Paillard 768

[Richardgr]

Still spending time on this old thing.

There was a loose connection by the last smoothing capacitor - 2 components were soldered together and floating above another conection point on the capacitor. There are quite a lot of these connections floating in the air, so was not obvious if this was how it should have been. I tried to match the parts to the hand drawn Radiomuseum schematic, and I just could not correlate the two views.

I have to be able to have a schematic, otherwise in my mind I am just plugging in and praying. I think it is important to take voltage readings, and have some understanding how it was designed. The one off Radiomuseum was extraordinarily difficult to navigate, plus the components themselves are not familiar to me, with their odd shapes and colour codes, so it has not been easy to get to grips with the circuit (which is not complicated!).

I have had a few attempts at drawing something. This first part is all the input stages, and I have simolified by ignoring switches. While drawing this I noticed that there is no power to the plates of the 58 tubes on the original schematic. The place where I have the loose connection is where a resistor, missing on the original schematic, R8, links the B+ to the coils to the 58 plates. R8 measures 1.7K but has a red body, black tip and yellow spot = '2' + '0' + '0000' or 200k? I measured it in situ.

The output stage is simpler. Everything is identified and linked to the new or Radiomuseum schematic, so I am ready to power up with the rectifier.

The schematic is fairly similar to other radios of the period, but I just do not understad what VR1 is doing. What is the point of varying the potentials of the cathodes of the two 58's?

[Herald1360]

There doesn't seem to be any AGC on the set so VR1 will be acting as a manual RF gain control- more resistance will bias back the variable mu 58s and reduce their gain. AGC would do similar by applying negative dc volts derived from the rectified carrier to the grids of the 58s. It's an odd circuit for a late thirties set!

[ms660]

Are you sure the 200k isn't the detector valves anode load (2A6)?

What's the resistance of the tuning indicator?

VR1 looks to be a RF/IF gain control, is it ganged with the gram switch?

Lawrence.

[Richardgr]

Yes, VR1 is also the gram switch, so I think it is working like a mixer. The other end of the pot has a wire disappearing into a coil, which isn't in the schematic.

I measured the resistor that is paralel to the tuning indicstor out of circuit and it was 200k, or therabouts.

I fired it up tonight. I'm starting to suspect there is an issue with the power transformer. First time I powered up I had the dim bulb, which started to shine dimly around 200V, then got brighter up to 235V, then there was suddenly some fizzing sounds from the transformer, and the bulb flickered brightly. I presume this is not a good sign ;-/

The fizzing from the transformer is not consistent or repeatable, so I decided to give it a whirl with the 80 rectifier in place and the field coil of the loudspeaker in circuit. It didn't seem that the rectifier got up to temperature, but I still had the bulb in the circuit.

Next up I decided to try it with the full complement of bulbs. I had hoped that there would be some noise from the speaker, even if it was a hum. I could just manage to get a very quiet 'rustle' from the speaker when touching the anode of the 2A6 driver tube with a test lead.

So back to the drawing board.

Firstly, how do I get confidence in the transformer, or diagnose if there is a short and where? I don't want to take the bulb out of the circuit until I am sure the transformer os serviceable.

Secondly, I just cannot work out how the 2A3 is biased. The field coil has a resistance of 1250R, and drops maye 10 volts, but there is no real power being consumed by anything. I should be measuring -45V on the grid, but I don't see that.

As luck would have it, I have stumbled across a second chassis. This one is a 70TR, and is just a little bit later than the 768. The rectifier and output tube are the same, but this one has 6v tubes (6k7 x 2, 6L7, 6Q7). It is incredible how much more modern the 70TR looks, even though it was only about 1 year later. It has a familiar switch assembly for changing channels, but it has an odd see saw arrangent between two coils, where a knob lifts the core of one coil, while lowering a core in another coil.

I think I will power that one up with no tubes to assess the transformer, and then decide what to do next.

[ms660]

Originally Posted by Herald1360

There doesn't seem to be any AGC on the set so VR1 will be acting as a manual RF gain control- more resistance will bias back the variable mu 58s and reduce their gain. AGC would do similar by applying negative dc volts derived from the rectified carrier to the grids of the 58s. It's an odd circuit for a late thirties set!

