Pye 350/C "All Electric Three" 1929

[6.3volts]

Originally Posted by HamishBoxer

Just clean out and put in a new replacement,while trying to keep original look.A set of above age ,no way should it be showing very often Yellow modern caps.

Yes I agree about the yellow caps. If I can't re-stuff a cap I'll paint it brown.
My main block can is largely empty now with the size of modern replacements.

[6.3volts]

Originally Posted by Mullard_Nut_74

I have seen a lot of reference to "re stuffing" caps lately.
Do you just drill holes in the originals clean them out and hide new caps in there to keep the original look? Or do you actually repair and refurbish the originals?

Would anyone ever bother drilling the innards out of a black or brown waxy and hiding a new cap in there or is it not worth the effort.

Just interested in the practice but not likely to do it myself.

Neil

My view is that the effort is rewarded particularly in older sets. One thing I like to do is manufacture the cardboard capacitor boxes mounted above the chassis in pre 50s radios. They were invariably removed when new caps were mounted under the chassis. I suspect this was visual proof to the customer that the work had been done.

[6.3volts]

Update on the main block.
First pic shows the sections as revealed with the base unsoldered. 91 years since they last saw daylight.
Pic 2 shows the sheet of insulating material I bonded to the case. The original terminal plates were live on top and had been set back from the case with a thick layer of pitch.
Pic 3 shows 4x1uF @650v (c 12 & 13) ready to work. I changed my mind about the electrolytics for c10 & 11 and used the same 1uF as original.
Pic 4 is the completed block with c1 & 2 waiting for surgery.
Due to the lockdown I have no suitable brown paint so used the black for now.

Next will be trying the rectifiers to see if they will live again. If not I'll hide some modern ones behind them.

[HamishBoxer]

Looking good.

[6.3volts]

Continuing with the HT circuit I start by assuming that the rectifiers are going to be (at best) unreliable. They are not new and look to be corroded. This set has been damp for many years and has corrosion all over the chassis. Anyway I decide to try them out just to see what is left of them and with a lash-up I run them up with a current limited variac expecting to see them slowly fail and heat up with reverse current before taking out my sacrificial capacitors.
After some time and much turning of my variac I arrive at pic1. This voltage is way above the 166 quoted by Mr Thornton so for now I am thinking there might be a future for these stacks of copper-oxide (I think) rectifiers. In all I left them running at 200 volts for a day with a drain of 20mA and they remained happy.

Next the dropper. Due to previouse experience with very fine wire from the early 20th century I am not expecting much in the way of reliability from this. All the high resistance sections are open and given that the other sections are involved in providing bias to the valves I decide to abandon them as to risk one opening could risk the life of the three valves. Just for fun I leave the intact sections running for a night with 4mA passing through them and in the morning they are open.
I'm going to use the body of R1 to mount 5w resistors to replace the lost windings. This will be seriously less expensive than rewinding the sections with nichrome and given that it's all out of sight in the case I will be happy to know that this important part of the radio is going to be dependable.

Can anyone suggest why the set's designers went for this voltage doubling setup rather than a straight rectification from a higher transformer voltage? Could it be the cost of a rectifier valve justified this arrangement? I'm curious.

[AC/HL]

I read somewhere that Armstrong used this arrangement to reduce strain on the transformer. Others manage OK, so that may be an assumption too far, but on a 90 year old example?

[6.3volts]

Here is the ht system assembled and running with a 16mA load. Ripple is showing one division with my scope set to 0.1 v/div. The voltage on C10 is 223v which is considerably more than Mr Thornton reported. Could be his caps were not as good as he thought or maybe the condition of the rectifier stacks could be involved.
Does anyone know if these voltage doubling circuits are inherently better at dealing with ripple?

[Herald1360]

Not particularly. The version in this set does produce 100c/s ripple unlike the common terminal variety with 50c/s and has a reasonable transformer utilisation factor.

[PJL]

It's definitely set to the correct mains primary tapping? That does seem considerably higher than the Mr Thornton's voltages but I am inclined to think yours is probably nearer the original figure. The AC/P datasheet shows anode voltage=200V and dissipation=4.0W but Thornton's figures for the AC/P are 155V and 1.2W. The HT regulation will be dire so it will probably settle at quite a bit lower when the HT is raised a bit.

