Tek 475 vertical noise

[SingedFingers]

Got an interesting problem with my bargain Tek 475. Unfortunately there is a lot of vertical noise which is resulting in very fuzzy traces. I immediately suspected the power supply so I measured the ripple on all rails with my true RMS meter and it is well in spec. So I stuffed it in X-Y mode and centered the dot and lowered the intensity and adjusted the focus and astigmatism as best as possible. Turns out there is a vertical line 1/5th of a div high. The horizontal is pin sharp but the vertical is crap. The fuzz appears on both channels so I did the Tek 465 trick of tightening up all the screws to ensure a good ground; this works very well in some cases. Alas no difference. the above was done with the channels grounded at the scope and the case on and the unit warmed up for 30 minutes. There is no near RFI as far as i can tell and nothing that interferes with its Philips colleague.

Then I noticed that if I drop the BW limit to 100MHz, things improve. Then if you drop it to 20MHz it's actually quite good but not perfect. If I sit it next to my Philips PM3217 that is absolutely pin sharp but the Tek is still way out.

Can anyone suggest any diagnostic steps or information that can help this issue at all. I'd rather use this than the PM3217 as the bandwidth is much higher and the trigger isn't made of poop.

Any help much appreciated

Edit: for ref this is a very late model 475 that has the new bridge rectifiers and no carbon comp resistors in it (thank feck)

[Andy Watson]

Just checked my 475. If I set the channels to 2mV per division and 200MHz I get about 1/5 division noise - but I expect that. All other settings give a crisp display. There is some evidence of the EHT oscillator breaking through at higher intensities. This scope hasn't been used in anger for 15 years and is, as far as  I know, still operating on its original capacitors.

[SingedFingers]

I'm getting the same on the 1v/div setting though which makes me think this is post-attenuator.

I fully expect noise at 2mV div, although there is virtually none on the 2mV/div on my PM3217.

[Paul Moir]

You said it's only vertical in xy mode?  Does that mean it's only on channel 1? 

[SingedFingers]

I can see it on both channel 1 and 2 when sweeping and grounded. In X-Y mode it is vertical only.

This eliminates specific channel involvement. I think it's after the channel switching as this is single beam. Also the BW limit is after that at the rear of the vertical amp.

Edit: Actually I'm wondering if this is the leaf switches on the BW limit switching array. Going to go and clean them with IPA and paper...

[David Hess]

200 MHz and 1/5th of a division at 2mV/div?  That is about the level of noise I get with my 7904 with 7A26 (200 MHz) and 2445B (150 MHz) so I think what you are seeing is normal.  Some of the apparent noise is from the spot size which is higher than some contemporary instruments because Tektronix used CRTs with scan expansion meshes for higher vertical deflection yielding higher bandwidth.  The bandwidth limit on the 475 should have a very visible effect on displayed noise which you also observed and I see the same thing on all of my oscilloscopes.

Your Philips PM3217 might not have a scan expansion mesh (The PM3217 schematic seems to show one but it is equivocal.) giving it a clearer CRT and its lower 50 MHz bandwidth will have less noise anyway than a 475 at 100 or 200  MHz.

[Paul Moir]

I believe SingedFingers says he's getting about 1/5th a div at 1v, or about 200mV of noise in reply 2.  It must be coming in somewhere after the multiplexer.  But to be sure, is it even on all the attenuator settings?

[SingedFingers]

Back again. Some good news. The cleaning has improved it to reasonable levels. Looks like it was the old dirty switches problem. There was that noise on all attenuator settings which was consistently the same 1/5 of a div to start with. In 10mV div and greater at 200MHz BW there are negligible noise now and the trace is a dot again in X-Y mode. 2mV/div is noisy but that as David Hess says, this is expected really.

I cleaned the areas marked below where there are leaf switches relating to the bandwidth limit function.



It's nowhere near as good as the PM3217 for sharpness still but that unit I have was stored for longer than it was used.

I paid literally nothing for this and it came with a 250MHz probe so I'm happy now

[Paul Moir]

Good stuff.  Don't forget though that with great bandwidth comes great noise. 

[tombi]

A couple of times I grumbled about the fuzz when compensating the probes on mine and later realized most of it goes away if I just ground the probe ground. On my other scope (DSO) this isn't a problem.

Maybe as others have suggested the EHT is noisy on these.

Tom

[David Hess]

A couple of times I grumbled about the fuzz when compensating the probes on mine and later realized most of it goes away if I just ground the probe ground. On my other scope (DSO) this isn't a problem.

Maybe as others have suggested the EHT is noisy on these.

Tom

This is a problem during probe compensation with all high bandwidth oscilloscopes whether the ground lead if it is long is connected or not.  If you use a coaxial connection to the probe tip for the compensation signal, then the noise goes away.  Some old oscilloscopes used a BNC connector for the probe compensation output so a BNC to probe tip adapter could be used.  Others just had a probe tip socket for the compensation.  Later only a little hook for the probe to grab onto became common and those have the problems with noise.

[SingedFingers]

I miss that on my old Tek 453 actually. There was a BNC for the cal output. You just touch the probe inside it and lean the ground ring on the edge of the BNC shell without worrying about the ground strap.

I've left the scope running for 3 hours solid now with a 10MHz signal off an xtal oscillator straight into CH1 with the delay sweep on a rising edge, 200MHz BW limit and it's spot on now. The fuzz has decreased to reasonable levels.

Super happy with this. I've got 5 scopes now, 4 of which are Tek, 3 of which work and this is the bee's knees of analogue units without having to lug 7000 series units around

This just popped into my head from Linear AN47:

[David Hess]

I've left the scope running for 3 hours solid now with a 10MHz signal off an xtal oscillator straight into CH1 with the delay sweep on a rising edge, 200MHz BW limit and it's spot on now. The fuzz has decreased to reasonable levels.

I have noticed before that if an oscilloscope is left in storage for a long time, then the CRT sharpness and brightness is impacted and improves significantly with hours to days of operation.  My guess is that this has something to do with ion contamination inside the CRT which is swept out as it is operated.

This just popped into my head from Linear AN47:

J is typical of the finger probes described in the text.  Note the ground strap on the third finger.

[SingedFingers]

Good points there. I think there was some moisture in it as well.

I love that appnote

[tggzzz]

Good points there. I think there was some moisture in it as well.

The user manual for one of my DVMs has a section describing how, after it has been left in storage, you should make an "oven" so that it can cook itself over the course of a day.

I love that appnote

I have several paper copies. Maybe I should sell one for an obscene amount.

[SingedFingers]

Cor real paper copies. I miss them.

I was working for a defence contractor in the mid-late 1990s and I used to intercept all the outgoing (to the skip) databooks and appnotes. Unfortunately in a random moment of complete stupid in about 2002 they all went to the tip. Massive regrets there.

You could probably get £20 a copy for AN47 I reckon. Easy money.