T1154 Refurbishment

 The condition of this example is not pristine, but it's already been tastefully cleaned up by its last radio ham owner. I believe most if not all the front panel hiding the valves is a home-made replacement. The plastic dial settings label looks the part and has 4 frequency bands rather than the three for the standard Model N. As the wavechange switch also has 4 positions I guess this must be the way it was made and is really a different type fitted with the wrong identification plate?

Versions A, B, J, N = 3 wavebands; 200/500KHz, 3.0/5.5 MHz, 5.5/10 MHz

Versions C, F, H, K, M = 4 wavebands; 200/500KHz, 2.5/4.5 MHz,4.5/8.7 MHz, 8.7/16.7 MHz

Versions D & E = 3 wavebands; 200 - 500KHz, 2.5 - 4.5 MHz, 4.5 - 8 MHz

Version L = 3 wavebands... 200 - 500KHz, 1.5 - 3.0MHz, 3.0 – 5.5 MHz

The T1154 wthout an identifying letter was similar to the A version.

See the table below...

 200-500KHz

 A

 B

 C

 D

 E

 F

 H

 J

 K

 L

 M

 N

 1.5-3.0MHz

-

-

-

-

-

-

-

-

-

 L

-

-

 2.5-4.5MHz

-

-

 C

 D

 E

 F

 H

 -

 K

-

 M

 -

 3.0-5.5MHz

 A

 B

-

-

-

-

 -

 J

-

 L

-

 N

 4.5-8.0MHz

-

-

 -

 D

 E

-

-

-

-

-

-

-

 4.5-8.7MHz

-

-

 C

-

-

 F

 H

-

 K

-

 M

-

 5.5-10.0MHz

 A

 B

-

-

-

-

-

 J

-

-

-

 N

 8.7-16.7MHz

-

-

 C

-

-

 F

 H

-

 K

-

 M

-

 Mobile ground use with R1188

-

-

-

 D

 E

-

-

-

-

-

-

-

 Aluminium Al, Steel Fe

 Al

 Al

 Al

 Al

 Al

 Al

 Al

 Fe

 Fe

 Fe

 Fe

 Fe

 High speed launches + trainers

-

-

-

-

-

-

-

-

-

 L

-

-

 Click stops (multi=M, single=S)

 M

 M

 M

 M

 M

 M

 S

 M

 M

 S

 S

 S

 

 

 

 
 

 So here's the puzzle, presumably the previous owner fitted the wrong indentification plate because changing the wavebands would have been a bit tricky to say the least. See my NOS real "N" version S/No 73350.

The steel chassis with these ranges and lots of click-stops makes this a "K" version does it not? In fact it's said this model was replaced in service by the "M" version having only a single click-stop.
 

 Everything that should be present on the chassis seems to be there and continuity checks indicate the various chokes are OK. I usually check all the resistors first. It's hardly likely that the large green ceramic ones are anything but perfect but measured as below.

 RESISTOR

 CODE

 VALUE

 MEASURED IN SITU

 R6

 10W/1045

 75Kohm

 50Kohm

 R5

 10W/1045

 75Kohm

 49Kohm

 R9

 10W/1047

 5Kohm

 5Kohm

 R10

 10W/1048

 350ohm

 334ohm

 R4

10W/9820 

 20Kohm

 20Kohm

 R3

 10W/1043

 50Kohm

 38Kohm

 R7

 10W/1046

 14Kohm with tap at 2Kohm

 69Kohm with tap at 2Kohm

 Because of parallel circuit parts resistors often read low, but they shouldn't read high unless there's a large capacitor affecting the ohmeter (and if you have the patience the reading will drop to the correct value). Clearly from the above results R7 is suspect so I'll need to examine the circuit diagram and remeasure it. Probably some oxidation on the resistor connection... drat.. not oxidation... the 12Kohm resistor is open circuit.

How can this have happened? Well, it comes as no suprise it's a design weakness. The resistor would normally be fitted with the tap sticking out towards the rear to aid connection of the wire going to S5. This wire is used to arrange for the CW position to increase the power to the PT15s. The user documentation warns against the problem of the tap shorting to the outer case so presumably this is what happened?

