Home made transformer

[Tim]

Hi all.

For some reason I have never had a go at making or re-winding transformers, but
have seen some other people doing it on u-tube so decided to have a go.
I have a couple of donor transformers that I can re-purpose. It looks reasonably straight forward once one gets ones head round the maths( never been my strong point)

Just a couple of questions though

One transformer I want to re-wind is for an Oscilloscope. One winding is 950 volts, which then gets doubled for the EHT.
This seems to be a problem area in other scopes using the same techniques from the same manufacturer.
I’m thinking of re winding the transformer without this winding, and making a dedicated tx just for the 950 volts.
Is this worth doing?

It doesn’t need to be big as required current is very small. That creates problems in itself of course. I could make it bigger to aid cooling but then my losses might be more than the actual power used.
Guidance please!

[Ed_Dinning]

Hi Tim, insulation is one of the bugbears of transformer design, especially if you sell it on it has to meet all sorts of regulations; even if only for your own use.

For some transformers, where the sec winding "floats" it needs good insulation between the low end of the HT winding and the pri & core.
This was a feature of the so called "admiralty pattern" transformers.

For a small dedicated transformer you will have a high t/v figure and lots of turns as well as lots of insulation (grade 2, double insulated wire helps). It will be innefficient, but the actual losses will be low anyway.

If possible the best method may be to use a slightly larger transformer and put your HV winding on last. Plenty of interwinding insulation and possibly about 3mm od "margin tape" to space the HV winding away from the bobbin cheeks.

Good luck, Ed

[Diabolical Artificer]

Having wound the odd tfmr with HV secs I've found it's essential that you treat the wire with kid gloves so the enamel doesn't get damaged,you need to make sure there are no kinks too. This is easier said than done if your winding a very high voltage low current winding IE thin gauge wire, go up a gauge or two if you have the space to allow for losses due to resistance, it's also easier to wind.

The problem when a novice winds a tfmr is finding out some of the specs of the materials your using & EHT materials are very specialist and usually V expensive. IE you'll find no voltage/insulation breakdown specs for tfmr copper wire, only heat and chemical specs. Try to use class H poly wotnot coated wire. I use kapton tape for insulation,it sticks well and is thin.

I'm about to wind a 1000VA 1500v toroid and have looked into things a bit, but I'll still be winging it to an extent. I'll wind several smaller voltage secondary windings, say about 170v AC making the windings aren't too tight and I'll coat with air dry tfmr vanish as I go, I'll also be testing as I wind. I'll be testing for pri/sec insulation as well as insulation between each wndg and each turn up to 10Kv.

If your going to use an EI core it might be best to use a bigger than needed core and bobbin, calculations no matter how carefully done have a habit of being wrong. Your first tfmr wind probably won't go to plan, you haven't five tfmr winds say, under your belt; any job is part calculation,part experience and gut feeling.

Lastly if your rewinding an old tfmr with a primary already wound you can work out if the core is roughly suitable by saying your efficiancy will be 90%, IE 90% of it's VA rating. Check the wire gauge of the primary to be sure with a micrometer, there are lots of amps, resistance per gauge X on line, use the "enclosed" "chassis" specs, not free air obviously. If you use an un-speced tfmr, IE no VA rating there is a free windows simple tfmr calculation software,see here - https://320volt.com/en/transformer-c...cin-hesaplama/ that'll work this out for you. A toroidal tfmr is good here, as there is no max space as on a bobbin and you can work out an unknown toroid core by weighing it and doubling it,then looking up commercial tfmrs of the same weight.

Am sure there are a few tings I've missed,oh yes....using an unknown tfmr core of laminations is always a gamble as you have no BH curves, that said you can still wind a decent tfmr for a non critical application and wind custom jobs with the specs you want ,with as good as if not better specs for less money than a commercial jobbie.

Good luck, Andy.

PS, I have a few scope tfmr's if decide not to bother.

