I'll try to keep this brief, but this is an end to end diagnosis and repair description. Maybe this will help someone in the future who has a 465B or similar scope with a similar problem? Experienced techs will laugh at my ignorance, a chuckle which I both accept and invite. :-)
This post concerns the diagnosis and repair process of a Tektronix 465B, SN 37xxx, suffering from a display that was horizontally compressed and shifted far to the right. When operating, the display showed only 3 divisions trace width. In addition, the trace start could not be forced to go anywhere left of the centerline on the graticle. (THIS IS A CLUE!) All the typical sweep functions and vertical amplifier functionality appeared o.k. Just a horizontally shifted and compressed display. (THIS TOO IS A CLUE) The Beam Finder was not centering the beam either, but it was going center and just off the top of the screen. I noted this as a possible "other problem" as well as a symptom.
This scope had been working perfectly. It is my backup scope. Fired it up after a 6 month inactivity in climate controlled storage. No particular reason for failure noted
REPAIR PROCESS
ONE) Since it had been stored, I did the usual "clean the switches/exercise the switches". Nothing changed
TWO) Opened up the scope and started taking measurements. No board removal required. Just remove the case and face the scope screen to the right so it matches the orientation of the diagrams. I was pretty sure the PSU was right, thanks to relatively recent recap by my favorite scope tech, my brother Mike - but I checked PSU test points on A4 board anyway. Examine Figure 8-7 TP 4340, 4437, 4338, 4339 and 4439. All these voltages were exactly right. They are printed on the board there. I did not check CRT HV, since the beam was bright and sharp.
THREE) Upon visual inspection of the A4 board, I found a couple of suspect looking tantalums which I shotgunned in there because I could - but to no avail. C4441 in particular looked bad. Screen still offset. Noted a couple of resistors on A4 that looked like they were running warm as well. R4368 and R4369, and noted them for future attention. Checked Axial Electrolytic C4461 by lifting one end. LCR meter said cap was o.k., so I put it back as is.
FOUR) Checked the following trim-pots on A4 board. NOTE: IF YOU ADJUST THESE YOU WILL HAVE TO RECALIBRATE SOME - Move with discretion and note positions so you can at least get them back close.
XGain control, R4381 - no effect on horizontal width, i.e. inoperative (THIS TOO WAS A CLUE!)
X10 gain R4275 - working but off to the right as expected.
X1 Gain R4274 - working, showed a very small adjustment range (as expected with a horizontally squashed display)
Note that these Bourns trim pots are known to fail. If they fail, your display will go black. Ask me how I know. :-) But mine were fine this time.
FIVE - FIRST PROGRESS) I finally got smart and ran a diagnostic test I got here on TekScopes I/O. The test is simple. Put the scope in X/Y mode and supplied the front-panel calibration oscillator to the x/y inputs. This is a good tip, posted on this forum somewhere, and repeated here. I quote below the exact language of the test routine.
"""""Put the scope in x-y mode (to bypass all the sweep generator circuits) and drive both x and y with the calibration source on the front panel (it's just a handy signal source). If you a.) still see width compression there, the problem is downstream of the sweep generator. If you b.) no longer see width compression, then the problem lies before the deflection amplifier.""""""
I observed condition a.) Therefore, something wrong "downstream" of sweep generator itself, in the horizontal amplifier post sweep circuit. Now I finally knew where to start looking.
I checked the +15 and -8 voltages at the front panel position indicator potentiometer and observed that they seemed reasonable when the pot was turned. That put me firmly on the A4 board itself. I re-read Section 465B manual section 3-19 and 3-20, and the description of the "Input Paraphase Amplifier" / "Gain Setting Cascode Amplifier" and associated driver transistors. I looked at the A4 board physical layout, Figure 8-7 in the manual. I looked at the A4 Schematic which is #10, and is missing from some PDF versions of the manul, like the one at Tek Wiki. So download the file called "missing pages" there at Tek Wiki. :-) Note that the A4 board schematic is also scattered around in various other places on other board drawings. It is a little hard to follow as a PDF, IMO. Paper might be easier for this one. A bit of a head scratcher, but it is all in there.
