Cossor 916 lives again.

[Graham G3ZVT]

Three months ago David (FERNSEH) kindly sent me a LOPTX from a KB EV30 so I could try it in my Cossor 916, that had been out of commission for a couple of years.

Here is the original transformer in it's metal can, before it failed



and here is the donor organ in place




The circuits for the two sets (segments shown here side by side) show the first problem, the KB transformer is one continuous winding whereas I needed the scan-coil section isolated so I drilled a hole and screwed in a small self tapper to act as an extra terminal and separated the winding as shown by the red marks



I got a healthy spark at the anode of the EHT rectifier, but its heater did not light and consequently no picture.

In desperation I tried a solid state diode from a microwave oven, it worked, but the EHT regulation seemed poor.
I noticed in one KB model using this or a similar transformer (but not the one in the circuit above), a U25 (2v 0.2A heater) is used and not a SU61/EY51 (6.3v 0.09A heater) could that be my problem?

So I ordered a U25 and it worked, I found by experiment that putting a 100pf pulse capacitor across the primary improved the width and EHT

Line linearity could be better, but apart from that the picture is as good as it was.



But I was not satisfied with that, its that "O" word, originality. I really wanted the old transformer can re-stuffed with the new component. It was a tight squeeze , I lined all four sides with polypropylene sheet cut from food containers. There is no hint of discharge anywhere.

I also disconnected two of the unused width tapping winding at the top of the terminal board and extended two connections to them that would otherwise have been inaccessible in the can.

Original on the left, new on the right


I tried it with the lid on for an hour, but the windings were a bit too hot to touch for my liking so I am keeping the lid off.

[Heatercathodeshort]

Nice job but I would keep that KB transformer open to air. No point in straining a 70 year old LOPT just to preserve appearances. They produce 10kv in the KB EV30 but the Cossor 108K crt should hold up OK. John.

[beery]

Great job,
Nice to see a late 40s/early 50s Cossor on here rather than another Pye or Bush.
Poor regulation with a microwave oven diode is probably because it is not designed to operate above 3kV or at 10kHz. The good thing about the U25 is that it looks like an EY51. I once put a Mazda U25 inside the bulb of a broken Marconi U16, but don't tell anyone...

Cheers
Andy

[FERNSEH]

I'm delighted that the transformer substitution has worked well.
The operation of this type of line timebase is explained in the service manual for the GEC BT2147 series of TV receivers.
At the start of the scan the line output valve is cut-off. Conduction starts half way along the forward scanning stroke. Current in the deflector coils causes the spot to move from approximately the centre of the screen to the right. At the point the where the blocking oscillator is triggered by the sync pulse the spot has reached the right side of the screen. The line output valve is immediately cut-off by the negative charge in the sawtooth forming capacitor. The spot immediately flies to the left as the anode current ceases. The collapsing field in the the deflector coils causes current to flow in the reverse direction at their naturel period of resonance reaching an almost instantaneous maximum and completing the flyback of the spot to the extreme left side of the scan. The current having reached a maximum starts to fall and the resultant reversal of potential across the coils causes current to pass through the efficiency diode to charge the boost capacitor and maintain a boost potential across it. The spot is now moving from the left of the screen as the current through the diode resulting from the collapsing field, falls to zero. By this time the line output valve has been cut off for nearly half the scan has started to conduct and the whole cycle commences again.

The Cossor 916 employs two valves developed by the company, the 185BT and 27SU. The choice of loctal valves is unusual in a UK made TV set, 7C6 and 7Y4. The 6AM6 RF pentode is said by some to have been developed by Cossor although others believe it was Mazda.

DFWB.

[Graham G3ZVT]

Originally Posted by Heatercathodeshort

Nice job but I would keep that KB transformer open to air. No point in straining a 70 year old LOPT just to preserve appearances. They produce 10kv in the KB EV30 but the Cossor 108K crt should hold up OK. John.

I tried hard to resist John, but the temptation was too strong
I had to remove approx two layers of overwind to get to a break, plus the tuning cap across the primary seems to vary the EHT (50-100pf) although I can't measure the voltage directly, I get the feeling it is not too far from the 7kV specified.

Originally Posted by beery

Great job,
Nice to see a late 40s/early 50s Cossor on here rather than another Pye or Bush.
Poor regulation with a microwave oven diode is probably because it is not designed to operate above 3kV or at 10kHz. The good thing about the U25 is that it looks like an EY51. I once put a Mazda U25 inside the bulb of a broken Marconi U16, but don't tell anyone...

Cheers
Andy

Actually Andy the diode spec was 12kV at 350mA!
https://datasheetspdf.com/pdf/668803/Hio/CL01-12/1
I expect it is the frequency, the above sheet seems to imply they are not even suited to 60Hz mains, but maybe I'm reading too much into it.

Originally Posted by FERNSEH

I'm delighted that the transformer substitution has worked well.
DFWB.

Thanks once again for the transformer David, and that explanation.
When I first got it to work with the solid state diode, I was greeted with an unexpected frame collapse. I could see the frame osc was hard on with practically no voltage on the anode but it took several sessions for me to realise what was going on. The frame osc half of the 6SN7 had a grid to anode short, well in the order of several hundred ohms "short".
I just wasn't expecting that!
I zapped it with 200v from a transformer and it's been absolutely fine since.

Just as well seeing how much some 6SN7's go for these days.

I think the fault would have cleared itself, or never occurred if the valve was operated with more conventional voltage, in this relaxation oscillator the anode only sees 20v.

Any ideas about the stretched line lin on the LHS?
My chassis does not have a linearity control, but there is another circuit Trader 996/T13, also using the 108k that uses this interesting arrangement.

I have tried that on my set without much luck. I've also tried to make the width coil saturateable with a couple of magnets, but they had no effect.

[FERNSEH]

In my post about Cossor 916 line output stage I suggested the operation of the circuit might be similar to the GEC BT2147 series. But I've had a rethink about the way the Cossor line output stage works. Rather than acting as a switch and only conductive on the latter part of the forward scan, the 185BT line output valve amplifies the sawtooth waveform in the normal manner. The damper diode doesn't contribute any part of the forward scanning process but still provides extra HT for the line output stage.
If that is the case then the shape of the waveform applied to the grid of the line output valve can be adjusted to correct any non linearity of the forward scan.
The attached circuit diagram shows that the line drive can be adjusted by trying different values for R47 which is the HT feed resistor to the blocking oscillator.
R47 could be substituted by a 500Kohm variable resistor in order to determine the optimum value.
Also, check C45 (1000pF) which is the sawtooth forming capacitor, small changes in value of this capacitor might correct shape of the drive waveform. Essentially, introduce a degree of pre-distortion to correct the horizontal linearity. R61 is the flyback "sit-up" resistor. The value of R60 across C45 could be altered to change the shape of the waveform.

DFWB.

[Graham G3ZVT]

Thanks David, next opportunity I get to put the set on the bench I will try both of those and report back.