Tektronix 2236 delayed sweep jitter

[alexanderhiam]

I recently scored a Tektronix 2236 on Ebay, and for the most part it seems to be in great shape. The one problem I've found is that the delayed sweep circuit seems to be introducing a lot of jitter. I'm trying to measure the rising edge of a 1kHz square wave, and the scope doesn't have the intensity range to see the trace with a small enough time base, so I'm using the delayed sweep. When I lower the B time base enough to where I should be able to get a good rise time measurement, the jitter is so bad that it's impossible to take a measurement (about 100ns total deviation on 50ns/div). The same signal has no noticeable jitter on my digital scope with the same time base, nor is it noticeable on the channel 1 trace, so I'm guessing that it's being introduced by the delayed sweep circuit. Any thoughts?

[jahonen]

Where is your 1 kHz coming from, are you sure that the signal itself does not contain excessive jitter?

Regards,
Janne

[grumpydoc]

Where is your 1 kHz coming from, are you sure that the signal itself does not contain excessive jitter?

I think Alexander already said "The same signal has no noticeable jitter on my digital scope with the same time base, nor is it noticeable on the channel 1 trace"

Edit: having said that I just thought again about what you're doing - when you said there was no jitter on channel 1/the DSO, how did you check? Photos might be useful.

Manuals here if that's of any use.

[grumpydoc]

Sorry, posted quickly this morning before heading to work - I've now had chance to think properly  about what you describe.

Ok so you have a 1KhZ signal (the 'scope cal output?) going into one channel on the A timebase. You want to look at the rise time of the square wave so you're using the delay timebase and B sweep to focus in on one edge?

You have several possible sources of jitter.

One is the delay time jitter - this is specified for the 2236 as 0.01% of the total available delay time which comes from the main timebase - basically one sweep. Assuming you have this set for 100us/div (so you see one cycle of your square wave) this corresponds to 100ns. If you have the A timebase set slower - say 1ms/div then the jitter could be 1us. So the delay circuit could introduce the level of jitter you are seeing without being broken or out of spec.

Then there is the jitter in the signal. For a 1kHz square wave just 0.01% (again) jitter would be 100ns. If your signal is generated from an RC oscillator it could easily be this high (probably higher) and most 'scope calibration outputs are fairly simple RC oscilators. If you have looked at this with the A timebase and 1-10 cycles on the screen on either your 2236 or DSO you wouldn't see this level of jitter.

Finally there is the jitter in the trigger point. This will depend on how fast the square wave edge is and whether there is any noise. It should be smaller than the two sources of error above but could be significant if the edge is rising at 20ns or slower.

Usually rise time measaurements are taken on the triggering edge to avoid this sort of thing. That's the reason that analogue 'scopes have delay lines to allow you to see the triggering edge at fast sweep speeds. You said the trace won't go bright enough for you to see it - have you looked at why that is? It might be a worn out CRT but it's probably worth checking the basics and making sure all the 'scope PSU rails are at the correct voltages.

The service manual in the link I posted should help you check the 'scope out.

[alexanderhiam]

Thanks guys. The 1kHz signal is coming from a GW Instek SFG-1003 function generator, so I would think the jitter it introduces should be pretty minimal at 1kHz.

grumpydoc, I think I must be seeing that delay time jitter. Not only does it come out to the right number (I probably should have looked into that before posting, oops), but when I go up to a 1MHz square wave with a 100ns/div time base there's no noticeable jitter on the delayed sweep.

I just tried again to look at the triggering edge with a fast time base on channel 1, and it seems to be a lot brighter now than it was yesterday for some reason , but I still can't quite see the whole rising edge. It seems the delay line is not quite delaying as much as it should. The service manual describes a 90ns delay (3-7, Delay Line Driver), but i'm only seeing about 40-50ns, which is not quite enough to see the full edge on that 1kH signal. Here's the trigger view on the same 1kHz signal at 50ns/div:



I got excited when everything seemed to be working great at first, but I guess I'll have to do some work on it after all!

[tekfan]

I have the same scope and I can say that it is exactly the same. What you see on your scope does not mean it is broken or not calibrated correctly.

Using free running delayed time base with the delayed timebase switch set a lot faster than the main timebase switch will introduce a lot of jitter and the B trace intensity control won't allow you to increase the brightness of the trace nowhere near as high as you would like. There is the option of using triggered delayed timebase to elliminate jitter but that wouldn't help you in this case where you're trying to display a square wave.

The option bet as you already found out is to just use the main timebase with a very fast sweep setting and relying on the delay line to show you the rising edge of the waveform. The delay line is a long loop of wire inside the scope. The signal is fed into one end and it comes out the other 90ns later. You can't do anything to adjust this delay. The delay line provides a delayed signal to the vertical amplifier so that in the meantime the unblanking circuit has time to turn the beam on (it can't do it instantly when it receives the trigger pulse). When observing a waveform whose risetime is longer than the delay time of the delay line this is what you see. The trigger circuit triggered on the middle of the rising portion of the waveform. You can set the trigger level control more anticlockwise so it triggers earlier. This should bring the entire rising edge on the screen.

[alexanderhiam]

Yup, hadn't thought to actually measure the rise time; it's 100ns. Can you tell this my first analog scope?

[tekfan]

Search w2aew's channel on youtube. He has a lot of very useful videos on how to operate an oscilloscope and using it to measure waveforms. You can also read the manual for your scope. Tektronix always included an operator section in the manual that teaches you exactly how to use it and get the most out of it.

In my oppinion the 2236 is the ultimate analog scope. Compact and lightweight and has the multimeter/counter/timer which makes exact frequency voltage and timing measurments without any external gear. The multimeter is even autoranging can do resistance up to 2Gohm and has a true RMS AC voltage reading.

[alexanderhiam]

Thanks for the tips. The scope did come with the original manual in perfect condition, so guess I have no excuse not to give it a full read through.

Yeah, I thought the counter/timer/multimeter system would be more of a gimmick, especially with the small display, but it's come in handy just about every day since I got it last week. It definitely feels like I chose the right scope!