LOPT design

[renato_enca]

Hello. Some time ago I started to design a valve television, and now I faced the problem of designing/finding a suitable LOPT to work at 625 lines, 90º deflection angle and 8kV of EHT. I want to use a PL36 as the output valve, a PY88 as booster and a DY802 as EHT rectifier. I also have a ferrite rod and plenty of 0.05mm enamelled wire. Does anyone know how to wind the LOPT, or things like the turns ratio, inductance, etc?

[Andy Green]

Somewhere I have some practial Television articles from the late 1950s which, as I recall discussed home made LOPT's - I will take a look tomorrow, and scan and post the article if it appears relevant.

[Ed_Dinning]

Hi Gents, there were some articles on LOPT design in Mullard tech pubs in the 50's and in Electronic Engineering.
I am presently trying to produce a "modern" LOPT for the Bush TV22. This transformer has a laminated core and a primary that fits inside the secondary tube. It can use a PL38 as the op valve. One of the other forum members is presently trying the unit out and I will report later on how it went.
This new design uses Pyramid winding of the EHT section, not a wave wound Pie due to the difficulty of obtaining the correct wire.
The design can be simplified if 18Kv doides are used as then no highly insulated heater winding is needed. The turns on this winding are many more that would be expected from the transformer equation due to the peaky nature of the waveform and is best derived by measurements with a HIGHLY insulated true RMS DVM and a resistor in place of the heater (or you can just use Si).

In design terms most of the power from the valve goes to the scan coils. Some is used for EHT and the boost rail is recovered energy.
Generally therefore the valve Ra should be matched to the scan coils impedance through the turns ratio of the transformer. (This is not an audio application so Ra can equal Rl, consult valve curves to determine Ra)
The turns can be estimated by using the transformer equation:

Np= Vp/ 4.44B*F*Ae.
The figure 4.44 is for a sine wave and the frequency should be 10125Hz. Take B as 0.8Tesla (or lower if you want lower transformer losses). Ae is the cross sectional area of the core in sq METRES.
This is an engineering solution and may offend some purists but should give you a working design. A ferrite core will need a B of 250mT or less.
Then knowing the approx volts across the anode winding it is easy to calculate the boost tap and any other taps that are required.

There's your starter for 10! let me know how you get on.

Ed

[ppppenguin]

Originally Posted by Ed_Dinning

In design terms most of the power from the valve goes to the scan coils. Some is used for EHT and the boost rail is recovered energy.

I've never sat down and worked out a line timebase to that level of detail but I'm slightly surprised that the scan coils are the major power user. Obviously in an idealised energy recovery timebase the only power used is for the EHT. In a practical timebase power is dissipated in the resistances of all the inductors, iron losses in the cores and switch losses in the valves.

Probably easiest to get it going with a silicon EHT rect initially. You can always add a valve later. The heater winding was a common source of failure. In the absence of the correct cable the inner of solid dielectric co-ax was often used to make a new heater winding. If you don't have a true RMS meter you can roughly estimate the heater voltage by running an identical rectifier from known DC and comparing the brightnesses of the 2 heaters. I've no idea how accurate this would be.

[Andy Green]

Here are two articles, (parts 5 and 7 of 'Scanning and Synchronisation) from January 1958 and the other March 1958 Practical Television, which discuss winding your own LOPT. I don't have the February edition to hand, (not sure if I have it at all)although I will have another look for it tonight.
Apologies for the quaility of the scans, but I think you can read the article OK!

[newlite4]

Jeffrey, I was'nt aware that the heater winding could fail since it is only a few turns of insulated stranded wire. What could cause this failure?
Neil

[ppppenguin]

Originally Posted by newlite4

I was'nt aware that the heater winding could fail since it is only a few turns of insulated stranded wire. What could cause this failure?l

Probably heat. I'm sure the real telly menders in this forum will be able to give a more authoritative answer.

[Ed_Dinning]

Hi Gents. The early heater windings (TV22) were well spaced from the EHT and the coupling was poor, hence the empirical method for setting the voltage.
In later sets it was on top of the EHT winding and much fewer turns (sometimes with a series resistor) but there was a high level of insulation required (EHT volts plus max neg swing on rectifier anode). Once you exceed about 6KV there will always be a slow growth of resistive "leaders" through all insulation, ultimately leading to failure.

I used a simplistic approach for the valve loading in ignoring the EHT load. It should however yeild a workable solution.

The article look very interesting. There are also some designs published in Mullard Technical Communications fron the early 50's onward. I have some later copies, mainly on CTV design and SMPS; does anyone have earlier copies. They would form a great design resource.
There is also the Philips tech library books and Mullard "Radio Valve Manual" (Basic design for a 625 line ccir TV receiver) from 1958.


Ed

[renato_enca]

Many thanks for your answers. After doing some maths, I could estimate the turns necessary to wind the transformer. The ferrite rod I have is a long MW antenna rod (without the coils, of course). However, I have some questions:

1- While doing the calculations, should I consider Vp the total voltage in the primary (total boost voltage, before the EHT overwind), or just the supply voltage between the PY88 and the PL36 tap? (the 230V +B voltage), and then calculate the turns ratio to the other voltages?

