Tektronix 475 repair assistance

[bd139]

This has appeared in various places around the Internet. Feel like I need to consolidate it all on the best resource for such things, which is EEVblog forum by the looks!

I scored an apparently dead Tektronix 475 with DM44 module the other day for a respectable £29.

I'm going to enumerate the identified problems with it and what has been done and resolved so far.

Problem 1 - 15v supply out. RESOLVED!

This was a duff tant on the interface board on the output of the 15v regulator. All rails are in spec now. I will do a wholesale tantalum swap out job over the next few days. There are quite a few to replace!

Problem 2 - Occasional HV drop out. RESOLVED! NOT RESOLVED AGAIN FINALLY RESOLVED!

Every 2-3 minutes the display jumps in brightness and then returns to normal as per the following video.



Cleaning it appeared to make the problem go away for a bit but it's back again suddenly. I've opted to replace all the film capacitors and diodes in the HV section as a couple of caps look slightly mottled/cracked (like shitty RIFAs about to explode). They are being replaced with ceramic equivalents.

Note: postponing HV refurb as the issue may have been cured by another tant on the 50v rail going short periodically and self-clearing.

Confirmed it was the 50v rail tant that was doing this. It doesn't do it now that has been replaced with a low ESR electrolytic.

This occurred again and I found another dead tant on the 15v rail that might have actually been doing it to start with. Grr. Replace them all!!!

Problem 3 - chop doesn't work. RESOLVED!

As yet undiagnosed. Shows channel 1 in chop. Hopefully not the channel switch IC gone phut and just a dirty switch or something! Not a big priority at the moment. Drifted resistor in chop oscillator bias circuit and duff ceramic decoupling capacitor causing VCC to be around 4v which broke the startup condition for the oscillator. 7400N's are possibly hooky so I'm ordering a couple as replacements anyway.

Problem 4 - fine horizontal pot broken - IGNORE

Just going to try and replace this if I can find one or live with it. There's enough resolution in the coarse pot for now.

Problem 5 - at fast sweep speeds, the trace starts moving to the right - RESOLVED

This one is a real PITA. I've had another scope on the timebase and that appears to be working fine. Horizontal amp seems fine as well as it has full deflection and bandwidth in X-Y mode. However as you increase sweep speed from 1us to 50ns/div, the trace compresses and moves right. Scoping the timebase shows that the sweep return is much shorter for these timebase settings than others and if you turn the brightness up fully you can see the retrace so I'm wondering if it's Z-axis related. Also looks a little blurry in delayed sweep mode when the trace is intensified. This is possibly CRT blooming or on it's way out (is it?). Any hints welcome!

After much arguing I spent until 3AM working this out. Unplugged the z-axis cable to poke a scope in the hole, plugged it back in again and gone. Ahha we have a problem. So I resoldered each end of the z-axis cable's sockets and hey presto it works. Looks like one of the terminations was iffy and had an accidental filter in it.

Problem 6 - dirt and corrosion - RESOLVED!

Slowly working on this. Corrosion is minor and appears to be limited to the front cast panel. Everything else is cleaning up nicely. All the controls are in really good condition and the bits that are cleaned leave it looking like it's brand new!

Any help or suggestions on any of these issues would be appreciated as this is a fine cosmetic specimen; would be nice to get it up to spec again.

State after I fixed the power supply:

[tggzzz]

Problem 4 - fine horizontal pot broken

Just going to try and replace this if I can find one or live with it. There's enough resolution in the coarse pot for now.

Classic; can be caused by resting the scope on it when removing the cover. The spindle is forced inwards, separating it from the clarostat's rotor.

You can try pulling on the knob while rotating, to see if that re-engages the spindle and rotor. If it does, you can consider putting a coilspring between the two knobs, to provide the equivalnt of a gentle "pull".

The clarostats are available at the usual places; if not, I have a 311-1192 that I might be prepared to part with.

