The TF1020A and TF1152A Wattmeters

  I inherited two of these wattmeters and can't say if they ever worked properly after the last date on their labels, but it wouldn't surprise me if they were drummed out of service when they became unreliable. There are several versions of this useful equipment. Without the "A" is a 75 ohm version. The 1020 is a high power model and the 1152 a low power model. Also worth noting is the fact that there are two versions of the TF1020A. The basic model with /1 can measure 100 Watts and 50 Watts and the higher power meter which is coded /5MI (which is mine) and can measure 150 Watts and 300 Watts. It appears that both types use exactly the same power resistor mounted in a special enclosure which provides excellent matching up into the UHF region. Both types use a similar power sensor which provides a small DC voltage which is proportional to the heating effect of the incoming AC power. In the case of the higher power meter the sensor is shunted by a 22 ohm resistor which bypasses about two thirds of the output voltage, leaving a third driving the metering circuit. The higher power meter has a mains operated blower for cooling the power resistor.

I spotted a model TF1020A/4MI on Ebay which was a 75 ohm 150W/300W.

The first picture shows the complete TF1020A/5MI equipment with base-mounted blower. As my experiments were initially on this high power model, I'll leave the low power model till later..

 

  As you can see there's no shortage of labels, dating from 1964 to 1992. I particularly like the one about the RF connector. They even underlined OMHS.

Does the anchor label mean it was owned by the Royal Navy? I guess not as a rear seal plainly mentions RAF.

 

 The instrument has a large 50 ohm resistor with a working length of about 20cm, tapped at 2cm and wired to a metering circuit using a thermal sensor.

Firstly the parts that drive the meter on the front panel. The sampling connection at the 50 ohm dummy load is a metal band set at about 5 ohms from ground which connects via a 220 ohm resistor in series with a small coil having about 4 turns tuned by a dust core. The coil connects to ground via a 22 ohm resistor and also connects through the glass encapsulated device to ground. This device is a thermal sensor and has two output wires, one of which is connected to ground via a 15 turn high frequency choke and the other via a similar choke to the meter circuit. These will have an inductance of around half a microhenry. Two adjusting pots and a selection switch connect the meter to either the sensor choke or to the sensor choke via an additional 5 ohm wirewound resistor mounted across a small grey coloured component, probably a VDR.

The monitoring circuit is therefore a potential divider feeding the thermal sensor which connects to the meter via two switchable circuits. One circuit connects directly to the thermal sensor and the second to the thermal sensor plus a 5 ohm resistor. The meter is marked 850uA and the two potentiometers for setting the scale accuracy are low value wirewound devices of 25 ohms for the 150W range and 10 ohms for the 300W range. I suppose the coil in the potential divider will increase the impedance of the leg carrying the 220 ohm resistor as the input frequency increases. From the table above, if this coil is 0.5uH it will add an extra 300 ohms at 100MHz, however this ignores the capacity of the circuit and in reality the coil will act as a resonator with stray capacity to maintain the voltage divider accuracy at higher frequencies. From the table above a stray capacity of around 3pF will resonate the coil at something over 100MHz. This coil can be preset as it has an adjustable core. The RF chokes will stop stray RF from getting into the meter circuit.

Interestingly the wattmeter will work at DC and at 50Hz as well as at radio frequencies. If 100 volts DC is applied across the input socket you will draw a current of 2 Amps through the 50 ohm resistor dissipating 200 watts. At the tapping point there will be a voltage of 10 volts.

This voltage feeds the sensor circuit divider whose output voltage is (220+22)/220=1.1 volts. The resistor tolerances are 2% for the 220 ohm and 5% for the 22 ohm making the output 1.09 to 1.103 volts. This is applied to the thermal sensor for which I don't have a spec sheet. However, if the output from the sensor is say 0.1 volt this will feed the 830uA meter via the circuit resistances. If these, including the meter resistance are 125 ohms. the current will be 660uA.

Having worked out the circuit I did some more tests.

Firstly, the two potentiometers for adjusting the meter readings were poor. Although the tracks looked pristine the end readings were bad, intermittently showing a few ohms instead of zero, but by using the resistance range in the centre of travel they were OK. The fact that only an extra 5 ohms is used to double the power reading is interesting. The implication is that the pots will have a fairly coarse effect on the readings (and this is what I discovered later).

