Cossor 1039 Oscilloscope

[WaveyDipole]

I have been working on restoring an old Cossor 1039 oscilloscope which I had intended to use as a "display" for an Octopus type component tester. Work has been progressing well. The wax paper capacitors and two electrolytic capacitors have been replaced. One of the smoothing caps had actually failed and was dragging the HT supply down by 70V or so. The original selenium rectifier was already missing and silicon diodes had been substituted. Both EHT and HT voltages were high so suitable resistors were added in series to lower the voltage close to specification. On initial testing the CRT came to life and showed a trace at which point operation of controls could be verified.

The running trace is not particularly bright and the BRILL control needs to be set pretty much all the way up. The entire visible adjustment range is limited to the last quarter of the rotation and the instrument is best operated in subdued light. I didn't expect it to be particularly bright, but had hoped for better. Resistors R24 and R34 have both been replaced as they were high and out of tolerance. Germanium diode (MR1) was leaky and was replaced with a silicon 1N4148. These replacements may have had a slight impact on improving brightness. Preset RV8 is at more than 3/4 adjustment and the remaining adjustment range makes little difference to display brightness. I have yet to check R29-R30 but the X-Shift and Y-Shift controls work perfectly fine and I don't see that these resistors going high would have a profound effect on CRT brightness. The state of C23 and C24 is unknown, but unsmoothed EHT voltage is 950VDC, smoothed at 900VDC and although slightly higher than spec, are close enough and well within the 1kV spec of the CRT.

The running trace does fill the CRT width and a sine wave of over 100kHz can be displayed before bandwidth is affected and the amplitude starts to drop. A lot of fiddling is required of both the SYN and VEL controls to get a stable trace and multiple triggering is evident much of the time. It takes a bit of patience and it works, although not as important to me as X-Y mode. I didn't expect the SYN control to have much effect without a signal on the SYN terminal but it does so I am not sure whether it is operating correctly.

In X-Y mode a dot is displayed as expected and the trace is much brighter and adequate, however deflection of the axes is disproportional. I am using a 1kHz signal at 10Vp and according to the manual DC deflection should be about 10V/cm. The Y control is set in position 1 or 2 (direct to Y plate or through the input capacitor) and timebase control in X-Y mode (position 1). The Y axis deflection is about 2cm and consistent with the signal but X deflection almost half of that. I tried injecting the signal both into C11 while disconnected from switch S2.B and directly to the X plate but the trace width remained exactly the same.

I doubt I can do much about the overall CRT brightness and it will suffice, but is it normal for the CRT deflection to be dispropertionale in this way? Some compensation could be built into the octopus tester I suppose, but it would be preferrable for the deflection to be proportionate on both axes. he display type is quoted as type 24D which I understand is similar to a 3RP1.

[WaveyDipole]

Just an update about the SYN control. I found the information on page 5 of the MkII manual to be helpful. Mixing in a bit of the Y signal by connecting the Y and SYN jacks and operating the SYN control does seem to help considerably in obtaining a stable trace. The scope will display a stable sinewave of up to 1MHz at less than 100mV. It seems that SYN does work as expected. I just now need to understand the reason for the multi-triggering when the SYN input is not used and whether there is anything that can be done about the deflection between X and Y axes being significantly different.

[WaveyDipole]

I did a further small experiment with the X-Y deflection. I centred the spot and checked the X1 and Y1 plate voltages. They were 0V as expected. I then moved the spot to the outer edge of the CRT in both directions and both axes. The Y axis required ±5V for full deflection in both directions, whereas the X axis required ±7V for full deflection in both directions. Clearly something is not right here?

[barrymagrec]

I don`t think that the X and Y plate sensitivities are normally the same.

[Radio Wrangler]

The X and Y plates are not in the same place along the tube.

The Y plates get the electron beam first and the angular deflection they make subtends a greater movement of the spot on the screen.

The X plates come next and they have to be larger to accommodate the deflection from the Y plates, so they have more capacitance as well as lower deflection sensitivity.

In an oscilloscope, having more sensitive Y plates means the Y amp doesn't have to work so hard or swing so far. This helps tha bandwidth. The lower capacitance deflector plates help too. The X deflection doesn't need the same bandwidth normally, and so the X amp can be made slower and wider swinging.