Problem 5 - at fast sweep speeds, the trace starts moving to the right

This one is a real PITA. I've had another scope on the timebase and that appears to be working fine. Horizontal amp seems fine as well as it has full deflection and bandwidth in X-Y mode. However as you increase sweep speed from 1us to 50ns/div, the trace compresses and moves right. Scoping the timebase shows that the sweep return is much shorter for these timebase settings than others and if you turn the brightness up fully you can see the retrace so I'm wondering if it's Z-axis related. Also looks a little blurry in delayed sweep mode when the trace is intensified. This is possibly CRT blooming or on it's way out (is it?). Any hints welcome!

Might this be related to the DM44? See what happens if the DM44 is in volts/ohms/temperature mode.

[bd139]

Yeah same trouble with my last 475. Seems to be a big failure of the device putting that knob there. Or letting whichever knob did it out of prison

I can pull it towards me fine and it will engage. I might drill the rivets out and see if I can repair the pot first. If not, I’ll look for a replacement. I won’t steal yours unless you really want rid of it as I know how rare they are.

With respect to the horizontal and the DM44, that’s a good question! This appears in horizontal A only mode however so I suspect it is related to something else. Worth a look and trace the circuits though.

Edit: oooooh tempting https://www.ebay.co.uk/i/282694103060

[bd139]

I finished cleaning it:

Before:



After:

[bd139]

Debugging chop circuit now. After some poking around it appears that the chop oscillator, a simple SN7400 multivibrator doesn't start by itself when you put it in chop mode. The moment you touch pin 9 with the scope it springs into action!

Circuit in question here:



More debugging later. Will write the voltages down and then pull and check all the parts.

[tggzzz]

Debugging chop circuit now. After some poking around it appears that the chop oscillator, a simple SN7400 multivibrator doesn't start by itself when you put it in chop mode. The moment you touch pin 9 with the scope it springs into action!

Check the solder joints and that the resistors are in tolerance and wiggle them to see they are not cracked. Use some DeOxit or equivalent to ensure IC leads are making good contact.

[bd139]

Good plan. I’ve yanked the IC quick and checked the bias voltages and they are well out so there’s a dicky resistor or two by the looks. There’s nearly a volt difference between pins 9 and 12 which makes it slight more stable than it should be.

[SeanB]

It's a 7400, so if the rest checks out change it, but beware, use only a 7400, not HC, LS, ACT or anything other than bog standard TTL there, as this relies on the input bias current of the gate to also provide part of the bias.

[bd139]

Ok I found one drifted resistor which I have replaced, I have touched up all the joints, swapped the IC out for another new SN7400 (not LS) that I had and it still won’t bloody oscillate without a kick.

Grr!

[bd139]

Found t! Bloody ceramic decoupling cap!

[bd139]

FMH it just blew up again. No HT at all!

[bd139]

C1428 across the 50v line was short. Who the feck puts a 50v tant across a 50v line? No derating at all!

Have replaced it with a Panasonic 15uF 63v low ESR electrolytic. Wonder if that was causing the drop out as per problem 2.

Really can’t wait until my pile of replacements arrives next week!

[tautech]

C1428 across the 50v line was short. Who the feck puts a 50v tant across a 50v line?

Many, as manufacturers from way back then didn't know any better.

For a bit of gear that you intend to keep I'd go through the BOM and schematics with a view to rate higher any Tants where possible.....especially so if you have in all apart. When rail voltages are 30+ the prefered derating of 50% becomes challenged and while probably fine you might want to explore other options like electrolytics with a ceramic in parallel. Implementation is key to deciding what to substitute with, sometimes it's only remote bulk capacitance while in others you also need the speed of Tants so there you need to add a ceramic too.
Today electrolytics are in much smaller packages and often just a low ESR version will do the job in which a Tant was used.

[bd139]

Fair point!

I've ordered a whole bunch of replacement parts for this. Anything which is close to the signal chain is tantalum, suitably derated. Anything else is low ESR electrolytic. I'm not derating by 50%. I've picked 25v for the 15v rails etc. It's "close enough" without increasing the price too much. New units are high quality AVX part.

The replacement BOM cost more than the scope did so far, which I was expecting. Budget is set at £150 to get this up to spec and the expenditure is £65 so far so I'm winning pretty well so far.

