Noob vs broken Agilent 33220A

[aheid]

Hi,

Got an Agilent 33220A from a friend, he said it was tossed out and probably needed repair. Sure enough, it boots with a self-test failed 606 error code:

606 - Self-test failed; cross-isolation interface.
This error indicates that the cross-isolation interface between the main processor (U101) and Synthesis IC (U501) has failed, or that the synthesis IC itself has failed

The display comes on and menu is functional, but it complains about some PLL being unlocked (from what I gather that's from the add-on card), and the signal output is dead.

So, I started up by checking voltages at semi-random spots, and quickly found that the test points marked +15V and -15V on the analog side had just 3-4 volts. A bit of schematic reading later and further probing I found that the isolation transformer is most likely all good (I measured ~15V across each of the rectifier diodes CR1302, CR1303, CR1309, CR1310), but something around the voltage regulators for the +/- 15 rails was not good. The 5V rail looked good.

I poked some further and finally found that the filtering cap C1308 on the output of regulator U1305 was basically a dead short across, unlike all the other caps I measured (they were all in kOhm range).

I'm a complete repair noob so I'd just like to verify that the next logical step would be to desolder capacitor C1308 and see if it improves things?

I've attached the relevant part of the schematics as well as a few pictures from the board.

[coromonadalix]

remove the shorted capacitor, take new ohms measurements, if the short is officially gone, you found your little buggar,   once removed you can power up your unit and see if the voltage return to normal values

[aheid]

Thanks for the quick response.

I tried that and unfortunately the cap was perfectly fine 

So, I put the cap back, and pulled out the voltage regulator instead. Hooking it up to my lab supply with some test leads and a dummy load it is not giving any output. So that one seems dead at least.

The only issue now is that I don't have any 7815's around, and the local store only stocks up to 7812. Doh!

I got one of those LM317 modules with a trimpot on it, so I'll try to hook that up with some leads to see if that improves things.

[aheid]

Aaand I just thought of checking the resistance between ground and the vout point of the now-missing voltage regulator... 0.1 Ohms still. So the hunt continues!

[Chris56000]

Hi!

If you want to get it going with local bits rather than waiting for fleabay to arrive, get a 7812, a 2.7V zener diode and a 1u 50V electrolytic, wire the zener in series with the middle (earth/gnd) pin, with the zener cathode (marked end) to the reg gnd pin and anode to chassis/0V on your psu, adding the 1u across the zener.

Alternatively your LM317T will be fine but note the pinout differs from a 78xx, and you'll also need a TO–220 isolation kit to ensure it's tab is isolated or you'll only get a volt out of it!

Chris Williams

[aheid]

Thanks for the tip Chris, I might just do that

So after a lot of hunting, I got some better lighting for my bench, and I discovered this little bugger. It's 0 Ohm to ground on both sides of that thing, and it looks like it's been toasty.

So I guess this is the culprit. The issue I have now is that the I have no idea which component in the schema this is. I can't find the diode it is connected to anywhere in the schematics (or any of the other diodes that look just like it), and I can't seem to trace the paths on the PCB

[aheid]

Doh, I thought that big thing was a diode, silly me   

It's a capacitor of course. Found it in the schematics now.

[aheid]

Removing the capacitor was not enough

I removed the 7815 and 7915, both were fried (did not regulate) when I tested them out of circuit. I've not replaced them yet as I still measure a short between the +/-15V output points and AGND.

So how does one go about identifying where the short might be?

[Berni]

Yes clearly the ferite got hot from the current.

If you have taken out all the big black tantalum capacitors and there is still a short there then you probably have that U902 dead.

To double check take out the capacitors and the small ferites. This disconnects that part from the rails so they should get to 15V and you can then measure the chip to see if that's indeed shorted internally.

[Kleinstein]

To find a short in the supply, there are two techniques:
1) provide some current (e.g. from a lab supply) and carefully (e.g. 10 µV resolution) measure drop over sections of the traces to follow the current.
2) separate the circuit at something like filtering ferrites into parts.

Sometimes a thermal camera can help if the voltage is going all the way down to 0.

Tantalum caps are usually among the first suspects.