Tektronix 465B needs help

[Seg]

Yesterday was the first rainfall the Bay has seen in months, which just happened to fall on HSC's big Saturday parking lot sale. Wet cardboard, components, DLT tapes and various other crap galore:



Anyway, against my better judgement I indulged my 465 lust and picked up an as-is Tektronix 465B for $45 USD. (40% off, )



The poor girl is a bit sick:



Looks like the horizontal board has gone wrong. I've got some service manuals to read. 

[Bryan]

Hmm, wonder if the CRT is damaged. Scope was dropped.

[ddrl46]

Hmm, wonder if the CRT is damaged. Scope was dropped.

I doub't he'd get full deflection if the crt was dropped.

[grumpydoc]

Hmm, wonder if the CRT is damaged. Scope was dropped.

I doub't he'd get full deflection if the crt was dropped.

The internal glass supports can break producing odd traces.



(above image from http://www.sphere.bc.ca/test/tek-parts/tek-info.html)

One clue is that the distortion will depend on the vertical position of the trace - move it up and down and the shape will change.

[PaulAm]

I'd look at the power supplies and check ripple values.  Bad power values can make those scopes do very strange things.

[Seg]

First order of business, the fan:



The original fan must have died so someone mickey-moused a replacement on the back, with its own power cord and switch. The impeller is gone and the fan board looks a bit cooked. Someone must have been taking that "Do not remove covers" tag seriously. Wire got broke off the switch and needed to be soldered back on. If I ever get this thing working I'll see about wiring it up inside to the power switch.

I wonder if the fan seized and no one noticed until some parts cooked and died. It appears to have seen some service before...

Tektronix did in fact publish a good general "how to troubleshoot analog oscilloscopes for noobs" guide: http://www.sphere.bc.ca/test/tek-parts/troubleshooting-scopes.pdf

Power rails look okay excepting the -8v power rail was at -6.2v, and now seems to be down to -6v as I measure it again. That's not right. Dug through the service manual for guidance, found something about checking resistor R4202, which looks okay but it is connected to what I assume is supposed to be the -8v power regulator, Q4102 but it seems to only have 1.5v going in and 0.9v coming out. What would it be doing with that low a voltage? The other regulators seem to be for +55v and +15v. And that's all I had time for today.

Still waiting on some scope probes to show up so I can check ripple with the LeCroy. Looks like they just passed through New York last night, so they took the long way around the globe for some reason...

Spent a lot of time fiddling with the knobs and feeding in test signals. It appears to sweep and trigger just fine. Seems to be some kind of 60hz noise bleeding in to both the horizontal and vertical. What is supposed to be a dot is far from it.

The beam really doesn't want to go the right, except for that one spike. So I'm thinking there's some problem relating to the right deflection plate. Wires look connected to the tube near as I can tell. I suppose physical damage to the tube is possible, but thumping the thing around on the table doesn't seem to effect the trace in any way so I hold out hope that it's electrical...

[Seg]

Well, now that I have probes and a working LeCroy, it is obvious that the input filter cap (C4419) on the -8v supply is doing nearly nothing. Oscilloscopes are handy.

[PaulAm]

When you pull that cap out, there are 4 pins that need to be desoldered, 3 from the can.  Tek often used the can as a jumper, so you need to bridge 2 or 3 of the ground pins together.

It is possible to fix without pulling the trigger board, but you get much better access if you do.

It's very easy to lift the traces on the board; use a good vacuum tool if you have one.

[Seg]

So, an excuse to go check out Al Lasher's Electronics in Berkeley. Picked up a 4700uf cap, the biggest you can easily get apparently. Also an NTE194 transistor and a fuse, to replace victims of my inept probing...

With solder wick, rosin and a cranked up soldering iron I managed to get the old cap out without much damage. Managed to tear out the plating of only one hole, of course it was the only one that mattered.

Piggy backed a 470uf cap on for good measure, which should give me 5170uf total, exceeding the original spec:



Replacement in place:



Since I ripped the hole plating out, I had to solder in a jumper to a nearby diode, mirroring the trace on the other side and thus reconnecting both sides:



And look at that, a recognizable trace!



Still a bunch of noise in the vertical. Ripple is still too high, 30-70mV. Knobs are a bit flaky, looks like I need to clean some contacts.

But that's all for today.

[Seg]

Okay, probing around trying to track down the remaining noise, I noticed it was affected by where my arms were positioned. Turned everything off around me, the noise dropped a bit. Yup, external noise. Tried putting the cover back on and bam, a nice sharp trace!



Oscillofun gives me a nice cube:



+55v +15v +5v all have no discernible ripple. +110v has 6mV ripple which is in spec. -8v has 10mV ripple which is still out of spec, with a half sine waveform.



What would cause this, post regulation? Near as I can tell the bridge rectifier is working. Looking at the schematic, I checked out everything involved in -8v regulation and it all looks okay, except maybe the tantalum cap C4203...

But that's all for today as my work table just collapsed.