Initially I connected an external power supply (actually several small supplies in series) and set the voltage to 100 volts. This drew 2 amps through the 50 ohm load resistor and I expected to see 200 watts indicated, however the indication was only 87 watts. I then set the voltage to read 70.7 volts which should register 100 watts on the meter. The meter showed 50 watts. I then connected a small pot across the 220 ohm divider resistor and adjusted it so the meter read 100 watts with RV2 set half way to avoid the rough end track. This divider modification increases the voltage across the sensor.

Disconnecting the pot revealed it was set to 218 ohms so I fitted a 220 ohm resistor across the existing 220 ohms. Setting the input voltage again to 70.7 volts gave me around 85 watts on the meter, but adjusting RV2 gave me exactly 100 watts. Clearly the difference of only 4 ohms has a significant affect on the readings. Switching to the 300 watt range, I set RV1 to read 100 watts. The instrument should now read RF power correctly on both ranges.

What was the problem? Well without details for the thermal sensor it's difficult to say exactly, but the instrument was reading low. The original 220 ohm divider resistor was indeed 220 ohms so it wasn't that. It's possible the 22 ohm resistor had gone low? It's also possible the 850uA meter is faulty?

I'll try the wattmeter now that it's modified and see if it's stable.

 

 Below.. the thermal sensor looking a bit like a 955 triode

I found this information about the thermocouple

Ultra High Frequency, Vacuum Thermo-couple, Insulated to 100V DC, Heater Current range 25ma, Max overload 37.5ma
Heater resistance 25 Ohms, Couple Resistance 3-4 Ohms, Couple output 16-20ma

Ormandy & Stollery LTD.
3 Victoria Place
Brightlingsea
England

 

 Below is a circuit diagram of the TF1020A found by Fred, VK2FM

 

 View of the load resistor in a shaped aluminium enclosure for matching.

 

 Looking down the load resistor to the tapping point

 

 Load resistor clamped at earthy end showing tap for sensor

 

 Next, I need to test the wattmeter to see if the settings have changed and then to confirm that it responds to RF power in the same way as DC.

I'd also like to see what its frequency response is like to see if it meets its spec of DC to 250MHz. The mechanical design is rather strange and I guess this is to make its response to RF more or less uniform. At the lower end of the load resistor the aluminium screen is very close to the surface of the carbon surface and bracing strips imply the shaping is pretty important.

First power up the wattmeter with 70 volts DC. Read slightly low so twiddled RV2 to read exactly 100watts.

Hooked up the 2 meter transceiver, an FT480 + Microwave Modules linear and saw the Bird Model 43 meter reading 100watts but the TF1020A read only about 35 watts or so. RV2 couldn't reach anything better than this so either the wattmeter is reading RF differently to DC or the Bird wattmeter is very optimistic.

Here are the pictures of the experiment

 DC TESTS

 

 

DC power reading=100 watts...

 

at 71 volts

 Current at 71 volts is 1.45 Amps= 103watts

 

 RF TESTS

 Linear amplifier output 100 watts

 

 At 145MHz

 

 And the TF1020A reads about 35 watts (At least it's better than the 25 watts before I modified the TF1020A)

 

 and reverse power from mismatching about 20 watts.

 

 Could the Bird wattmeter be reading too high as 20 watts of relected power seems too much?

I have two of these so I swapped them over using the same insert which is 100MHz to 250MHz 100Watts

This time I got 90 watts forward and zero watts reverse instead of 100 watts forward and 20 watts reverse (the first Bird Wattmeter proved faulty and worked normally after repair).

The TF1020A read 42 watts. This is getting odder and odder. Admittedly I changed the various interconnecting cables and adaptors because the first Bird uses PL259 and the second N-type. Could the anomalies be partly due to a faulty cable?

 AC Tests

 As there are several possible reasons for the power reading discrepancy including the response of the TF1020A to VHF, I decided to test it with a low frequency. The simplest of course is a test at 50Hz so I found a few low voltage mains transformers and connected these to my variac so that I could easily alter the output voltage. Because the windings are connected in series one needs to get the phasing right. If two windings, say each of 12 volts are connected in series the result could either be 24 volts or zero, so by trial and error with the transformers powered up I arranged the output voltage to be the sum of the windings. I used two 14 volt windings and a 24 volt winding plus a 36 volt winding to produce about 88 volts. One connection was made to the centre pin of the N-connector and the other to the chassis of the wattmeter. Now by adjusting the variac I could set the reading on the wattmeter to conform to powers up to 154 watts.

