Cossor CDU150 rescue

[Chris55000]

Hi!

Yes, either C4(+) or tag 12 will do whichever is easier to get to!

The diode is not a permanent fixture – it's only to protect the PSU units from reverse voltages, etc., as a temporary measure whilst you carry out tests!

1N5404 diodes are 3A forward current rating and will do fine for this purpose – you only need one diode in series with the positive lead – your negative end can go straight to 'scope chassis!

Chris Williams

[MelJon66]

I had another attempt at using the variac as a temporary supply but it was no more successful. In fact the fuse in the connecting lead from the variac blew with FS3 removed from the +50V PSU. It was at switch on so it may just have been a surge. I suspect the variac is too rough an instrument to be used like this so I won't be trying it again. Certainly the Cossor does not like a.c. supplies being routed inside.

The bench power supply proved to be far better than the variac. The horrible curved and dual traces that appeared even on ground settings with the variac connected have gone and the trace is bright and sharp. It still fades from left to right but it's a steady fade and there are no bright spots.

Anything between 55V and 70V (I didn't go any higher) on the bench supply and the +50V rail is rock steady so I can run it at 60V from one dual bench supply and it works perfectly. It draws a steady 320mA.

I will have to put the Cossor aside for a few days at least as I have other stuff to do but I will be posting a wanted request for a donor Cossor CDU150 that hopefully has a good transformer. If anyone knows of a possible source please let me know.

Thanks for your interest.

Cheers
Mel

[barrymagrec]

Fading of the trace on Electrostatic tubes can be decoupling caps in the cathode circuit.

I found this more than once on large HP XY displays.

[MelJon66]

Greetings All,

The Cossor CDU150 ‘scope is back on the bench today.

I am seeing some major differences in the outputs between the two channels at low sensitivity settings. Ch2 is the culprit as the displayed waveform amplitude drops disproportionately when switching from 2v/div to 5v/div sensitivity. The waveform becomes distorted, unstable and hard to trigger on. 10v/div and 20v/div sensitivities are similarly affected.

The problem is in the Ch2 attenuator. At all sensitivity settings from 5mV/div to 2v/div the attenuator resistances for each channel are very similar but at 5v/div the attenuator resistance for Ch1 is 1M Ohm and Ch2 is over 6M Ohm. I suspect a bad contact on one of the attenuator rotary switches or a resistor gone high. The attenuator circuit and values is shown in the schematics attached in my first post if anyone is interested. This shouldn’t been difficult to resolve but access to the attenuator is difficult on this machine as it is fully enclosed and I can’t even get in with switch cleaner. I have tried working the rotary controls to clean the contacts but it has done no good so a full strip down is required.

I have noted the attenuator resistances at each sensitivity setting and they do not increase linearly. There is a general increase in resistance as the sensitivity reduces, which makes sense, but some ranges have slightly lower resistances than the more sensitive range before them. Both channels are like this. I assume the values are selected to suit what the pre-amp requires so I will just check that the resistances are as given on the schematics. If anyone can explain this non-linearity I would like to understand it? The manual describes how to adjust the variable caps to tune the high frequency response so I will follow those instructions.

Before I embark on this strip down does anyone have any ideas on a simpler fix or any advice on a good method to tackle the job? The manual says removal of the attenuator PCB’s J1 and J2 “should not be attempted” (Page 4.4.1). I don’t think I will have to remove them but I can’t see how much access to the resistors I will have when I get in there so I suspect a trap for the unwary.

Any advice from anyone who has done this job before would be very much appreciated.

Cheers
Mel

[buggies]

If I have interpreted the circuit correctly, (someone will correct me) the attenuator switch is drawn in the 20V position and the nominal 1M input resistance for the top 3 ranges (5, 10, 20V) is provided on CH2 by:
R28 100R
R14 1M
R29 47R
R15 1k
This should narrow your search a bit. Probably someone overloaded the input and cooked one of the first three.
Sorry - I don't know the machine so no idea on reaching them.

[MelJon66]

Hi George,

Thanks for your input.