[Electronpusher0]

How wedded are you to keeping original?
Another option is to wind the scope transformer without the HV winding and use a separate switch mode supply for the EHT.

Peter

[Wendymott]

I may be wrong but..... Surely mains derived EHT is not considered safe these days...if you have a problem of insulation (short) etc.. the impedance of the input is quite low and thus "high" currents may flow before any fuse fails.. The other obvious thing is electric shock...... I was always told.... keep away from the EHT....
Surely the safest way is to use as suggested above.. a dedicated high frequency eht generator, where the impedance is much higher, less current available to "kill" you.
A Mono TV Lopty or similar could be used safely.

[Tim]

Hi chaps.
Thanks.
The original transformer let out the magic smoke, so I may not be able to strip it down and use the existing primary. Manual says mains consumption is 100VA so I may buy a physically similar transformer, strip off the secondaries ( to work out the turns per volt) then wind on my desired secondaries, minus the 950 volts( 6,3v, 210v, 13-0-13, and 8-0-8) then repeat the process to provide the 950 with a separate tx( assuming there is space inside) both using a split bobbin.
The suggestion of using an SMPS is also a good one, ( and one I had considered)there is a German chap on u-tube who has a video about Cold Cathode lamp supplies. One of those would probably do the trick. Not sure if it, and it’s power supply would fit inside, but I could always bolt it on the back.

[Tim]

Scope is a Gould 4035 DSO

[Radio Wrangler]

With 50Hz derived EHT, you can never easily get as many volts out of a transformer as you'd like, so you resort to Cockroft-Walton multipliers. At 50Hz, this means a number of high voltage, quite large capacitors. These are getting hard to find, many contain PCB type oils (liquid cancer someone called them) and they store fatal amounts of charge.

High frequency EHT generators (with only small capacitors needed) have a lot going for them.
Use of the flyback boost technique reduces the number of turns/volt dramatically as well

I built scopes with mains EHT when I was a kid. Somehow I survived.

David

[Tim]

It’s an 80’s scope I think. It seems to be a souped up Gould OS300 with cursors & DS facility added.

[mole42uk]

There's an intersting article about the same problem here:

https://www.eevblog.com/forum/testge...ed-trsnformer/

[Tim]

Thanks Richard.
I had exactly the same problem. Mains input fuse blew instantly even with all secondaries
disconnected. I had a light bulb in series but it still smoked just as faulting was completed.
Judging by the pictures the TX has one bobbin & has layers wound on top of each other. Not a good option to start with when one has a high voltage winding. I may make a split/ divided bobbin to help things a bit.

[Tim]

On to my next question.
I wouldn’t have thought the EHT would requires much current. Probably microamps I would have thought sizing the secondary wire/size of a transformer for two or three milliamperes should be fine.

[kalee20]

Originally Posted by Diabolical Artificer

using an unknown tfmr core of laminations is always a gamble as you have no BH curves...

True. Although, if you design for 1.3T flux density, you're unlikely to have any problems.

Or: Wind a test winding of a couple of hundred turns, doesn't matter too much as long as you know the number. Apply a low voltage, measure the current drawn. Keep increasing the voltage and measuring the current until the current starts to increase disproportionately (the thing will probably start humming quite loudly then too). Reduce the voltage to 85% of this 'knee' voltage, divide by the turns, and you have your volts-per-turn figure.

The other points: Handling the wire like you're wearing kid gloves is important. Kinks are a no-no, unless it's a trial winding only, which will be run under supervision. Sizes below 40SWG (0.125mm) are tricky to handle and you really need a shaft through the wire reel centre, on smoothly turning bearings, with a spring-loaded snatch-absorbing arm and pulley, to avoid breakages. Finally, wind evenly, so that each layer is filled - you don't want turns of wire from a late layer slipping down to an earlier layer, or else you'll have many turns-worth of volts which the thin enamel will have to withstand.

Testing: if you have an audio amplifier, feed it with 100Hz sinewave and use it to drive your transformer, into a suitable winding, so that it's running at about double the voltage. If it survives this, it's good to go.