The circuit diagram shows asymmetric drive to both the Y and X plates. This can make the deflection non-linear to a visible extent, and also affect brightness and focus as the spot is moved from the centre. Better circuits use long-tailed pair balanced amplifiers to drive both Y plates and both X plates.

David

[WME_bill]

Cossor 1039, mk2.
The CRT 24D is the same as Mullard DH7-91 and Etel 3AFP1. With flat screen, 2.75" diameter, and 20 & 12 v/cm sensitivity on X and Y. The 3RP1 you mention is 3" diameter, and much less sensitive, at 33 & 24 v/cm. You would find it difficult to even fit it in the case.
The Cossor 1039 mk1 uses the Cossor 23D, which is our old friend, the war time ECR30 / VCR139A / CV279 /CV1588, 44v/cm on both X and Y plates.
The manual for the mk2 quotes the EHT voltage as -930v after the voltage doubler, and -870v at the top of the brilliance control using an Avo7 (1ma). Be slightly higher with an Avo8 (50ua), so the voltages you quote sound right. But well within the maximum of 1.5kV
If the EHT is correct, and the brilliance limitation pre-set R8 is fully clockwise at minimum resistance, and you still have a dim trace, then I suspect you have a tube getting near its end of life.
Happily, those CRT DH7-91 seem to appear frequently offered for sale. The tube datasheet is on Frank Philipse's tube date site. Else. I can post it.
Your difficulty with the sync & timebase suggests that you could check the capacitors C7 (brightup),C8 (transitron feedback).
Also C6 (sync pulse), as altering the sync control with no signal should have no effect on the TB.
Your X-Y sensitivity direct to the tube plates sounds about right with the tube used. Y is about twice the sensitivity of the X plates.
wme_bill

[WaveyDipole]

Bill and RR, thank you for the detailed replies. Evidently the information I found online comparing the 24D to the 3RP1 was plain wrong. I did think the comparison slightly odd as the tube is larger in diameter as you point out and it has different characteristics. That's why I used the term "similar" in my post. I was trying to avoid implying that the 3RP1 is a direct replacement and you have now confirmed that it definitely is not.

Obviously erroneous also was my assumption that the deflection on both X and Y would be the same given the same voltage applied to the plates. Thank you for educating me. It makes sense that with the X plates being positioned further from the gun than the Y plates, the deviation effect would be more pronounced on the Y plates. I was aware of the differences between the Y and X input circuit which is why I tested with C11 disconnected from S2.B as this would mean having only the 22nF cap in series on both inputs. Pre-amplifier stages in more sophisticated scopes can adjust calibration to precisely match the graticule and compensate for this Y vs Y plate deflection difference in X-Y mode, but this circuit being much simpler, has no such adjustments or compensation so this is all starting to fall into place. Aside from the rather low brightness, the CRT is working as might be expected. I suspect Bill is correct about the CRT getting towards end of life, but hopefully there is some life left in it yet. In the meantime, it might be worth my while to look out for a DH7-91.

The tube information and reference to Frank's site are very helpful. I had not come across that resource before. I will look up the data on there. Thanks.

Its an interesting point you make about the AVO 7 vs AVO 8, Bill. I am using an AVO 8 Test Set No.1.

Since I have had no prior experience of using a VELocity control, I have no frame of reference so cannot be sure whether what I am experiencing is normal for this type of scope, but I frequently see multiple traces (usually 3, possibly 4 cascaded) or two overlapping traces at 180deg phase difference and getting a single clean trace can be a bit of a fiddle. Applying the Y signal to the SYN input does stabilise the display though so that aspect, at least, does seem to be working. Nevertheless, I will check out the suggested components. C7 is one of the two large ceramic 2kV caps which I have left alone. C8 and C6 are tropical fish types.

[WaveyDipole]

These photos show some of the interesting effects I am getting. The photos are a bit soft due to the subdued lighting. The symmetry is a bit suspect as well. The curves seem to be tighter towards the right had edge of the CRT.

[Ed_Dinning]

Hi Wavey, I might have a spare tube for this scope if you feel you need one

Ed

[Radio Wrangler]

Regarding the multiple traces... do a bit of reading-up on synchronising and triggering oscilloscope timebases.

The Miller Transitron style circuit you've shown can be synchronised to an extent, but it is naturally a free-running circuit. It got consigned to the dustbin when triggered timebase designs came along. The triggered timebase was invented by Howard Vollum. He offered the patent to Bill Hewlett and Dave Packard. But they were too busy with government contracts for microwave stuff, so they advised him to set up his own business, and they helped him. He called hs business 'Tektronix' and the rest is history.