AGC is taken from the detector output in the original schematic and in the OP's schematic, 1meg>0.05uF>>IF>>200k>0.05uF>>RF.

Lawrence.

[ms660]

VR1 is a gain control for the RF and IF amplifiers, the cathodes of those valves are connected to the bottom of the RF and IF valves g2 potential divider (30k & 25k) via the 400 Ohm resistor, that ensures maximum cathode bias at the minimum gain setting which is what's needed when using a gram pickup.

Lawrence.

[Richardgr]

Hi Lawrence, thanks for your assistance!

I can see what is wrong with the bias for the 2A3, I think. It is a sort of back-bias arrangement, and I should have a polarised capacitor there. Currently I have a 1uF MKP.

I have a variac controlled DC supply up to 400V, so I think I can test without being dependent on the power transformer. The hard part is having so many uncertainties! It could have been dropped, the tubes could be worn out, the transformer seems to make a random fizzing noise, and the speaker is quiet. Need to eliminate some things.

One thing I noticed when testing the transformer is that the heater voltages will not get up to 2.5V, even when I gave the PT more than 235V (I can go up to 250 on the variac). Plus none of the tubes were more than vaguely warm when I tried with them in place.

[Herald1360]

Originally Posted by ms660

AGC is taken from the detector output in the original schematic and in the OP's schematic, 1meg>0.05uF>>IF>>200k>0.05uF>>RF.

Lawrence.

So it is! At least I wasn't wrong about VR1....

Originally Posted by Richardgr

Hi Lawrence, thanks for your assistance!

I can see what is wrong with the bias for the 2A3, I think. It is a sort of back-bias arrangement, and I should have a polarised capacitor there. Currently I have a 1uF MKP.

It matters not here whether the 1uF is polarised or not. The MKP is fine and could be used either way round, a polarised capacitor would have to be the right way round!

[Richardgr]

Originally Posted by ms660

Are you sure the 200k isn't the detector valves anode load (2A6)?

The 200k for the 2A6 is ringed. The other resistor, which is brown with a red spot and a red tip ('1' + '2' + '00', 1.2k?) was floating above the point it is soldered now. Without it the 58's would have no anode load, as far as I can see.

I am not the first interloper in this radio, so it is good to have the other chassis as a reference.

[ms660]

Yes, the bias for the 2A3 is developed across the loudspeakers field coil winding as per Post#7, the 1uF capacitor across the bias supply doesn't have to be a polarized type (electrolytic) it's outer foil should be connected to the ground line (chassis) in order to minimize stray hum pickup.

If measuring the heater voltage remember that the heater winding is center tapped to ground which means the heater voltage measurements should be done between the heater tags on the valve holder and not with respect to ground.

If the secondary winding voltages are notably low with full mains applied and the secondary's disconnected then the transformer is faulty.

EDIT: Post crossed.

Lawrence.

[Richardgr]

Here are the final bits of the schematic, and some pictures of the similar chassis I picked up. The case of the other chassis is fairly ropey, and it is quite large for the size of the chassis. It is incredible how much more modern theother one appears, even though there is only about 2 years between them. Those metal tubes are not much bigger than noval tubes.

The loudspeaker looked identical, but turned out to have a resistance of 2500R, whereas the other one is 1300R. However on the positive side I do have 5 matching knobs!

[Richardgr]

Originally Posted by ms660

Yes, the bias for the 2A3 is developed across the loudspeakers field coil winding as per Post#7, ...

If the secondary winding voltages are notably low with full mains applied and the secondary's disconnected then the transformer is faulty.

I noted your previous pointer to the mechanism for bias, but my lack of understanding is the electronics part. I don't have a feel for what B+ will be, or how many volts are dropped across the loudspeaker coil. Do I treat it as an inductor (reactance) or a resistor? It is 1.3k, and I suppose there could be 100ma across it? Say HT before LS coil 380v, so 130v drop and 250v after?

