The Gerry Wells TV- continued

[Niechcial,Steve]

I last wrote about this on the Forum about a year ago and have only just now gone back to it with some good progress made and some interesting problems remaining .With no circuit diagram to work from, no idea of the original performance, and no idea who else might have 'got at it', it's hard to know what standard of performance to aim for. It is certainly a very basic set, lacking for instance flywheel line sync or any form of boost/width control.
Thanks to David and Dave who helped me identify the IF strip as one from a Sobell/Gec ST 125 I have been able to sort out the low gain/ poor contrast which was due to a high screen feed resistor in the IFs and resistors gone high in the contrast network- I replaced everything of 1M or higher for good measure. The contrast control now operates over the range you would expect. However there is a lot of video buzz on sound. I know a lot of these early D/S strips were prone to this. Is this one of them, and can anything be done to improve it?
When contrast was restored I noticed I still had very bad line tearing from the direct sync arrangement. There is a sync separator on the IF strip but Gerry doesn't use this. Having checked values of resistors around the PCF80 sync separator Gerry had built, I could see nothing obviously wrong. There was a 33K resistor inserted between the output of the PCL84 video amp and the grid of the PCF80. Removing this completely solved the problem of line tearing- but read on.
When this set came to me it had barely adequate width and the first 1.5 or so inches of scan on the left of the picture cut off. On replacing the boost diode, the picture shot out to an over scan of about 1.5 inches each side and a boost voltage of 900+ volts! I reckoned the Thorn 800 LOPT fitted would not last long like this. As I say, there is no boost/width reduction so I cut down the line drive to the PL36 by means of a series resistor. This has brought the width down to normal and the boost back to around 700V but the cut off of the first 1.5 inches of scan on the eft remains. See attached pic (this is taken in a mirror so left/right reversed)
I then scoped the line drive to the PL36 alongside the line drive to the output valve on another set on the same channel. The other two pics attached show the two traces separately and then superimposed (upper trace normal, lower trace Gerry). The scope is set to 10ms a division horizontally. You can see from this that the normal set starts the forward scan when it should do, 10 ms or so after the sync pulse. Gerry's set starts it early, about 5ms after the sync pulse, and the first part of the scan is of a steeper slope. According to my logic, starting the scan too early should result in part of the black level before the border castellations on the test card being displayed, but in fact the opposite is true. This makes me think that for some reason the output stage is being biased off during the first part of the scan. When I feed in the line drive from the other set, the scan is normal. I have already tested the resistors around the line osc. which is a simple multivibrator. The caps are RS silver mica ones and so are unlikely to be at fault. I think this is almost certainly a design limitation. I plan next to try and fiddle with the scan shape, but I would appreciate anyone's ideas please. Incidentally, the rest of the line lin. is good, even though there is no corrective inductor or sleeve in the scan coils.

[FERNSEH]

The first part of the scan comes from the contribution from the boost diode.
The line output valve should be non conductive during that period.
The whole line drive waveform should be negative going, below the zero line so to speak.
The final part of the scanning stroke comes from the current through the line output valve.
The line drive waveform should be -100V P - P.
It is possible to monitor the current through the line output valve by inserting a 1ohm resistor between the cathode and ground and use the 'scope to measure the voltages across the resistor.

DFWB.

[Niechcial,Steve]

Thanks for that David. I must be getting old to forget that the PY provides the first part of the scan! So perhaps the problem is the opposite of what I first thought. The PL36 is conducting too early and damping the PY before it begins to conduct fully itself? That fits better with the waveform. The PL already has a cathode bias resistor so I will check across that and see what is happening.

[FERNSEH]

Hi Steve,
What type of line oscillator is employed in the set, a multi-vibrator or blocking oscillator?
We know that the sync separator is a PCF80 so it is assumed it is the conventional type which conducts on the positive going sync tips and provides high amplitude negative sync pulses at the anode.
Perhaps the sync pulses are being applied to the wrong part of the oscillator?

DFWB.

[Niechcial,Steve]

Hello David. It's a multivibrator with the sync applied in the normal way to the pentode grid. I have already scoped the flyback against the sync pulse and can confirm that the flyback is being triggered correctly by the leading edge of the sync pulse. I will look tonight at what is happening to the PL cathode current

[Niechcial,Steve]

Sorry, I meant triode grid- it's an ECC82 osc.

