Marconi 2022 2955 OM345 OM360 Replacement

[jdavis6809]

Marconi 2022 2955 replacement for OM345 OM360 Feedback please

Please let me know what other Marconi Instruments use these devices so I can check they are compatible to be replaced.

Does anybody know what the 1db compression point for the OM345 ? (not in the datasheet I have it is blank ).

Does anybody know what the 1db compression point for the OM360 ? (not in the datasheet I have it is blank).

Replacement for OM345

I have a Marconi 2022c which was faulty found three OM345’s faulty (bias voltages were miles out)

I did find a supply for the OM345’s from someone in South Africa wanting £25 each (not Philips modules) but decided to make my own. The AB1/2 board uses 7 OM345’s which would have been a total cost of £175 to replace all 7.
On the internet there was a design using a MAR6 but this has too much gain with bias current of 12mA having to replace resistors on the AB1/2 board. This is the wrong MMIC device to use for a drop in replacement.

I have made one prototype board with a MMIC and two 1.2k resistors in parallel to prove that the gain was 12dBm at a supply voltage of 12v. The bias voltage is 5v and the bias current is 12mA for a 12v Supply.
This led to building three prototype boards to put in the Marconi 2022c in IC positions IC8 ,9 and 10 on the AB1/2 board assembly .
Tests proved that the unit was working correctly also made no changes to the EAROM values.

I would like to explore the idea of making a PCB and supplying a built and tested modules please give me feedback so I can see that there is a demand to produce the OM345 replacement module. This would be a drop in replacement with no other modifications.
The overall height will be slightly higher due to the thickness of the plastic of the 0.1” header. This module will be far more reliable than the OM345’s

The contents are
1. PCB size approx 13mm x 8mm x 0.6mm double sided with ground plane
2. MMIC
3. 1.2k smd 0802 resistor x 2
4. 0.1” header x 5 round pin or flat pin no sure yet


Replacement for OM360

No tests carried out yet but I do have a MMIC in mind to get 24dBm gain. One problem faced with the replacement module is the bias voltage. In the Marconi 2022 the output load is open collector which feeds an inductor 2u2 (L26) to a 12v supply. Thus not a direct drop in replacement when using a 5v bias MMIC.
In the case of the Marconi 2022
There are 2 options
1. At the 12v supply end of the 2u2 inductor L26 insert a resistor in series with the inductor L26 so to give 5v bias.
2. Remove 12v supply end of the 2u2 inductorL26 and feed 5v from the nearest point on the board (in case of the Marconi 2022 there is another 2u2 inductor L33 nearby that is fed from 5v).

I would like to explore the idea of making a PCB and supplying a built and tested modules please give me feedback so I can see that there is a demand to produce the OM360 replacement module. The overall height will be slightly higher due to the thickness of the plastic of the 0.1” header. This module will be far more reliable than the OM360’s

The contents are
1. PCB size approx 26 mm x 8mm x 0.6mm double sided with ground plane
2. MMIC
3. 0.1” header x 10 (with 2 missing pins) round pin or flat pin no sure yet

Regards

John Davis
(RF design engineer by trade) Location UK

[Pamphonica]

Hi John,
this is certainly a perennial problem with the Marconi equipment of this era. I have replaced quite a few myself, with original NOS OM-345, mainly in 2022-series sig-gens.

Distel used to stock OM-345s but are now sold out.

The Marconi forum on Yahoo has had a lot of discussion about possible replacements using MMICs. I don'y know if you have checked the posts there out.
This included one supposedly fully-tested module from a chap signing himself "Steve in Cornwall" for about £10. But he does not seem to answer emails or posts on the subject now.

There is definitely a demand, if they can be made at a reasonable price.

I will PM you my email so you can send me details off this forum.

By the way, if you do offer modules via this forum, make sure you use the For Sale section and give an exact price. Sales via this forum are strictly regulated, to protect forum members

Good luck - these replacements are definitely needed!

All the best

Jeremy

[Radio Wrangler]

There's no rocket science on one of those parts. A small PCB and a few SMT components should fill the bill. I think the OM345 was a single transistor with a dozen passives. Yes, if things were being built on alumina you could make your own resistors, but you still needed transistors and capacitors. I think the active device involved was awfully close to a BFR93, if I remember from when I de-capped one of the beasties. Nowadays we take that sort of performance for granted straight on the PCB.

David

[G0HZU_JMR]

Yes, why not take one apart and copy it?

You could use 0402 and 0603 parts for the RLC parts and a decent SOT-23 device like the BFR93 or something from NEC.

Looking at the bias point of 11.5mA and the 99dBuV spec for the 3 tone DIN test I'd expect the P1dB to be fairly weedy.

I'd guess +3dBm in a 75R system but I have never played with one.

What RF output power level do they run at in the Mi2022?

Note that to describe it as a replacement for the OM345 you would have to prove it had similar S parameter behaviour to the original and also that it meets the DIN test for IMD.