There's a slight mystery here though, because there's no trace of the wire from S5. As it can't have just disappeared someone has either hidden it away or cut it off from S5. What's the effect of shorting the tap to ground I wonder? Well, with the HT at 1200 volts the resistors R4 and R7 would draw around 38mA. R4 would dissipate about 28 watts and R7 about 17 watts. Surely, neither of these dissipations would result in damage to the resistors?

What would happen if R7 did go open circuit? The screen grid voltage of the pair of PT15s would be higher than planned. Exactly what it would be is complicated but it would certainly be higher than the designer planned, being essentially dependent on the screen grid current at the particular operating point of the valves. Looking at the PT15 spec: if the supply voltage for G2 attempted to place 800 volts (which is more than double its 300 volt rating) on this electrode it would draw excessive current (say 2 x 25mA for the pair of valves) and would result in a voltage at G2 of 200 volts or an input of 5 watts. So R4 will provide a measure of safety and hopefully limit G2 to less than its maximum rated dissipation. However, until I test the valves their quality must remain a worry. 

 

Above, on the RHS is resistor pair R7/R8. The tap should be connected to the nearest wafer of the yaxley switch bottom left, but has been cut off.

 Next, I need to solve the mystery of the missing lead.... I checked the circuit diagram and looked at the mode switch. The switch tag in the CW position had been soldered but the wire was missing. A neat cut indicated sidecutters. Is this example suffering from damaged PA valves and did the last user try and hide the evidence? What's the option if the PT15s are beyond hope? My collection of valves list some 814s, one of which (with a UX5 socket) might provide a suitable replacement for the two PT15s? Let's hope it doesn't come to that!! Then I had another thought... the seller had removed the case to show pictures of the inside so I wonder if he cut off the flying lead which maybe had not been connected to R7/R8?

I bit the bullet, wheeled out my AVO valve tester and removed the four valves from the T1154. Surprisingly, both the PT15 and ML6 valves are included in the AVO manual. For some reason the PT15 mA/V setting was blank. I set the tester to the appropriate voltages and switch setting and tried the first PT15. Pressing the Test button showed no movement on the meter... oops.. my fault as the top cap lead was plugged into the G1 hole. Setting this to "A" produced excellent results with the mA/V dial set to about 3.5. The second PT15 produced exactly the same results... well into the green area of the AVO meter. Next, the pair of ML6 valves. After readusting the settings both valves produced good green area results.

After searching my junk boxes I failed to come up with a 12K + 2K ceramic resistor. That's not to say I haven't got one because my workshop is a bit chaotic. Looking at RS products I quickly found some chassis mounted resistors that will be OK. These are Arcol aluminium-housed parts, size HS50 (50 watt rating) these have a voltage rating of 1250 volts and insulated to 2500 volts RMS so can be bolted down to the T1154 chassis. I've selected 10K+1.8K+2.2K which will mount in a space about 8 inches long. I could either fit these to an aluminium plate suitably secured to the chassis or directly to the chassis itself, depending on cosmetic considerations and undo-ability if the correct part ever appears.

 Looking ahead... advice mentions the block condenser (C21) in the corner of the chassis. This is secured by three 4BA screws so maybe I'll whip it out, check it's capacity and stick a high voltage across it, although, second thoughts, I'll replace it because C21 (and C5) connect to the control grid of the oscillator valve so need to be in first class condition otherwise a DC leak would mess up the bias of V1 via R11 and via R12 the bias of the two PA valves. Drilling out the contents of the block would seem to be the ideal solution then fitting modern high quality parts.

The date of June 1944 is presumably just before this transmitter was made?

 When it comes to condensers I have very few problems because, many years ago I spotted on Ebay a stock of surplus capacitors. I won these, I think for around £5, and when they arrived I was amazed to see the entire stock of capacitors from Greenweld, a local electronic parts supplier in Southampton. Their total weight was around 100 pounds and consisted of every type of capacitor under the sun, including loads of NOS NATO codified parts in sealed bags, with some varieties dating back to WW2 vintage. There are so many that I've never had time to properly sort them out. I remember visiting Greenweld many years ago to buy some bits and pieces. I was aware at the time the difficulty the lad had finding the parts on my shopping list. Alas, no doubt their stock was too disorganised and so abstruse for the company to survive.