[Ian B]

McMichael also used a similar voltage doubler arrangement around this era. I asked Gerry Wells once what advantage there was in it. He thought the transformers were simpler to wind , not having a centre tap. It occurs to me too that I believe there was a valve holder tax - and a metal rectifier isn't a valve!

Ian Blackbourn

[John M0GLN]

A scan from the 1930-31 'Catalogue of Wireless' from,

Northern Radio and Cycle Supplies,
61 Skinnergate,
Darlington.

John

[6.3volts]

Originally Posted by PJL

It's definitely set to the correct mains primary tapping? That does seem considerably higher than the Mr Thornton's voltages but I am inclined to think yours is probably nearer the original figure. The AC/P datasheet shows anode voltage=200V and dissipation=4.0W but Thornton's figures for the AC/P are 155V and 1.2W. The HT regulation will be dire so it will probably settle at quite a bit lower when the HT is raised a bit.

Thanks for your thoughts. Yes it's set to 250v and secondary voltage is 131.
I note that Mr thornton had concluded that the six caps were all 1uF and that he "tested" them on a Wheatstone bridge. I wonder if the bridge would have been able to account for leakage or ESR and what with a bit of "rounding", I wonder quite what he actually had inside that sealed block. I stopped considering mine when initial tests had them passing more than 20mA at 250vdc. One guess might be that his caps were quite poor (as were mine, every one in the set) and that they were reporting circa 1uF each. In several pictures I've seen online the blocks were all mounted with the 12uF legend at the back as viewed from the front of the chassis so he would have been unaware of the true value of the six caps inside.
Then we also can speculate on the condition of his rectifier stacks.
I measured reverse current in my stacks at 1.1mA and 1.9 mA at 200vdc. when first removed from the set.

[6.3volts]

Originally Posted by John M0GLN

A scan from the 1930-31 'Catalogue of Wireless' from,

Northern Radio and Cycle Supplies,
61 Skinnergate,
Darlington.

John

John, thank you for that. Quite an investment in 1929! No hint of headphones so I think maybe another week's wages (or more) for those. I can only wonder how many sets they sold at that price.

[6.3volts]

Originally Posted by Ian B

McMichael also used a similar voltage doubler arrangement around this era. I asked Gerry Wells once what advantage there was in it. He thought the transformers were simpler to wind , not having a centre tap. It occurs to me too that I believe there was a valve holder tax - and a metal rectifier isn't a valve!

Ian Blackbourn

Ian, I can only wonder at how and why a valve holder tax was supposed to be good use of paliamentary time.
Maybe some crusty mind decided it was a fitting replacement for the window tax.
Perhaps it might be another tool in the fight to maintain the correct social and class divisions and prevent these new fangled things getting into the wrong hands. I expect Lord Reith would have barely repressed a smile.

Jonathan.

[6.3volts]

Just a Quickie showing new life being restored to caps 3,4 and 5. This was not easy to remove and involved much wire pulling and manouvering.
At this stage I had not been able to remove any knobs which meant the front panel was not coming off. The steel grubscrews were quite happy where they were and no amount of discussion could persuade them to shift. This was not helped by previous historical attempts to ruin the slots and even heat-soaking them all day as in pic4 with penetrating oil failed to shift them.
Historical attempts to lever them off had only served to splinter off the finely moulded edge of the central knob and it looked like in its present condition it would spoil the set. Lucky for me the phantom srewdriverer (technical term) had not tried very hard. The shaft is part of a large but fragile switch and serious force would had rendered it into many pieces. In the end I firmly clamped the shaft just behind the front panel with mole grips and levered the panel forward to push the knob off its shaft. Fortunately the reaction knob levered off after much heating to expand it and the volume control dismantled in situ to allow the shaft to withdraw through the panel.
In the end all three grubscrews had to be drilled out and oversize 2BA screws fitted. The central knob had already been oversized but was able to accept a 0BA screw and still be structurally strong.