My theory is that the route for the 20v/div position is R13 (47R) - R28 (100R) - R14 (1M) - R29 (47R) - R20 (10R) - R21 (750K) - R30 (100R) = 1.75 Mohm.

The trouble is that it that is not what I measure on Channel 1 = 1.19 Mohm, which I think is probably correct as it seems to work properly. The difference might be explained by the effects of R15 and R22 which take the signal to ground and seem to form part of a voltage divider which will reduce the overall resistance. I think Thevenin's theorem would solve this but I can hardly remember it let alone apply it.

Then again, my thinking might be complete rubbish! I am sure someone will come up with the right answer.

Cheers
Mel

[buggies]

Hi,

Just a quick note as about to sleep!
You can ignore the other resistors you mention - the path I indicated is just from the input to ground and is just over 1M. Any further stuff must be in parallel with this and will reduce the resistance. You mentioned 6M which is impossible if the first resistors (and switch contacts) are in circuit from input to ground.
G'night...

[MelJon66]

Thanks George,

I can see that the circuit is attenuating the signal at the same time as maintaining a constant 1Mohm input impedance to ground, as specified for most oscilloscopes. Looking at it from the input impedance point of view I can also see that for a 20V/div input routed through R28 that the second part of the circuit is in parallel to R15 so can only reduce the impedance to ground, but can’t take it below 1Mohm.

I can also see that my approach of adding up all the resistances along the signal path is wrong. When I measured the d.c. resistance across the attenuator from input to output when on the 5mV/cm setting I got 157 Ohms. The resistances in series are R13(47) + R20(10) + R30 (100) = 157 Ohms so that works as a series of resistors because there are no connections to ground along the way. Is it better to think about the other lower sensitivity circuits just as circuits that develop a voltage across a resistor that is proportional to the input? In the case of the 20V/div selection this would be the voltage across R22.

Cheers

Mel

[buggies]

Yes - 157 Ohms would be right on the 5mV range.
The final voltage is across R73 1M to ground on the amplifier input - with each section of the attenuator providing its own reduction when required.
I have sketched the 20v circuit - attached - but don't get too bogged down when you have a 6M error to trace! It should still be in the 4 (starred) resistors/wiring/switch of the first section as these are used on th 5-10-20V ranges where you have the fault.

[golexica]

Originally Posted by Chris55000

Be very careful if you do that because the oscilloscope will be LIVE to the mains supply if you try and use a Variac direct – you MUST also include a suitable mains isolation transformer between the mains and the Variac., or a suitable low voltage Stepdown transformer with isolated secondary windings between Variac and 'scope!

Or worse, it would be live for a very short time- oscilloscopes voltages are ground referenced, so the secondary of the transformer inside them is also going to be ground referenced, so trying to power it directly with an autotransformer while it was still grounded, would create a short to ground, and much excitement

[MelJon66]

Many thanks to George for his input and drawing of the attenuator circuit. I am much clearer about what is going on now.

A fellow forum member has very kindly offered me a transformer for this ‘scope so I am carrying on with all the repairs so that it will be ready to go when the transformer arrives and will allow me to complete the work.

The current task is with the attenuator controls so I have now stripped down the vertical pre-amp module which was pretty straightforward, despite my initial reservations. Why the manual doesn’t cover this and issues dire warnings about removing PCB’s J1 and J2 is a mystery so I will cover it here in case it is useful for others in future.

The vertical pre-amp module has to be removed first, which requires some desoldering as correctly described in the manual. When the module is on the bench it then says remove the inverter knobs, desolder the earth braids and remove the shafts. Desoldering the braids and removing the shafts wasn’t necessary as when PCB H1 is released it can be moved back far enough to allow the inverter shaft ends to be pulled clear of the front panel when the knobs are removed. The gain and position knobs do have to be released from the front panel to allow the front of the PCB to be moved upwards. They are connected by flying leads so it is no problem. Three soldered connections to PCB H1 then have to be removed, pin 18, pin 19 and R38 resistor from its solder tag. The 6 screws holding PCB H1 can then be removed and the PCB hinged upward at the front pivoting on the flex connections coming in at the back from PCB H2. PCB H2 does not have to be touched.

