Line output transformer substition Decca DR123.

[FERNSEH]

At last I can can write about a line output transformer substitution that really works.
The week I acquired this impressive looking 23" dual standard Decca TV set from Gateshead only a few miles from my location.
The DR123 is a posh version of the DR1 which was introduced in 1965. The chassis layout resembles the BRC 850 series, the upper deck takes care of the signals and the lower deck the timebases and power supply.
It's likely that this set has not been used for forty years so the standard test procedure was implimented, just plug it in.
It wasn't long before it was noticed that the PY800 boost diode was glowing red hot. To keep the story short the line transformer is short circuit.
The line transformer has an additional winding to supply a negative flyback blanking pulse to the first anode, it had shorted to the main winding. The transformer is a plug in affair, easy to remove for replacement. I disconnected the offending winding, replaced the transformer, switched on the set and this time some results, a tiny spark at the anode of the DY86 EHT rectifier. It seems that the transformer is faulty anyway. I'm never going find a replacement, so what trying a BRC 1400 jellypot? After all both sets have 110 degree CRTs. A jellypot was wired in and results came in no time, plenty EHT and full scan on 625 lines, except there was frame collapse, another job to do later. The old transformer was removed from the plug board and the jellypot fitted using an adaptor plate. A number of wire links added and cuts in the PCB print had to be done in order that the BRC transformer will work in the Decca circuit. The method of switching the scan coils between 405 and 625 differs between the two makes.
The results are impressive, well worth all the effort and the troublesome line output transformer used in the original design has been replaced with a much more reliable part. Very does rarely a jellypot fail.

DFWB.

[HamishBoxer]

Looks great, agreed I never had a jellypot fail either in all my 50 years of servicing.

[Heatercathodeshort]

Another enforced marriage by way of yet another 'adaptor plate' All that is missing is the shot gun..

Well done David! Those Thorn jelly pots were very versatile. I never kept one in stock bacause they never failed. Must be the most reliable line output transformer ever manufactured. I did have a couple of 3000 EHT jellypots fail but just the two in all the years the chassis was around.
Just before colour started in 1967 I had a Decca SRGTV777 in the living room. It had the transistor radiogram chassis incorporated in a lavish cabinet. Sounded fantastic! I had two LOPT's fail in two years. Just as well they were plug in. Not a good period for Line output transformers, not even for Plessey that had in the past, maintained an outstanding reputation for reliability in everything they manufactired. They did improve considerably by the late 60's. Regards, John.

[FERNSEH]

Hi John,
Guess what? It's got pekalit knobs.

DFWB.

[Heatercathodeshort]

Nurse!

[FERNSEH]

The set has developed a new fault. The picture has shifted over to the right with foldover. The active video has in a sense moved over and there is a grey area at the left. A line timebase phase error. The DR1 chassis has a flywheel sync system, an ECL80 triode pentode is employed as the line frequency and phase discriminator.
The triode section inverts the negative going sync pulses from the "F" section of the PFL200 and feeds them to the screen grid of the pentode section. The output from the line oscillator is integrated and sent to the pentode control grid. The output at the anode of the pentode is filtered and supplied to the ECC82 oscillator as a control voltage. Along with the picture shift there is a degree of picture instability, even on the test card generated from the Aurora. It's really wobbly on VCR signals.
Time to use the scope to find the fault. The two 2mfd electrolytics in the discriminator have been replaced.

405 lines: Of all the people that have seen set working on 405 lines none have made any comment about the line structure. If it is pointed out to them the answer is: "I never noticed the line structure until it was pointed out to me"
And remember this is a large screen TV set.

A picture of the modified timebase board, note the wire links and cutouts in the PCB copper lines.

DFWB.

[Focus Diode]

The LOPT conversion is an excellent achievement. Not only was I impressed by David's craftsmanship but also how well the set worked with the 1400 LOPT and tripler in situ on both systems, as if the Decca was designed to take this transformer.

Sorry to hear about the new fault condition, probably my least liked of all fault conditions.

Cheers,
Brian

[Heatercathodeshort]

Strange one David. Not the most pretty of line sync circuits and odd that it uses such out of date valves as the ECL80, not the most stable of valves at the best of times.
Have you checked that thermistor in the suppressor grid of the ECL80? The one in the Pye V14 gave problems giving some odd symptoms. A fixed resistor as a test should rule this out. Small thermistors of this vintage tend to go unstable. A squirt of freezer may question its innocent looks.
Looking at the circuit there is little else to go out of tolerance.

Of course it may not like its retirement location. Gateshead is a long way from it's maker in Battersea, South London and it's probably a non smoker.
Very smart receiver. John.

[Studio263]

Lots of high value resistors in the ECL80 triode grid circuit, I'd go straight for the 2M2 one that goes up to the HT supply (R108?), I'd be quite surprised if it was still the right side of 5M...

