Marconi TF2507 RF Wattmeter - UK Vintage Radio Repair and Restoration Discussion Forum

TonyDuell · 24th May 2021, 4:38 pm

The other instrument I've got is a TF2507 wattmeter.

It seems to consist of 3 detector units (1W; 3W/10W, 30W/100W ranges) linked to dummy loads. The output of the appropriate detector goes to a module labelled 'DC amplifier' and that drives a moving coil meter.

The input impedance, not surprisingly, is 50 Ohms, but there's a '75 ohm converter' clipped to to the top. This may be nothing more than a 25 ohm resistor (there is a switch on the instrument that you flip when you use the converter, I guess to compensate for the power lost in said converter). It's in a large heatsink that appears to be liquid-filled, I assume some kind of transformer oil so if I do open it up, I will do so outside in a suitable container to catch the oil.

There is a warning on top of the main instrument that the detectors and loads contain beryllia, so I've not looked at that part of the instrument. I have extracted the DC amplifier and opened it up, it's discrete transistors on a glass fibre PCB.

Again I can find nothing about it on google.

Granitehill · 24th May 2021, 5:59 pm

I don't know the TF2507, but it sounds like a later version of the TF2503 which I have.
The 2503 is 30/100W and has a conical shaped thin-film resistor, cooled by fluid inside a large finned heatsink. According to "Marconi Instrumentation" vol 12, the liquid is "A fluorocarbon liquid coolant up to 40 times as efficient as conventional transformer oil". This is pretty well unobtainable these days I think, so best to keep it safe. If it's the same stuff I used for cooling klystrons, it's non-toxic and just looks like water, but is a lot heavier. Pretty expensive also, I think.

I think the beryllia is probably in the 3/10W load which was used in the smaller TF2502. In this case the load was a thin film deposited on a beryllia substrate, about the size of a quarter watt resistor. The substrate was set into a small finned heatsink.

These were pretty upmarket meters - £493 in 1969 is a lot for an absorption load !

TonyDuell · 24th May 2021, 6:10 pm

There are 3 separate input connectors, one for the 1W range, one for the 3W and 10W ranges and the last for the 30W and 100W ranges.

Each connector links to the 'detector', there being 3 of these, different colours for identification. The detectors have 2 RF connectors and a screend cable to the range switch (I think), the last being the DC output representing the power. Of the 2 RF connectors on the detectors, one as I said goes to the front panel input socket. The other to the load. The loads for the 1W detector and 3W/10W detector are small round heatsinks on the back of the chassis. The load for the30W/100W detector is a much larger heatsink, but I am not convinced it is liquid-filled. The 75 ohm converter, a separarte heatsink unit with an RF connector on each side, does contain liquid, I can hear it sloshing about.

The warning on the instrument about beryllia seems to imply it's found in all the detectors and loads. I'd want to read a service manual before taking that side of things apart anyway.

TonyDuell · 25th Jul 2021, 7:37 am

I've now got it apart on the bench. It splits into several modules quite easily :

The 3 detector units. These are similar in principle to the one in the TF2502, and use thermocouples to sense the temperature rise in a resistor in series with the RF signal. There's also a diode probe which is used for the 'peaking' function I think. The casings come off easily (4 screws), there is a warning about 'beryllia cement' on them. I am not going to dismantle any further unless I have to attempt to repair a fault. Each detector has a coaxial input from the front panel connector, an N plug output to the load and a 3 pin connector to the front panel wiring. The 3 pins being the ends of the thermocouple and the diode output.

The 3 loads. These are cooled non-inductive resistors, chosen so that the total of load + sense resistor is 50 ohms (or so...). The 100W load is liquid-filled, I unscrewed the end cap, saw it was full of liquid, and quickly screwed it back on again. I suspect all contain beryllia but as they are essentially sealed I am not worried.

The front panel. This carries all the controls, fuses, mains connector, calibration pots, meter and a pseudotagboard full of passives. The pseudotagboard is a piece of FR4 or similar with tags on it with resistors soldered between them and wires going off elsewhere. The 'pseudo' part is because there are PCB traces between some of the tags on the underside. There's a socket for the cable from the amplifier on a bracket on the back of the panel and one for the power supply on a flying lead. This part is complicated and painful to work on!

The power supply. Transformer (21V secondary according to the label), 4 diodes, capacitor. Nothing odd about it.

The amplifier. A flat metal can with a cable that plugs into the front panel assembly. Inside, discrete transistors on a PCB. Not looked at that yet.

TonyDuell · 1st Aug 2021, 9:58 am

Still battling with this instrument.... I now think I understand how it should work so I am replacing obviously defective components (a broken switch on the panel, a leaking capacitor in the amplifier, etc) then I'll power it up and see what the real faults are.

One problem... The mains input connector is the sort of thing commonly found on military equipment. This one has 3 pins and a bayonet lock. I do not have the cable-mounted socket to fit it. And I am not sure I like it much anyway as the socket contacts may well not be touch-proof, the metal shell of the socket is going to be a pain to earth, etc. What do people here do about such connectors? I think the small Bulgin one could be fitted without too much work but it would look out-of-character if that matters.

