Marconi 2022C fan problems

[G0HZU_JMR]

I've recently been playing with my Marconi 2022C sig gen on the bench mainly to try and clean up the spurious levels seen on the synthesiser. Some of my findings may be of interest and also I'm looking for hints and tips to cure one main problem. This problem is that the top VCO range (353-499MHz) has a lot of close to carrier spurious signals on it. As you can see by the plot below, this 2022C does not look good on a spectrum analyser. The 2022D sig gens at work look much better with no trace of this problem.

I've traced the cause of the problem to the onboard fan. I'm not sure if the fan bearings are worn or if the supply cleanup circuit that feeds the fan has a dead filtering cap. If I stop the fan, the spurious signals vanish. So far I've not proven if the interference gets there via the +5V supply or if it is simply due to microphony. I suspect it is microphony as my 2022C seems especially prone to microphony when lightly tapped with a screwdriver. I may try adding some RTV in places to damp and large components in the VCO area.

I don't know if the fan needs replacing or maybe it could be mounted on a thick rubber gasket to stop the vibration affecting the whole chassis. The fan doesn't sound noisy and the amount of 'tremble' felt on the chassis (due to the fan) seems fairly typical to me.

Has anyone else encountered this issue? Is there a common fix for this?

[G0HZU_JMR]

I think I've found out why VCO 2 (353-499MHz) is so prone to microphony. I've taken out the RF board and there is a lot of oxidation on the metalwork that forms the lower RF pockets under the VCOs. This was at its worst in the area underneath the place I found the last solder splash so the heat from the soldering may have accelerated the oxidation or maybe it was from contaminated solder flux. It meant there was an area under VCO 2 that couldn't make a reliable RF connection to the metalwork.

To make things even worse I found another solder splash on the gold layer on the PCB underside. This one was much more proud than the one I found on the top side. It was over by the 160MHz VCO so probably not the cause of any extra microphony at 354MHz. This gold layer has to be clean (no oxidation) and flat (no careless solder splashes!) in order to achieve a good and consistent RF ground when it mates up to the metalwork around the pockets under the VCOs. There was also a wire repair that probably replaces a broken track under the VCOs and this was left unsecured so it could wobble with vibration. Hopefully my 2022C will have less microphony when I clean it up and put it back together. I'm going to RTV some of the large components in the VCO area and this will take a day or so to cure.

However, it could be that none of the above issues are the main cause of the microphony problem so I'm still wondering it it's possible to rubber mount the fan (assuming it isn't already rubber mounted).

[Radio1950]

Gday Jeremy,

Interesting.

I was given two M2022A s, fixed both, gave one to a friend.

Later, mine had a noisy fan.

I opened the Sig Gen case up to see what fan it was.
I think it was a Swiss made Papst?
Forgot to take photos; it always happens.
And my notes just say "Swiss 5 Volt 40x40".

But, my M2022 always ran cool, so did the other one.
So, I tried reducing the fan supply volts by one Si diode, and think I left two diodes in.
Dramatic noise improvement, much quieter
I think I put a drop of oil on the bearings.
The M2022 doesn't run any hotter.

Today I tried an experiment with two 10 MHz sources, one my M2022A, fed to a 6 dB resistive splitter as an adder, then to a diode, monitoring output with the CRO.

Didn't notice anything really, but it isn't much of a test, and I don't have a Specan.

But what I did notice with this setup was that I can confirm your microphonics on mine, when I tap the case.
Probably ceramic capacitors somewhere?


On an other note, I replaced all my Front Panel switches.
Really worthwhile.
.

[G0HZU_JMR]

Hi R1950, thanks for the reply. I've managed to prove it is microphony by unscrewing the fan from the chassis. The interference stops when I do this.

My fan looks to be the original Micronel 5V fan. These are really expensive to replace. It's a shame the sig gen is so prone to microphony. Mine is much worse on the top VCO range and this covers about 353MHz to 499MHz. The lower VCO covers about 250-352MHz and it is much less microphonic.

It seems odd that Marconi chose to use printed VCOs directly on the large RF main board. Usually this would be done using little daughterboard VCOs.