WHERE I SHOULD HAVE STARTED: At this point, based on the results of above and especially advice from the Scope Whisperer (my brother Mike L., here on the TekScopes group) I scoped the waveform checks #87-91 from the manual. You'll find the manual's page of waveforms right after the A&B Timing Switch Diagram #9. For physical location of the test points, check Figure 8-7 A4 Interface Circuit Board, location i 5. You'll see it there. U4269. That's where you take those readings for waveform 87-91. That 14 pin DIP integrated circuit.
These waveform checks are scope measurements of the expected waveform performance of the Input Paraphase Amplifier and the Gain Setting Cascode Amplifier mentioned in section 3-19 of the manual. These are both part of a 14 pin DIP, common part CA3046, Tek part 156-0048-003, schematic devices U4269A-E (it is a "five transistor array" on a 14 pin dip). If you haven't already, you should hope yours is in a socket. :-) If it is not you are probably pulling your A4 board. :-/
PROGRESS! I quickly observed that something was wonky there at U4269, as about half of those predicted sawtooth waveforms were missing or way wrong. I checked the various discrete driver transistors downstream of the 14 pin DIP and they were faithfully trying to reproduce the wonky signals from U2469. I looked upstream and it looked kind of o.k. to me as well (I'm no expert). So the CA4036 Chip was very suspect. Fortunately, mine was SOCKETED! :-) yea!
PARTS ACQUISITION: I searched my pats bin - no dice. I looked online in the usual places, and quickly saw that true Tek "selected" parts were not real common on eBay, etc. Some were rather pricey.
You are looking for a CA3046 when you look for a replacement. Note that this is the number for the "improved version" in scopes S/N B03920 and above. For Older scopes like mine they specify another similar device, Tek P/N 156-0179-00, but you can ignore that as the Tek 156-0048-03 is "improved" and drop in substitute. Note that all of these CA3046 devices apparently have a fairly poor reputation in the Tek community, even though Tek was "selecting" them. Makes you wonder how bad the unselected ones were! :-) Tek specifies 04713 Motorola as the manufacturer, and a mfg. P/N of SC77161B, which is less useful today, but might help a few of you sort through your parts bins.
Anyway, as I looked around plenty of non-selected modern manufacturing versions of the CA3046 Device were available at low cost. It is used in a lot of stuff. :-) I hopefully opined that new production CA3046 are probably better than 1970's vintage of the same...so that was an option even though they are not "Tek selected" (haha). Another option might be the CA3045 which is a "ceramic version" of the above. But I would have to wait at least two days and we can't do that, can we? :-)
So another option for me was to harvest one from a donor scope. I have a 2445B parts mule scope. On a lark, I examined the 2445B manual and noticed that the 2445B ALSO used this very same device...at least five of them - same exact Tek P/N 156-0048-03. When I went to my mule, I discovered that a yet more "improved version" of CA4036 was used in the later serial numbered 2445B scopes, namely device type CA4336. I examined my parts mule. Sure enough, five CA4336 there on the A1 board of the 2445B. Sadly they were not socketed. In a stroke of good fortune, I noticed one of them, U550 was right at the edge of the board and accessible. I broke out the solder sucker and wick and managed to carefully remove U550 without ruining it (I hoped)
I inserted the rescued CA4336 chip into my 465 and turned it on. No magic smoke, just a normal display again. Bottom line CA4336 = CA4036 for this application.
Thanks to my brother Mike for the great help, and to whomever published that X/Y diagnostic test tip here on the forum more than 10 years ago.
No idea why my original CA4036 chip failed. As I said, they are apparently weak by design. I'm might rescue a couple more CA4336 off my parts mule, just in case.
Now to fix the beam finder and a quick calibration.
Keith
coolblueglow
