In 1962. I was 12 and my sister Sue was 10 when Dad bought our first TV. We were probably the last house in the street to get one!
Sue and I watched the Beatles arrive in Sydney – live – on that TV. I remember when the LOPT failed under warranty and the service technician had to take the chassis away to fix it.
It was a 23 inch Pye with the T20 chassis and it had a console cabinet of a design that I never saw anywhere else. Which is surprising because I worked as a TV service technician for 10 years while I was studying engineering at night. I must have seen tens of thousands of TVs.
The short-lived T20 chassis was a vast improvement on the T18 and the even shorter run T19. It was basically a point-to-point wired version of the PCB-based (and very successful) T21 that followed it. The T21 was to run for many years, featuring in the legendary Pedigree TV (Australia's Philco Predicta but without the reliability problems). It made sense to give the circuit a good trial before committing it to a PCB. The T20 and all later designs were “flip-down” chassis. Very easy to work on, service techs loved them.
So when my friend and fellow collector Pete in Albury sent me this picture of what is undoubtedly a very rare model TV, just like the one I grew up with and looking to be in original condition, of course I had to have it!
Pete packed it onto his ute for the 6 hour drive to Sydney.
When it arrived and we took the back off, I was disappointed to find it had the T18 chassis and not the (admittedly quite rare) T20 chassis. The CRT looked like it was a relatively new replacement and it tested healthy, so that was some consolation.
The T18 was probably the most British-inspired TV design made by Pye Marrickville. Not surprising because the chief engineer at the time was a Pom. It also appeared just after the Credit Squeeze of 1961 that was the death blow to more that half the TV makers in Australia.
Pye was one of the survivors but the T18 was a “poverty model” - no money was available for new tooling. You can see this very clearly in this TV. The chassis metalwork is the same as that used for the W201 series from 1957, which was designed to be mounted alongside the CRT with the controls down the right-hand side. A pair of 6N3's still sit atop the mains transformer at a time when everyone else was using silicon diodes. The rest of the chassis has more holes than parts, clearly showing it to be a ruthless cut-down of the earlier design. More about the implications of this later.
Anyway, time to replace a couple of critical paper caps and bring it up on the dim bulb. The original filter cans formed up nicely with no drama.
A few tweaks and we have a raster, snow and sound!
But alas, and not entirely unexpectedly, we are losing width after about 5 minutes running. Opened the cage and found a version of the LOPT that I had never actually seen working before. I must have replaced hundreds of these, back in the day! Easily recognised, the Improved transformer had a different terminal panel design. This was not one of those!
So, before condemning it, I replaced the 0.1uF inter-winding cap (I'd already replaced the boost cap) and tried again. No joy. The thermometer, reading the primary winding temperature, tells the sad story.
With nothing to lose I soaked the thing with WD40 and let it sit for a month.
This time, it ran for about 10 minutes, raising my hopes but ultimately failing the same way.
There is a bit of a back story around these strange-looking LOPTs.
As a result of the 1961 Credit Squeeze, Pye found themselves unable to buy stock from their previous deflection component supplier, Rola. However, they did have a huge, mothballed WW2-vintage plant that made tropic-proofed ignition coils for aero engines. There were coil winding machines and a large quantity of DCC (double cotton covered) winding wire. So the decision was made to resurrect the plant and make their own LOPTs.
I once visited this plant when it was operating. Housed in one of those WW2 aircraft hangar-style buildings made from laminated timber was a huge gas-fired caldron with a winch that lowered trays of transformers into the boiling bitumen, where they sat for a time to force out any trace of moisture in the windings. It was hot, unpleasant work and there were fire extinguishers everywhere. It would have made a good movie set if you were making a movie about what it was like in hell!
The mistake that they made initially was to use bitumen, when the plant was intended to use “Berry-Wiggins Compound” (look it up). When they eventually started using this stuff (imported from the UK), they made reliable transformers. This plant continued in operation up until the early '80s making LOPTs for all Pye's colour TVs made at Marrickville. I never had to replace one of these. They looked ugly but they worked.
Anyway, what to do about my dud LOPT?
I had a Philips NT3101A LOPT that had been pulled from an HMV parts chassis. Fortunately, this was a replacement and was the -A variant which Philips developed for HMV who ran their line output stages unregulated. Since the T18 was in this category it was a good match.