The block condenser C21 also houses C8. The manual says that C8 is 0.5uF for the N type transmitter (as opposed to other types which have C8 as 0.25uF), the same value as C21. This is a little odd because both condensers (C8 being 0.25uF and C8 being 0.5uF) are coded 10C/2053. I'll stuff the old block with a pair of new capacitors, whatever are the closest standard values. C26 is also mentioned elsewhere as being suspect. I'll check this and see what it looks like.

Not just now though as I've just counted the jobs in for repair and there's 15 plus a bunch of non-urgent ones... That means 15 local lifts out of action... Let them use the stairs. Later that day, as I'd completed three jobs and ordered parts for three more I decided I deserved some time of my own so upended the T1154 and removed the 10C/2053 block. It's clearly marked 0.5uF nearest the labelling and 0.25uF at the bottom pair of tags. The 0.5uF condenser measured 0.28 ohms, virtually a dead short and the other condenser measured 0.32uF. I fitted a set of 0.22uF high voltage capacitors giving me 0.44uF and 0.22uF then bent back the flap I'd cut on the underside and refitted the block.

I noticed that the Jones plugs use pins very similar to standard 6.3mm fastons. As the pins are slightly thicker I bent back the prongs on the faston sockets slightly and these fit OK. I have a large number of brass faston receptacles and I've just ordered some insulating rubber boots for these at 3p each. The carbon mic sockets take standard 4mm banana plugs so I'll make up a short lead with a pair of these, together with a free jack socket into which I'll plug Chris' carbon mic which is fitted with a quarter inch jack plug.

I did a little more work on the T1154 the next day. I was delayed starting because I had an early delivery of parts followed by a lift engineer who dropped by with some circuit boards for repair. I think they're actually for air conditioners and when I checked there were 27 in all, including 19 of one type. I tested one of these and it seemed OK, as were half a dozen more so I put them on one side, reassembled a couple of boards received last week with new parts and then turned my attention to the T1154. The exercise today is to replace the 12K +2K power resistor which is open circuit.

I think any of the green high power resistors could be swapped for the types below. To improve power and voltage handling you may need to select two or three to be wired in series.

I made some calculations on heat dissipation and selected 50 watt chassis mounting types which meet all the design criteria (see below). These arrived today from RS with some lift parts. My three favourite suppliers are Farnell, RS and CPC. There's been lots of competition between these three over the past year. Free postage is currently the theme but RS have now imposed lots of minimum order quantities to claw back postage costs so I shift between them and Farnell who sell most stuff singly.

Firstly, I identified a space in which to mount a suitable resistor heatsink. I then made it from a piece of scrap aluminium and fixed the three resistors. You can see that I had to choose standard modern values rather than the early wartime values but I don't expect much difference in performance in practice. Resistors are mounted (with 6BA screws) using a special heat sink grease. This is recommended in the resistor spec.

 

Adjacent to the original resistor mounting clips is an ideal position for the new substitute. There are three countersunk 6BA screws under the top lip of the chassis which I removed and used to hold the new heatsink. If I decide to fit an original resistor I can readily remove the new parts.

To improve clearance I bent the clips slightly inwards (the lower clip sits at 1200 volts).

 

This picture shows the new resistor fitted and wired up. The tap wire had been missing. A word on voltages.. The HT is circa 1200 volts so this figure must be bourne in mind when making changes or fitting new parts. The new resistors have a rating from element to case of 2500 volts RMS and a resistor rating of 1250 volts so should be OK. I used 1000volt connecting wire sheathed in an extra grey cotton cover. The top pink wire is at ground potential. Note: since carrying out this resistor substitution I've got a replacement tapped resistor which I'll fit.

 

Then preliminary testing.....

 I connected the tidied-up power supply to the T1154 and watched the filaments come on in the PT15s. Not very bright because the voltage at the front terminals was only a little over 5 volts. The internal transmit/receive relay had clicked over (supposedly to Receive) and the main mode switch seemed to do the right things. Applying the HT and connecting a wire across the key connections seemed to do nothing in particular, except that the transmit/receive relay clunked to transmit.

