Marconi 2022 AM Modulation Problem

[Wendymott]

Hi Guys. Looks like the OM345 problems have been solved by your good selves, however my 2022 has just stopped AM Modulation. FM is ok and the RF output is OK. Is it another of the OM345's causing problems? Or some other device?
I seem to think I have seen a thread on this before, but I cant find it. Your assistance would be much appreciated.
Regards
Wendy G8BZY

Moved from this thread: https://www.vintage-radio.net/forum/...d.php?t=100840

[G0HZU_JMR]

Hi Wendy
The AM system operates in two separate ranges:

LF to 62.5MHz
62.5-1000MHz

Has your AM failed on both ranges or just one?

[Wendymott]

Hmmmmmmmmmm let me check. I know it was ok on the LF range. I will come back to you Jeremy.

Thanks, Wendy G8BZY

[Wendymott]

Hi Jeremy. AM Modulation OK on lower range. None on upper range.

Regards, Wendy G8BZY.

[G0HZU_JMR]

Hi Wendy
The AM system and the ALC system for (correctly) calibrating/levelling the output on the 62.5-1000MHz are strongly linked. So can you confirm if it is possible to control the RF carrier level in small steps on this upper range?

i.e. set the sig gen to AM mode and the RF output carrier to something like 0dBm at 70MHz and then press

'DELTA' RF LEVEL' '1' 'dB'

and then use the up and down keys to see if the level from the sign gen changes in (roughly) 1dB steps on a scope.

i.e. click it down 6dB in 1dB steps and see if the voltage on the scope steps down nicely to be about a half what it was at 0dBm.

This test will prove the ALC feedback circuit is in basic good health and hopefully proves the generator can control and level the output to the calibrated ALC target for each of the 1dB steps. On the 62.5MHz to 1000MHz range the AM system exploits this ALC system in order to generate its AM. So this is a good initial test to try and it can help to narrow down where the problem is.

[Wendymott]

Hi Jeremy, will do and come back to you. Bench a bit full at the mo, will do it Monday.
Thanks for your assistance

Regards
Wendy G8BZY

[Wendymott]

Hi Jeremy. I am in the very fortunate position in having two of these generators, one I bought at Xmas, so I have a good reference. The faulty one is stepping in 10db steps only, not 1db steps as suggested. I have a feeling this goes back to the OM345 IC9 that I replaced with a MMIC, unless there is another suspect.
I am also fortunate to have a 1Gig spectrum analyser with tracking generator, as my retirement present to me. So I am not short of good measurement instruments.
Again thanks for your interest and help
Regards
Wendy G8BZY

[G0HZU_JMR]

It sounds like the ALC system has gone open loop on this range.

Usually when this happens in such a control system the RF level slews to a level way too high or, if there isn't enough RF drive in the system to reach the calibrated target for the ALC then the RF level will sit some way below where you expect it to be.

So can you measure the level with the generator set to 0dBm at 70MHz and see how far away from 0dBm it is? Is it too high or too low?

The system is designed to have a fair bit of excess RF gain so if the ALC is working it can throttle back the RF level using a diode based attenuator (just before IC14) and achieve exactly 0dBm.

But if there isn't enough RF gain to reach 0dBm even with the ALC attenuator set to minimum attenuation then the ALC cannot begin to add any low level AM modulation.

I don't know if you know how the AM/ALC works but have a look at the diagram below. The ALC system will settle when the RF level at the output causes the diode peak detector to produce a DC voltage at the error amplifier that matches the target voltage at the reference input to the error amplifier.

When you change the power in 1dB steps the controller inside the sig gen sends digital data that changes the DC target voltage from the DAC that gets fed to the error amplifier. So if this Vtarget was reduced then the sig gen level at the RF output will fall by the same amount as long as the system is healthy.

I think the ALC is at fault and the AM will probably be OK once you sort out the ALC issues.

