Marconi 2388 1GHz RF Probe Investigation

[G0HZU]

This old Marconi 2388 1GHz active RF probe is part of the stash of test gear I saved at a recent clearout at work. I've finally got round to checking if it still works and if I can modify the power supply interface so it will work with all of my spectrum analysers.

It does have a very good spec and I'm hoping that it hasn't been overloaded and toasted by some idiot at work.

Sadly, the kit is not complete but it does still have two new spare tips. If it works, I'll make the important accessories by hand. eg I'll make an ugly grounding sleeve and an earth spike. I've already had a basic attempt at making a sleeve as you can see in the image.

I'll post up some pictures of the insides later once I get it apart. I did the same thing with an old Marconi TK2374 probe a few years back although it only works up to about 200MHz. Hopefully, this probe will mean I get decent performance to 1GHz.

[KJDS]

I last took one of those apart in 1991

[G0HZU]

I last took one of those apart in 1991

For some nostalgia, here's a couple of images of the insides of the other end of the Mi 2388 probe.

In case anyone else wonders what this bit does, it is a screened box when complete and it plugs into the analyser RF input port via an N type connector on the underside. It acts as a junction to allow the RF signal from the probe to connect to the analyser but it also is a junction point for the DC power feed. This feed normally comes from the front panel of the 2382 analyser and the little blue AA1 connector is used here.

The probe is designed to be used with the Marconi 2382 spectrum analyser. We got rid of the 2382 analyser about 10 years ago so I'm hoping that this probe has sat neglected (but not abused) for all of this time as it won't work with any other analyser because of the custom power connections it requires.

The colour coding on the wiring looks a bit odd for +15V and -15V but I'll check another Marconi manual to see if the pinouts are correct in the schematic.
You can see that the schematic shows that two of the connections are bridged together in the probe power lead at connector PLA but clearly this one isn't wired the same way as the green wire does nothing and no two pins in that little blue connector (marked as PLA in the schematic) are bridged together if I test with a meter/buzzer.

Hopefully, I just need to connect it up to +15V and -15V with a ground return and it will work. I'll try this later this evening

[Fraser]

I have a couple of these probes. The schematic was in the manual and I was able to use it with other spectrum analysers by using a separate dual rail power supply. The positive and negative rails were simple to create using linear regulators. My SA's also have the required power rails available from their 'Probe power' connector.

Good probes though, with the built in variable attenuator. I use them with 3GHz Advantest SA's. My kits are complete so I can likely supply details of any parts you are missing.

Fraser

[G0HZU]

Thanks for the info. I can remember using it with the grounding sleeve when the kit was complete and there were various adpators with it. Sadly, these are long gone. I may try searching for them at work in some of the RF adaptor storage boxes.

There is some good news though... I removed the original power feed and wired up to external supplies and it all seems to work fine I've included the series 120R resistor for the -15V rail and this now lives inside the probe body.

I'm quite pleased with my first bodged attempt at an RF ground sleeve and spike. The metal I've used is a special springy alloy and it fits and grips very snugly. I have to twist and pull it quite firmly to get it off as the probe outer is ribbed under the sleeve area. I recall that the original sleeve spike had a ferrite bead near its tip. Do yours have this?

I've tested the probe for response up to 1GHz at various attenuator settings and it seems fine. See below for my fudged connections. Ultimately, I'd like to refit the power lead as per original and I'll do the same as you in that I'll find a way to connect to some clean supplies via the original power connector. The crude wire mods I've done are all reversible so I can get it back to original very quickly.

I'll obviously have to insulate my homebrew earth sleeve on the outside. But for now I'll just wrap some insulating tape around it

I'll do some linearity and input impedance tests on it at some point in the next few days to prove it really is all OK.

[vk6zgo]

I last took one of those apart in 1991

For some nostalgia, here's a couple of images of the insides of the other end of the Mi 2388 probe.

