Philips GM5650 oscilloscope very ill.

[Clay1905]

Hi folks,

Some time ago I was given this little oscilloscope. It was fairly ill then, producing a squiggly Lissajous style trace with no input. Since then things have degenerated, and all that the 'scope produced was a little green glow hard left of the screen. I understand this is the 'default' condition if the time base isn't working properly.

I decided to do a capacitor transplant, eliminating all the ancient wax and paper types, as well as the smoothing electrolytics from the power supply.
Tests revealed that one of the valves had a dead pentode section. It has been replaced, as well as a handful of resistors that got a bit carried away with themselves too.

My ministrations were well received. Now the green glow is no more.

Further tests show:

The time base seems to be working. I can see the waveforms at the junction of the DC isolating capacitors for the horizontal CRT plates. On the HV side of the isolating capacitors I find about 390VDC on both plates.

The vertical amplifier is good too (I think). I find about 380VDC on both vertical plates, (no signal) indicating the spot should be vertically central at least. Until it disappeared, the green glow was centred about the horizontal diameter.

The CRT circuit appears to be OK too. The voltage divider for getting the various grid voltages has a positive input at the 'top', and a negative one at the bottom end.

I think this is rather looking like the CRT has become faulty. I'd like to be a bit more certain though. I'm somewhat out of my depth here, and would like some advice on what, if any, further definitive tests I could perform to confirm or rule out a dead CRT.

I think that's about all I can add at this time.

Thanks,
Clay.

[Diabolical Artificer]

Don't know the scope in question, a schematic would help but in a nutshell a CRT is just a complicated valve, so if you have +1000v up on the anode, -1000v ish on the cathode and the grid is about -20v in relation to the cathode you should have a trace or spot, that is roughly speaking, anodes are usually anywhere up to 10kv, hard to check without an EHT probe or similar. Also the X & Y plates have to read zero volts in relation to each other else the spot can be hiding.

Therefore if you can pull off the X & Y leads, it may be hard to do without damaging the CRT, then check you have about 20v between cathode and g1,you should have a spot. This is obviously generally speaking without reference or knowledge of your CRT.

Andy.

[TonyDuell]

Can you see the CRT heater glowing at the back of the tube? Be careful if you measure the heater voltage, while it's only about 6.3V AC (as you'd expect) between the ends of the heater, the heater is normally at much the same voltage as the cathode wrt earth so expect -1000V DC or more. You need a voltmeter that can stand that sort of voltage between input and ground.

[factory]

It might be hard to see if the CRT is glowing due to the shield, also looks rather cramped for taking measurements at the CRT socket.
The webpage here suggests the base of the CRT may be loose from old glue, be careful if removing it to help prevent the wires inside the base from breaking off.
http://www.glensstuff.com/gm5650/gm5650.htm

The first oscilloscope I bought had a dead CRT, I found the heater open circuit, after removal if was apparent the CRT vacuum had failed as the getter had gone white, it took a while but I did eventually find a replacement CRT for it.

David

[MotorBikeLes]

As we all know, the best thing for fixing a scope is another scope. however there are some shortcuts.
X-scan. You think it may be working? Set your scope to its slowest speed. Connect your avo or other analogue meter across the two X-plates, DC range, 250v. If the T/B is running, you will see one side increasing whilst the other decreases, observable by the meter pointer movement.
T/B NOT running? But is X-amp OK? Leaving your Avo leads alone, can you switch to X-Y mode, or turn off T/B (often with T/B knob extreme Anti clock, or sometimes fully clock.
Y-amp. Move the Avo probes to the Y-plates. Rotate Y-shift fully each way. You should see the voltages rise and fall together (opposite direction of course) just as for the X-deflection.
Does the CRT have a high voltage anode connection? - easily visible at the screen end of the tube. If so, its voltage should be between 3 and 25KV -- CAUTION.
If no HV connection, then the cathode is likely to be up to -1.000v, but WITH an anode connection, the cathode voltage could be between 1Kv and 5Kv negative. All voltages with respect to chassis.
In a dark room, the CRT heated filament should be easily visible.
Answer those questions, and you are well on your way to a diagnosis.
Les.