I knocked up a quick simulation and found I needed to include some SMD inductors to get the flatness across 40-860MHz but I managed to get close to the -60dBc DIN spec for the 3 tone test in a 75R simulator and also I managed to get fairly similar smith chart plots. The P1dB when tested in a 50R simulator was close to +3dBm but this was all based on my guesses for the component values as my datasheet for the OM345 only gives a schematic but no component values.

[G0HZU_JMR]

If you want to stick with a MMIC then maybe consider the VAM-7+ or the RAM-7+ from MiniCircuits.

They are meant to be biased at 22mA but you might still get close to 12dB gain at 12mA bias. Certainly worth a try.

The VAM-7+ is in a smaller and cheaper package but the gain is slightly less than the RAM-7+ below.

http://217.34.103.131/pdfs/RAM-7+.pdf

My slight concern with this approach is that a darlington based MMIC won't offer quite the same linearity. But at least they are easy to fit on a tiny PCB and they should be unconditionally stable with a decent RF layout.
I guess it depends how they get used in the sig gen. Do they just see a steady cw signal or can it be AM modulated?

[davidw]

It's taken me a few days to remember where I have seen one of these devices in use.
A trip into my loft revealed this. It's an old Antiference TV Aerial Amp.
May be worth looking out for similar bits of kit at boot sales,rallies etc.
Just a thought.

[jdavis6809]

Originally Posted by G0HZU_JMR

If you want to stick with a MMIC then maybe consider the VAM-7+ or the RAM-7+ from MiniCircuits.

The tests I have done show that the max output from the OM345 in a 2022 sig gen is -10dBm

I am aware that mmics are not linear if over driven or near the P1db point.

The tests I did show that CW, AM ,FM and phase mod modes were OK.

The devices you quote are no good, 22mA for 12db of gain that translates to 22mA x 7 devices = 154mA. For the OM345 it is 12mA x 7 = 84mA. Also if driven at 12mA the gain will be a lot less than 12dB.

Not a good idea to add more current than needed (not sure the power suppy can take it).

I have done the PCB layout

The plan

1. Place order for the components and PCB.
2. Build 7 boards and fit in my 2022C and fully test (log the amplitude EAROM values). Log all the amplitude calibration points before and after adjusting the EAROM values.
3. Offer 7 boards to someone else for free at no cost (for 2022) who is willing to do a full change and log the results.

4. Offer 2 boards to someone else for free at no cost (for 2955) who is willing to do a full change and log the results. (I have checked there there are 2 devices used in the 2955 if not please let me know).

5.After 4 and 5 concluded offer boards for sale.

On this forum when items are offered for sale does evebody on the forum get emailed?

When I post a message on this forum I get no email responses why?

Regards

John Davis.

[Station X]

Originally Posted by jdavis6809

On this forum when items are offered for sale does everybody on the forum get emailed?

No.

Originally Posted by jdavis6809

When I post a message on this forum I get no email responses why?

I assume you want to receive a notification email when someone replies to a thread?

Select "User CP" at the top of this page then select "Edit options". You'll find an option "Default Thread Subscription Mode". Change the option to suit your requirements.

[G0HZU_JMR]

Hi John
I suggested the RAM-7 as a possible contender for the OM345 because it has such a flat response up to 900MHz. If it could be biased at 13-15mA I would expect it to achieve about 12dB gain but I've never tested one. However, the P1dB would probably sink to something like -1dBm.

My guess would be that it's heavily based on the old Avantek MSA-0735 so probably has similar characteristics.


I have no experience of diving inside the Mi2022c and I don't know how much this would affect the operation or how much it would stress the PSU. I'd expect most 12mA MMIC amps to have a P1dB around 0dBm and an OIP3 of somewhere about +12dBm so I think this would be slightly worse than a genuine OM345. Can you locate and measure a healthy OM345 for P1dB etc?


Note:
I did have a quick look at the schematic for one of the Mi2022 boards and spotted 8 of these amps on the RFP board. Should you be offering a full kit of 8 amplifiers rather than 7 if you want to test in every slot? Or is one of them a no fit or option? At a casual glance they all looked like they were needed to me.

Regards
Jeremy

[jdavis6809]

2022 has

OM345 x 7 (12dB)
OM360 x1 (24dB)

Regards

John Davis

[G0HZU_JMR]

The reason I ask is because the parts list and the schematic both list OM345 qty as 8 on the RFP board.

There is also a single OM360.

Have you missed one or is one of the OM345s not fitted or linked out in these generators?

I quickly scanned over the RFP circuit and spotted 8 of them (plus an OM360) because I was interested to see which of them carried amplitude modulated signals.

[jdavis6809]

Hello,

Yes. Stand corrected. there are 8 OM345 on the board (2 under VCO screening metalwork)

Regards

John Davis.

[jdavis6809]

Update.