[FERNSEH]

Focus Diode called around to see how progress was going on with the Decca DR123. Well, we can report that after replacing a number of resistors in the line discriminator circuit there is marked improvement in lock-in range and stability but the picture is still off-set to the right. The cathode resistor R172 (8.2Kohms) of the ECL80 line discriminator had risen to 15Kohms. The VA1039 thermistor in series with R172 checked OK, 680 ohms cold. All the other high value resistors associated with the ECL80 have been changed including R109.
The screen grid decoupling capacitor of the line output valve has been replaced, the result of doing that is even more EHT.
Meanwhile the CRT seems to be getting better, it's almost good as new. it's recovering after a long forty year rest.

DFWB.

[FERNSEH]

Originally Posted by Studio263

Lots of high value resistors in the ECL80 triode grid circuit, I'd go straight for the 2M2 one that goes up to the HT supply (R108?), I'd be quite surprised if it was still the right side of 5M...

It had risen to 3 Megohms.

DFWB.

[Focus Diode]

As David says problems persist with the picture moved over to the right. There's also a wasp-like hum which develops, sound remains hum free. I couldn't help thinking something may be amiss in the PFL200 sync separator side although I'm probably totally wrong. In addition to the components replaced David also tried another ECL80 which made no difference as expected.

Performance on a moving picture now looks really good. One could happily view 405-lines on the 23" screen. A stunning set and the only known d/s Decca around, at least in working order.

Cheers,
Brian

[FERNSEH]

The 405 scan coil coupling capacitor C133 was 0.25microfarads, I've replaced it with a 0.47 microfarad component. This action has improved the horizontal linearity. Somehow the set doesn't like the signals from the Aurora, the picture is still off-set to the right and yet the 405 signals from my own build converter produce correctly centered pictures.

DFWB.

[Focus Diode]

That's looking much improved. It didn't occur to me the Aurora could be the cause of the symptoms on the set. Certainly the performance when fed from your own converter look absolutely fine and without the wasp-like hum problem.

Cheers,
Brian

[Heatercathodeshort]

Sounds a really daft one David but I'm wondering if the power supply feeding the AURORA has a smoothing/regulator problem. Worth substituting another supply just for a quick test. John.

[FERNSEH]

Hi John,
good tip. I've tried another power suppy for the Aurora but the results are just the same.
First picture shows an off-sceen image on 405 lines, the second on 625 lines.
Note the displacement to the right on 625. The reason for this is all to evident after scope analysis. The waveform at pin 1 of the ECC82 the line oscillator shows us that the pulse is 18 microseconds, this is too wide, it should be far less than that, I'd reckon on 6 to 7 microseconds should be the correct figure.
I'll check the values of resistors R130, 132, 139 and R140. It's likely they have risen in value.

DFWB.

[FERNSEH]

Replacing those resistors has effected an useful improvement on 625 lines, although the picture is still slightly off-set to the right on that standard. An extra bonus on 625 is that those curious striations have disappeared.
The pulse at pin 1 of the ECC82 is now 15 microseconds on both standards.
R130 (680Kohms) had risen to 1.7megohms. Way out of tolerance. The others, R132, 139 and R140 were reading high and well out of tolerance.
Decca describe the line flywheel circuit as a coincidence detector which indeed what it is. That pulse on the anode of the ECC82 appears as a sawtooth waveform at the control grid of the pentode section of the ECL80. This ramp waveform serves as the reference to be compared with the high amplitude +ve sync pulses at the screen grid of the ECL80.
The Decca DPT schools set has the same timebase circuit.
Might be well worthwhile comparing the line sync circuits of the dual standard model.DR101.

DFWB.

[AC/HL]

Two off-topic posts moved here: https://www.vintage-radio.net/forum/...d.php?t=104900

[FERNSEH]

The off-set picture problem has been solved. It turned out that R110 a 150Kohm resistor in series with the grid of the ECL80 triode section had gone high. Actually, R110 had been replaced but unfortunately I hit on a faulty resistor, it's value read over 600Kohms!. That component has been replaced and now the picture is correctly positioned on both line standards.
The line drive waveform at the grid of the PL504 on 625 has the correct 60:40 mark space ratio but on 405 it is something like 27:73, hardly ideal. The pulse at pin 1 of the ECC82 multi-vibrator can be stretched by increasing the value of R130 (680K) but this would involve fitting an additional system switch. An easier solution is to shunt R140 (56K) with a 27K0hm resistor, this action produces the correct mark space ratio for 405 lines.
Reference to the attached circuit diagram on post #6: you'll notice that the slider of S104 is connected to the cathode of the ECL80. In fact in my set the slider is connected to ground. It was a simple task to wire in a 27Kohm resistor to the hot end of R140 and the 405 position of S104.
The attachments show the revised circuit diagram and an oscilloscope trace of the line drive on 405.
The modification has improved horizontal linearity on 405 lines.

The set had developed a low contrast fault, in fact the contrast control had no effect. This was traced to R67 (2.2Megohms) having gone OC. R67 is connected between the slider of the contrast control and the AGC line.

DFWB.

[Focus Diode]

I had great pleasure in visiting David and seeing the set yesterday. The results from this set look absolutely superb, as can be seen from the photo taken with David's kind permission. The more critical observer may notice some stretching on the left. Don't forget, the original scan coils are in use and the PL504 is the same as found fitted to the set.

I think we should all congratulate David in his fantastic achievement in getting this rare set working so well.

Cheers,
Brian