The first check I made was 100 watts, setting the input voltage to 70.7 volts on an RMS reading voltmeter. Having set the wattmeter to work correctly on DC I expected it to work on 50Hz but alas it did not. The reading was low. Having already learned that the resistor feeding the thermal sensor has a significant effect on dial reading I connected a 5kohm pot across it and carefully reduced its value until I got 100 watts on the scale. Finding the resistance of the pot was around 490 ohms, I substituted for the original 220 ohms plus the 220 ohms dictated by DC tests, a new resistor combination of two 180 ohms in parallel=90 ohms. RV2 needed tweaking slightly to adjust the dial reading to exactly 100 watts with 70.7 volts AC input because 90 ohms wasn't the precise value required..

Summarising: the original resistor was 220 ohms, the resistor for accurate DC was 110 ohms, the resistor for accurate 50Hz was 90 ohms.

Because there's a 22 ohm resistor across the sensor the potential divider outputs for the resistor combinations are 0.091; 0.166 and 0.196 or expressed as percentage power available at the load resistor tap, 9.1%, 16.6% and 19.6%. This means that the power sensor now needs twice the power input to give the same power reading since its last pass date.

Now, I needed to know if the linearity of the wattmeter scale was OK. Setting the variac to supply 40 volts resulted in a scale reading of 32 watts, 50 volts gave me 50 watts, 60 volts gave me 71 watts and 80 volts gave me 128 watts. These are all fine so the sensor is supplying good outputs at all the power inputs tried.

Next, I switched to the 300 watt range, and after adjusting RV1 all power readings were consistent.

Next, I connected 100 watts of 145MHz to the N-connector hoping to see a scale reading of 100 watts. Instead it read 45 watts. Now this isn't as bad as it sounds if expressed in dB. If the wattmeter is specified at around 120MHz or so at -3dB then 45 watts is about -3.5dB which would be about right, however I understood the wattmeter was good to 250MHz not 120MHz. I decided to measure the SWR of the wattmeter as this should give an indication of it spec falling off at higher frequencies.

 New resistors. The 22 ohm resistor shunting the sensor is underneath

 

 SWR Measurements

 To make SWR measurements you need to measure the forward and reflected power into the 50 ohm dummy load within the wattmeter. To do this I used a Bird Model 43 wattmeter. I have two of these, and one had been giving odd results so I fitted the plug-in insert to the good one and got 100 watts from my linear at 145 MHz with only a tiny reflected power. Before the reflected power had been too high to make sense. To be more certain of the results I used a Welz dummy load specified at 150MHz to compare with results from the TF1020A.

144MHz: Welz 1.43, TF1020A 1.15

145MHz: Welz 1.40, TF1020A 1.16

146MHz: Welz 1.38, TG1020A 1.17

These results are exactly as one would expect if the TF1020A is indeed specified better than the Welz.

Could it be possible then that the TF1020A is an excellent dummy load for VHF, but as an accurate wattmeter only accurate to say 100MHz and beyond this only good for comparative measurements such as tuning up a transmitter?

Next, I'll need to carry out tests at HF and sort out a couple of anomalies, one of which appears to be an intermittent toggle switch. This is the switch selecting 150 or 300 watts which sometimes does not cleanly operate because it's wetting current is less than half a milliamp..

Both potential problems were resolved. The switch was fine and the 850uA meter read 860uA full scale deflection.

I then hooked up a TRIO TS120S transceiver via a Heathkit HF SWR/Power meter to the TF1020A. Switched to CW and adjusting the carrier control enabled me to see exactly 100 watts on the Heathkit meter and an identical reading on the TF1020A, initially on 3.5MHz then on 7MHz, 14MHz, 21MHz and 29MHz. So the TF1020a appears to be working OK at HF. There's a small tunable coil in series with the TF1020A sensor potential divider and I guess this can be used to peak the drive to the sensor at the point where its sensitivity starts to drop off, but from the measurements I made this must be around 100MHz?

The tests were useful as I now know the limitations of the TF1020A and its a jolly good dummy load. Mounted on the underside of the wattmeter is a mains driven cooling fan so if I need to carry out extended testing this can be used to keep the wattmeter cool.

 View showing blower

 

 and underside

 

 Before ending here's a picture of my TF1152.

 

  As you can see this is a TF1152A/1 and can measure power in two ranges viz. up to 10 watts and up to 25watts.

Later I'll open it up and inspect the interior then carry out some tests to determine it's accuracy

Return to Reception