My attempts to damp the VCO parts did seem to improve things a bit but when I refitted the large inner screened lid over the main RF board the microphony was bad again. I'm going to try a rubber gasket on the fan but I don't think it will help much. I also held the fan loose in my hand and it does feel a bit unsmooth but nothing unusual. I think it is the fact the VCOs are so microphonic that is making this problem worse than it should be. Maybe there is a component on the high VCO that is faulty or it could be a microphonic ceramic cap as you say. I checked all the soldered joints and they were OK.

Probably my best hope is to buy a smooth running 5V fan and try it in the chassis.

[Radio1950]

What about running temperature tests with sensors placed inside and left for a few hours.
Then, disable fan and redo temperature tests.
Remove fan if satisfactory.

Paint case cover insides with matt black.
Maybe make new covers with perforated sheet metal.

These sig gens are just so useful because of small size and ease of setup.
Maybe not for exacting work, but ok for 99.9 percent of home tests.
.

[CambridgeWorks]

Hi Jeremy. Years ago when I occasionally worked on Racal 9081 or 9082 Sig gens, I found the residual FM could be improved by removing the main vco board and cleaning up the contact areas of both the PCB and the casting, then replacing, making sure the screws were tight. I used one of those fibreglass bristle pens. Amount of improvement would vary, but was worth doing.
Rob

[G0HZU_JMR]

Hi Rob, There certainly was some pitting and oxidation in the metalwork under the VCOs. I think it made a slight improvement after cleaning it all up. However, the really good news is that my bodged rubber 'gasket' has worked really well.

It's only an experiment so far but I glooped about a 3mm deep 'gasket' using Loctite 5145 and left it to cure for 2 days. I just tried it and the microphony has almost gone. Certainly not a problem now. However, it might get worse when the 5145 really sets. I'll try and make a proper rubber gasket out of some squishy rubber material. The 5145 should just peel off cleanly.

I'm still curious as to why the 353-499MHz VCO is significantly more microphonic than the 250-352MHz VCO. There may be something wrong in the VCO area. Maybe a dodgy ceramic cap as R1950 suggests. I keep meaning to do a microphony test on one of the 2022D sig gens at work to compare against my 2022C.

[G0HZU_JMR]

Note that I have just checked that the 5145 gloop hasn't got into the fan and stopped it from rotating! I just had the sinking feeling that maybe that was why it looked so good now. But the fan is spinning just fine. A good result! I can't feel any trembling on the chassis now either when the fan is running.

[G0HZU_JMR]

I think I may have put the fan back in with a 90deg rotation because I see that the manufacturer's sticker is at the top now. Could the rotation also be helping I wonder? Maybe that could help with smoother running. It seems too good to be true at the moment.

[nemo1956]

Hi guys. That's very interesting.
I have 2202 that I'm going to start working on.
At the moment all I get when switching it on is all the front led lights come on and that's it apart from the fan.
Going to be fun I'm sure.

Nemo.

[Radio1950]

Jeremy

I had my M2022 out and on, all the day.
I had mentioned previously that I dropped the fan volts on my unit by two silicon diodes, to reduce my fan noise to virtually nil.

I measured the M2022 fan exhaust temperature today as 35C.
Case temp was 32C
Ambient temp 28C (eat your heart out).

For consideration and interest.

[Radio Wrangler]

Those fans can also radiate strong AC mag field components as well as mechanical vibration.

Solder or corrosion disrupting continuous ground gonnection to screening boxes etc can leave slots to leak RFI in or out, but also they spoil low frequency screening as well.

These little sig gens were cost-competitive units and were a bit simpler than they could have been, resulting in lower loop bandwidths, and so leaving VCOs open to external influences (electrical, magnetic, and mechanical)

I suspect a clean-up or re-plating of the grounding surfaces will help, as will a better fan.

David

[Dickie]

Don't forget that the 2022, like other professional gear, will have been designed to operate to full specification up to an ambient of at least 40C, so reducing the airflow at lower ambients should not be a problem. (Though it won't help the long term reliability, of course).