It was just a matter of wiring the Philips transformer the same way it was in a Philips chassis and working out how to fit it, physically.
I pulled the EHT bucket off the Pye transformer bracket. It was filled with wax around the 1S2 socket. Simple, I'll just sit it upside-down in the oven with a catch tray underneath. 140 degrees C should do it. But after 15 minutes, nothing had happened, so I set it to 160 and set the timer for 10 minutes while I did something else.
Came back to find the plastic insulator bucket had melted into a puddle on the catch tray but the wax was still intact! Hard to believe but the picture shows it.
OK forget that idea. The chassis had a large round hole under the bucket position, obviously left over from an earlier design. Searching the kitchen I found an Aldi travel cup that was a snug fit in the hole. It had always leaked anyway so it was no great sacrifice to cut its top off and use it. Some M4 nylon screws and nuts, 2 plastic angle brackets and an M4 tapped spacer for a new B9A valve socket and job done!
At this point I decided to rid the chassis of the remainder of its paper caps and some well-out-of-tolerance resistors. The chassis had already had a full complement of mustards fitted to the vertical circuits many years ago.
Time to fire it up on the test CRT.
OK, I need to reverse the + and – on either side of the boost tap.
Now looking good, but with these old TVs there is always an assortment of niggly little faults. And so it proved to be with this one. Stay tuned!
Thanks Frank.
I'm just writing up the next chapter in this saga. Will post soon.
Running the chassis for a few hours on the test CRT revealed the following faults:
-
Marginal width and poor linearity with the raster off-centre.
-
Intermittent sideways jump and tear. Sometimes this would not happen for an hour or so.
-
Black level wandering and contrast changing with picture content.
-
Poor AM rejection in the sound, especially on high-contrast scenes.
-
Snow in picture, tuner AGC voltage near zero.
-
Height and linearity controls at end of travel, still too much height.
-
fault was the horizontal S correction capacitor. I had fitted a 150nF cap because the Philips circuit has 180nF in this position and the T18 had 100nF. Changing it to 100nF and adjusting the centering tabs resolved this one. I guess the tabs may have been out because there was a paper cap in that position previously and it may have been leaky.
-
Someone in the distant past had been looking for this one, as evidenced by the number of parts that had been replaced, including the original selenium double-diode by a pair of mid-60s vintage germanium diodes. The 6BL8 (ECF80) was newer than the rest of the fleet. And there was an 8uF 500V electro bypassing the cathode of the reactance triode.
So I synched the scope to the composite video output from the modulator and went searching for movement. The reference signals were jumping around a bit so I took a guess and replaced the 560pF micas that formed the lower half of the two reference dividers. No, that wasn't it! After going down a few rabbit holes my attention was drawn to the 5.6pF cap that couples the sync into the AFC diodes. And that was it! How do you test a mica cap that you suspect has silver mica disease? Hook it up to 200 volts or so with a voltmeter in series to ground.
-
Watching program content on this TV was not a pleasant experience! On dark scenes the contrast was reduced dramatically. I didn't recall the T20 being like this so I wondered if this was a fault or a design flaw. Looking at the circuit it was obviously the latter. The video OP stage G1 was half AC coupled to the detector using C18 and R18. This was not so in the following T20 chassis or the T21. Analysing the effect of this, I could see that APL changes would affect the AGC and this was borne out in measuring. Why would anyone add AC coupling inside the AGC control loop? It just made no sense. I will remove it and change the cathode resistor to the 100 ohms used in the T20.
-
Poor AM rejection – this design uses a Foster-Seeley discriminator. The Foster-Seeley has no AM rejection (unlike the very similar Ratio Detector) and because the T18 is a cut-down design it lost its limiter stage that existed in earlier models. It has been replaced by a biassed diode limiter of dubious efficiency. By running the modulator at maximum audio gain it is just acceptable. I will investigate re-wiring the Foster Seeley as a Ratio Detector.
-
The snow has gone, the tuner AGC voltage is back! This happened slowly, of its own accord. I suspect the 6FC7 (ECC89) RF amp may be recovering from a long period of hibernation. I don't have a replacement for this valve, it was only ever seen in the very compact UK-sourced tuner that Pye used between 1961 and about 1966, when it was replaced by a turret tuner sourced from EIL / Astor.
