4191A Impedance Analyzer E-12 narrow range problem (899-900 MHz) (AD Converter?)

[coppercone2]

My 4191A IA has a problem where after calibration (which passes), it will flag E12 when it is set to a narrow range of operation between 899 and 900 MHz.

I see the manual has a procedure for trouble shooting E12 (normal IF signal was not produced when IF signal frequency was locked to 100KHz) , but the narrow band of the fault made me think its maybe a good idea to post this, I though that might be a little unusual.... (8-35)

https://www.keysight.com/us/en/assets/9018-05401/user-manuals/9018-05401.pdf

I am waiting to fix my Rigol 1052E (the keypad has the same problem, needs new conductive surface on the membrane as my Wavetek 395 that I fixed), so I can't do the tests right now on the equipment that is convenient for trouble shooting.

Does anything unique come to mind for a failure that occurs in only a 2MHz region for a instrument with a 1GHz bandwidth?


I almost want to ignore it because its such a narrow problem, but I figure I should fix it.

Also, screw RigLOL, I can't believe their keypad broke down after only a decade. And horizontal encoder is dodgy too

I wonder if it just needs adjustments, I did replace a PCB I remember now... maybe its just slightly out of spec some where.

[m k]

Does anything unique come to mind for a failure that occurs in only a 2MHz region for a instrument with a 1GHz bandwidth?

Only thing that came up here is some sort of a digital presentation.
But at least direct hex multipliers are nowhere near of anything.

[coppercone2]

nah I did a board swap and when I tested it I only went by increments of 20 or so, jumped over the problem. I think there is a procedure for adjusting some stuff after PCB swap

[coppercone2]

this unit blew a shaffner

[coppercone2]

well I got a replacement card. (A4 is said to be possibly related)

The operating range decreased, it now had e12 from 890-905 MHz.

I noticed the old card was dusty. Alot of dust under the IC, came out when i shot it with spray alcohol.

I decided to rinse is under the sink with the sprayer head, then 30 seconds in ultrasonic in Alnox, followed by rinse with hot water spray head and then cool filtered water. Then a 30 min bake out at 78C in PID oven.

I put it back in. Works across the range. I think it was a dust problem. The replacement card I got was really dusty too. I also applied deoxit gold to the PCB card edge connector pins.

Guess i wasted some money on a replacement card. Maybe.

In these units check any cards that are on the edge (exposed to alot of air flow)

this is some digital board BTW, like the poster in this thread guessed.


and of course after an hour, I tried again to experiment with the new fixture, it does not like the band again. I need to replace the multiplier card maybe....

or could it be something with capacitor ESR. It was warm when I put it in there.

I suppose I should heat up the card again, and replace it. see what happens

[coppercone2]

mm I half assed it a little bit but it looks like at 900MHz one of the test voltages is out of spec. should be 11.4 +-0.3 but its like 0.2v less then it should be

[coppercone2]

ah the trim pots were able to fix this problem. Sequence is important with them.

But when I went over A21, I did find a ceramic memory chip with a lead that was not installed. It was bent into a U shape hidden behind the thingy!

the unit had some intermittant operation before, I wonder if that is it.

I don't think I did it because I had a bad experiance with ceramic IC once, I don't touch them to resocket. I did resocket all of them with deoxit gold though. One of the processors or something (looks like a expensive chip) had pretty tarnished leads. I rubbed them good with a chamois wipe with alcohol then deoxit. That one kinda stuck itself in the holder. I love how smooth they go in after deoxit. I imagine its cutting something like a little endmill thats greased up

The Pots were quite a bit out, I guess I should go through the entire adjustment procedure.

I did not touch them because I figured a proper instrument should not be able to be put in a error state by adjusting a pot! I guess not! It looks like this procdure requires complicated equipment like spectrum analyzers and oscilloscopes. Where can I find those?! i would need some kind of well stocked laboratory. does anyone ever adjust this shit?! thats only there for the lawyers right?