2- To minimise insulation problems, should I wind the EHT overwind to supply, for example, 3000V and use a tripler to give the required EHT? I don't know how to wind the overwind as described in the articles that Andy Green posted, but I'd like to know if it is possible to do the overwind with layer winding, insulating plastic coating and insulating tape. It's always a risk...however, the one who makes an LOPT, makes two or three...and the EHT is not very high...

3- Is there any X-ray emission hazards if I use a EHT rectifier with this sort of voltages?

PS: If anyone wants, I may send the diagram of the expected taps of the LOPT (or the desired entire output stage).

[Ed_Dinning]

Hi Renato, a quick reply.

No X-rays at these voltages
Tripler is the best idea
Pyramid winding should be OK, say:
first layer 200 T, 2 layers polyester type 56 tape (Yellow transformer tape)
2nd layer 180T, 2 layers tape, layers symetricially spaces
3rd layer 160T etc, etc

Wave winding needs a special machine (not too difficult to get) but requires enamelled and cotton or silk covered wire (very hard to get)


Ferrite rod may not be the best material as high powers are involved and aerial rod is optimised for low power RF. Lams may do or look for some power ferrite parts
Ed

[renato_enca]

Hi everybody,
here's the scan of the experimental line output stage
A friend of mine gave me an old book talking everything about television. Unfortunately, it doesn't describe the line output transformer, but does describe the output stage and how things work.
It has a full schematic diagram of a working valve TV set, but the components have no values, so I gave my own values to the resistors on the diagram
This diagram was therefore transcribed from the book.

PS: HT+ = 230V

[renato_enca]

oops...after studying the origin of the CRT (an old Atari SM124 monitor...no good use for it, it has the supply transformer o/c), I discovered that the EHT is 12.5kV. Now THAT may give insulation problems...also, it needs 600V on the CRT's G2, and the focusing is electrostatic (don't know the voltage). The CRT model is 310KGB (or 310KA).

[renato_enca]

after some fiddling, i've built the line oscillator, and using Ed Dinning's formula I calculated the windings for the line transformer. after sealing every pin, I connected the line output circuit as described, and, with the fingers crossed I applyed power to the circuit. after some seconds of heater warm-up, a line whistle began to appear and the EHT lead began sparking (albeit a small spark), but after 1 minute the boost cap (a 0.1uf ceramic) began smoking and ended up giving a bang. the HT (250V) suddently dropped while it was burning, but it was so fast I didn't have the time to pull the mains cord. any clues? the valves' plates (PY88 and PL86) didn't go red, so it may be just the boost cap under-rated for the purpose, I think.

PS: The LOPT was wound in an aerial ferrite core.

[Ed_Dinning]

Hi Renato, Design details (comprehensive) on their way to Andy for scanning. It sounds like it almost worked, but I suspect that the boost cap was not up to it, a 1250V component is needed here. Possibly best to use a variac and bring the set up slowly, monitoring the voltage on a scope through a x10 probe.

Let us know how it works with a new cap.
If ypu were just using a ferrite rod, and not a closed magnetic circuit effieciency would be low and the waveform would be very peaky. The Bush TV 22 was a bit like this but was run at a much higher flux density than your ferrite rod.
What flux density were you aiming for? A ferrite rod probably has a saturation density as low as 100mT (if you can find it quoted for that type of material).

Best regards, Ed

[renato_enca]

good news...a friend of mine gave me this LOPT former...it was from a Philips set, which CRT had the heater o/c. While dismantiling the LOPT, I realized that the turns were not made with ordinary wire, but with aluminium foil wound concentrically over the former. Between each turn was a thin plastic layer to insulate turns. No doubt those LOPT's were so reliable. (I still own my 15 year old colour Philips, model unknown, the CRT is perfect)

[Ed_Dinning]

Hi Renato, foil winding is not a new technique, but has been used for many years in SMPS design, especially where high currents and low volts are concerned; I must admit I had not thought of it in a LOPT.
The other technique you might like to look at is the "diode split" method pioneered I think by Philips.

Ed

[renato_enca]

afternoon valve TV experiments report:
after winding the LOPT in the new former, the circuit has been reconnected. now, and after placing a new boost cap, the circuit powered up very well, and made a LOT more line whistle than the previous effort, and the EHT lead(used the coil from the old LOPT, before dismantling) began sparking about half an inch to a screwdriver held in proximity. however, the PL36's plate began to glow red. maybe it's best to wait for the design details.

[Ed_Dinning]

Hi Renato, it may be a mis match between the valve impedance and the scan coils, or it could be that the DC conditions on the valve are not correct. This may be best addressed by adjusting the G2 voltage. Short tem running of a power valve like this will do little harm.

Ed

[Andy Green]

Originally Posted by Ed_Dinning

...copy of the Electronic Engineering article from 1958 on the design of LOPTs with booster diode winding; it has lots of good information. Ed

OK here is part 1 of the article - apologies for the slightly scratchy quality, I will try and get part 2 a bit better - It is quite legible though.

[Andy Green]

Ok, here is part 2 of the electronic engineering article.
I will have to read it and see if its principles could be applied to design a 405 line LOPT for a modern colour set. Any thoughts anyone?