you probing a tube's gate, voltages&frequencies&manuals be damned(deflection plate: practically a gate) in the "restoring functionality phase".
Now, the (insert reward here) question: What causes the BIG FLAT intro on the right picture?<sup>*



*</sup> See if is the same on the input: collector of V1413/V1414/GND(dual trace AC, 10:1 probes are OK )

[Robert763]

Vince,
If the x10 expands the detail of the trace, but not the overall width of the horizontal display something must be limiting the maximum voltage on the X plates or the CRT itself is limting the sweep. Does the peak to peak voltage (or RMS measured with a DMM at 10ms/div sweep)  across the plates increase when you adjust the gain or activate the x10 switch?
If it does without the overall trace width increasing then the tube must be limiting the sweep.
Another check is if the X shift can move the edge of the trace off the screen.

Robert.

[BU508A]

I wasn't fast enough ...

[m k]

Tube question.

How accurate is a filament voltage tolerance?
Circuit is a transformer direct 6.3V connection and input is specified as 220V, nowadays the wall here is around 240V.
Heated thyratron is 884 and manual offers a tolerance of +/-10%.
The device is an adjustable power supply.

[Qw3rtzuiop]

I wasn't fast enough ...

It was a fake offer. I saw it multiple times in the last few weeks

[tonyalbus]

Welcome to my Keithley collection 776 )
with 2.4 and TCXO option 

[BU508A]

What is this yellow thing next to the Schuko-Steckdose?

[capt bullshot]

Clips for ESD straps

[AVGresponding]

What is this yellow thing next to the Schuko-Steckdose?

Antistatic grounding point by the looks of it.

[alm]

It contains studs to attach ESD mats and wrist straps to ground, like this: https://uk.rs-online.com/web/p/esd-grounding/1229159

[McBryce]

What is this yellow thing next to the Schuko-Steckdose?

https://www.esdshop.eu/catalog/product/1538-ESD-grounding-plug-3x10mm/

For those who don't "believe" in the cordless versions! 

McBryce.

[Vince]

PHILIPS SCOPE REPAIR

Thanks for the help Robert and Zoli.
I have no time to work on the scope this evening... will probe/scope the thing tomorrow I hope, based on your advice.

Still, I can reply at least partially to your comments...

The idea of 100MΩ||4pF(typ.) sounds much better to me then 10MΩ||13pF(typ.) when you probing a tube's gate, voltages&frequencies&manuals be damned(deflection plate: practically a gate) in the "restoring functionality phase".

Oh I see, I missed you point sorry ! I though you were being overly cautious with probing, to keep me from getting zapped, or from zapping my scope's front-end ! ...but in fact you were thinking of probing influencing the measurement ?!  Well II have an easy way to tell whether the probing was having any negative impact on the probed signal : just look at the trace on the scope to see if probing it made any change at all... and it did not.
Still, it was a good point...

Now, the (insert reward here) question: What causes the BIG FLAT intro on the right picture?^*

I turned the X shift control fully CCW so the plate voltage is clipping.... do you mean it's not normal ? Uh.... maybe you mean it's not normal to clip THAT early ? I mean the amplifier is fed with 180V... OK,  there are two 36Volts Zeners in series with the supply, so the plates can't possibly see more than 100V or so grand maximum eh ?  Well stupid me, the schematic does tell you tel voltages... +98V for the upper amp and -85V for the lower one...I will measure that tomorrow. So you mean the plate voltage should be allowed to go much higher/lower than it does, before clipping ?

Oh I have an idea ! Let's say we consider the upper amp.  Look there is a 100nF cap, C1416 across the two Zener diodes. It's in series with R1446, which is 356kohms, tied to the -180V supply. Maybe that cap leaks a bit, and it makes the resistor drop more voltage than normal, hence constricting the plate voltage ?  Probably makes no sense, sorry it's 22H15 and I am really tired right now, going to bed in a minute !
Hmmm....

^* See if is the same on the input: collector of V1413/V1414/GND(dual trace AC, 10:1 probes are OK )

Will do  !

If the x10 expands the detail of the trace, but not the overall width of the horizontal display

yes that's it .

something must be limiting the maximum voltage on the X plates or the CRT itself is limting the sweep.

Yes I am starting to think like that as well... the trace of the length is so very independent of absolutely ANYTHING that it's the only logical conclusion I guess...

Does the peak to peak voltage (or RMS measured with a DMM at 10ms/div sweep)  across the plates increase when you adjust the gain or activate the x10 switch? If it does without the overall trace width increasing then the tube must be limiting the sweep.

Will do tomorrow...

Another check is if the X shift can move the edge of the trace off the screen.

Nope, X shift  has no effect on the trace length.

Maybe a short video is worth a thousand words so here it is !



1) At first, input signal is grounded so we see a flat line, with the main time base enabled only.

2) Then I turn on the delayed time base as well, so you can see that it too, is restricted to the same left and right boundaries...

3) Then I play a bit with the delayed time base, sorry...

4) Then I use the X shift control on the front panel. First I slowly turn it fully CCW, then back the other direction til fully CW, then back to centered.
You can see that I can shift it way more to the left, than I can to the right, it's severely asymmetrical.