[factory]

The webpage I linked to gives the CRT as a Philips DG 7-32, this is apparently a low-voltage CRT according to the Philips datasheet, the webpage also mentions the Mullard version is a bit too short to fit this oscilloscope.
https://frank.pocnet.net/sheets/152/d/DG7-32.pdf

I do hope it's OK, as it's now an expensive CRT no thanks to timepiece builders.

David

[Clay1905]

Well,,
Thanks for so many replies.
A schematic and CRT data sheet are attached below.

First, the good news. I had the 'scope on in the dark to see if I could detect any amount of glow from the filament. No, one can't. But a faint glimmer of green on the screen gave hope. After a little twiddling, I managed to get the spot back to hard left, on the diameter. At this stage I can't say why the spot went away, whether or not it was some intermittent fault.


Applying 6VDC to the BU3 & 4 input terminals causes the spot to jump upwards, so the vertical amplifier is at least kind of alive.

The time base circuits seem to be pretty much out of commission. I have probed various places with an oscilloscope where I would have expected some sort of signal, but not a sausage. All I can find is background mains noise. (This had me fooled earlier) None of the control grids in the valves on the right of the CRT in the drawing have anything that looks remotely sawtooth like. Nor the anode or cathode of B5.

Applying the same 6 volts to Bu7 causes the spot to jump a little, but doesn't cause the spot to change position, regardless of polarity. It does cause the spot to reduce in diameter just a little. I think this is the external horizontal position input terminal.

Now I've gone past the limits of my knowledge. Any suggestions on how to proceed in diagnosing the time base department please?

Clay.

[Clay1905]

Originally Posted by factory

The webpage I linked to gives the CRT as a Philips DG 7-32, this is apparently a low-voltage CRT according to the Philips datasheet, the webpage also mentions the Mullard version is a bit too short to fit this oscilloscope.
https://frank.pocnet.net/sheets/152/d/DG7-32.pdf

I do hope it's OK, as it's now an expensive CRT no thanks to timepiece builders.

David

Thanks David,
I had hit on this advice too. What the author doesn't seem to have noticed is that the bracket that holds the socket is adjustable by about 12mm (1/2"). Philips saw it coming, but not the use of the CRT for making otherwise pointless clocks.
As I've posted though, my tube is still (thankfully) alive.

Clay.

[MotorBikeLes]

I take it there is no instruction book for this? Can we have a photo of the front panel?
Is Bu7 an input or output?
Are all the supply voltages (RHS of CRT in circuit) correct? All the +2 and +3 voltages.
Les.

[jamesinnewcastl]

Ah

You have to congratulate the author of the circuit diagram. It must be a talent to draw something in such a convoluted way! The mains voltage selector is sheer genius!

Cheers
James

[Clay1905]

Originally Posted by MotorBikeLes

I take it there is no instruction book for this?

The only instructions I can find are in Dutch. A few snippets are intelligible.

Ah, I think the schematic might have been produced in a bit of a hurry. There are a few errors or omissions.
R57 (B6 grid) should have +2 on the upper end. Likewise B6' anode. (This is from a later version schematic). I have also included a bill of materials, so values of components can be found if needed.

Can we have a photo of the front panel?
One is in the attachments, even as we speak.

Is Bu7 an input or output?
Umm, I think it's an input, but don't quote me. I say this because it seems to be to influence the voltage on the grid of B5, and that directly affects the horizontal deflection plates.
Bu1&2 appears to be to disconnect the trigger signal from B2, and replace it with some external signal. I can make out 3V to 100V in association with Bu2 in the instructions. So, 6VDC was applied here too, with R4 in both extreme positions. No response whatsoever on the screen.

Are all the supply voltages (RHS of CRT in circuit) correct? All the +2 and +3 voltages.
A few things. Firstly, I'm not sure what the nominal voltages here are. I'm finding +390VDC at all +2 and +3 points on the schematic.
But because R's 80 & 81 are quite different values I would have expected some voltage difference too.
The -55V output from B9 is high (-43VDC). Don't think this is a cause of present issues though.

Clay.