In case anyone else wonders what this bit does, it is a screened box when complete and it plugs into the analyser RF input port via an N type connector on the underside. It acts as a junction to allow the RF signal from the probe to connect to the analyser but it also is a junction point for the DC power feed. This feed normally comes from the front panel of the 2382 analyser and the little blue AA1 connector is used here.

The probe is designed to be used with the Marconi 2382 spectrum analyser. We got rid of the 2382 analyser about 10 years ago so I'm hoping that this probe has sat neglected (but not abused) for all of this time as it won't work with any other analyser because of the custom power connections it requires.

The colour coding on the wiring looks a bit odd for +15V and -15V but I'll check another Marconi manual to see if the pinouts are correct in the schematic.
You can see that the schematic shows that two of the connections are bridged together in the probe power lead at connector PLA but clearly this one isn't wired the same way as the green wire does nothing and no two pins in that little blue connector (marked as PLA in the schematic) are bridged together if I test with a meter/buzzer.

Hopefully, I just need to connect it up to +15V and -15V with a ground return and it will work. I'll try this later this evening

Dear old "Macaroni" is sometimes a bit cavalier in their attitude to colour codes.
The TV Transmitter at my old workplace with had three colours:-

White,Pink,& Black.

They were all supposed to have identification sleeves.
The problem was,that the sleeves perished & fell off,& Pink or White wires carrying a reasonable amount of current got quite warm,& turned into Black wires!

[G0HZU]

I remember seeing a similar colour code and sleeve system in Marconi's radar products from many years ago.

I think I'm going to make a second type of grounding sleeve for this probe because this one shows some unwanted peaking up at 700MHz and this will be due to the inductance of the skinny ground/earth spike I've used.

I'm going to make one with a conical shape at the tip and a much shorter spike for those occasions where I want to use it up at the top end of its range.

But I plan to use the existing sleeve for most things up to a few hundred MHz because it is going to be easier to manage. I also managed to find a tiny ferrite ring suitable for VHF+ use and I found that this helped damp the peaking when I slipped it over the ground spike.

[joeqsmith]

The table surface looks well loved!   

[G0HZU]

The table surface looks well loved!

Yes, that's my old and battle scarred ESD mat

I took a few response plots of the probe when fitted with my lashup sleeve and ground spike. These show the response up to about 1GHz by monitoring Port 1 of the VNA with a 50R termination. The probe is attached directly to port 2 of the VNA. The source is really flat on this VNA when tested with a power meter so this basic plot just shows a combination of the source flatness (typically <+/-0.1dB over this range) and also the response of the probe. So the response is dominated by the probe rather than the VNA.

You can see in the probe0 plot there is a lot of peaking caused by the inductance of the ground spike.  In the Probe_Ferrite plot I added a small ring of ferrite over the leg of the ground spike and this reduces the peaking but spoils the response at the top end.

I know I said I'd make a conical sleeve for the probe but I've been too lazy/busy.

But see the probe_tape plot to see what happens if I add copper tape to the sleeve to fudge the equivalent of a conical tip. Note that I had to use memory store to grab this probe response  because I couldn't keep the tape in place and use the screen dump UI at the same time. I needed three hands for this. So the final plot looks dimmer because I stored the trace to memory and then took the screen dump at my leisure.

The response is very  good up to 1GHz. So I can see the benefit this will bring and I'll try and make a real one that doesn't look too ugly

Looking at the plots I can see I should have locked the rotary attenuator built into the tip because it looks like it has rotated slightly between plots as I messed with the tip and the sleeve. This explains the slight changes in absolute level between plots. Sorry about that...

[Andrejka]

Hello

I also have this probe, but readings is approx -25dBm below specs. Can You suggest what is toasted inside? Maybe GaAs transistor inside tip?

Thanks

Andrej

[LR]

Hi,

good you help me with a manual/schematics of this probe?

Thanks in advanced
L.R.