5) I meant to then finish the video by turning the x10 MAG sweep so you can see it too  is restricted to the same trace length, like every thing else but.... sadly this knob it broken again ! Bloody switch won't engage... it's kaput. I need to pull the X shift pot out of the scope so I can work on the x10 switch that's mounted on the back of it.... trouble is that this pot is very hard to get to, so I don't think I will find enough motivation to remove it....

OK 22H30, I am really... really... 

See you tomorrow for some probing...

[syau]

It contains studs to attach ESD mats and wrist straps to ground, like this: https://uk.rs-online.com/web/p/esd-grounding/1229159

Planning to DIY one as it is only 1Mohm to ground 

[syau]

Welcome to my Keithley collection 776 )
with 2.4 and TCXO option 

Keithley sold all their RF product to Keysight as I recently found out when I looks for software for my Keithley RF Power Meter

[Zoli]

PHILIPS SCOPE REPAIR
...
I turned the X shift control fully CCW so the plate voltage is clipping.... do you mean it's not normal ? Uh.... maybe you mean it's not normal to clip THAT early ? I mean the amplifier is fed with 180V... OK,  there are two 36Volts Zeners in series with the supply, so the plates can't possibly see more than 100V or so grand maximum eh ?  Well stupid me, the schematic does tell you tel voltages... +98V for the upper amp and -85V for the lower one...I will measure that tomorrow. So you mean the plate voltage should be allowed to go much higher/lower than it does, before clipping ?
...

Now I've got infected¹ by your long posts, so bear with me:

Short estimate of the voltage excursion on the lower side:
1. V1423 fully closed, all the current(5mA) from V1426 goes thru V1431(75V)|V1429(36V) to the plate X1(180V/365k=~0.5mA, the current thru R1447 can be ignored) =-175+75+36 =-64V Minimum
2. V1423 fully open, no current from V1426 goes thru V1431(75V)|V1429(36V) to the plate X1(R1447 supplies the current to V1431&V1429) =0+75+36 =+111V Maximum
This is a theoretical estimation based on the schematics voltages; as exercise, estimate the top excursion 

The expected(normal) deflection behaviour would be:

the saw-tooth moves up/down when the X position changed, no clipping, preserving his shape; after all, position is just an offset applied to the deflection saw-tooth.

DC level of the clipping - good indication of what's sick:
1. >9V negative side don't have enough power to pull down, or positive side draws too much current
2. <9V positive side don't have enough power to pull up, or negative side draws too much current
9V=Common emitter on V1419&1422; Oh, and V1419(BSX20) is NPN - schematics mistake, you should request refund
The clipping level close to a voltage in the schematics even can give you hint exactly where the problem lies(example: 0V - change V1423; BC shorted).

DP^* will help you a lot, since the signals(AC wise) on both sides should be identical(sorry 'bout the previous confusion of symmetry)

Short explanation of the outputs:
V1419/V1422: common emitter amplifiers, responsible for the current gain for the whole output stage
V1418/V1423: common base amplifiers, responsible for the voltage gain for of whole output stage together with V1417/V1426(current generators)
So, let's take a look on the on how it works on the active/low side(DC wise):
The input signal gets in V1422 base, the amplified current goes out on the collector straight in the emitter of V1423, which provide the drive voltage of the plate X1 together of V1426|V1429|V1431|R1447.
All the strange voltages, asymmetrical resistors/zeners are in fact ensuring that the outputs are centred, using the already existing voltages in the different parts of the scope; in fact, the whole output design is very symmetric, no matter what the schematics suggest.
It could had been done more straightforward way and centred around 0V instead of 9V? Yes, at the expense of two more voltages(±24V), and level translation; actually, I have a feeling that it was adapted to the already existing voltages in the later stages of the scope design/development.

¹At least I've get infected with something...
^*Dual probing, you filthy perverts

[VK5RC]

What triggers me is that you have outlets/power chords to turn them all on at the same time 

Oh yes and i am also 'living on the edge' 

That is a fine collection - both quality and quantity 👍

[Ice-Tea]

R&S CMS-33. Trying to get the hang of them. Started testing and it didn't behave how I thought it would. Problem is I'm not quite sure how it is supposed to behave. Solution? Compare...




So I thought I'd start out with something simple. Use the 50MHz or so reference from a power meter, see if it is detected on the inputs using the counter functionality. Results are not good. Top left unit works on RF1 & 2. Bottom left RF2. Top right RF. Bottom right none at all.

Yes, I was switching between inputs with the bottom left key on the screen.

So, this is pretty gastly. The odd thing is at least one of these must have come directly from verification/calibration I think. So, I got this nagging feeling there's something I'm simply missing. That's a lot of damage, especially odd on RF1 which is supposedly rated for 125W.

Update: RFout is good on top right and bottom right. Dead on bottom left and 3dB off on top right.

So I have zero units without issues.

Update 2: did a reset on the top left one. 3dB offset is gone. Sounds like the source was unterminated and the reset put the  50 back in place? Then why can't I find a setting for that