Author Topic: Tektronix 475 weird traces both channels (Read 5979 times)

MadModder · « **on:** December 31, 2018, 04:32:56 pm »

I have a Tektronix 475 that worked very well when I bought it in 2007. After a year all trigging stopped working. Resolved by reseating (and deoxit sockets+component legs) a few tranistors in the trigger circuit.
After a few more years it got weirder, and I had no room on my bench to fix anyuthing. It has now been sitting on the shelf about 5-6 years non-working.

Attached is a photo of the traces. Both channels, no input signal. The set V/div does nothing.
But when feeding a signal in to either channel, the signal can be seen following the trace pattern in the photo. Very weird. And the V/div does alter the height of the trace. But the weird pattern stays the same.
V position is working for both channels.
And if I change the trig mode in any way, both traces disappear completely. Can then only be seen when pushing beam finder. If I turn the oscilloscope off and wait half a minute and turn it on, the traces are back.

Here is a film clip when I change sweep time and V/div back and forth.

I'll go through the tips and tricks I found in the other 475-threads of course, starting with cleaning contacts and then check the voltages.
I'm mostly wondering if someone else has had this exact problem.

mzacharias · « **Reply #1 on:** December 31, 2018, 05:36:01 pm »

I would definitely look at power supply issues (open filter cap or caps mainly) first.

pbarton · « **Reply #2 on:** December 31, 2018, 06:13:43 pm »

Well, I find that very unusual. When scopes fail its normally the timebase, OR, the Y amplifiers. However, your video shows that the fault affects both timebase and the Y axis, at the same time. A very rare fault!
So we have to think of something that is common to both. At first I thought a power rail, common to both timebase and Y amplifiers. But then I thought what if your CRT X and Y plates had an internal short?
That may explain the combined timebase/Y amplifier symptoms, as seen in the video. If true you would need a new CRT to fix the problem! Can you meter between the two pairs of CRT X and Y plates to eliminate this possibility?

You could try disconnecting the two CRT X plates and re-run the tests. Then re-connect these and disconnect the two CRT Y plates, and re-run the tests. Do you get what you would expect to see?
Any sign of a ±45 degree sloping line, would indicate an internal short between an odd pairing of the X and Y plates.
Re-running your video, I can already see lots of waveforms with a tendancy to be inclined at ±45 degrees!
Just look at the video start screen (you don't need to click the video run button) ±45 degree sloping lines everywhere. Symptoms also visible in the initial attached .jpg
That's what you get when the timebase ramp waveform (X), inadvertently gets applied to (one of) the Y plates.

Dacke · « **Reply #3 on:** December 31, 2018, 08:28:32 pm »

The few times I've seen strange traces like that on the 465/475 it's been bad caps in the power supply.

Armadillo · « **Reply #4 on:** December 31, 2018, 08:42:15 pm »

From the strength of the beam, the power supply appears OK.
The problem is quite obvious I thought?,

it's a joy to troubleshoot.

PaulAm · « **Reply #5 on:** December 31, 2018, 09:30:19 pm »

Definitely check the ripple on the low voltage power supplies. You'll probably find at least one of them with 5-10v of ripple on it.

This is a very common failure on the 4xx series. The filter caps dry out and fail open. The good news is this rarely causes other components to fail.

The bridges are another thing to check.

Rules 1-5 for troubleshooting Tek 4xx scopes: CHECK THE POWER SUPPLIES FIRST! Repeat ad nauseum. Nothing works if the power supplies aren't right.

David Hess · « **Reply #6 on:** January 01, 2019, 02:44:46 am »

The electrolytic input capacitors to the power supply have worn out. As long as you have it apart, change the output capacitors as well and check the rectifiers.

Armadillo · « **Reply #7 on:** January 02, 2019, 11:28:03 am »

This is rather unique problem here, so waiting for your discovery and next publication.
Worth to clear the table and make space for it.

MadModder · « **Reply #8 on:** January 02, 2019, 04:02:58 pm »

I did clear my lab bench yesterday.

Had some time to spare today so I opened the case and jotted down the voltages:

Rail	Voltage	Ripple

+105/160	135.85V	40mV but rising slowly
+110	110.4V	50mV
+50 unreg	64.9V	2V_PP ugly triangle wave ripple
+50	49.79V	50mV
+15	14.97V	30mV
+5	4.1V	3V_PP ouch...
-8	-7.97V	1mV
-15	-14.96V	1mV

But not as bad as the suggested 10V ripple!

I noticed that CR1462 (bridge rectifier for -8V) and C1472 (big cap for -15V) have been replaced. Broken traces around that cap is fixed with wires.
I guess next step is to dismantle the whole shabang and take the board out. *sigh*

Armadillo · « **Reply #9 on:** January 02, 2019, 04:25:53 pm »

I think you need bigger space for this unique problem.

Watch at 40 sec, the trace moving to the left, unblanked, timebase looks ok too.

Obviously you are on alternate mode, its tracing alternate ~~curves~~ [edit] retrace lines... hehehehehe

Can't take all the joy out of your hunt....