[FERNSEH]

Hi Steve,
A cross coupled multi-vibrator or the cathode coupled type?

DFWB.

[Niechcial,Steve]

Yep- cathode to chassis direct

[Niechcial,Steve]

cathode coupled

[Niechcial,Steve]

I have been playing around with the bias on the output valve and the oscillator timing. The best compromise solution I came up with was to return the line drive to how it was, and instead to lower the HT to the line output stage with a 330R resistor. This has brought the boost and width down to something reasonable and has got rid of most of the start of scan cramping. I am now onto the last problem which is an horrendous video buzz on sound- one of the worst I have heard. I will look for a fault, but I know a lot of these early d/s IFs were prone to this on 625, There is plenty of sound coming through but the buzz is almost as loud. All the coli cores that it might help to twiddle are locked solid. I may have to resort to trying a 6mhz ceramic filter.

[Niechcial,Steve]

I got very lucky on the sound. There should be a 50 mfd decoupler (C128) for the anode feed to the EH90. It wasn't there. Looking at the wiring I would say that it was probably fitted elsewhere on the GEC chassis for which this chassis was intended, and got overlooked. Anyway, fitting it completely cured the fault.
So the set is finished with only significant change to the original design- the inclusion of the 330R resistor in the HT feed to the line stage. When I've got it cleaned up and turned round on the bench I'll take some off screen photos and as soon as possible it will go back the the Dulwich museum.

[Niechcial,Steve]

correction- should read 'only one significant change'

[FERNSEH]

Hi Steve,
In order the line timebase stabilising loop can operate the BRC 850 line output transformer was designed to supply extra energy over the normal requirements for scanning a 110* CRT. The set you are working on doesn't have that circuit feature.
Even as late as 1963 many TVs relied on for width adjustment simply by having a tapped resistor in series with the anode of the boost diode.
Examples are the early versions of the Pye model 11 and the KB VV series dual standard receivers.
Fifty years ago an Ultra Bermuda 950 series TV came in for repair. On one of the chassis rails somebody had written "died in 1969 of too much EHT". Usual story the owner couldn't get it fixed. Of course the fault was in the line timebase stabilising circuit.

Does your set have the "S" correction capacitor in series with the scanning coils?

DFWB.

[FERNSEH]

Looking at the circuits of the line timebases in the early Pye model 11 and KB VV10, the three width control resistors in series with the boost diode anode HT supply have a value of 68ohms .

DFWB.

[Niechcial,Steve]

Hi David. Yes I was aware of the HT reducing method of controlling width. In fact I have not so long ago restored one of those KBs. This LOPT however is at earliest from a Thorn 850 which does use a VDR stabilisation and control network. I guess it was one of the first to do so. The tube in Gerry's set is also a 24 inch 110 degree one. The resistor I had to put in is over twice the value of those on the KBs so it is a mystery where all the excess energy is coming from. Perhaps different scan coils to those for which the LOPT was intended, or perhaps Gerry has mixed 405/625 windings in a way not intended? I will check to see if there is S correction. If there is, the cap will have been changed because I did all the paper jobbies.

[Niechcial,Steve]

I've just re=read your post David and I see what you mean. But surely the energy for the stabilisation loop would not have been that much extra?

[Niechcial,Steve]

On a related thought of dating this set, does anyone remember when all 3 channels became available on UHF in the London area. I'm guessing 69?

[Heatercathodeshort]

15th November 1969. J.

[FERNSEH]

Hi Steven,
The earliest sets with the 110* CRT didn't require much more line scanning energy than the 90* CRT models. The much thinner neck of the 110* tubes helps to reduce the energy requirements. In fact many manufactures continued to use the PL81 line output valve.
In 1959 Ferguson was the first UK manufacturer to make a TV set a stabilised line timebase, this was in the model 546T. The line output valve was a PL81.
I do remember that the unfortunate Ultra set mentioned in post No.13 had excess width and EHT well over 20KV. Fixing the stabilising circuit tamed things and reduced the EHT to 16KV.

DFWB.

[FERNSEH]

"In 1959 Ferguson was the first UK manufacturer to make a TV set a stabilised line timebase, this was in the model 546T."
Correction: In 1956 Peto-Scott made TV sets which employed a stabilised line timebase.
Some sets used an OA71/81 instead of the diode section in the PABC80.
Perhaps this circuit could be introduced into the TV set under discussion?

DFWB.