-
The Height and Linearity control range issues were easily sorted. Wrong value resistors, one by a past service technician and the other one which looked original, but didn't match the circuit.
I was surprised by the number of original Mullard valves in this chassis, especially by the 6AL3 damper, the 6BM8 (ECL82) audio amp and the two 6N3 HT rectifiers, all of which are original. That would have to be some sort of record, especially for the 6AL3 which tended to have a high attrition rate, especially due to cracked glass just after rain for some reason I've never been able to discover! Mullard valves in Australia were made in the Philips' Lane Cove (Sydney) plant and were generally long lasting. That is, unless they were abused by leaky paper caps.
OK, time to put it back in the cabinet.
Back in the box.
No raster. Turned brightness down and slowly back up. Aha, there we go! A dim and blooming picture. It looks like the original 1S2 (DY87) won't drive the big jug.
But it did before. Why not now?
Probably because I used the heater loop that came with the new LOPT. It is resistive wire, not copper with a series 1.5 ohm resistor like the old one. Maybe just a tad too long.
Let's just try a new bottle first.
I had two NOS 1S2As in stock. These were branded RCA and Sylvania but were both actually Philips of the no-getter-flash type, made in Holland.
On warm-up, the first NOS 1S2 erupted in a purple internal fireworks display. Oh well, it happens.
When the same thing happened to the second one I was taken aback.
A rummage through a cardboard box of hundreds of old chassis pulls I picked up in a throwout some years ago got me a couple of Philips Lane Cove made 1S2s. They looked like they were from the late 50s, going by the nickel plating on the anode caps. Also no getter flash. So without much hope I plugged the first one in.
Success! A nice bright picture with no sign of blooming.
I've been asked, why persist with 1S2s when you can use selenium stick rectifiers? My answer to that is I much prefer the more aesthetically pleasing startup of the TV that the indirectly-heated 1S2 gives. By the time the EHT appears everything else has warmed up and both timebases are locked and stable. The picture just zooms out onto the screen.
The width just makes it to the sides of the CRT with nothing to spare. I could move the 6CM5 (EL36) anode to the “220v stabilised” tap on the LOPT to fix this but the boost is already running at 834 volts with HT at 220 volts so let's not tempt fate! I might twiddle the magnets at the sides of the yoke to see if I can straighten the edges. It's a hot night and everyone is running their A/C. The mains reads 230 volts at the moment so that's probably as bad as it's ever going to get.
I'm not happy with the focus, especially in the middle of the screen. Moving the focus off the G2 connection to ground results in a picture that's sharply focussed over 90% of the screen, only falling away a bit at the extreme edges. Good enough.
There is a dark patch in the middle that can be seen on a white raster – see the pic of the opening credits of a Yes Minister episode. (Those black wiggly streaks in this pic are due to a bad connection on the balun letting in noise from the plugpack powering the modulator).
It looks like an ion burn, except that, of course, the 23HP4 has an aluminized screen. It can also be seen with the TV off under bright daylight conditions. I hate to say it, but it looks like it may be a defect during the laying down of the screen in the factory. Surprising it got through inspection. It's subtle, you don't see it on most content, but you know how it is, once you notice it you are looking for it all the time! Not much I can do about it though.
I added a channel indicator lamp with its holder and rubber grommet. They were completely missing from this set. The hole for it looked like there had never been a grommet there, ever. Another bit of cost cutting?
I used a 6.3V 250mA MES lamp with 15 ohms in series to keep the heat down in deference to the nearby back-painted plastic. I washed the knobs and the control panel plastic. Also added a gold insert to the channel knob and re-painted its numbers.
This must have been one of the last of the 10 channel tuners. I remember when I was a kid that our T20 definitely had a 13 channel tuner because we could receive Ch3 and 5 from Newcastle as well as Ch4 and 5A from Wollongong – the latter through the back of the antenna. It was great for watching the football when it wasn't being broadcast live in Sydney. That would not have been possible with the older 10 channel tuner which didn't tune Band 2. Not relevant these days of course.
I'm going to pull the chassis again and do the black level and sound mods I mentioned previously.
Looks like I may have to replace the discriminator transformer because it does not have the necessary centre tap on the secondary. I hope I can find the ratio detector I pulled from that KGH chassis some years back. Failing that, the two HMV parts chassis that Pete has in Albury both have ratio detectors.
