Author Topic: Rohde & Schwarz SMIQ03B with Error 110 Output unleveled; IQMOD: ALC failure (Read 25852 times)

sorenkir · « **on:** January 28, 2016, 01:31:48 pm »

Hello,
I have this generator with the error in the subject (no photos yet).
It is cosmetically like new, very heavy, built like a tank, but very high hour count: 88640h (10 years) and attenuator toggle count more that 10M!
The RF output is very distorted (seems like clipped and wrong amplitude).
I will soon try to find the fault with the help of the 4 Service Manuals found on KO4BB's Manuals Repository http://www.ko4bb.com/manuals/index.php?dir=Rohde_Schwarz
Unfortunately, these manuals are missing the schematics.
Does anyone have them or know where I could find them ?
Thanks,
Michel.

rastro · « **Reply #1 on:** January 28, 2016, 11:17:29 pm »

I have not seen any CLIPS for this unit in the public domain. Have you try running the calibration routines in volume 1 of the service manual. They will should provide error messages if they fail. However you will need to first the first level password. There are 3 levels of password protection for the SMIQ models. They are as follows:
PASSWORDS
LEVEL 1      123456
LEVEL 2      250751
LEVEL 3      120155

There are also digital test points available at the user interface.

-rastro

sorenkir · « **Reply #2 on:** January 29, 2016, 10:19:32 am »

Hi,
Thanks for your help.
I have tried the test points of IQMOD, the only one to be out of range is 2122 (Control voltage burst element / Power Ramp).
Maybe related to back panel input "POW RAMP" ?
I will try to follow the path from this input into the IQ Modulator...
Michel.

rastro · « **Reply #3 on:** January 29, 2016, 05:12:04 pm »

Michel,

There is also a troubleshooting write up on an SMIQ06. It's in German but you can probably pick and paste into google translate.
http://www.bymm.de/documents/27/SMIQ06B_6GHz-Modul_V1_11.pdf

-rastro

sorenkir · « **Reply #4 on:** January 29, 2016, 09:30:51 pm »

Thanks rastro, but this note applies to the 6GHz option board, which I don't have.
I will continue my debug but I'm afraid it will not be very easy: double sided board, no silkscreen, no schematic except the block layout on the cover and a short description in the Service Manual...

rastro · « **Reply #5 on:** January 30, 2016, 03:47:44 pm »

Michel,
I agree it's going to be challenging to troubleshoot.

There is also a users group you can try - although there is not a lot of activity (compared to the hp users group).
https://groups.yahoo.com/neo/groups/Rohde_Schwarz/info

Also the manual gives a parts list and X/Y coordinates in mm for components on the PCB's. If I remember the (0,0) coordinate is lower left corner on the front of the PCB and mirrored for the back of the board. At least it list might be helpful if looking up datasheets for the components.

-rastro

sorenkir · « **Reply #6 on:** January 30, 2016, 04:22:05 pm »

Thank you for your support!
First thing, I think, is to build an extender: DIN41612 Type B 64pins connectors male/female & flat cable + 2 x SMA-SMA cables + 1 x MMCX-SMB cable (why so many connector types on these boards: MMCX, SMA, SMB & SMC ?)
Then I will be able to probe the effect of this "POW RAMP" rear input which goes to an analog switch DG419. Maybe this switch has been damaged by an over-voltage on the input (no visible input protection on the board) --> measurement 2122 out of range, whatever value on the input.
To be continued...
Michel.

sorenkir · « **Reply #7 on:** January 31, 2016, 07:47:28 pm »

Troubleshoot Sunday !

No extender for the moment, so I took some time to watch Shahriar's excellent repair video of an Agilent E4433B which has an unleveled error:
https://www.eevblog.com/forum/testgear/video-teardown-and-repair-of-an-agilent-e4433b-esg-d-synthesized-signal/

It gave me inspiration and I examined the final stages of my SMIQ with more attention (and the help of the Service Manual).
It seems that the principle of operation is similar: one direct path above 450MHz and an heterodyne path below (LO @ 2.4GHz).
Then, the final stage is composed of one MMIC amplifier NLB-400 (marking N4) followed by the same HFET than in the Agilent: SHF-0189 (marking H1). The Service Manual says this final stage is made of 3 HFETs, so it must be a different hardware version...
The ALC detector is after this final stage.

By removing the front panel, I managed to gain access to the upper side of the board and have been able to probe the signal path: it appeared the signal which is correct up to the MMIC, is not amplified by the SHF-0189 but its amplitude decreases through it!
The bias on the input is -2.8V as expected but on the output, it's only 0.7V, when I was expecting about 7V. The bias chain, driven by a BSP31 transistor is only 1.5V at the starting point.
I removed the transistor and tried to power the HFET with an external power supply but when I injected 180mA, which seems an acceptable maximum, the voltage was clipped to 1.6V.
The BSP31 is OK but I think the SHF-0189 is defective.

Now I will try to find one, and then, the hard part will be to desolder the defective HFET, because it stands in a small cutout into the shield (see photo). If someone has an idea ?

