[SOLVED] Noise around CW carrier spectrum analyzer Advantest R3465

[hafrse]

Hello,

I have a problem with my spectrum analyzer, some times it displays the signal (for example the calibration signal of 30Mhz) with extensive noise occasionally.

I have connected the 21.4 Mhz IF output to another spectrum analyzer (R&S) and watched the signal at 0 span at the moment it started to act with extensive noise but the IF output is very clean!

The calbration signal is clean at all times, the spectrum analyzer consits of these blcoks:

1. RF Block with all 3 Mixers
2. Synthesizer Block
3. IF Filter Block
4. LOG A/D Block

The 21.4 Mhz IF signal is the output from the RF block.
When I measure the outpout of the IF Filter Block which goes to the LOG board at 0 span, the amplitued gliches a litte, also visible on 0 span on the R3465 (the faulty analyzer).

What I can conclude here is that I have a problem in the IF Block itself since the 21.4 Mhz output of the RF block is perfect.
Is this correct ? or could I have missed something here?
Attached a picture showing the noise when it appears.

Many thanks for any help!

George

 

[Warpspeed]

I have very recently bought an old 3361 from e-bay, which has yet to arrive.

But I have been avidly reading through the enormous 500+ page service manual these last few days, so its all pretty fresh in my mind.

There is a lot of circuitry involved with switching the various bandpass filters in and out in the IF block.
Have you tried selecting different bandwidth filters ?

 

[hafrse]

Hello, Yes I did, noise was still there , it is not related to a specific filter section, I also recapped all the electrolytes on this board, can be some noisy oscillator or input/output amplifliers within the IF filter board (?)
George

 

[Warpspeed]

Could be many things, but all the filter and gain switching in the IF block is done by switching on and off small silicon diodes.

If one of those was intermittently leaky it could conceivably create a lot of noise in the biassed off condition. The same diode might work fine forward biased in the on condition.

So I was wondering if there is a specific condition where the noise is not there ?

A few judicious squirts with a can of freeze spray might provoke some action.

 

[hafrse]

Tony,

Thanks for the information, You are right, the behaviour is like this, if I switch on the analyzer and IF the noise is present, it will fade away after about 15 min. and the analyzer is ok.
I tried to switch on and off the analyer many times last week, I can tell that about 1-2 times out of 10 the noise appears on switch it on, then when on , the noise fades away .
George

Picture of the IF board

 

[Warpspeed]

The more I think about this, the more likely it could be one of those switching diodes, and there are rather a lot of possible candidates to choose between.

Just reading through the service manual, these spectrum analysers must be about the most intimidating thing I have ever seen from the servicing and repair aspect.

 

[hafrse]

There are no schemtics available for this analyzer.
I have applied hot air to the IF board yesterday , waited unitil it become completley cold and after that I could not recreate the noise!

 

[Warpspeed]

All you can really do now is wait for it to become noisy again.
Arm yourself with a can of freeze spray, and give each diode on the IF board a very small and very local blast.

Some kind of masking device to confine the cold to one component may also help.

But I doubt very much if later models have changed all that much except perhaps in the RF deck and the software. I am guessing that there is not really that much that could be changed in the IF section.

An earlier IF board might be similar enough to at least help you navigate around various parts of the IF, and get a better understanding of what it all does.

 

[hafrse]

Correct, I will wait and see what happens. The problem is that I do not have access to the board when it is installed, I need a jumber adapter board or cables to do that when the noise appears. I have the schemtics for the R3271, the IF filetrs are different, it has filters down to 10Hz and uses FFT (digital) for narrow bandwidthes as well, it uses externa 25Mhz from the synth. unit and internal mixers and osc.
Could this type of noise be comming from the intenal oscalltors on the IF filter board ?

 

[Warpspeed]

Could this type of noise be comming from the intenal oscalltors on the IF filter board ?

I really do not know.
I do not yet even have my own spectrum analyser yet !
Should arrive in one to two weeks.

That noise looks to me like something breaking down, classic sudden high leakage current.
It might be ESD damage somewhere, but my bet is still that it is one of those pesky switching diodes, as I have seen very similar symptoms before with leaky switching diodes in some Hewlett Packard equipment.

Heat sensitivity is another suggestive clue pointing in that direction.

 

[hafrse]

Today, after many power up attemps, still no noise. Will watch and see in the comming days.
anyhow, thanks for the information I got, when the problem arises again I will try to build connection ribbon cables with sockets and work with the board while the noise exists.

 

[Warpspeed]

I think that one of Murphy's laws states that:

If you are totally prepared with extender cables and freeze spray, ready to pounce at a moments notice, the fault will never return.

 

[hafrse]

Hello,

Problem is solved! after many hrs, the problem was in the YIG oscillator RF output to the RF deck! the YIG RF output is soldered directly to the RF deck PCB and was loose. Thats explain why it worked when worm etc...

 

[Warpspeed]

Well done !

My own spectrum analyser has arrived and although pretty old, it works perfectly.
Its something I have always wished for, and it has opened up a whole new world.