I was thinking I might also implement the improved beam current limiter circuit that Philips used, it has a more sharply defined limit threshold.
I like to be able to actually watch vintage movies and TV shows on my old TVs. So to improve this experience I'm happy to make minor changes to TVs that should have been in the original design – and often appeared in subsequent models. It might not please the purists. But if you were restoring a classic car that you planned to use as a daily driver and it had drum brakes but the following year model had discs, what would you do?
While the chassis is out I might take the cabinet outside and give the finish a rub-down with turps as recommended by Vintage Pete. And hopefully free up those siezed Shepherd castors.
Fantastic write up Ian, thanks for taking the time to provide such an extensive article about this TV. It's a rare treat for us in the UK to follow and get a glimpse of an Oz set being repaired. We used to have a member from NZ who would share his native TV's as they were being repaired, equally interesting. The final result with your TV shows it was well worth all your efforts, and a very stylish TV it is too.
These sorts of threads remind me of the early days of VRAT when we have a huge number of members documenting B&W 405 TV repairs, including pre-war, bar a remaining few, for the most part they've all abandoned us now. You've set the bar high for the rest of us when fully documenting our repairs.
Any plans for a follow-up project? Over here, it's getting to that time of year when many of us used to contemplate a winter project to see us through the long dark nights. However, as your neck of the woods approaches summer, I guess you be out and about enjoying the long days and fine weather instead of being locked up with another TV.
CrustyTV Television Shop: Take a virtual tour
Crusty's TV/VCR Collection: View my collection
@irob2345 Mean level AGC and the loss of the DC component in the video were common in the UK sets until colour and 625 negative modulation were introduced. Rank Bush Murphy (RBM) started adding a ‘Black level clamp’ to there mono sets around the middle of the 1960’s, superb pictures but we had lots of complaints about it. Customers were used to the screen not going dark between scenes and many didn’t like it.
Eventually they got used to it but at first it was a problem.
Frank
Great work, and very interesting to see a telly from the other side of the world!
Regards,
Lloyd
Not quite finished the Pye yet but yes I do have another project coming up. It's a rare and very stylish oiled teak high-end monster 3 in 1 from HMV (the brand was owned by EMI here), made in the late 60s. It's 25 inch and uses a version of their Decal valve chassis. Video, like all Oz HMVs, is no-compromise DC coupled with no beam limiter to muck it up. By the mid 60s HMVs mostly had stabilised line stages so the raster size stays stable with rapid APL changes. The TV sound goes through the solid state stereo amp to some generous vented enclosures. I serviced a couple of these when they were new and the sound is truly impressive.
https://www.radiomuseum.org/r/hismasters_y1_d3_y1.html
although mine is the later Y2 version, different stereo amp.
$800 was a LOT of money in '68, you could buy a new car for not much more than that. But that cabinet is beautifully crafted.
No FM radio here in '68, Band 2 was used for TV! The FM band was not fully cleared until the late 70's.
I did the DC coupled video mods (as per the T20) and implemented the improved beam limiter. Had to change a couple of resistors around the Contrast circuit to bring its range back to where it should be, probably because of the cathode resistor change.
Wow, what a difference! It's actually quite pleasant to watch now!
I was going to add horizontal retrace blanking but, with the stable black level, it doesn't seem to be necessary any more, the grey "pole" is no longer visible under normal picture settings.
Pulled the Foster-Seeley coil out to investigate the feasibility of adding a centre tap to the secondary winding. But then I thought, maybe I should check those OA81 germanium diodes. Sure enough, one of them measured 4k ohms one way and 15k the other. So in went a couple of 1N4148s, assembly put back in the chassis and aligned.
Much better!
I've never had a problem replacing germanium diodes anywhere with silicon 1N4148s. They always work just fine. My theory is that the 0.6v of a silicon diode vs 0.2v forward drop for a germanium is compensated for by the lower bulk resistance of the silicon diode.
I now have one last Shepherd castor to unsieze the wheel on. The pintels on both the stuck ones responded to some persuasion with a big spanner and a vice, followed by some oil. The solidly stuck wheel responded to heat but got stuck again when it cooled down. So I boiled it in turps for about 15 minutes (outside on the BBQ side burner, fire extinguisher at the ready!). It's had time to cool down now. I just pulled it out of the turps, gave it a few twists and it's running as good as new.