To be continued...
Michel.

rastro · « **Reply #8 on:** February 01, 2016, 06:55:24 pm »

Michel,

Sounds like you made some very good progress. The solder connection on H1 center pin looks a little questionable - but this could just be the picture.

As far as removal:

I'd probably leave the shield soldered in place to avoid PCB damage.
Cutting H1 out and de-soldering remaining leads may be the best approach ... otherwise....
1. You are going to need a lot heat on that tab of H1 and its large pad. Probably needs a direct solder iron connection using the largest a tip that will fit.
You may need to apply hot air to the 3 legs at the same time - if you have equipment and hands.
2. I may also be good to mask adjacent components with Kapton tape to protect them from head or migrating if their solder gets too hot.

-rastro

sorenkir · « **Reply #9 on:** February 27, 2016, 10:35:47 am »

Hi,

Thank you rastro for your advice.

The SHF-0189 I have ordered (to Utsource, good service, two week delivery from China to France) have arrived. I had ordered 3, just in case (which has proved to be a good decision)...

I managed to remove the defective chip with my dremel, did not succeed to completely desolder the ground pin, but I removed enough metal to be able to solder the new chip. I had to remove the shield above the IC, to gain access to the pins.
The first soldering attempt failed, I broke the case while trying to bend a pin.
I did a much better job on the second try, see photo.
It is not perfect, but it is not easy because of the very small size of the IC (~4mm width) and the adjacent components are very close.
I did not bother to rebuild the shielding.

And voilà: No more errors, self calibration passes, output signal is in spec.

Just the backlight intensity is a bit low, due to the high hour count. I can live with that, but if anybody has an idea ?

While the beast is open, I have moved the back connectors to the front: quite easy, all the cables, even the semi-rigid one can go back or front.
In the meantime, I have upgraded the firmware to the latest available version 5.93HX on R&S website (lost the self calibration tables in the operation, but not a problem now that everything works, it is rebuild in a few seconds)

You know what? ...

Michel.

rastro · « **Reply #10 on:** February 27, 2016, 04:35:36 pm »

Hello Michel,

That's great news! Thanks for posting pictures of the repair. If you suspect any RF noise issues later you may be able to fill in the cut out shield by soldering in some thin copper sheet. It's probably not an issue.

Regarding the display:
I saw some talk on the yahoo HP blog that someone successfully fixed his HP and RS system by replacing the Cold Cathode Fluorescent Lamp (CCFL). Various sizes are available on eBay. The CCFL inverter would be another candidate - universal replacements are also available.

Also thread in eevblog: https://www.eevblog.com/forum/testgear/rohde-schwarz-fsiq-7-signal-analyser-screen-not-working/

Regards
-rastro

sorenkir · « **Reply #11 on:** February 28, 2016, 10:29:43 am »

Hi rastro,
You are right, replacement CCFLs can be found on the Internet.
Accessing the 2 CCFLs seems easy but requires to fully disassemble the display and I am not sure I want to disturb this glued ribbon cable...
Michel.

rastro · « **Reply #12 on:** February 28, 2016, 03:22:05 pm »

Hello Michel,

Yeah, they don't make it easy. I would also be hesitant to put stress on the ribbon cable connection. Those connections seem pretty delicate. I appears you have to bend the tabs to remove the panel also.

I'm not sure but I thought I saw some information that some of the models/units of RS displays are interchangeable. Maybe a CMD test set that is obsolete. I would probably be hard to be certain.

Regards
-rastro

sorenkir · « **Reply #13 on:** March 04, 2016, 10:22:12 am »

Exact! I have a another german device (Wandel & Goltermann SF-60) that uses the same display, but it has a few black lines, probably due to the glued ribbon cable...
May be I will try to fix it one day, as Dave does in a video for a Fluke Scopemeter.
Michel.

rastro · « **Reply #14 on:** March 04, 2016, 11:10:27 pm »

Michel,

Good detective work finding a comprable display. Too bad it's not 100%. I saw Dave's video. It was a good attempt but with mixed results - there has to be a better way.

Some further thoughts on display repairs:

1. I wonder if one of these "Home Sealing Impulse Heat Tool" could be hacked to reflow the solder or reactivate the glue on these troublesome LCD ribbon cables? It looks like these have a heating element riding on a spongy heat resistant pad - so it may provide better contact with the faulty area. The heat element is small enough that you could work on one area at a time. They are certainly cheap enough to experiment with. I have not seen anyone else try this - who knows...
Example:
Mini Bag Sealer Home Sealing Machine Heat Tool Impulse
http://www.ebay.com/itm/Mini-Bag-Sealer-Home-Sealing-Machine-Heat-Tool-Impulse-Food-Packaging-Fashion-/252052218343?hash=item3aaf7ba1e7:g:rhMAAOSw3ydVyBUh

2. If you go the route of changing the CCFL it may be possible to add a strain relief where the cable meets the PCB. On either side of the cable there is a hole in the PCB. Maybe a threaded hook on both ends to clamp down a metal strap and thin piece of rubber in contact with the cable.