I checked the HT and it was present at about 600 volts at the PT15 anodes, but no movement at the left hand meter. My PSU showed about 30mA HT current, but this could be consumed by the oscillator and modulator valves.Odd, looking at the circuit diagram, below, I'd have expected some PT15 HT current to flow? I'll check a few things. Having seen 600 volts at the anode of the PT15s maybe their screen voltage is missing? Maybe their control grids are too negative, cutting off the anode current?

There's a few other possibilities as well. I'd found the HT fuse was blown. Not just a break from metal fatigue, but a break from what must have been a short-circuit. Maybe the anode current meter is open-circuit? I'll have to mighty careful. I normally test circuit boards using transistors operating at no more than 24 volts or so. Occasionally I repair drive units using 3-phase mains and with 800 volts DC powering IGBTs so I tend to be careful when handling these and the same rules apply with a powered T1154 chassis.

This morning I measured the resistance of a few things. The ammeter in the anode feed twitched when I checked it on the diode test range of my multi-meter so that's probably OK. Next I checked the HT fuse holder which was OK. Then between the PT15 anode and the top end of the HFC which seemed OK then from the HFC to the ammeter and saw about 50 ohms which isn't quite right but not too bad. Then from the HT fuse to the HFC which was 26Kohms. That's odd because the circuit shows R16/R17/R18 being 10 ohms and 2 x 19.5 ohms. Not so fast though because AP2548A says that a mod to the fuse circuit was advised, moving the fuse to the HT negative feed. Was this done in my example? Sure enough this equipment has been modified by placing the fuse in the HT earthy feed, instead of the original live feed. I then checked the PT15 screen connections and found this goes via R4 to the G2 pins of both valves.

 

 Before Christmas (2015) I'd decided to replace the 6 volt regulator in my home-brew power supply with one that would run cooler. The series pass design using an MJE2955 needed a fan to cool it down. I also decided to add a 6 volt lead-acid battery located close to the T1154. This would provide a better regulation for the 6 volt heater supply and reduce the demand on the new regulator. As the new PSU parts weren't due for a week or so I added the battery and tried the T1154 once more. The PT15 filaments were brighter but the mag feed meter (M1) was still stuck at zero.

I found a high wattage 3kohm resistor and touched it to the earthy side of the milliammeter. It went from zero to 210mA proving the meter was OK. The PSU HT current went from 25mA to 75mA. That's odd. The mag feed meter must be calibrated wrongly?

I then touched the resistor on the anode of one of the PT15 valves. The mag feed meter showed 80mA. Very odd? Looking at the circuit diagram above shows two HT feeds to the PT15 valves via R16 plus either R17 or R18. Presumably the meter is calibrated to take into account the two feeds? For example, if the mag feed meter has a resistance of 10 ohms and R16 is 10 ohms, then shunting the meter with 20 ohms will half its reading because the combined current flow through R16 and either R17 or R18 will divide equally. A reading on the mag feed meter of 300mA is essentially only 150mA through the meter and 150mA through either R17 or R18.

However, shorting the earthy side of the mag feed meter to ground will include a shunt of both R16 and R17. A meter of 10 ohms will now be shunted by 20 ohms so will read low by a third.

Turning to the PT15 anodes. These are fed via HFC1 which will have a DC resistance, thus shorting an anode to ground will reduce the short-circuit current from that seen when the meter is shorted to ground.

Switching the T1154 to transmit in any setting of the mode switch reveals zero HT current on M1 so there's a problem somewhere, maybe in the biasing of the PT15s or their G2 voltage. G2 is derived from a potentiometer comprising R4 (20Kohm) with R7 (12Kohm) and R8 (2Kohm) to ground.

One possibility is G2 decoupler, C13 (0.01uF) is short-circuit. This would place close to zero volts on G2 and R4 would draw 30mA from say 600 volts. This would neatly explain the PSU standing current of circa 50mA. Well, I checked G2 and it was 213 volts with the anode at 528 volts, however G1 read minus 142 volts, hence the PT15s are cut off.... Where's this negative voltage coming from? Well, for starters HT minus isn't grounded. The PSU feed meter indicates about 50mA and the HT minus feed goes via R10 (350 ohms) and R9 (5Kohm) to ground. This will produce a bias value of about 150 volts. This lines up with what's stated in the EMER.