24th May 2021, 4:38 pm	#1
TonyDuell Dekatron Join Date: Jun 2015 Location: Biggin Hill, London, UK. Posts: 5,208	Marconi TF2507 RF Wattmeter The other instrument I've got is a TF2507 wattmeter. It seems to consist of 3 detector units (1W; 3W/10W, 30W/100W ranges) linked to dummy loads. The output of the appropriate detector goes to a module labelled 'DC amplifier' and that drives a moving coil meter. The input impedance, not surprisingly, is 50 Ohms, but there's a '75 ohm converter' clipped to to the top. This may be nothing more than a 25 ohm resistor (there is a switch on the instrument that you flip when you use the converter, I guess to compensate for the power lost in said converter). It's in a large heatsink that appears to be liquid-filled, I assume some kind of transformer oil so if I do open it up, I will do so outside in a suitable container to catch the oil. There is a warning on top of the main instrument that the detectors and loads contain beryllia, so I've not looked at that part of the instrument. I have extracted the DC amplifier and opened it up, it's discrete transistors on a glass fibre PCB. Again I can find nothing about it on google.

24th May 2021, 5:59 pm	#2
Granitehill Tetrode Join Date: Mar 2019 Location: York, North Yorkshire, UK. Posts: 95	Re: Marconi TF2507 RF Wattmeter I don't know the TF2507, but it sounds like a later version of the TF2503 which I have. The 2503 is 30/100W and has a conical shaped thin-film resistor, cooled by fluid inside a large finned heatsink. According to "Marconi Instrumentation" vol 12, the liquid is "A fluorocarbon liquid coolant up to 40 times as efficient as conventional transformer oil". This is pretty well unobtainable these days I think, so best to keep it safe. If it's the same stuff I used for cooling klystrons, it's non-toxic and just looks like water, but is a lot heavier. Pretty expensive also, I think. I think the beryllia is probably in the 3/10W load which was used in the smaller TF2502. In this case the load was a thin film deposited on a beryllia substrate, about the size of a quarter watt resistor. The substrate was set into a small finned heatsink. These were pretty upmarket meters - £493 in 1969 is a lot for an absorption load ! __________________ John G4FDD G-QRP 431

24th May 2021, 6:10 pm	#3
TonyDuell Dekatron Join Date: Jun 2015 Location: Biggin Hill, London, UK. Posts: 5,208	Re: Marconi TF2507 RF Wattmeter There are 3 separate input connectors, one for the 1W range, one for the 3W and 10W ranges and the last for the 30W and 100W ranges. Each connector links to the 'detector', there being 3 of these, different colours for identification. The detectors have 2 RF connectors and a screend cable to the range switch (I think), the last being the DC output representing the power. Of the 2 RF connectors on the detectors, one as I said goes to the front panel input socket. The other to the load. The loads for the 1W detector and 3W/10W detector are small round heatsinks on the back of the chassis. The load for the30W/100W detector is a much larger heatsink, but I am not convinced it is liquid-filled. The 75 ohm converter, a separarte heatsink unit with an RF connector on each side, does contain liquid, I can hear it sloshing about. The warning on the instrument about beryllia seems to imply it's found in all the detectors and loads. I'd want to read a service manual before taking that side of things apart anyway.

25th Jul 2021, 7:37 am	#4
TonyDuell Dekatron Join Date: Jun 2015 Location: Biggin Hill, London, UK. Posts: 5,208	Re: Marconi TF2507 RF Wattmeter I've now got it apart on the bench. It splits into several modules quite easily : The 3 detector units. These are similar in principle to the one in the TF2502, and use thermocouples to sense the temperature rise in a resistor in series with the RF signal. There's also a diode probe which is used for the 'peaking' function I think. The casings come off easily (4 screws), there is a warning about 'beryllia cement' on them. I am not going to dismantle any further unless I have to attempt to repair a fault. Each detector has a coaxial input from the front panel connector, an N plug output to the load and a 3 pin connector to the front panel wiring. The 3 pins being the ends of the thermocouple and the diode output. The 3 loads. These are cooled non-inductive resistors, chosen so that the total of load + sense resistor is 50 ohms (or so...). The 100W load is liquid-filled, I unscrewed the end cap, saw it was full of liquid, and quickly screwed it back on again. I suspect all contain beryllia but as they are essentially sealed I am not worried. The front panel. This carries all the controls, fuses, mains connector, calibration pots, meter and a pseudotagboard full of passives. The pseudotagboard is a piece of FR4 or similar with tags on it with resistors soldered between them and wires going off elsewhere. The 'pseudo' part is because there are PCB traces between some of the tags on the underside. There's a socket for the cable from the amplifier on a bracket on the back of the panel and one for the power supply on a flying lead. This part is complicated and painful to work on! The power supply. Transformer (21V secondary according to the label), 4 diodes, capacitor. Nothing odd about it. The amplifier. A flat metal can with a cable that plugs into the front panel assembly. Inside, discrete transistors on a PCB. Not looked at that yet.

1st Aug 2021, 9:58 am	#5
TonyDuell Dekatron Join Date: Jun 2015 Location: Biggin Hill, London, UK. Posts: 5,208	Re: Marconi TF2507 RF Wattmeter Still battling with this instrument.... I now think I understand how it should work so I am replacing obviously defective components (a broken switch on the panel, a leaking capacitor in the amplifier, etc) then I'll power it up and see what the real faults are. One problem... The mains input connector is the sort of thing commonly found on military equipment. This one has 3 pins and a bayonet lock. I do not have the cable-mounted socket to fit it. And I am not sure I like it much anyway as the socket contacts may well not be touch-proof, the metal shell of the socket is going to be a pain to earth, etc. What do people here do about such connectors? I think the small Bulgin one could be fitted without too much work but it would look out-of-character if that matters.