Test Equipment: Oscillocopes

[Brian516]

Good, thanks. It would be nice to have some good images of the board layouts. As you can see, the ones in the pdf really are bad and hard to read.

When I change the zoom to 100%, everything seems perfectly readable.  If you don't have anything visible showing your % zoom, try the "view" menu. Depending on your program, it might be under zoom or magnification, and turn it up some until it's readable.  If you are using Adobe Reader, go to the "View" menu, -"show/hide" - "toolbar items" - "select & zoom" and then check "zoom value" and the two above it.  That'll add it to the toolbar so you can change the zoom much easier.

That seems odd. They must be very tiny. I think the smallest ones I've got are 100 nF. A 68 uF 16V one is about the size of a large pea. This is the great advantage of tantalums, in addition to their fast response: they are really physically small for the capacitance value.

Yep they are pretty tiny.   

The very idea gives me the willies. I would definitely _not_ recommend that you put any kind of compound on anything. Clean it off the coax connectors!
The black oxide layer is silver oxide forming on the silver plating of the connectors. Cleaning carefully can remove this, but you don't want to remove the silver plating by being too aggressive with sandpaper!

Will do, sir. Will clean it all off with alcohol.  I didn't rub the sandpaper on it much at all, I didn't even take off all the silver oxide layer. I just rubbed it gently around the contact one time to clean off any debris.

I have been thinking all along that the regularity of the glitch and the steady rate no matter the scope settings seemed like an arcing problem. You may have found the trouble.  The one from Allelectronics should be fine. Don't worry about the tolerance, you are now dealing with modern components that will be much closer to the printed value than those ancient ones in the scope. C1476 appears in the schematic on page 228 of the .pdf SM and appears to be an ordinary decoupling cap for the Emitter of Q1478. This should be a "nothing special" capacitor and the one from Allelectronics will work perfectly here.

Well now that I know that it doesn't need to be anything special, I will check my boxes for a quick replacement! I most likely have one of those, as I remember vividly having quite a large number of .1uf caps of that type.  lets just hope that I have at least one that's 200V or more! Then if that was the problem this whole time, I can just get that one done and over with, leave the other two for now, and give 'er a shot!   of course I'm still going to check the rail voltages before and after.
One thing I did notice, which would go hand in hand with this bad cap:    after it warmed up the first time, it started with the noise. I let it cool for a while and then powered it back up, and it was still there.   Therefore saying the damage happened when it warmed up the first time round, and didn't go away, so this little cap arcing issue that melted the lead could very well be my problem. let's hope it's my only problem, and the other two caps are just cosmetic issues! (I'll still replace them, though. If it was a part that was more than a dollar and was only cosmetic, I would leave it. But I can get 2 or 3 good quality tant caps for a dollar....)

The Q1478 transistor is a PNP silicon, and 2n5401 is the generic cross for that one, should you need to replace it. This is a HV silicon amplifier and the NTE equivalent is NTE288. Cheap and easy to find if you need to replace it. Check that associated resistor R1476 also for off-value.

Will checking the Hfe on a DMM actually tell me if a transistor is bad?  I know sometimes I may not be able to find the datasheet for the original cap, but the generics should still have close to the same Hfe range, right?  I will def check the resistor as well.  I checked a bunch on the preamp board and at the inputs, and they were fine.

http://www.reprise.com/host/tektronix/reference/find_transistor.asp

That is a great tool to have!  thanks for pointing me to it! 

This is where a Variac can come in handy. For general troubleshooting reference: if you have a device that is blowing the main fuse, for example, you remove the fuse and put an ammeter across the fuse contacts, then use the Variac to slowly run up the line input voltage from zero to your local mains voltage. If the current draw indicated on the ammeter approaches 90 percent of the fuse value.. STOP!
If you suspect a severe power supply fault, you can do the same thing, while monitoring the suspect power supply with a voltmeter. Again, if the output voltage of the monitored supply begins to look like it is going to exceed the spec value + tolerance, then STOP increasing input voltage from the Variac.

However I don't think your problem fits in this category. The scope seems to be working, basically, at least so far. Once these suspect caps are replaced we can go further into the various features and functions to see if there is anything else wrong.

Well unfortunately I do not have a variac yet.  Closest thing I have is a dimmer switch, and there's no way I'm going to use that on any equipment! haha  different when using for some oddball experiment, like I did for the mini Kapanadze before I had a clue...

I never did see what the CH2 trace from the calibrator looks like. I'm presuming it is also mostly OK but with the same glitching as you've shown on the CH1 trace.

If the power supply rails are OK, or if I need to replace that cap before running it, and have a replacement, I will make another vid with both CH1 and CH2.  Rob, the guy I got it from, said that there was something minor wrong with CH2.. and I remember that the trace was almost flat (had no vertical amplitude) when I first turned it on before I did any cleaning or anything to it.  Remember? I should have but didnt make a vid of it, but that was when I had a ton of noise from the switches, and the trace was also very, very dull, even when I had the intensity turned way up. ( don't know why it was like that, but whatever cleaning I did fixed it, cuz in the vids I made it was nice and bright.)

If for some reason your copy of the manual is unable to be zoomed in on, let me know and I will scan the diagrams that we are dealing with for you.

