AnsweredAssumed Answered

Repair of a HP8753C searching for schematics

Question asked by maikhermanns on Feb 22, 2007
Latest reply on Aug 13, 2007 by JohnMiles

Like • Show 0 Likes0
Comment • 14

Hello altogether,

At first something about me, at the moment IÂ´m still studying Microwave and RF engineering and going to write my diploma thesis in a couple of months.

IÂ´m a (glad) owner of a HP8753C NWA and have problems with the signal source.
The output power is not levelled at all, in the lower frequency range (about 10MHz) it decreases below ca. -30dBm and the PLL circuit fails. With increasing frequency the output power increases, too. (from -30dBm @ 25MHz to -4dBm @ 3GHz)
If the source is used in CW mode the power level menu(-10..+20dBm)has no influence on the output power.
I have the field service manual from the HP Website (500pages), but it contains no detailed schematics. I'm afraid that I have a problem with the source board and that I will need them to fix it. Is there any possibility to get these schematics? (I just want to get this unit running for my personal use (HAm..hobby) and Im afraid for the next years a collection of the circuit diagrams will be necessary..)

best regards

Maik

No one else has this question

Outcomes

14 Replies

Dr_joel Feb 22, 2007 4:26 PM
Mark Correct
Correct Answer
Probably a dead Q4 inside the source micro-circuit, which is an NEC9002 GaAs FET. If this is it, you'd have to have a replacement source module, as the transistor is eutectically (sp? ) attahed to a molybdenum carrier, epoxyied to a larger carrier. You'd have to have a wire bonder to replace it.

If you've got the service manual, it will tell you how you can test some of the internal source nodes to detect if this is the problem. If the source detector node on the cavity oscillator has a large signal, then it's really likely to be the source power amplifier.
Like • Show 0 Likes0
Actions
maikhermanns Feb 28, 2007 2:43 AM
Mark Correct
Correct Answer
Hello Joel,
many thanks for the answer, but due to the directly bonded components Im afraid of a "DIY" repair of the source module RF assembly.
First I really want to check if there is everything ok with the rest of the Source Module (Bias supply, Level control loop ..)

So my first question : is it possible to get a detailled schematic of the whole source assembly ? (If not directly from Agilent - do You know a dealer who has them?) ( I also would sign a NDA document..in the worst case just the bare scematics would be more than ok... my email: maikhermanns _remove me_ @maikhermanns.de)

Is there still an exchange program for the source or the RF Part of the Source Module existing? (if not - are there other companies repairing them ?)

I already did some measurements,

In the "SRC Tune Mode" the output frequencies are about 100MHz too high (I've measured the YIG with an external counter).
The YIG Driver Test (16) shows a good control signal, shape as in the manual, but the voltages differ partially about 300mV.

And my last question: is the RF Part of the source module build in 50Ohm Stripline technology and what is the amplifiers gain (just round about) - in the worst case I think of connecting an external broadband amplifier module istead of the NEC9002 Amp using a little bit of semirigid cable.
But the Ceramics PCB looks really "interesting" with all its laser trimmed thickfilm resistors...
(dont worry - this solution would be just for my personal use..non commercial)

regards Maik
Like • Show 0 Likes0
Actions
dgun Feb 28, 2007 10:17 AM
Mark Correct
Correct Answer
I'm sure anything of that vintage is based on a thinfilm or duroid microstrip transmission medium. Otherwise, it would be microstrip or stripline on PCB, but I don't think the 8753 had any RF sections on PCB.

The 8753C is no longer supported, so no exchange parts are available. However, the 8753E is still supported. If the needed parts happen to be compatible with the equivalent 8753E parts, you might be able to get something.

Any schematic that is not part of a service note or user manual is restricted or confidential and likely requires management approval before release. Try calling the service center and escalate.
Like • Show 0 Likes0
Actions
maikhermanns Mar 12, 2007 3:45 PM
Mark Correct
Correct Answer
Hi Daniel and Joel.
I just dismatled the RF Section, everything on ceramic substrates. I've checked the amp with te mixer and the coupler removed, using some selfmade adaptors. The amplifier is ok. (app. 40dB gain, very flat,measured from 10MHz to 1.3GHz)
But I'm afraid that the main fault is in the mixer, I measured the output power of the mixer behind the lowpass filter, and it is less than -35dBm.
(in the case of a open power control loop I would expect app. -15dBm)
The drivers are ok, too.
(3.8GHz @ + 14dBm ->pin attenuator --> +6dBm)
(YIG about + 16dBm).
Do you have any ideas how I could fix the mixer? would a good "normal" RF Mixer work? Any ideas which one to take?