What you need id for the timbase run to be started sychronised to the same point on repetitive incoming waveform. If this doesn't work, you get each redrawing of the trance not lining up on top of the last one.

David

[AdrianH]

The waveforms are as I would expect after building a very basic scope myself. You have to have the X sawtooth osc running generally slightly slower buy a division of the incoming Y waveform frequency, this gives the sync circuit a chance to trigger the sawtooth and lock at the same, either rising or falling voltage on the incoming waveform. Clear as mud I know, was never very good at descriptions.

The velocity control is doing varying the speed or repetition rate of the sawtooth, the synch triggers it at some amplitude of the Y voltage. Think it is covered to a degree in Foundations of wireless 8th ed, on Time bases page 421.

The non linearity of the waveform I believe can be caused by the fact that the deflection plates are not being symmetrically driven, in that one plates is not decreasing in voltage as the other increases as you would get by see-saw driving of the plates.

If you look at the spec sheet for the DH7-91 there are different tolerances for deflection distortion for the two systems with asynchronous being higher. In fact of you check the specs they will only quote for 72% of the useful scan area as you get closer to the edges of the crt the distortion will get worse.

Adrian

[AdrianH]

Just had another, probably more simple thought of a reason for the distortion of the waveform as it moves to the RHS of the screen and that is linearity of the sawtooth waveform, it could be flattening out, i.e. not a perfect sawtooth.

Adrian

[WaveyDipole]

Thank you. I found some information about the Miller Transitron circuit here:

http://www.r-type.org/articles/art-135.htm

Most interesting. Having hooked up the anode and g3 of V3 to another scope, I got the waveforms shown. The first photo is a free running trace, the second is with a 1kHz signal applied and with sync.

As Adrian suggested, the ramp is not linear which would explain the distorted symmetry of the displayed waveform.

I noticed another problem though. When the velocity control is adjusted, I expected the timing to go from slow (fully anticlockwise) to fast (clockwise) with the waveform gradually tightening as the control was rotated clockwise. Instead, it goes from slow to fast to slow with the waveform compressing in the middle of the range and widening close to either outer limit of rotation. While this does correlate to and goes some way to explaining its behaviour (it would sometimes almost sync at either end of its range), this doesn't seem correct behaviour?

When I have a bit of time I will investigate further.

[WaveyDipole]

Further investigations show a number of resistors gone high in the Miller Transitron timebase circuit but the key discovery was that the installed VELocity control pot is not only faulty but and incorrect part has been fitted at some point.

Once slackened off the chassis, a marking of 5000Ω was revealed on the pot body but the circuit diagram indicates that it should be a 2MΩ part. Perhaps someone mistakenly thought is was a 5000kΩ (5MΩ) pot?
Clearly a 5kΩ part would not be a suitable replacement for the 2MΩ pot. Its track resistance, with one end disconnected from circuit, actually measured just over 100kΩ and 2-3 megaohms between wiper and track except at either end where contact is made with the end stop where it would read 100kΩ again. It seems that the track must be fried somewhere along its length and the odd resistances are likely the result of soot residue.

Whatever the case, the pot needs to be replaced. I do have a 2M2 pot to hand but the thread diameter is rather wider than the aperture which I prefer to to widen to that extent so will have to source something a little smaller. In the meantime though, I hooked it up temporarily to test and its a totally different experience. Whereas previously waveforms were an un-synchronised free running blur most of the time, stabilising only very occasionally with a lot of fiddling, they are now stable and can be viewed, contracted and expanded with ease. Occasional artefacts do still occur while adjustment is being made, but as indicated by David and Adrian's comments these can now be accepted as normal.

[WaveyDipole]

I was able to open the GAIN pot and clean it but it still behaves a bit erratically so will probably need to be replaced. I found a Morganite 2MΩ pot on eBay which replaced the VEL pot.

I have noticed that the brightness seems to fade a little after maybe 4 or 5 minutes once the scope has had time to warm up a bit. Is this a normal? Or is it indicative of a CRT problem, or some other thermal effect on a component somewhere?