Still, you've made pretty good progress on this instrument!

Regards
-rastro

jgorsk · « **Reply #15 on:** August 25, 2016, 09:30:44 am »

I also have a faulty SMIQ with the 110 error (Output unleveled; IQMOD: ALC failure) but the output stage amplifier seems to be working.
I suspect some filters between the frequency divider and tha output amplifier stage.

Attached are the spectra of the -10dB signal at 450MHz and 460MHz. The amount of harmonic tones at 460Mhz (and above) is way higher
than specified in the SMIQ documentation.

Could someone do me a favour and post the spectra of their SMIQ generator at -10dB and 450 and 460MHz please?
I'd like to see how those signals look like on a working generator.

rastro · « **Reply #16 on:** August 25, 2016, 12:57:36 pm »

Hello jgorsk,

Which model of SMIQ do have (3,4,6??). It will probably make a difference. The 4/6 go through an additional PCB.

-rastro

jgorsk · « **Reply #17 on:** August 25, 2016, 01:53:29 pm »

It's the SMIQ 02 model

rastro · « **Reply #18 on:** August 25, 2016, 04:02:00 pm »

I've included the spectra for an SMIQ06 set to 450MHz and 460Mhz. I don't know if this will help since it has additional RF paths.
- rastro

jgorsk · « **Reply #19 on:** August 25, 2016, 07:39:42 pm »

Thank you very much, that's what I needed to see.

jgorsk · « **Reply #20 on:** October 11, 2016, 07:14:16 pm »

I managed to find a few faulty components. The FET at the output from IQCON board and a few amps in the 300MHz
vector modulation signal path were faulty. CW, analog and digital modulations work now but there are still a few errors left.
They are
-313 calibration memory lost; LEV_PRESET
-313 calibration memory lost; VECTOR_MOD
-313 calibration memory lost; LEVEL_ATT

LEV_PRESET calibration fails with 'IQCON no convergence' error message,
VECTOR_MOD and LEVEL_ATT calibrations fail with 'Reference voltage oot (6.9 ... 12V)' message.

In vector modulation mode with no I/Q signals applied the carrier level is about 20dB below the setpoint.
Is the carrier supposed to be suppressed in this mode or should it be as set by the level setpoint?

analogRF · « **Reply #21 on:** February 19, 2019, 12:28:56 am »

Quote from: jgorsk on October 11, 2016, 07:14:16 pm

I managed to find a few faulty components. The FET at the output from IQCON board and a few amps in the 300MHz
vector modulation signal path were faulty. CW, analog and digital modulations work now but there are still a few errors left.
They are
-313 calibration memory lost; LEV_PRESET
-313 calibration memory lost; VECTOR_MOD
-313 calibration memory lost; LEVEL_ATT

LEV_PRESET calibration fails with 'IQCON no convergence' error message,
VECTOR_MOD and LEVEL_ATT calibrations fail with 'Reference voltage oot (6.9 ... 12V)' message.

In vector modulation mode with no I/Q signals applied the carrier level is about 20dB below the setpoint.
Is the carrier supposed to be suppressed in this mode or should it be as set by the level setpoint?

did you ever managed to fix the 'IQCON no convergence' error?

zmetzing · « **Reply #22 on:** June 28, 2020, 07:01:35 pm »

Schematics for the 1084.9800.01 IQ Modulator board, among other things, are located here:

http://eww.ko4bb.com/getsimple/index.php?id=manuals&dir=Rohde_Schwarz

I'm currently diagnosing my SMIQ 03B's failure "ERROR: CW: ALC_ON voltage oot" during CALIB ALC TABLE. The SHF-0189Z final has been replaced, as a shotgun-debugging start. (That was a booger to get out. Hot air @ 375C plus my Metcal on the back tab.) Didn't fix the issue.

Extender cards seem to be unobtanium, so I'll take the advice of others here and pull the front panel so I can have limited access through the front of the machine.

zmetzing · « **Reply #23 on:** June 28, 2020, 08:02:02 pm »

Looks like the output is fairly low, so it points to the output stage.

For reference, my 8568A's 20MHz -10dBm reference signal into a Iwatsu 470 MHz analog scope (50 ohm input) shows 200mV p-t-p at 20MHz (correct).

However, the SIMQ's output shows only 100mV p-t-p at 20MHz when set for +10dBm. This is -16dBm.

Handy table here for dBm -> mV (50 ohm) here http://wera.cen.uni-hamburg.de/DBM.shtml

(Yep, I'm a noob at this forum software)

charlyd · « **Reply #24 on:** July 08, 2020, 07:12:49 pm »

can this help you:


EEVblog Main Site	EEVblog on Youtube	EEVblog on Twitter	EEVblog on Facebook	EEVblog on Odysee

Author Topic: Rohde & Schwarz SMIQ03B with Error 110 Output unleveled; IQMOD: ALC failure (Read 25852 times)

Share me