#1901

Episode 1901
which module is the culprit?
Be a Patron: / imsaiguy

[@greengrayradio1394]

Hi from Norway! Yes, I have been struggling with my 3006 for quite some time, and it partly works now. I changed some 20 bad Sprague 10uF 25V (?) tantalums inside most of the plug-in modules, some hidden under components in the point-to point wiring inside, real fun! And, the two separate grounds in the PS, also great fun.. I borrowed a 3001 to test modules from, it also had shorted tantalums. Mine works on only some freq ranges, I got frustrated and put them aside for a while.. I will definitely look at your findings, good luck!

[@MrXenon1977]

I´m silently screaming for thermal imaging... Especially dead tantalums are that easy to find that way.

[@jspencerg]

Getting IR visual very difficult with shiny metal.

[@ats89117]

@@jspencerg You're looking at the wires...

[@MrXenon1977]

@@jspencerg Of course you can't see through the metal cans. But e.g. the tantalum on the perfboard could have been checked with the very first switch on. And wires are often very worth to look at: The one carrying the short current often is easy to be recognized.

[@jim9930]

The easiest method to troubleshoot a supply rail short: measure the voltage drop along a wire (or circuit board trace) away from the power supply: I x tiny R in millivolts. You don't even need the internal supply - any current limited low voltage {so you don't turn on any silicon junctions} will work. Also works with AC current source, ie 24vac transformer with 240 or 24 ohm with anti-parallel schottky diodes to limit output voltage. Great way to find which branch on series connected copper traces has the most voltage drop, therefore short. (sometimes only way) I would expect the Exar 2206 is a mod for variable frequency modulation instead of stock 1kHz.

[@andymouse]

:)

[@lc79tourer26]

You beat me to that suggestion, I have traced out what has been pulling supply rails low using that technique with much success myself. Also I have used a current limited external supply when internal one was not up to it. Another possible method that I have not tried as the above technique worked, would be trace the current path with a sensitive hall effect device. Anyway a great suggestion and maybe imsai guy will give it a go??

[@jim9930]

@@lc79tourer26 Yup... sometimes the old dog tricks are hard to teach new dogs

[@greengrayradio1394]

Exactly what I did also

[@maenishi]

You have one or more shorted tantalum capacitors inside modules, this is a nightmare

[@qutube100]

Do any of your oscilloscopes (hameg usually do) have a ct component tester on them if so how do they stack up against the genuine curve tracer!

[@jeromekaidor7254]

Did you fix it? If not... I have one of those. Same problem with the 18V supplies. It turned out to be tantalum bypass caps inside the RF cans. Real PITA, but simple enough. It's pretty tight inside those cans - in one case, I gave up, clipped the cap inside the can, and soldered one to the mother board.

[@nickcaruso]

I always thought the paint was to detect warrantee violations! 🙂

[@iowahank1]

At 6:25 when you wiggle that board it appears that the wire that is attached to the resistor and goes straight up to the solder point doesn't wiggle. Kind of like that joint might be floating. It's really hard to tell but something you might just double check.

[@centheiatrust9153]

That perf-board is NOT Wavetek original. I have several Wavetek generators and have never seen anything like that in any of them. The real clue, I think, is the hand written component designations. I suspect the factory waveform generator quit working and this is somebody's cluge repair job.

[@jspencerg]

How did wavetek defend their choice of using these cans for everything? Better performamce, reliability? ITC used this construction method on their oscilloscope, which I have. I've never seen a more abandoned vintage brand. No schematics available or repair anecdotes.

[@jim9930]

Double shielding. Outer case + inner cans/coax jumpers, loss has to be greater than output at lowest level from step attenuator. ie local oscillators and desired frequency < -140dBc Also the reason for solid coax.

[@jspencerg]

@@jim9930 Those results would make some sense. Do the units' spec's support these gains?

[@jim9930]

@@jspencerg It would be useless for measuring SINAD of sub microvolt receivers without double shielding. Magnetic & Faraday shields are necessary to get those kind of loss numbers - you can be sure they used a nearfield broadband antenna to sniff for leakage. H loops and E field

[@gretalaube91]

Take a picture of the cable rat's nest! ...quick! - ask me why.... There's also a Molex that feeds the rails to the front. ...yank. see... 73's W3IHM