[ortek_service]

I wouldn't expect the brightness to normally reduce after a few minutes or a CRT problem to cause this.
I didn't have this problem with a similar Cossor 1048? Oscillograph, that did mostly-work (Despite me having to wash it all out and dry it, as had been sat in someone's garden for months and was full of spiders etc.!). The only issue I had was one of the traces was off the screen, and someone had been at the shift control circuitry.
> 30years ago, I did have to strip-down an old Cossor (1039?), that was being dumped, as I couldn't carry the large & heavy parts home. I recall the red threadlock paint being very-permanent!

So probably worth measuring all the DC voltages to the CRT grids etc. (+ AC Heaters voltage) at switch-on and after a few mins, to see if any are varying. It does sound most-likely a resistance is changing when warm.

[WaveyDipole]

Thanks. Well that seems to eliminate the CRT. I have replaced some resistors but will check the remaining ones for thermal drift.

[WaveyDipole]

An investigation of the circuit around the CRT shows that voltages are very stable. The EHT voltage, other related DC voltages across various resistor and pot chains, as well as the CRT heater AC voltage were monitored over a 10 minute (600 second) period starting from cold. Test were carried out using a DMM with logging capability via sigrok smuview and results recorded on a graph. A HV probe was used to measure the near 1kV EHT voltages. Lower voltages were probed directly. The graph lines for all tests were remarkably flat with the EHT smoothed output showing a slight surge of 10V at power up. The CRT heater AC supply surged to 9V at power up but settling at 6.5V approximately within a couple of seconds. Any other very minor variations could probably put down to the fact that mains power is unregulated.

I have attached the EHT and CRT heater graphs. Please note that due to the use of the HV probe, on the EHT graph the meter units are in volts and the graph units in kV not mV.

The effect isn't terribly bad, but when brilliance is set to a comfortable level, after 3-4 minutes it then fades a little into the rather faint zone but doesn't actually disappear.

[John_BS]

Couple of possibilities:
leakage in C7

thermal drift in either RV7 or RV8.

John

[WME_bill]

Cossor 1039 Mk2.
May I summarise what you have found. It is important when fault finding in something as complex as a scope to concentrate on one thing at a time.
The trace fades a little after a few minutes use. Not during the trace. This suggests a heat or voltage stress effect around the CRT feeds.
1. EHT voltage dropping. Between the accelerator anode p4 and the cathode p1. Have you checked the voltage actually between these pins. This will confirm the brilliance pots and slider settings, RV7 and RV8.
If you have a digital (10Mohm) meter , use it as preferable to your Avo as you want as low a current as possible. Connect up the Avo (on 100v range, or better 250v range) with the unit switched off. Then switch on. Preferable not to touch the meter, as it is all floating at 800v or so. Watch the readings as the trace fades. You are looking for perhaps 10 to 20v negative change.
2. Tube emission going. I would not expect this to drop in use and then come back later. If it drops, then it will stay dropped. So unlikely.
3. Grid p4 and Cathode p1 voltage changing. This is difficult to test, as it is all high impedance, and with a tube current of 1uamp, the 50 ua of your Avo will pull the grid-cathode voltage all over the place. If you are trying to measure grid to chassis voltage, then grid will swing vastly positive, and probably the ensuing current destroy the tube.
Try connecting the Avo between pins p4 and p1, when switched off. Set at 250v range to keep the current drawn down, and expect 10-20V variation as it fades.
You have changed the diode MR1, so that probably exonerates that.
What about C7, which carries the positive moving bright-up pulse from the screen of the transitron to the tube.
If that is leaking slightly, then the CRT grid will decline negative during the trace, making the trace get dimmer towards the right (does it?). And after a time under voltage stress (it apparently does). Change it. Voltage rating must be 1000v or greater. Value not critical; anything over 1nf (0.001) enough for this test. If you operate the sweep, keep the trace speed high, as a smaller capacitance will result in the trace fading towards the end of each sweep as the capacitance discharges the square pulse of sweep duration produced by the screen of the transitron timebase.
4. The two resistors around the CRT grid should be alright, as they are under little voltage stress: R13 - 27k, R24 -470k. Might be worth changing the 470k. Also cleaning around the pins and CRT base.
5. You do not say if the Focus changes as the trace fades. That could suggest the EHT voltage chain is varying. Though you have apparently checked that.

I have a 1039 Mk1, so cannot fully cross check what results you should expect.
Let us know the resuts of these extra tests.
wme_bill.