regards from Germany

Maik
Like • Show 0 Likes0
Actions
dgun Mar 12, 2007 5:38 PM
Mark Correct
Correct Answer
Probably a beam-lead dual diode mixer like something you'd find here. I couldn't find the actual part number.
Like • Show 0 Likes0
Actions
maikhermanns Apr 9, 2007 4:09 AM
Mark Correct
Correct Answer
Hi all, FYI I've repaired the RF Section of my 8753C.
The fault was definitely the mixer diode. I ordered some Avago HSCH5512 Dual Diodes and "attached" them to the Board. (the old diode was epoxied, I had to cut/grind it out carefully...) The next interesting thing was the bonding of the new diode. Because nobody around had a wire bonder (not in my university and some RF Companies I asked did not want to do it either) I build my own from an drill stand with an thin soldering iron (0.4mm tip modified to app. 0.1mm diameter) , an old 50x microscope mounted to a special support to see the "Thermocompression Bonding" -> and some diodes and several hours later the analyzer worked.
So don't get worried if your source section fails - it can be repaired- with some Luck even with DIY equipment.

I just recalibrated the source related things - no errors anymore and a flat source Power output level.

My question for today - Is it possible to do the power level correction with an normal power meter instead of the GPIB one connected to the analyzer?
regards Maik
Like • Show 0 Likes0
Actions
RayH Apr 9, 2007 9:49 AM
Mark Correct
Correct Answer
hmmm. I always thought a GPIB power meter IS a "normal" power meter...?

If you want to level the 8753C using a analog feedback loop, look at the rear panel external AM ("EXT AM") BNC connector. A + or - 1 volt signal here will give you approximately +/- 8dB of level control.

This was used with the 8347A amplifier, which had a detector output. You might find a way to make it compatible with the analog recorder output of a power meter - but I have never tried to use it for this purpose.

To someone of your ingenuity and persistence, I'm sure the solution will be trivial...
Like • Show 0 Likes0
Actions
Dr_joel Apr 9, 2007 11:12 AM
Mark Correct
Correct Answer
"
Hi all, FYI I've repaired the RF Section of my 8753C.
The fault was definitely the mixer diode. I ordered some Avago HSCH5512 Dual Diodes and "attached" them to the Board. (the old diode was epoxied, I had to cut/grind it out carefully...) The next interesting thing was the bonding of the new diode. Because nobody around had a wire bonder (not in my university and some RF Companies I asked did not want to do it either) I build my own from an drill stand with an thin soldering iron (0.4mm tip modified to app. 0.1mm diameter) , an old 50x microscope mounted to a special support to see the "Thermocompression Bonding" -> and some diodes and several hours later the analyzer worked.
So don't get worried if your source section fails - it can be repaired- with some Luck even with DIY equipment.

I just recalibrated the source related things - no errors anymore and a flat source Power output level.

My question for today - Is it possible to do the power level correction with an normal power meter instead of the GPIB one connected to the analyzer?
regards Maik
"

Congratulations on fixing my mixer (that was my FIRST design back in 1983 when I joined the company, and it was used, practically unchanged, up through the 8753ES and the first version of the PNA...I probably have a few prototypes hanging around if your fix ever fails). It's odd that the diode failed...almost never happens (I wonder if it was some sort of degradation of the beam lead, maybe thermal stress). Originally, that microcircuit was the most complicated every built at the time, and we were really afraid of the warranty costs, so those ceramic modules (Thick Film, by the way, built in Colorado Springs), were designed to be user-replaceable. I'll bet you could still get the mixer module, 5086-7424 is the part number. I just checked and it's still being used in the low-band portion of the 50 GHZ PNA!

I'm not sure about what you mean by "power level correction" but if you mean the factory calibration, then yes, it is possible to do that without a GP-IB power meter, by fooling with the source EEPROM values. They are a bit tricky to setup, though, and I'd have to dig through some old design guides to remember what they do. The internal source calibration service test sets them up by iterating through the power levels and finding the best values for detector offset, level, etc. However, almost all of those quantities are set by the detector diode on the coupler/detector, so the calibration should still be good.

By the way, did you replace the black block on top of the diode (sometimes on top, sometimes on the balun arms). This was used to fine-tune the balun and get better spurious response. Probably won't see the difference except for some mixer measurements.

If you mean the "user" source power cal, that still requires an external power meter connecdted by GP-IB
Like • Show 0 Likes0
Actions
maikhermanns May 8, 2007 6:44 AM
Mark Correct
Correct Answer
Hi Joel, really intresting that it is your mixer design.