[6.3volts]

Having persuaded the front panel to come off I investgate the volume control which looks like a very low value wirewound. Mr Thornton tells us that his was 70K and was working very much at one end of its travel.
Mine looks like a 50ohm pot but it measures 4.2meg with many open sections on its length. This makes no sense as I can see the wiper making contact with the element. Even less sense can be gained by reading 25000 ohms printed on the body of the pot.
It turns out to be a carbon resistance element pressed onto a wirewound runner that takes the wear.
The first pic shows the wire former with the carbon on paper element pressed into contact by a tight bakelite collar. Corrosion between the wire and the carbon track has rendered it open in sections and unreliable. Unfortunately the collar splits as I remove it and the carbon on paper track decides to read off scale on my meter.
Pic2 shows the carbon track and the wire former. The wire turns have been cut through on their outer side except at the ends so as to provide a means of making multiple pressed connections to the inside of the carbon track.
The track has failed completely and the pressed collar has broken so there seems little prospect of any rescue.
A week later a 25K wirewound pot drops through the door and I find that, with a bit of crafting, I can make it fit so as to use the original screwed terminals and knob. Pic3 shows me turning the shaft to suit the original knob. Pic4 shows my idea for mounting the control. The clear insulating sheet is to prevent the screw heads from grounding on the faceplate. Pic5 shows the finished item.
Now given that the construction of the original seems a little unreliable after many years and that mine was clearly marked 25000 ohms, I'm wondering if Mr Thornton had the same control in his set.

[6.3volts]

After some time spent trying to work out why nothing was working after the first stage I took the switch (S2) apart and found that some of the contact fingers under the armature were bent down away from making contact. Why? How? Maybe someone thought this was a good idea? They are not easy to get at and are placed so as to need deliberate action to push them down. The finger on the left in pic1 is clearly bent down. several others showed signs of bending.There is no weakness in the metal and having bent them into shape they are perfectly reliable. The upward bend on the far right top contact is factory and is a crude but effective way of ensuring "break before make" on S2g the PU switch.
Two more pics show the armature construction and the last pic shows how I did my best with the badly damaged skirt of the wavechange knob. I mounted it on a rod and, turning the chuck by hand, used a very sharp tool to scrape away the material back to a concentric solid edge. I blended the ends of the cut with a tiny file and am pleased enough with the result.

[6.3volts]

Having repaired the switch I looked at the intervalve audio transformer.
It was no surprise to find both windings open and my six other spares from that era had the same problem.
The unit in the set had a 2.2meg grid leak across the secondary and at first I wondered why. My guess is that the secondary opened up and this "fix" allowed it to stagger on with capacative coupling between windings.
Given that the original was disposed of I decided to rewind the one in the set and at first I tried to save the wire as it came off. After several breaks within the first thousand turns I abandoned this idea and cut the lot off.
I have often wondered what the thinking was with these interstage transformers. (ideas welcome) Why so many turns? They always fail and so rather than obtain and use such fine wire I used some from stock which was 3.5 thou thick. (original was 1.8! This meant a lot less turns and I had no idea of the original ratio so I wound it at 4:1 and looked on this as an experiment with no idea if it would give good results.
Pic1 shows the stripdown and pic2 is installed ready to work.
That fixes most things and I've already had the first stage amplifying a signal so next should be some sound.

[PJL]

You are dealing with valves that don't like much of a load so higher primary inductance will give better LF response and then if you want some gain you need to wind 4 times that amount for the secondary. I also rewind these by hand and 12000 turns for a secondary is not unheard of.

[6.3volts]

The chassis is more or less ready to make a noise. I have a very badly corroded front panel and need to strip it (pic1&2). Pic4 shows that under the paint the legends for the controls are stamped into the metal, but I'm not sure why. I had this on a wartime Marconi communications receiver but those were very deep and designed to be filled with contrasting paint after spraying and this worked very well. The stampings on this Pye set are very slight and are covered by the original printed letters. Maybe a change of design? The back of the panel is good and I've left the paint on as a reference for the colour for when we are allowed out to get the paint, (pic3). In the meantime my plan is to spray the front with etch primer and assemble everything.
I've added a fuse to the mains input for safety and protection, (pic5).
Hopefully the set should soon make a noise in my pair of BBC headphones.
Can anyone suggest a way of producing the lettering something like the original?