Happy Hunting..

MadModder · « **Reply #10 on:** January 02, 2019, 04:54:50 pm »

You devilish you...

No I'm in chop mode there, as can be seen by the depressed CHOP button. That's why there are a bunch of "squares" at the bottom.
I tried the alternating mode too later, and single channel, but it makes exactly the same pattern, but without the cut up line.

[edit]
No that is not why there are chopped up lines. It's still chopped up in single channel mode.
Damn it, why is it retracing in that weird pattern while it is drawing the curve?

Armadillo · « **Reply #11 on:** January 02, 2019, 04:58:22 pm »

No, serious, waiting for your discovery and finished video of this not easy unique problem. heheheheheheh

MadModder · « **Reply #12 on:** January 02, 2019, 05:53:15 pm »

More fun weirdness. This time with channel one only. No alternating, no chopping. But it's chopping it up in its own way anyway...

time/div 1ms, 1Hz

time/div 0.5ms, 1Hz. Why is the pattern mostly gone here?

time/div 1ms, some arbitrary frequency around 1.3kHz

Oh well. I'll buy a handfull new caps some time to start with, after I've gotten the board out for examination. Later. Some time.

In Vacuo Veritas · « **Reply #13 on:** January 02, 2019, 06:01:52 pm »

Is that a waveform from Treez latest LED driver?

MadModder · « **Reply #14 on:** January 02, 2019, 06:05:56 pm »

That would be a quite bad driver. A really annoying 1.3kHz buzz

PaulAm · « **Reply #15 on:** January 02, 2019, 07:43:10 pm »

You can buy or make little adapter boards so a standard snap cap can fit the weird 4 pin pcb footprint of the oem caps. That makes for a nice repair. Tektronix sometimes used the can as a jumper so you need to connect the 3 ground pins together.

tggzzz · « **Reply #16 on:** January 02, 2019, 07:58:16 pm »

Fix the 5V line first.

If there are no short circuits (5V is 4,1V no 0V), then a dirty technique is to solder tack a suitable value good capacitor across the relevant PSU lines, in parallel with the suspect caps with high ESR. If the problem behaviour is diminished, indentify and replace the faulty caps.

james_s · « **Reply #17 on:** January 02, 2019, 10:14:58 pm »

Wow that is a *really* weird symptom. I do agree that fixing the 5V supply is the obvious place to start. No point in troubleshooting further until that is solid.

MadModder · « **Reply #18 on:** January 03, 2019, 10:40:56 am »

@PaulAm, yes, Iv'e been looking for the triangle type adapter but the ones I find are bloody expensive.
Guess I'll just roll my own.
And the Harwin pins look neat. I'll have to hunt those down.

Armadillo · « **Reply #19 on:** January 03, 2019, 10:54:13 am »

Is it difficult to salvage the BASE from the old cap?

MadModder · « **Reply #20 on:** January 03, 2019, 11:10:55 am »

That has crossed my mind. I'll try to loot the old caps first. But mounting a new cap on that base isn't going to be pretty.
The replaced cap though... it is hanging in the air with a zip tie around the other caps, and wires going down the holes.

PaulAm · « **Reply #21 on:** January 03, 2019, 12:46:29 pm »

Search on ebay for "Tektronix 7000 capacitor" and you'll see a set for $12. There are also gerbers in a 7000 repair thread on this board. I have a postscript file for printing to pnp blue if you want to etch your own.

It's easy to damage the pcb when pulling the caps, so I wouldn't recommend reusing the base.

You had zip ties? I'm impressed; I've seen them just hanging in the air

Edit: That 7000 kit only has 1 triangle adapter , try item 273254508468, that's a kit of 5 and it's only $10.

MadModder · « **Reply #22 on:** January 03, 2019, 01:28:21 pm »

Ya, I saw the $10 one, but I need more than 5 if I want this properly done, and $14 shipping here, and our postal service adds VAT to all that, plus $8.27 for VAT handling fee for non-EU packages. That makes it $50.77

Armadillo · « **Reply #23 on:** January 03, 2019, 02:57:25 pm »

Like many others, you don't only own one scope.

That amount $50.77 for just the bases, can already buy a new used scope.

Furthermore, its so heavy, so boxy, takes so much space, and after repairing it, more than likely you are going to put it aside to collect dust, example it don't meet modern days measurement needs.

Just float it or drill holes on the salvage base, nobody is going to look inside.

Well, that's just my opinion.

MadModder · « **Reply #24 on:** January 03, 2019, 04:53:55 pm »

Yes, that is what I’m going with to start with. The bodged on cap though... Alright I ziptie it back when done.

I got the 5V cap out now. Only one of my instruments showed something, once. 255µF 0.14kΩ.

Now all instruments say theres nothing there. It’s shot alright.


EEVblog Main Site	EEVblog on Youtube	EEVblog on Twitter	EEVblog on Facebook	EEVblog on Odysee

Author Topic: Tektronix 475 weird traces both channels (Read 5979 times)

Share me