Looking further at the EMER I found a TX/RX relay test was advised if the transmitter failed to work. All the contacts were blackened so these needed cleaning was the first thing I found. A contact resistance test showed that nothing made a low resistance contact in either the RX or TX setting. Next a coil check. This produced odd results (all bar one coil was close to zero ohms) so I need to retest the coils, although I see the valve filaments are across Coil A so that would read zero (instead of 12 ohms) unless I unplugged them. Coil C seemed OK at about 2 ohms but Coil B showed zero ohms which looks wrong.

Ater cleaning the contacts and re-reading the EMER I discovered that the high bias for switching off the oscillator and PA anode current was achieved by passing HT negative current through R9 which is bypassed by the transmit relay, contact set 2. However it only does this if Plug B, pin 14 is grounded. Presumably this is done by whatever conects to plug B.

Once I'd grounded Pin 14 and keyed Pin 13 to ground the PT15 valves drew plenty of current. I switched on a monitor receiver and tuned this to 3.6MHz. I was then able to tune the T1154 to 3.6MHz. Connecting a 100 ohm dummy load (a 7 watt 100 ohm resistor was conveniently lying on the bench) and turning on my spectrum analyser showed plenty of output. Twiddling the output tuning condenser and fiddling with the tapping switch produced some interesting results. Without fine tuning the condenser harmonics were very obvious but a fine tune resulted in mostly the fundamental getting larger and the harmonics almost disappearing. I took several pictures of the traces and when I'd finished the dummy load resistor was quite warm so I reckon I was getting at least 5 watts output. The pictures below show measurements made using my high impedance probe so the actual RF readings are not significant.

Below: The first view of the RF output of the T1154 into a 100 ohm dummy load, tuned to RED BAND 80 metres and with the PA untuned

 

 Next below: As above but tuned to BLUE BAND Range 1

 Below: Now with the VFO tuned to 40 metres

 And finally, below as the previous picture but with the PA tuning adjusted using the tapping switch and the output tuning control.

The second harmonic is now 50dB lower than the fundamental signal.

 The next day I tried some more experiments. Switching to the lowest (MF) range I found my oscilloscope display showed an awfully distorted RF output signal. Basically a sine wave with flat tops so my guess is the HT supply is too low. Although I'm providing 600 volts input to the T1154 there are large resistors used to develop the HT from this for the VFO and modulator. To provide the correct HT for these parts of the circuit I really need to increase the HT voltage to the design value.

The two HF ranges were OK, giving decent sine waves. My scope said the 80m signal was 3.57MHz at 23VRMS. As my dummy load is 100 ohms the output will be about 5 watts RMS.

 

 

I noticed a mod record plate screwed to the inner wall of the PA compartment. This gave the number "10D1587 S/No 8099(8)". This explains the puzzle over the plate on the front panel. As you can see it says "10D1588" which is an "N" version having two HF ranges. "10D1587" is an "M" type with the correct three HF ranges. Clearly a previous owner fitted the wrong identification plate.

 The LT regulator kits arrived from China today. These have all the parts needed to construct on its tiny circuit board a regulator based on the LT1083CP. I built and tested it then connected it in place of the home made regulator using an MJE2955. I was able to set the output to 6.3 volts and run 6 amps with only about 0.1 volt drop at its output connector.

I then made a small modification to the HT power circuitry increasing the maximum stabilised output from 625 to 700 volts. The next step is to swap the feed to the rectifiers from the 550 volt to the 620 volt tappings. Currently I'm getting 725 volts into the regulator circuit (a drop of 52 volts from the peak AC value). Stepping up to the 620 volt tapping should give me around 820 volts into the regulator and a maximum stabilised output of about 800 volts. Currently I'm using a zener chain in my power supply of 340 volts so stepping up the maximum voltage needs another 100 or so volts. The zener chain is used to set the adjustable output voltage range. There will be 11 steps of 33 volts plus the increased 450 volts in the zener chain will give me a switchable output from 800 in steps of 33 volts down to around 470 volts.

The next step will be connecting everything up and seeing if the T1154 will transmit successfully on the ham bands....

Next step... commissioning the T1154

Return to Entrance