Bias should be -43v. I get the potential divider of the 2 x 100K, then there is another potential divider of 100k and 500k, with capacitor providing a reference to ground, so 130v divide 2, then 5/6ths of that as bias = 50v?

Then I presume only around 100V will be dropped across the LS coil, to get bias in the ballpark?


I have totally isolated the transformer, and used my Fluke meter on the megaohms setting, and I can not detect any resistance from the windings to each other or to ground. All resistances look reasonable, and there is no smell or heat with the transformer when powered up, so I will assume it is probably OK. Maybe there was a loose connection in the voltage selecter when it crackled before. I'll remain observant!

How do people feel about transformers that are getting on for 90 years? Do they degrade with reasonable use?

I noticed there is a lead that comes from inside the windings to ground. It is not the CT for the HT winding; this other one goes to ground, after the smoothing caps. I presume it must be some sort of ground for the transformer? The transformer is isolated form the chassis on rubber grommets, so that points to some care in the build quality and safety of the radio.


Finally, I removed this little blighter. What is it? The colour code is brown body, black tip, and red and yellow dots on the body, and it measures as 0.5uF or 28Mohm? It is R17 on my PSU schematic, and should be 100K. That would be a reasonable explanation for the lack of bias, but it would be interesting to know what it is meant to be.


Finally finally ;-) Does anyone have a simple ESR tester they can recommend? My conscience has been pricked after reading the thread on replacing capacitors, so I ought to test what I have.

[ms660]

So far as the bias is concerned for the output valve there's only one potential divider, the 2 x 100k resistors across the field coil, the valve data says -45 volts for typical Class A operation, that would mean a 90 volt drop across the field coil which would mean a load current of approx. 69mA.

The ground wire from the transformer might be from an internal screen in the transformer, if it is there should be no resistance to speak of between that wire and any of the other windings with it disconnected from ground.

So far as the component in the last photo goes, I've not come across that type of colour coding before but if it's in the circuit position you say it is then I would go for 100k.

Lawrence.

[Richardgr]

Thanks again, Lawrence.

On the other chassis I have I have 3 identical 100K resistors next to each other. This is what was on this one. Maybe they had to use up some odds and ends?

Some more questions:
1) What are C10 and X6 doing, the cap and the pot to ground from the grid of the 2A3. Is that a tone control?
(On the other chassis the connection of the pot wiper is to the B+ side of the OPT).

2) The third 100K resistor and the 1uF capacitor, R15 and C16 on my drawing, are they just decoupling/smoothing?

[ms660]

Yes, C10 and X6 form the tone control circuit.

Yes, R15 and C16 form the de-coupling circuit for the bias voltage.

Lawrence.

[Richardgr]

Some positive progress. I ended up replacing all the capacitors, and tidied up some of the wiring. There were too many connections in mid air for my liking, so I added a 4 tag board, collected some of the connections there, and simplified the wiring a bit as a consequence. Found a couple of errors in my schema along the way.

Was keen to power up, but caught a couple of howlers I had introduced. It always surprises me how sure I am that I have been thorough, there is always room for error. Some sort of colour convention in the wiring would help, so that is a lesson learnt.

I fired up with the bulb tester, rectifier out. Transformer seemed OK.
Rectifier in, and voltages looked promising, but the bias for the 2A3 is only set when it is conducting.
Then all tubes in with the bulb tester, and there was no smoke, but not a lot else albeit a small bit of hum on the speaker. Bias looked healthy, though, so the precious 2A3 should be safe.
So in for a penny, in for a pound, and all tubes, no bulb tester, and it all looked pretty reasonable. B+ is around 260Vdc, and 80Vdc are dropped over the voice coil, so bias around -40V. That is OK.

I got the hang of the control that varies the cathode voltage of the first 58, and had a healthy bit of static on the speaker. I rigged up 2 metres of hook up wire from the aerial, and managed to pull in some French station on one of the bands, and a bit of Just The Way You Are by Billy Joel. There was one more faint station, but I don't know how many stations to expect these days.