A positive note: the tant that just went short appears to have been causing the HV to drop out. Been staring at it for 20 minutes and no problems so far.

[Specmaster]

Hmm, never realised that you were having so many issues with this. It is probably a result of it being left unused for so long, I think it harms electronics to leave them unused for long periods, they start to decay and corrode, but you're doing a good job so far.

Considering how it was when you first saw it, it now looks a million dollars, great job.

I don't like the way that they seem to have cut corners with running those tants at their rated voltage, nothing lasts long if it constantly being driven flat out, are all Teks like that?

[bd139]

There's always many issues with Tek kit that has been left sitting around. You don't usually find them until you've been using it for a few days

Agree entirely. I think this was left in the guy's garage for a number of years. Some green flakey corrosion around the bezel indicates some damp as well. The main filter caps are all really good though; they're all sprague branded ones which tend not to die miserably even after 40 odd years! All of the major issues so far are failed capacitors so all the tants are being replaced in the unit. Just waiting for them to arrive now. Sometime this week apparently.

As tautech said, I don't think they knew any better back then. They worked as they were rolling off the production line and to the end of their expected service life absolutely fine. It's quite spectacular that the things even work at after this amount of time. I couldn't see my Rigol lasting more than a decade or so.

Edit: just to add, inside this scope was completely untouched. No visible repairs. It's the ones with the repairs that are problematic I find. You have no idea what quality soldering iron jockey has been inside it

[bd139]

Large bag of capacitors just arrived. Starting to regret this refurb 

22 to replace as priority! Rest (x53) will be shelved until I can be bothered and I'll do them in batches

Z-axis problems appear to be tantalum related as well. The VCC drops rapidly in a couple of places just after supply decoupling, probably due to increased ESR. Hopefully replacing them will do the job.

[tggzzz]

Z-axis problems appear to be tantalum related as well. The VCC drops rapidly in a couple of places just after supply decoupling, probably due to increased ESR. Hopefully replacing them will do the job.

Quick test: tack-solder a new capacitor in parallel, and see if the problem is improved.

CPC sells cheap tant beads.

[bd139]

Not any more they don't; I just bought them all. I'm joking obviously but I did get them from there though

I decided to skip the AVX ones from Tayda because I'm impatient and they take 2-3 weeks to turn up

Good plan - will do that.

[bd139]

Oh the b*****d thing has started dropping out again.

[Specmaster]

Debugging chop circuit now. After some poking around it appears that the chop oscillator, a simple SN7400 multivibrator doesn't start by itself when you put it in chop mode. The moment you touch pin 9 with the scope it springs into action!

Circuit in question here:



More debugging later. Will write the voltages down and then pull and check all the parts.

Was it C336 by any chance?

[Specmaster]

Oh the b*****d thing has started dropping out again.

Did you replace that tant that you thought might have high ESR?

[bd139]

Was it C336 by any chance?

Nope. Miles away from that on the schematic...

Oh the b*****d thing has started dropping out again.

Did you replace that tant that you thought might have high ESR?

I've literally just replaced 22 of the things and it's not doing it now (yet). I'm setting my digital scope on single trigger mode to see if I can catch any transients on the 50v rail. Fingers crossed!

[Specmaster]

Your digital one can save the image automatically I take it? I might have to invest one of them later on then.

[bd139]

It can indeed. You set it to trigger on single events and it'll quite happily just sit there until something interesting happens. In this case I'm connecting it to the 50v rail and setting the trigger point to 48v on a falling edge. With a bit of luck that will catch a power supply transient at the same time as the display problem. If it doesn't and it still does it, I'll work my way around all the rails and see where I end up. If it's none of them, I'll look at the HT again and probably rip it all out. There is a chance it's the HV multiplier at which point it requires a slightly extensive rebuild, one I'm not up for if I'm honest.

I power cycled the scope quick to trigger this so you can see what it would record:



You can zoom into that and analyse it after the fact (and screenshot it onto a USB stick which is really handy!)