The attenuator controls are under two covers beneath the PCB. With the PCB hinged upward at the front as high as necessary the two screws holding each of the covers can be removed. With the covers off the attenuators are easy to see but access to do testing and repairs is very restricted. To get better access undo three screws on each side of the module and two nuts underneath the module that secure the ground tags from the front input selectors. The module can then be lifted up at the front, leaving the channel mode and trigger selector wiring in place. The attenuator controls are now above the case sides giving excellent visibility and access for repairs. I have attached a photo of how it looks on the bench now.

With my CDU150 the problem was immediately obvious. Resistor R13 (47ohm) was burnt but still measured correctly. R28 (100ohm) was also burnt and measured over 5 Mohms. That was the cause of the poor sensitivity on the lowest sensitivity settings. I will replace these as soon as I get some time. I have checked all the other components and they are fine. The rotary switch contacts were all very clean but I have applied some contact cleaner anyway.

Anyone who has used one of these ‘scopes will know that the rotary controls are very stiff to turn, so much so that your fingers can get sore after a lot of use. These are military specification machines so they must be made for “real men”. While I was at it I spotted that each control has two very strong indent springs, each pressing a nylon peg into contact with a nylon gear. Two indent springs seems excessive so I removed the top spring on both channel controls. The controls are now much easier to turn and smoother to operate but are still very positive. If I ever need to do any work on the horizontal module I will do the same with the timebase control which is also heavy going.

Work is ongoing and I am sure I will have more questions.

Thanks for your interest.

Cheers

Mel

[MelJon66]

Hi golexica,

Sorry to be pedantic but can you expand on your comment please? I am keen to learn and safety is as good a place to start as any.

I am well aware of the "Don't float your 'scope" rule so I always have the device under repair plugged into my isolation transformer but not the 'scope I am using for testing. When my Cossor 'scope is the device under test and repair I do connect it via the isolation transformer but not the Iwatsu that I am testing with. I think that is all good practice.

When I was planning to use my variac as a temporary substitute for a burnt out secondary winding in the Cossor, Chris Williams reminded me that I should also connect the variac via the isolation transformer, which I did. Both were isolated but the Cossor was not being used as a 'scope so I'm sure that's okay.

But, when I first got the Cossor, which had not been used for many years, I gave it a visual check, fixed some obvious problems and brought it up to full voltage with my variac WITHOUT the isolation transformer. This was laziness really as I switched it in the porch so that any sparks and smoke wouldn't be released in the house and I didn't want to lug the heavy isolation transformer around. The Cossor didn't go bang and I am still here.

So, is the issue you have mentioned always the case or only in certain circumstances?

Thanks
Mel

[MelJon66]

Ahh, so you mean this only happens when the variac is being connected to the 'scope transformer secondary. But if it is all ground referenced why is that?

Cheers
Mel

[MelJon66]

Hi golexica,

Ignoring my misunderstandings in my last two posts let me try and get to the bottom of this.

When I was using the variac to replace the damaged secondary in the Cossor I connected it through my isolation transformer so that there was no mains earth return if I touched a live point with my Iwatsu ‘scope ground clip when fault finding. I always ensure the device under repair is floating and I did recognise that the variac had to be floating like the Cossor. Chris emphasized the point by saying that the Cossor would be live to the mains supply if I didn’t maintain isolation. I understand this and I use an isolation transformer as standard practice but I don’t float my ‘scope.

Okay so far but back to your point. “Trying to power it directly with an autotransformer while it was still grounded, would create a short to ground, and much excitement”. I always like to avoid too much excitement with electrical power so I would like to understand this so that I can avoid problems in the future. Is this because with a non-isolated variac connected on the secondary side of the Cossor's mains transformer, there are some components, capacitors perhaps, that present a short term low resistance path to ground at switch on and will draw a lot of current which won’t be limited by the ‘scopes mains transformer that is normally providing the power?

