Telequipment D66 Oscilloscope, - Incorrect Y-axis readings on one channel

[Jimbo]

The output from an RF amplifier that I was monitoring did not seem to be amplified as it should when comparing the input and output waveforms on my Telequipment D66 two-channel oscilloscope. When I checked each channel of the ‘scope separately against a known amplitude signal, I found that Channel 2 responded correctly at all frequencies, but at frequencies above about 100kHz the waveform displayed for Channel 1, [in Volts/cm on the screen] was progressively reduced as follows,

For a 500mV input, at different frequencies the trace displays are:

1kHz: 500mV
10kHz: 500mV
100kHz: 500mV
1MHz: 400mV
10MHz: 200mV

For a 50mV input, the corresponding figures are

1kHz: 50mV
10kHz: 50mV
100kHz: 50mV
1MHz: 40mV
10MHz: 20mV

…so the problem seems to be in some way ‘linear’

The circuit description in the Manual for the oscilloscope refers to the inputs of each channel being coupled to ‘attenuators comprising frequency compensated resistive dividers’, so should I start my investigations in this area or further downstream with the Y-amplifiers themselves?

I would be very grateful for advice from Forum Members who may be more familiar with the workings of the ‘scope, and if anyone has an original or a good copy of the Manual that they would be prepared to sell [or even loan for me to copy], I would be glad to hear form them. I already have the version that is to be found in several places on the Internet, but the circuit diagrams are poor and the PCB layouts are not in colour as they were in the original document.

Thank you for your help

Jim M0GJD

[Chrispy57]

Apologies if this question is way below your pay-grade, but did you check the same signal with the same probe into each channel? That point wasn't addressed in your post, but a faulty probe would be 1% of the effort involved with diving in to the front end of the 'scope chasing this fault.

Cheers
Chris

[frsimen]

Hello Jim

The D67 has four attenuator circuits, which are switched in different combinations to give the desired sensitivity. The 10mV position connects the signal straight through with no attenuation.

If you are using a X10 probe, make sure that the compensation capacitor on the probe has been set correctly.

Next, test the frequency response in the 10mV position. If it seems good in that position, one or more of the attenuator sections will need attention. If it is bad there, the problem will be on the input amplifier board.

You haven't stated which range settings you used, but you need to check a number of ranges, to determine which attenuator section is causing the trouble.

The first attenuator stage divides by 100 but is only in circuit for 2, 5, 10, 20 and 50 volts per division.

The second stage divides by 10 and is in circuit for 100mV, 200mV, 500mV, 10V, 20V and 50V.

The third stage is only in circuit when on the 50mV, 500mV, 5V and 50V ranges.

The final stage is only in circuit on the 20mV, 200mV, 2V and 20V ranges.

If there is a problem with the attenuator, it should be possible to identify the faulty section using the above information.

Do check the probe first, as Chris has suggested.

Paula

[Jimbo]

Thank you Paula and Chris for your helpful responses. No apology needed for making the point about probes, I should have made my test set-up arrangements clearer.

I am using an arbitrary signal generator that is built-in to a separate, second, oscilloscope as my signal source. Its output frequency and amplitude are set by controls on the separate 'scope and I am using a BNC 'Tee' connector with one leg of the Tee taking the signal to the Telequipment D66 oscilloscope with a BNC to BNC 50ohm cable, no probes, and no other 'x10' attenuator. The other leg of the Tee is connected to one channel of the second 'scope to monitor its own internal signal generator. The second 'scope has the advantage that it has a 'measurement' function that gives figured values to the waveforms on its screen so I have a comparison with my eye-balled measurements of volts/cm on the screen of the D66. [Why worry about a malfunctioning D66 if I have all this other clever kit? I hear you ask.... because since buying the new stuff I realise just what a fine instrument the old D66 is!].

At present my assessments have been with the attenuator on the D66 set to the 0.1 volt/cm range for the 500mV input and to the 10mV/cm range for the 50mV input, -simply because this should give me a 5cm high wave form that was easy to align with the graticules of the screen! However, now that Paula has explained it, I can understand what the manual means by the frequency-compensated dividers being switched 'singly or in tandem', so I will repeat the exercise and apply the same signal to different 'sensitivities' of the input attenuators.

I will report back tomorrow!

Thank you again for your help.

Jim M0GJD

[cmjones01]

Where is your 50 ohm BNC cable terminated? If there's no termination resistor at the "load" end of the cable, you'll get all sorts of strange readings as the frequency rises and standing waves are set up in the cable. That includes getting different voltages at different points on the same cable, as you seem to be experiencing.

Chris

[dave cox]

Chris is correct, a un-terminated 50R cable will give misleading results at higher frequencies. Worse still, using a 'T' connector, will compound that error as you will then have 2 un-terminated stubs that may even have different lengths

The D66 is only equipped with 1M inputs and to check the calibration you really need a 50R 'pass-through' termination. This should be connected DIRECTLY to the scope's input channel and the 50R cable can then make the connection to the signal generator (which should also have a 50R source). You can't use a T piece without messing up the impedance matching, and when it is unmatched you will get reflections and all bets are off.

dc

[Radio Wrangler]

At the frequency range the D66 works over, a BNC Tee with a 50 ohm load right on it will be as good as needed. It's what you have to do with scopes lacking a switched 50 Ohm termination when you can't find a pass-through terminator. It's one feature where the HP1740 beats the Tek 465 in the old 100MHz scope battle.

THe D66 is has a single beam tube and fakes two channels with chop and alt modes. The fact that one channel works exonerates the Y-plate drive amplifiers and at least one port of the beam switch. The Y plate drivers are where the herd to find high voltage fast transistors live. So this is very good news. It looks like your fault is in the small signal sections of Y-land.

Double check both channels have the same control settings (We've all done it, and we all know how embarrassing it feels, so double check) The poor bandwidth makes it look like it isn't just the variable sensitivity knob not set at 'cal'
Check both channels by moving the BNC connector directly from one channel to the other, in order to ensure there isn't finger trouble in the bits outside the scope.

Once you're through all this you just need another scope and work your way through from the BNC input to the input to the beam switch (is it done with diodes in this scope?)

You don't have far to go AND you can use a T to put the same signal into both channel inputs so you have a live working channel for immediate comparisons.

This makes it look like it shouldn't be too much work.

A straight forward approach should get you there. There aren't any fancy hybrid circuits, everything should be approachable. If it has dual input FETs for thermal matching then there are ways round that at some loss of DC stability.

David

[Jimbo]

Thank you Dave [Cox], Chris [CMJones] and David [Radio Wrangler] for your helpful advice.

I have been carrying out the checks in the manner that Radio Wrangler has suggested by moving my signal source lead with its BNC plug from one input channel of the D66 ’scope to the other. I appreciate that this is not an accurate calibration methodology but any lack of precision in the testing arrangement would be the same for both channels of the 'scope.

I believe that I have done most of the ‘finger trouble’ checks outside of the ‘scope itself and with the rotary controls on the D66 that require to be in the ‘cal’ positions, but I will go back and do it again just to be sure!

Thank you for the suggestion of feeding both channels with the same signal at the same time, it’s the obvious way of making the direct comparison. I had not thought of doing that before as I had been trying to maximise the size of the individual traces so that I could count squares on the graticule more easily.

Thanks again for all your knowledge and encouragement.

Jim M0GJD