[Tim]

Thanks Kalee. I was going to do a similar thing, do primary of (say) 500 turns. Then wind on a secondary of ( say) 50 turns. I was going to apply 24 volts or so, and see what the resulting output was. Hopefully I could then find out the volts per turn.
I was wondering how to make sure I wouldn’t draw too much current. Now I know!
I hope to knock up a test transformer to get the 950 volts just to see if the scope works. It was working until I turned it on one day and the fuse went. I’m going to check for shorts etc. beforehand.
I shall probably end up using one of those little EHV supplies for the permanent fix.

I’ll probably insulate between each layer as I wind. It’ll be a good academic exercise anyway. I may well do a few practice low voltage ones just to see if everything works out.

[turretslug]

Originally Posted by Tim

On to my next question.
I wouldn’t have thought the EHT would requires much current. Probably microamps I would have thought sizing the secondary wire/size of a transformer for two or three milliamperes should be fine.

Beware that the required current can tot up somewhat- whilst the CRT beam current itself may only be a few tens of micro-amps at full brighness, more is lost in the acceleration process to the gun electrodes, then the brightness/focus resistor ladder typically takes a few hundred micro-amps, the regulator's feedback sampling chain typically a similar current. Then there's the correction coefficient for the peaky current taken by a doubler and the doubling itself. I recently had Ed make up replacement for a Hameg 'scope with a transformer winding feeding a symmetrical quadrupler for EHT, I reckoned that the winding rating amounted to around 7mA in the end.

[Tim]

I might have a tx with 450-0-450 about somewhere, that I can add 50 volts to with another tx to try and measure current. If not I’ll aim for perhaps 10–15mA in my home made tx.

[Tim]

I have been watching a few videos on transformer winding and there aren't as many as I thought. There was one where an Asian chap had used fibre glass board(exactly like PCB but without the copper) to make a bobbin. It's an excellent idea, and I suppose you must be able to get it.
I have some Paxolin sheet in various thicknesses, including some at about 1.5mm, which I think would be fine.
Presspan has also been mentioned, which I think is a special kind of cardboard. If that is the case can I use "ordinary" cardboard, if varnished?

I have some stiffish stuff used by picture framers for borders.

Comments/expert knowledge welcome!

[Uncle Bulgaria]

I think I have read normal cardboard will do - it is after all insulative. Good varnish impregnation is a must, I think.

Presspahn is not expensive, however. Their shop also has the type of fibreglass boards you mention.

When I was investigating transformer winding I read Manfred Mornhinweg's website with great interest: ludens.cl

His 'practical transformer winding' page is here, and there are more pages on electromagnetics if you look through the website.

[Diabolical Artificer]

With homemade cardboard bobbins the problem is they fall apart if you try and wind too tight I've found, you have to be very careful. I tried using copper clad board soldered making sure to cut a gap but then faced the problem of SC windings that went through holes in the board. But for that varnished copper clad board makes very strong bobbins and is easy to solder,you just have to remember to insert a gap so it isn't a shorted turn.

Bobbins available here - https://www.ebay.co.uk/itm/Bobbins-E...QAAOSwC6dZqxIQ & here - https://www.ebay.co.uk/itm/BOBBIN-CO...QAAOSwTe5fPpKH

Andy.

[kalee20]

If you have an insubstantial bobbin, you definitely need a good mandrel to support it!

A hardwood block, accurately cut to size, and with end cheeks to stop the bobbin flanges spreading, is very useful. With this, you can get by with thin card for bobbins (though I tend to use Nomex sheet, 7 thou' thick).

One matter: after winding, it is quite likely that you won't get the flipping thing off the mandrel. But don't panic: put it in an oven, raise to 120°C, leave for an hour, take out, put in freezer (in a polythene bag to stop condensation), leave a few hours, repeat... At one of the extremes, you'll find it will be loose enough to push out.