If that is true then I hadn't fully appreciated the potential consequences. I have already decided not to use a variac in a situation like this again.

Thanks for your contribution, it is much appreciated.

Cheers
Mel

[Chris55000]

Hi!

I have offered the OP a replacement transformer for him to try and I will try and get it posted to him today!
.
Chris Williams

[Chris55000]

Afternoon Mel!

I would recommend the replacement of BR1 to BR4 on power supply panel A before you fit the replacement mains transformer, along with those axial reservoir capacitors on PSU panel A!

The 1.5A round W04 type will do fine for the CDU150!

Chris Williams

[MelJon66]

Hi Chris,

A very sensible idea and I have ordered the bridge rectifiers. If the caps you are referring to are C2, C5, C12 and C14 the only one actually fitted to PSU PCB A is C12, which is an axial. The schematic shows them all mounted on the board but C2, C5 and C14 are each one section of three different chassis mounted triple cans. I would have to change all the caps in those cans to replace them. On page 5.3 of the parts list it shows them as (50 200V. Do you know if (50 is old fashioned short hand for 50uF?

I will see if I can find some modern caps that might be small enough to squeeze in somewhere. There is a little bit of space between the capacitor cans and the transformer output tags.

Cheers
Mel

[Chris55000]

Hi!

Yes, a few British makers had a "de facto" standard of using "50" to mean 50uF on parts lists and circuit diagrams – it was commoner on the thermionic valve days!

I use Abacom's "sPlan 7" for my diagrams, which can print the "mu" Greek letter, so I use "p" for 0.1pF to 999pF, "n" for 1000–9999pF and "0.01" and up followed by "mu" for 10nF and higher – my method comes from the Thorn/BRC method of publishing circuit diagrams!

These double can 50 + 500uF electrolytic cans, do, for some reason, have a peculiar tendency to develop intermittent shorts between the positive terminals, however, if you make a simple rectifier with a W04 bridge feeding into a 47uF 450V electrolytic, with a 100R 2W resistor and a safety fuse of about 250mA on one of the a.c. I/P leads, you can use this from your Variac & isolating transformer to feed each 50–500uF electrolytic tag via a 100K limiting resistor to make a basic reforming supply which does seem to alleviate the trouble!

My personal preference for these 'scopes would be to buy a can–type electrolytic of 470uF of suitable size to fit the original clips and use this as the reservoir caps, and add 47u 160V axial or radial electrolytics (as best suits the pinout of PCB 'A') from the +120V, +12.6V, +50V and –50V stabilised o/p lines to chassis at the appropriate PSU earth line pins annotated on the diagram!

Chris Williams

[MelJon66]

Thanks Chris,

Yes, the cans are doubles, my mistake. That makes life easier as I can use one axial to replace C12 on the PCB and three radials across the d.c. output pins for the others rails. I think they will fit okay. That's good news because radials are so much cheaper . I can then bypass the 50uF sections of the cans leaving only the 500uF section in use so there will be no potential for shorts between the positive terminals.

I can replace the 500uF's later by fitting new radials in the same position as the old cans.

Thanks for your help.

Mel

[MelJon66]

Greetings All,

Just a quick update on progress.

The burnt out resistors in the attenuator section for Ch2 were replaced and Ch1 and Ch2 trace amplitudes now match perfectly on all sensitivity settings.

The only other problem I found in the vertical pre-amp was that when the gain was adjusted or invert trace was selected on Ch1 the trace disappeared off screen. It turned out be a broken wire link supplying +50V to the Set Balance potentiometer on the front panel so it was only getting a -50V supply. Easily fixed but it was an odd fault as the wire looked as if it had been rubbed all the way through and there was nothing in the vicinity that would have done this. It was probably something that happened when someone had it on the bench in the past.

Everything is now working well and I have two lovely bright traces. Photo attached. There is still a bit of fading of the trace from left to right at low brilliance settings but as another member has suggested, this will possibly be ageing capacitors in the EHT circuit. That's on the to do list after the transformer replacement and power supply improvements.

Cheers for now

Mel