I bought a Tektronix 2445A oscilloscope from eBay. It was a very good deal for a 4-channel 150 MHz oscilloscope. Maybe too good a deal, or so it seemed for a while...
The seller used the usual eBay boilerplate disclaimer for electronic gear — they had turned it on, no smoke came out, it lit up, but they "were unable to test further". It was guaranteed not to be "DOA" ("Dead On Arrival"), which probably means only that smoke shouldn't appear and it should "light up" when you turn it on, and nothing more.
The problem was that the scope would not focus! The very best, with the focus control turned to the end of its range, was as pictured. The trace was at least 1.5 cm wide, and the on-screen indication of vertical scale and horizontal timing was nothing but pale green smudges instead of readable characters.
I had not placed a bid on the item until after I had
found a PDF file of a service manual.
Places to find service manuals include:
— http://bitsavers.vt100.net/
— http://www.bitsavers.org/
Those were revealed by asking Google for something like:
tektronix 2445 "service manual" filetype:pdf
So, time to get to work...
The alignment procedure includes setting the trace focus as sharply as possible, and then using the astigmatism control (a small screwdriver-set pot below the display) to further refine the setting. Yes, the astigmatism control had some effect, but it seemed as if I needed another 90-180 degrees of rotation on the focus control to get the job done.
|
YOU SHOULD NOT READ BEYOND THIS POINT. ABOVE ALL, YOU SHOULD NOT DO WHAT I DID. IT IS VERY VERY DANGEROUS AND YOUR LAWYER WOULD NOT APPROVE. BUT IT DID FIX MY OSCILLOSCOPE.... |
I needed some Torx drivers — T-20, T-15, and T-10.
|
|
I removed the six Torx head screws and
then removed the rear panel.
Left — Before starting
Right — Rear panel removed
|
|
I tipped the unit up onto its back end, and very carefully lifted the oscilloscope out of its case.
|
|
At left above — the high-voltage power supply
at top, under the aluminum cover.
At right — after removing that cover —
four T-10 screws and one T-15.
Tracing the control voltage from the wiper on the focus control into the high-voltage power supply board, I found that the control had no easily measureable effect on the output of focus amplifier transistor Q1852. The output focus voltage was fixed at approximately -300 V DC. Further testing showed that the precision 332kΩ resistor R1854 had become an open circuit and now was a fairly precise infinite impedance....
| With R1854 opened — effectively removed — Q1852 was no longer correctly biased. And so that transistor was not controlling the output voltage as designed. At right is the circuit biasing Q1852, click on the picture or here for the full page from the service manual. |
|
|
|
Above is the view from the opposite side. R1854 is in the row of resistors between the white-topped transformer and the red potted HV circuitry.
I did not have a 332kΩ resistor handy, but I did have two 680kΩ resistors that I could stack in parallel. The original quarter-watt resistor was asked to dissipate about 0.19W continuously, and this was probably the cause of its failure. Since I would be using two quarter-watt resistors in parallel, I should have much more thermal leeway.
|
|
Above you can see my repair — the pair of 680kΩ resistors is stacked vertically in the original R1854 position. The repair sequence was:
And that fixed it! A friend later told me that he had had a nearly identical experience on a different piece of test gear. A precision resistor had failed to an open circuit and disabled the instrument's display. It isn't very clear why the design uses a precision resistance (in a circuit with multiple potentiometers!) instead of a more rugged device.
Vinny, N2LWN, bassist10@comcast.net, found my page through Google and sent me the following:
Hi Bob, I got one of these on ebay too. It seemed to work fine until I stopped using it for a couple of years and the lithium battery holding the setups wound down. Now for some reason the darn thing needs 20 minutes before I can see the error code on the confidence test. Fails test 5 code 44 trigger too positive main board failure.
After I escape this function via menu button everything seems to work fine. DMM display is clear, trace and readout clear in focus etc., but as time goes by it shuts itself off (2 hrs.). After it cools down a bit it seems to take even longer to produce the readouts, they appear shrunken and take longer to expand to normal size. I have no clue what might be causing this and the service manual is not clear about any specific adjustment/repair.
Does anyone have any idea about how to fix that? If so, please contact Vinny, and also copy me so I can add it to this page!
This is from Jeremy Linton, and it sounds like the fix to at least part of Vinny's problem described above....
I picked up a Tek 2445 from ebay back in 1999 or so. Its been great for intermittent use.
Anyway, last night I was monitoring a circuit and went to change channel 1 from DC to AC, when it clicked and the scale illumination light went out, all the indicator lights went out and the buttons stopped responding. The trace was still running though.
Power cycling yielded a test 05 fail 22 message. Out with the service manual and the bad news, it looked like the trigger control IC (hybrid) had suddenly become very sensitive to temperature. The test would pass if the device was off for a few minutes. Once warned up, it would act up by "crashing" at which point the buttons would stop working.
So, I pulled it expecting to find some dry ancient heat sink compound, thinking I could keep it cool enough to work while I found a replacement. Instead, I discovered the terminals were corroded to the point where I'm amazed it was working at all. A little 600 grit sandpaper and Q-tips with isopropyl alcohol got them shining again. Seems the problem is solved, the scope ran the test 5 loop for an hour or so without a failure.
I figured if that IC was corroded, the others might not be so good either. I pulled the preamps and a couple of the other IC's, cleaned them and put the whole thing back together. It appears that fixed some of the signal noise I was blaming on the cheap probes I've been using.
I'm not saying this is what is causing the 05/44 but I'm betting this is pretty common. Mine has been kept in pretty controlled AC for the last 10 years and before that was a school lab tool, meaning it probably sat in AC for the first 15 years of its life too. 25+ years is a long time to corrode. It's probably a matter of time before enough capacitance builds up on the terminals to cause a malfunction somewhere when the device heats/cools and the terminals move a little on the pads. So, this fix was just another variant on the old reseat the IC solution, but I thought it worth mentioning anyway.
|
||||||||||||
|
||||||||||||