But, for reference here is how it does the AM on this range:

The error amplifier is designed to be VERY fast acting so it can track very fast changes in Vtarget from the DAC. Even changes at an audio rate. So if you modulate Vtarget then you effectively AM modulate the sig gen because the error amplifier will faithfully follow the modulation peaks and crests and the sig gen will be AM modulated via the attenuator.

But there is a slight problem in that the (old 1980s tech) controller can't update the DAC quickly or finely enough to give a smooth modulation. So Marconi feed the AM to the Vref of the DAC and this directly modulates the DAC output really smoothly

So the controller sets the Vtarget dc level from the DAC to give 0dBm and the AM rides on top of the Vtarget voltage via the Vref input to the DAC. So you keep the RF level at 0dBm but also get AM for free.

Hope this all makes sense

[Wendymott]

Many thanks for the explanation and information. I will check Monday and let you know

Regards
Wendy G8BZY

[Wendymott]

Hi Jeremy. The output of the faulty 2022 is -12db compared with the good 2022.

Regards
Wendy G8BZY

[G0HZU_JMR]

If it is 12dB low then the open loop gain is probably lacking by >20dB in reality because the system will be designed with a fair bit of excess open loop gain.

eg to allow 100% AM modulation there needs to be an extra 6dB headroom for a start because the RF envelope doubles up in voltage on 100% mod.

So this sounds like a classic OM345 failure causing a >20dB loss in open loop gain. To narrow it down further can you check to see if the generator behaves itself above 250MHz?

eg try AM modulation on 270MHz. Also compare the good generator vs the bad one on the analyser for this test for correct RF level.

I can't read the fuzzy text on my schematic very well but the chief suspects will be the OM345 IC9 and possibly IC14. Probably the quickest way to verify this is to measure the dc and RF levels at IC9 at pins 1 and 5 to see if it has died.

It looks like your initial belief that IC9 could be the problem could well be correct

Just to clarify... if it suddenly starts behaving itself again above 250MHz then IC9 and the /2 and /4 section after it is likely to be the problem area.

[G0HZU_JMR]

Thinking about it a bit more...
It's unlikely to be IC9 because as long as it can drive the /2 and /4 circuits then these should work OK. As long as they are dividing correctly then the dividers will set the drive level and not IC9. If IC9 dropped too low then the dividers wouldn't work at all.
This assumes the generator really does produce the correct frequency of 70MHz (but at -12dB low) when set to 70MHz on the front panel.

It's still worth doing the >250MHz tests but it could be IC14 that has failed this time?

[Wendymott]

Hi Jeremy.
Thanks for the additional information.... I can confirm that the frequencies displayed are correct tested as per the output.....I can also state the RF output is still 12db at 260Mhz... lower than the good generator with NO AM modulation, as conformed by the good 2022.
You may remember I did change the IC9 OM345 for a MAR01 and that the output on the Lower range was restored, and that the displayed frequency is correct as per my Spectrum analyser. I will lift the 2022 down from its shelf tomorrow and investigate the output of IC14 and report back Tuesday.
If it has failed, I will try the "scraping" of the outer coating to look at the internal resistor.
Many thanks
Regards
Wendy G8BY

[Wendymott]

Hi Jeremy. The input to IC14, is........ using my analyser is 40db....the output on pin 5 is approx. 46db...I then looked on the collector of TR20 (Power device) and there is a decent signal..... WITH> MODULATION........Confirmed by switching on and off.......however there is NO Mod on the output coax terminal. Guess what........... it was a dry joint on the collector of TR20.............All up and running PERFECTLY..RF steps in 1db steps and mod 100% .... Many Many thanks for your assistance........
Kind regards
Wendy
G8BZY

[G0HZU_JMR]

Hi Wendy
That's good news! Glad you found the fault!

I did a quick paste together of the relevant circuit sections from two pages from the manual to show the schematic equivalent of the previous block diagram of the AM + ALC system on the upper range of the generator..

I guess you don't need it now but someone else might find it useful for future reference. So I've added it below

[Wendymott]

Great, thanks, I made a copy for future use. Fingers crossed there wont be a "next" time.
Case back on, back on its position on the shelf.

Regards
Wendy
G8BZY