[Clay1905]

Originally Posted by jamesinnewcastl

Ah

You have to congratulate the author of the circuit diagram. It must be a talent to draw something in such a convoluted way! The mains voltage selector is sheer genius!

Cheers
James

You ain't kiddin.

[Clay1905]

Whilst I've been waiting for replies here, I have been gradually working from right to left of the schematic to see that the actual chassis conforms with the schematic.
It didn't. The first picture below is of two wires that had nowhere to go. They are, effectively, the wire from B6, pin 6 (anode) to the first junction above, and from the same junction to B5' cathode and C24.
Once they were replaced, viola! a trace. Across about a third of the screen. It didn't last long, about two or three minutes. I think there may be something else not quite right like a cold joint. Moving the time base selector switch seems to make and break some connection.
Now, where's that deerstalker hat?

Clay.

It appears that pin 6 of B5's socket is loose.

[karesz*]

Clay,
Elektrotanya hase an SM too; in Dutch...
K.

[Clay1905]

Originally Posted by karesz*

Clay,
Elektrotanya hase an SM too; in Dutch...
K.

Thanks. I had a look at the link, and now I'm aware that there are at least three slightly different schematics out there. The variations are not very great.

Clay.

[Clay1905]

First off, the intermittent fault associated with the socket of B5 was actually a carbon composition resistor. It was of the type where there is a ceramic former, a film of sooty stuff, and a metal cap that presses onto each end of the former.
As I was extracting the socket for close inspection, the cap from one end of the resistor dropped off. A thoroughly cleaned socket, and a new resistor has sorted that out.
Now, at the prompting of a member, I've been looking at the values of the resistors in the CRT voltage divider. As it happens, they test as pretty good. However, I have found leaks to ground that appear to be leaking tag strip. Possibly other similar sorts of thing too. I'll have to have a go at getting them all cleaned up before I can say much more.

Seasonal felicitations.
Clay.

[Clay1905]

Hi,
At present the 'scope is reliably producing a trace. However, the trace is a little less than a third of the screen diameter. As in the picture I posted previously, the trace is offset to the left.
Tests show that all the components around B4, B5 and B6 are within tolerance. As are those in the CRT voltage divider.

Below is some data in the hope that it is meaningful to someone.

+2 measures as +375VDC
+3 measures as +390VDC
-55V measures as -47VDC
AC voltage on B'5 cathode 3V P to P
AC voltage on B5 grid 7.5V P to P
AC voltage on B5 Anode 7.5V P to P
AC voltage on B5 cathode 7.5V P to P
AC voltage on D2 -7.5V P to P
DC voltage on D2 280 to 260 VDC*
AC voltage on D'2+7.5 P to P
DC voltage on D'2 290VDC
All voltages with respect to chassis. DC voltages taken with 11Mohm VTVM. AC voltages are from oscilloscope peak heights (troughs).

During tests, I inadvertently left R44 disconnected from R42 & B'4 anode. This centered the trace. In this condition, the trace is about two thirds of the screen diameter wide.
* Also, when taking the DC voltage on D2, it initially reads 290 with the trace hard left, but drops over about 1.5 seconds to 260VDC, then the trace is the same as above.

It looks to me to resolve into two issues. One is the displacement of the trace to the left, the other being the short trace length.

Any suggestions as to where to look next would be greatly appreciated.

Clay.

[Clay1905]

Hello,

As a bit of follow up, in the end this 'scope needed a lot more resistors replaced in the horizontal circuits. Some were more that double their nominal value, others were literally coming from together.
In the end I decided to just replace the lot and be done with it.

Initially, the trace was about a third of the screen width. With some manipulation of the controls it's the full screen width now. There appears to be some distortion of the waveform to the right. I think this may be because the time base circuit is accelerating during the sawtooth generation cycle.

Clay.

[Clay1905]

The latest screen shot.

Image is of a 1KHz sinusoidal wave form. Looking pretty neat now.
Mostly the waveform was distorted by a poor solder joint making excessive resistance.

Clay.

[mole42uk]

That's a pretty good trace now - well done, persistence is useful (if I may mis-quote an old pal....)