Vanishing oscilloscope traces

[icon]

Hi

I know threads on this subject are two a penny, and the problem can't be fixed remotely anyway, so I guess I'm looking for moral support rather than practical advice, but here goes anyway.

I have this oscilloscope that I have about 20 hours invested in, fixing broken switches and troubleshooting what turned out to be a duff dual JFET. Now that's all fixed, most of the time it works fine, but if I flex or tap the PCB, sometimes the traces vanish. Ah well, I'll case it back up and hopefully that'll hold the shaky bits together. No such luck of course - doesn't work at all.

So, back apart it comes. The problem is, it doesn't trigger a sweep. X-Y mode works fine. I've traced the trigger signal as far as I can, but the point where it disappears is at a component in a large and complex feedback loop, so that doesn't get me very far.

I've examined every solder joint under a microscope. I've resoldered the main timebase resistor network rotary switch contacts on the PCB, just in case. I've previously cleaned the tiny little wafer switch contacts as best I can - but you can rotate the switch without the trace coming back anyway. Similarly, none of the other rotary or pull switches consistently cause or fix the problem.

I've gone tapping around both when it's working and when it's not working, trying to find a loose wire or component. Nothing.

The only thing I can think of is a tiny crack across a trace somewhere which is normally held closed but can be opened if the board flexes. How the hell do I find that? I'm out of ideas!

Bleat over
John

[G7PSK]

Put the board in the deep freeze, the crack should then open up and become visible, the other thing is get some engineers blue and rub that into the traces (works best if done after freezing) when you rub the blue off if there is a crack the blue will remain in it and show up clearly.

[PA0PBZ]

If something stops working you should be able to measure it. First try to find out what exactly is failing, are you sure it is the trigger or could it be the timebase? Then find a point where you can measure a difference and work back from there. Of course a service manual or just the diagrams will make it a lot easier. And another (working) scope would be very nice.*


^{* I remember having to fix my soldering iron when I only had one...}

[amyk]

Telling us the make/model of the scope would also be nice as then we might be able to follow along on a schematic and give advice on where to check.

(Did anyone else read the title as "Varnishing oscilloscope traces"? )

[icon]

Hi

I had a further play with this scope, with odd results. Luckily I do have another scope (the usual Rigol), without which fault finding would be damned hard! The manual for the scope - a Telequipment D1010 - provides schematics, sample waveforms at indicated points in the circuit, and a written explanation of how the circuit works. I've attached the schematic and explanation below.

I found that I could trace the trigger signal to 'A' at IC3, which is a dual flip-flop. By pulling the 'X-Y' knob out and pushing it back in, I could observe a single rise at 'E', followed by by a fall - but only one cycle. I was then poking around 'C' (at the blunt end of D76), looking at the Rigol's screen when I noticed that there was a repeating signal - a glance at the D1010 screen showed that it had burst into life. Well obviously I need it *not* working in order to fault find, so I flexed the board and bashed it with increasing force, trying to get it to fail. I gave up in the end and put it back in the case, convinced that would get to to stop working. No - bizarrely, it still works. What the hell?

BUT! In case after a few on/off cycles it packs up again, do please glance at the schematic and tell me where I might focus my efforts. As a reward, here's the old gal functioning:



Cheers
John

[vk6zgo]

Does it work on "external trigger",or "line"(mains freq)?

I don't know anything about Telequipments,but BWDs do this a lot.
The problem with them,is that the trigger is obtained from a voltage divider hanging off the vertical output to one of the deflection plates.
One of the resistors goes high,reducing the trigger level so the thing doesn't trigger.

You can still trace the triggering pulse down to where it disappears into a jungle of bits,but the real problem is insufficient amplitude in the first place.
You would expect it to be a permanent problem,but many times,it's "just on the edge"of failing.