Brian

[MileHigh]

Going back to the scope, there seems to be a problem with the raster scan.  So there must be test points for the horizontal and vertical rasters.  I am using a very liberal interpretation of the term "raster."  More specifically, there may be a three test points for the vertical raster.  The first one would be the DC vertical bias signal.  Then next one would be the input signal.  And the third one would be the DC bias plus the input signal.  Presumably the DC bias signal goes "off screen" during the horizontal blanking interval.  I don't know the guts of scopes enough to know if the beam is modulated off during the horizontal blanking interval also.  If it is there may be a separate test point for that, and that would form part of the "raster signal matrix."

When the scope is giving a good display at the slower time base, you should be able to scope (I realize you need a second scope to do this) something akin to a "TV signal" for the two rasters.  I am going to assume that during the horizontal flyback interval the vertical raster is supposed to be "down and off the screen."   Then when the time base is faster, something is corrupting the regular raster waveforms.  It looks like there may be a problem during the blanking interval.

If you did indeed observe a corruption in the raster waveforms, then you could trace the signal and follow it back through the schematic.

Higher raster speeds presumably mean more stress on semiconductor components.  Perhaps there is a transistor somewhere in the signal path that starts to croak from the stress above a certain frequency.

[Brian516]

PW and MH -

I know that you can't really see it, even with the close up of that .1uf 200V cap, but from the other side and various angles I can easily see that there is a split in the lead and a chunk melted out of the lead.   I tried to get a pic of it using some small inspection mirrors but I just can't get the angle. 
Either way that cap is getting replaced, and so is at least one of the cosmetically damaged tantalums, regardless of whether it is bad or not.  I'm perfectly capable of soldering in tight places.   

I am still just learning the ropes of equipment repair, so I appreciate all the pointers and pointing out what the proper order of operations and methods are for checking a piece of equipment. I will definitely be hooking a DMM up to the V-rail test point and ground and powering it on for a min, then letting it cool for a few min,  and checking the next point, doing so with both DC and AC settings. 

Yes, sandpaper and equipment don't really go together. I know not to even bother now. However, just for an FYI, it was 1000grit sandpaper, which feels more like a piece of paper than it does sandpaper.   I know not to touch the PCB or any actual components with it, as it would remove the protective layers.  All I use for cleaning is IPA, and Deoxit where absolutely necessary, like inside of pots or switches that aren't exposed to where I can easily get a camel hair paintbrush and IPA in, and just spraying with IPA won't do the trick.  I've also gotten the pointer from MH about using a tiny drop of WD40 and letting it sit for a couple days. What wasn't mentioned, and what I would likely do following these couple of days and a couple minutes of use to make sure it's done its job, is clean off the WD with some IPA.

Brian

[MileHigh]

the other two caps are just cosmetic issues! (I'll still replace them, though)

I would strongly advise against that.  If there are little trim pots all over the boards in your scope that typically means that the boards had to be tuned and aligned with the trimpots as part of the manufacturing process.  Changing a cap could adversely affect the tuning and alignment of your scope.  I am being cautious and very generic in my statement.

[picowatt]

Great post PW.  I also did not think the caps looked like they had any problems.  One suggestion for Brian when he checks the rails is to put the multimeter on DC and then on AC.  I believe that the multimeter on AC would pick up any glitching on a supply rail within limits.  However, careful because some "cruder" DC supply rails may have some AC on them which is normal.  I apologize because I am flying blind and haven't looked at the schematic.

In my days we would sprinkle tantalum caps around a board for the bulk decoupling and then put a 0.1 uF ceramic at the power pin for the chips.  I have to confess I didn't even know that they ware also good for HF decoupling, all I knew was that it was just standard practice to sprinkle them around.

It's interesting (fun?) to look at the "decoupling and regulation war zone" around a motherboard CPU socket these days.  I am 25 years out of the design game which is 100 years in non-tech time.

MH,

The AC voltmeter idea would be fine for rails, but if the issue is with the switcher or the sweep/HV supply, I'd want a decent probe and 'scope.

A bit off topic, but my favorite tantalum story is with regard to a PA mixing console a friend owned.

About 20 years ago or so, I visited this friend while out of town on business.  He was bummed because he had a 3 or 4 year old low cost, but nice looking, 28X4 PA mixer with a shorted +18V rail.  He took it in for repair but was told that due to the low cost method used to construct the board (all channels on one massive motherboard) it was impossible to isolate which of the literally hundreds of tants used at every IC for decoupling was shorting the rail.  He was told all tants would have to be changed at a cost more than the board was worth.  The shorted cap made the supply go into current limit so the entire rail was down at all measurement points.  Were I at home, I would have connected a constant current supply to the rail feed and used a comparator probe or a 6 digit voltmeter to follow the short (using the PC traces as a CSR and noting polarity).

Long story short, we took the board apart (every knob and control nut had to be removed), unplugged the supply from the motherboard and using a 12V gel cell (i.e., a high current source) and clip leads he had laying around, I attached the gel cell to the +18V rail for a few seconds and then BANG!  There went the bad cap!  The ugliest smelling/tasting curly trails of black "something" wafted downward as the room slowly cleared of smoke.  We replaced that single bad (and now exploded/mostly missing) cap and reassembled the board.  A thorough test showed all preamps, EQ's, etc, to be working just fine.  He was still using it eight years later when I was on another out of town trip.

PW