[When I was 11 I bought an HMV 5201 from a jumble sale, and rigged up an aerial to the nearest lamp post from my house, and could spend a whole day just traversing each of the wave bands; hundreds if not thousands of active broadcasters. Unfortunately I nearly garotted my dad mowing the lawn when my aerial drooped after a hot day, so had to make do with the loft after that].

I got adventourous at this point, and rigged up my bluetooth receiver to the pick up terminal, and had some music from YouTube at pretty healthy volume!

The question now is what to shoot for. The tone control is just plain nasty. The sound is very treble-y, nothing much in the bass department. Why is that? The speaker is substantial - what is it capable of?

Interested to hear what potential I should expect from a radio like this when it is set up, with the right voltages, etc. I will focus on the output stage for starters.

Also, after being on for a while I did think I saw a whisp of smoke from the side of the transformer. Am I kidding myself that the transformer is OK? The voltages were in the ballpark, but probably a tad lower than they ought to have been.

[Richardgr]

Just been listening to some field coil loudspeaker videos on YouTube. Some of them sound pretty good, so now I am intrigued what the potential could be. This set, with its 2A3 output tube, would have been designed for a bigger space.

I will experiment with the 2A6/2A3 output stage.

The signal strength indicator did not seem to be working, but it could just be down to bulbs. It measured 377R and has a 1.7K in parallel, that looks original but was not on the schematic.

The updated schematic is attached. There was an extra resistor, and a missing ground connection from a coil.

The other set I have acquired is totally original, so I will see if I can learn some of the tricks of the trade of testing and updating components, whilst keeping the restoration sympathetic.

I read through all the contributions on this thread again to make sure I had captured all the pearls of wisdom. I feel I have learnt a bit more since I started this project!

[Richardgr]

Since I am working on this thing while I am working away from home, I do not have access to all my tools, and a place where I can make a mess (filings, shavings, etc).

The case for the 768 is in two pieces; the bottom was broken off - and the sides were coming away from the front.

Looking closely at the damage, it is hard to see what has happened. Possible standing in a cellar, with a bit of damp. I think it was just constructed with a poorer glue, and insufficient supports for the joints, became weakened over time, and maybe received a sharp impact in one corner at some time.

Anyway, I manged to improvise for the lack of clamps with the straps from my roof rack, and it came out OK.

I have been ruminating on where to go with this radio. It is in a state now where I can plug it in, turn it on, and listen to a crackly radio station, or a bluetooth. Once I have attached the 3 wire mains lead and earthed it properly, fitted Y class caps where I have jst 630v film for now, then it is in a suitable condition to be passed on to another enthusiast.

I do not feel like I have done it justice yet - the speaker is very low-fi, and I am sure that is the way it is being driven. I also would like to hang on to the 2A3 tube, even though it could well be a tired specimen. The radio is not particularly collectible, and had been worked on in its life.

Purists please look away.

I picked up a couple of NOS 1619 tubes from a local auction site - £8 each - so I thought I would see how it goes using that instead of the 2A3. I have done a bit of research and the 1619 was a regular substitute for a 45, and with a B+ of 260V in this radio, it will be comfortably within spec here. It was actually almost a contemporary update, so is no the worst travesty.

I would like to get hold of some wafer octal sockets so I can swap out the UX4 one. They will be wired as triodes, and I will need to adjust the bias resistive divider so that it is dropped from 45V to a more suitable level - I think it should be 19V, but some research still needed.

Missing my oscilloscope. So useful for tracing signal levels.

[Herald1360]

Those straps (enabling the "house build" for the set) seem kind of appropriate!

[crackle]

Originally Posted by Richardgr

Does anyone know if it is OK to post a schematic from The Radiomuseum? Otherwise I will draw it again.

I have checked this with Ernest Erb and just had a reply that confirms the following:

It is OK to post parts of a schematic from Radiomuseum.org to illustrate a point, or a reduced version of the schematic not large enough to read specific detail.
But you should not post the whole of the schematic at full size.

See also forum rule B part 8 of this site;
Do not post or request service data, or links to service data, if it is for sale at www.service-data.com. The service data website funds this forum. You may post a small section from published data (such as a section from a circuit diagram or other diagram) to illustrate or help with a discussion.

Mike