Electronics > Repair
Rohde & Schwarz SMIQ03B with Error 110 Output unleveled; IQMOD: ALC failure
<< < (7/12) > >>
zmetzing:
I've made a first-order identification of a part which might work: HMC284AMS8G HMC270AMS8GE

Differences:

* The package is smaller -- not a deal-killer
* RF input pin is on wrong pin -- solvable with small wire
* Control inputs on wrong side of chip -- annoying, still solvable
zmetzing:
Appears that others have gone the same route with the HMC270 device:

https://www.eevblog.com/forum/repair/rs-smiq-ssw124-gt-hmc270-replacement-(_alc-failed_-error)/

Thanks, Roberto, for the helpful information!
zmetzing:
I replaced S2 (SSW124) and boy was that a bear. Another situation where hot air plus the Metcal got it out without damaging the board. Definitely not for the faint of heart. Note: In the pictures, I swapped the control lines, which was wrong. The lines should go straight across (2nd pad to top left pin, 3rd pad to 2nd left pin) for the RF to be routed properly. I didn't take a new picture, but you get the idea.

Partial success! The message "ERROR: CW: ALC_ON voltage oot" is no longer generated, and 7.4.4.5 test point 2108 is now ~3.3V, P54 is ~2V. I can trace the signal all the way to V144.

However, I'm now getting "maximal level oot" from the calibration routine, and the bias on V144 looks all wrong. The manual says the drain should have ~4.5V, whereas I'm measuring ~1.2V. Definitely not going to work very well without some more drain voltage. Drain current appears to be set by V6 (BSP31 device) along with some circuitry around it. I'll have to confess that I'm not quite sure what it is doing .. I'm not an analog engineer. V143 is a 5.1V Zener reference, R1301/R14 combine to a 14 ohm resistor, but I don't see what it's actually doing.

Note that I replaced this device with a "new" one from an eBay seller. I wonder if it is toast and drawing too much current from the current source (if that's what it is). I may try replacing it with the old device.

P.S. I replaced R1151 and R1150 with 1.5k resistors to decrease the negative voltage on SWITCH_A and SWITCH_B. The HMC270 part cannot tolerate much more than -5V.

P.P.S. You'll note that I didn't reconnect the signal path after putting the new switch in. Limiting the causes of down-stream failure is always a good thing when replacing components.
zmetzing:
"That's no small moon."

On the path toward a working SMIQ, we last learned that the bias on V144 was wrong.

While probing around on the bias circuitry, I found 0.8V at the junction of V143 (5.1V Zener) and R644. As the +7V5A rail was at the proper voltage, I was very suspect of the diode, and with good cause. Examining the diode under the microscope, I noticed that the cathode band was pointed at ground! This is definitely wrong. Also, it looks like this device was re-worked at some point, possibly in an attempt to fix V144.

I found a replacement 5V1 Zener in my SMT parts bin -- the last one in a lone ESD package, picked up by chance from someone dumping a load of remnant board build kits. It is a little lower wattage rating (375mW), but the bias circuit isn't drawing that much current (~40mA). Fitting it in, the bias on V144 is now correct, and it appears to be amplifying!

The LEV PRESET calibration gets a little further, but still no joy. The new message is "nominal level oot", which is different from the "maximal level oot" that I was getting before this fix.

We continue to follow the signal ...
zmetzing:
Getting very close here, folks. Re-reading the service manual pointed out that I should close up the resonator chamber M while running the bare board. Given that I don't have a cover handy, I re-installed the B-side shield with two screws over the M chamber and four in the corners.

With Level 2 Debug ON, CALIB -> LEV PRESET gets to the full bar graph like it will complete. but then says "maximal level oot" and the debug info says it failed at 2.8GHz. So, perhaps still not quite enough gain at the high end?

P.S. Telling the unit to "Continue calibration on errors" seems to get it through the CALIB -> ALL path, and the 110 error disappears! However, the IQMOD-CW LEV PRESET table definitely shows a trend on the high-end toward +128 numbers, which appears to be the maximum. Does that mean it was trying to add power (reduce attenuation) at the high end?

I'm considering changing R1226/R411, which are currently selection II, to selection III (edit:corrected). That should boost the overall gain of the final PA.

Would anyone be able to post a picture of their IQMOD-CW LEV PRESET table, specifically the > 2.2 GHz frequencies?

P.P.S. Even though my unit shows a front panel of "SMIQ03B .. 3.3GHz .. 1125.5555.03", attempting to enter a frequency over 2.2GHz (with no errors present in the system) says "Out of range". Did someone sell me SMIQ02B innards with a SMIQ03B nameplate? Or am I missing a setting?
Navigation
Message Index
Next page
Previous page
There was an error while thanking
Thanking...

Go to full version
Powered by SMFPacks Advanced Attachments Uploader Mod