At the moment I'm really happy that I've fixed it, because I just started writing my diploma thesis and the analyzer is really helpful to me. (especially because the Agilent 8510C in our university starts at 45MHz and I nedd the 13.56 MHz range)

I was begging all the time while I was searching for the failure, that it might not be the mixer, I was really afraid of the mecanical compexity of the module and already searched for matching mixers modules (MA/com or Miniciruits). Due to the fact that nothing really satisfied me, I decidet to fix it. I was really amazed about the 3 dimensional structure... The analyzer now has now done about 40 hours of work for my diploma thesis, but I'm still afraid that my bonding will not last forever. If you would have an old sample of these mixers I would be really happy.

Our Agilent subsidary here in Germany sais that they replaced the source module only completly and no spares (even the whole module) would be available.

I lost one of the ferrite rods, because I had to destroy the one glued directly onto the diode epoxy.

But maybe you can tell me some spur mesurement values / frequencies - just to check if its more or less ok. On my board there have been glued 3 pieces of this ferrite material.

best regards

Maik Hermanns
Like • Show 0 Likes0
Actions
wpmuckleroy May 14, 2007 7:03 AM
Mark Correct
Correct Answer
Our sister company (TXRX Systems) recently submitted an Agilent 8753C VNA for repair/recalibration. They continually recieved an "Phase lock lost--Check input to channel R." I traced the problem to the source assembly (p/n 08753-60003 and 08753-69003).

Since the analyzer is obsolete and no longer supported, I cannot get this assembly from Agilent. I know that this will remedy the problem because I pulled the source out of a functional 8753A and tried it in the 8753C. It worked fine.

After cross-reference of many parts and sub-assemblies, I discovered that the source out of an 8752C will work in the 8753C, but it does not level properly. I had three 8752C available as parts. To fix the leveling problem, I put an externel leveling amplifier (HP 8347A) between the source output and the s-parameter test set. It works well now.

Even though I was able to piece it together with odd parts, I'd like to return the 8753C to its former granduer.

Any suggestions???

Thanks,
Wayne P. Muckleroy
Chief Metrologist
Bird Electronic Corporation
phone: (440) 519-2249
Like • Show 0 Likes0
Actions
Dr_joel May 14, 2007 7:44 AM
Mark Correct
Correct Answer
You can try opening up the 8752 assembly and replacing the last circuit (which has the leveling detector on it) with the circuit from the 8753. There are about 8 screws in the lid of the package. Take the lid off and you will see all the individual circuits. Replace the last before the source output. Mind that the little "chomeric" conductive rubber pieces are properly aligned in their pockets (they tend to stick to the bottom of the circuits, and will crack one if it is not put properly into the pockets).
Like • Show 0 Likes0
Actions
DavidSudry May 16, 2007 10:17 AM
Mark Correct
Correct Answer
Hi Everybody,

I have repaired few source modules with blown output stage.

1. Scratch the blown output die with a sharp knife.

2. Install a SOT 89 Amplifier, EIC-008 ( DC-3GHz, 15db gain +24dbm output) instead of the output die. Need few 1206 resistors for DC voltage drop. Use the built in WB RF toroid as load.

3. Solder thin wires to and from the EIC-008 amplifier, you can use the on board coupling capacitors.

Regards,
David Sudry
Carmel Wireless
Shoresh, Israel
Like • Show 0 Likes0
Actions
JohnMiles Jun 6, 2007 12:09 AM
Mark Correct
Correct Answer
Probably too late to help anyone in this thread, but as it turns out, the schematics for the earlier 8753A/B models have come to light recently in the form of a CLIP document that was evidently never officially released ( 08753-90169, approx. 40 MB). I don't know how similar the 8753C modules were to the 8753A/B, but there may still be some useful clues in there somewhere.

Unfortunately the samplers' microcircuit was not documented in this package, nor was the 1349A monochrome display subsystem. Considering this is a 20-year-old document, we're pretty lucky to have what's there.

The link above goes to my personal web space, and will not be valid once the document has been posted to the hparchive.com (and hopefully agilent.com) sites.
Like • Show 0 Likes0
Actions
JohnMiles Aug 13, 2007 6:37 AM
Mark Correct
Correct Answer
Updated URL for the 8753A/B schematics volume:
http://www.hparchive.com/Manuals/HP-8753A-Manual-Service.pdf
Like • Show 0 Likes0
Actions

Repair of a HP8753C searching for schematics

Outcomes

Related Content

Recommended Content