The amplitude modulation is done after amplification via OM345's so has no effect.

PCB's and components ordered today.

Regards.

John Davis.

[G0HZU_JMR]

I've only skimmed over the circuit to see how it works but I think IC14 (OM345) is inside the amplitude level correction (ALC) loop for signals in the 62.5-1000MHz range and must therefore see the AM modulated RF signal when the generator is in AM mode and in the 62.5-1000MHz range?

[Stockden]

It's worth remembering also thats there's at least one OM345 in the 2018/2019 genny's

[G0HZU_JMR]

Originally Posted by jdavis6809

Does anybody know what the 1db compression point for the OM360 ? (not in the datasheet I have it is blank).

To try and answer this for you I studied the OM360 schematic in the Mi2022 and it looks to me that there needs to be 4V pkpk of RF at the output pin of the OM360 when that little circuit runs correctly in closed loop. It needs this much RF in order to get a detected 5V dc at pin 2 of the ALC error amplifier to close the loop correctly.

So whatever amplifer replacement you fit, I think it has to be happy to produce 4Vpkpk at the output pin (with reasonable linearity) and it also needs to be happy working in a closed loop level correction system because it will mimic a voltage source in closed loop.

[G0HZU_JMR]

So whatever MMIC you fit here for the OM360 will need to be able to cope with the baggage associated with the ALC loop and also the RF switching network around it without it all going unstable.

As most 55mA high gain MMICs (eg 24dB flat gain) are not unconditionally stable and also have poor reverse isolation you could find that a MMIC will cause subtle spurious problems if the OM360 replacement hoots on certain generator ranges. Most MMICs that aren't unconditionally stable hate having an RF short at the input.

Therefore, having looked at the Mi2022 schematic around the OM360 I think it would be safer to copy the OM360 internal circuit because the OM360 boasts very high reverse isolation (and presumably has a high K factor) and is going to be far less likely to go unstable in that environment

[jdavis6809]

An update

the pcb boards are here with all the other components

built 8 and fitted to Marconi 2022c (one at a time the retested instrument)

OM360 (24dbm) is still oiginal and has not be replaced

unit now soak testing

cailbration to follow (just amplitude eerom values)

at the moment none of the amplitude values are >0.5dbm in error from 10Khz to 1Ghz

next step to offer samples for free for some one to try out on a

1. marconi 2022
2. marconi 2995
3. marconi 2018 and 2019

see attached picture showing marconi AB1/2 Board populated with OM345 replacents

regards

john davis (uk)

[Louis Timmer]

Hi John,

A well needed exercise to keep 2955 and other Marconi Equipment going,
have a 2955A that needs at least 2 of the 4 OM345 or OM2045 to be replaced as I have the error messages indicating faulty ones. I am sure that many will welcome your effort as it is frustrating if one has a piece of equipment that cannot be repaired as the used parts are unobtainable....
As a pure hobbyist I did over the years accumulate a lot of mainly Tek equipment but do not have the expertise of a design engineer. Hope you find someone close by that can test your OM345 replacement in a 2955 Test set and you can make a note on my interest for 4.....73 Louis vk4bvy

Being new on this forum I also am restoring a Marconi 2305 modulation meter. Could not find any Philips OMxxx chips....but the frequency window
on this meter is broken and thus unreadable so needs replacement....any
member that can help with a spare LCD window from a 2305?

Louis vk4bvy

[G0HZU_JMR]

Sorry to keep sounding a bit cautious and negative but it's worth mentioning that whoever tests a set of these in a Mi2022 would need access to a spectrum analyser and a decent power meter and would also need to delve inside to measure the performance of the RF levelling system and also the AM quality.

The technical spec for these generators is already a bit poor when studied especially for things like harmonic levels and sub harmonics and it would be a shame if this got degraded and it wasn't spotted.

eg the quality of the RF levelling and the AM modulation will be affected by things like the harmonics and the sub harmonics at the generator output because they will influence the ALC action.

Also, swapping all the OM345s with replacements is bound to change the flatness and gain of the signal path across frequency and so this would require internal before and after testing to make sure the calibration levelling system was able to cope with a decent margin across all frequencies and across the full range of the levelling system. You can only test this by delving inside the generator to make a series of measurements on the levelling performance.

Having looked at the circuit I would expect IC3 and IC14 to be the OM345 devices that are the most performance critical.

IC14 is inside the ALC loop so it will be AM modulated on the 62.5-1000MHz range and I think IC3 will be driven very hard on the 500-1000MHz range. So if I had a 2022 and one of these two devices failed I'd be tempted to swap across a OM345 from another part of the generator and then replace the other OM345 with an alternative homebrew device.

Also, I remember that some 2022 generators have a lower output power capability. eg some do +13dBm max and some early models only do something like +7dBm max. I'm not sure how relevant this is but I'd expect there would be subtle differences in the levelling performance making comparisons a little risky.