DIY variac / General Science and Electronics / Forums

Forums

4hv.org :: Forums :: General Science and Electronics

« Previous topic | Next topic »

DIY variac

1 2 next

Move Thread

LAN_403

Kizmo

Sat Oct 27 2007, 02:22PM

Registered Member #599 Joined: Thu Mar 22 2007, 07:40PM
Location: Northern Finland, Rovaniemi
Posts: 624

Im looking for information about variac (aka. autotransformer) design, i need device which could handle at least 25 amps 0-230VAC. Mechanical construction and winding is not a problem but i didn't found any articles or other information about variac design. Things like how big toroidal core i should get and how many turns etc. I have one 1.9kVA toroidal transformer which could be used for this project. It's outer diameter is 200mm and height 85mm.

Suggestions?

Marko

Sat Oct 27 2007, 02:38PM

Registered Member #89 Joined: Thu Feb 09 2006, 02:40PM
Location: Zadar, Croatia
Posts: 3145

Finn built a nice variac, although it looked like great pain to do:

I would personally never have nerves to wind something like this and that many times.
Note that you need to pull entire spool through the middle EVERY time!

I wonder If this amount of boring work could be eliminated somehow...

I thought about, if you etched two PCB's in ring shapes, one with contacts and other just with interconnections, and simply use pieces of wire in and out soldered through boards, or something in that style. That would allow larger contact surface than copper wire.

Other thing is to use biggest possible core to minimize total number of turns you need to wind.

Now one thing I never understood about commercial variacs, they usually have a graphite contact which simply looks like it's way too large and shorting at least two turns together. Why is that alright? How doesn't it create a shorted turn?

Bauerb2

Sat Oct 27 2007, 04:27PM

Registered Member #973 Joined: Tue Aug 28 2007, 07:32PM
Location: Pittsburgh, PA
Posts: 66

yeah, i could never have the patience to do such a project. i already find it a huge pain to wind small 1 inch toroids, rather some giant one for a variac.

I have noticed as well the contact width covers almost 2 and a half turns on the variacs i have bought. Not sure why though..

I suppose making your own would be pretty much straight forward. The rotating contact may be hard to perfect, but as for the windings, it'll just take a while.

Best of luck to you.

-Andrew

101111

Sat Oct 27 2007, 05:02PM

Registered Member #575 Joined: Sun Mar 11 2007, 04:00AM
Location: Norway
Posts: 263

You also have the soilid state ones: Steven Connors Here and Here and a heavy duty one for his OLTC 2 Here

Marko also made a "dimmer" here

Carl Pugh

Sat Oct 27 2007, 06:31PM

Registered Member #1064 Joined: Tue Oct 16 2007, 05:04PM
Location:
Posts: 42

Don't try to build a variable transformer.

However if you do decide to build a variable transformer.
The brush acts like two very low quality diodes in parallel, shorting the two turns it contacts. Operate at about 0.8 volt per turn. For 230 volt this would be 287 turns. Volts per turn that can be used also depends on type of carbon brushes are made from. The highest volts per turn for brushes was 0.8 volt per turn when using Linde type AY carbon. This was many years ago and Linde may no longer make carbon brushes?
Use a silicon steel core and size it for 15 kg or use a larger core if required for mechanical strength.
Use silicon steel laminations, don't use a torroid.
Apply epoxy to hold conductors that brush rides on in place.
Sand epoxy and wire coating off where brush makes contact.
Coat where epoxy and insulation was sanded off wire with graphite coating.
Build mechanical assembly.

It's much more difficult to build a variable transformer than a fixed ratio transformer.

Good Luck

Marko

Sat Oct 27 2007, 07:08PM

Registered Member #89 Joined: Thu Feb 09 2006, 02:40PM
Location: Zadar, Croatia
Posts: 3145

The brush acts like two very low quality diodes in parallel, shorting the two turns it contacts. Operate at about 0.8 volt per turn. For 230 volt this would be 287 turns. Volts per turn that can be used also depends on type of carbon brushes are made from. The highest volts per turn for brushes was 0.8 volt per turn when using Linde type AY carbon.

Hey, need to solve this now!

I dug around but found nothing concrete, apart from this

So what's it? I never knew what kind of a ''diode'' could be formed out of graphite?

What's with anisotropic resistance?

It can't be that it's just high-enough resistance because I really think it would make the brush cook at several tens of amps we may pull through it.

Even a watt of dissipation in such small volume is problematic.

Kizmo

Sat Oct 27 2007, 08:55PM

Registered Member #599 Joined: Thu Mar 22 2007, 07:40PM
Location: Northern Finland, Rovaniemi
Posts: 624

Im thinking that maybe i don't make 'real' variac but instead i take large UI core (about 4kVA) and make one 230V primary winding, 46x 5V secondary windings in series and heavy duty switch to choose the output voltage. Same kind of thing can be found from cheap unregulated adjustable power supplies. It will be PIA to wind and wire but changes for success are higher. 5V steps are fine enough for controlling pigs or mot stacks etc..

Sulaiman

Sat Oct 27 2007, 09:56PM

Registered Member #162 Joined: Mon Feb 13 2006, 10:25AM
Location: United Kingdom
Posts: 3140

If it's just for pigs & mots couldn't you use a variable inductance (variable airgap)
or a saturable reactor in series - a lot easier than a variac or tap-switching I suspect.

mikeselectricstuff

Sat Oct 27 2007, 11:25PM

Registered Member #311 Joined: Sun Mar 12 2006, 08:28PM
Location:
Posts: 253

The thing about the carbon brushes is simply resistance - the carbon has a relatively high resistance, so the current flow from the relatively low turn-to-turn voltage isn't a problem. Spanning a few turns is simply to make it ride smoothly - if it were thinner than the winding pitch it would bounce up & down as it crossed each wire, and may also cause arcing as it came in & out of contact - being in contact with several turns means there will always be a few current paths from the wire to the wiper, so you never get the breaks that cause would arcing.

ISTR reading a while ago that graphite has different resistivity in different axes due to the atomic structure - I don't know if the effect is significant, and if so whether brushes take advantage of this by being made so the most conductive axis is perpendicular to the winding.

Even a watt of dissipation in such small volume is problematic.

But remember that it is intimite contact with a reasonable mass of copper (the windings and the brush holder) ,which will provide significant heatsinking

As for winding, you could probably improvise a toroid-winding machine with something like a bicycle wheel rim as the intermediate bobbin, which might take some of the pain out of the job.

In case you've never seen one, this video gives a rough idea of how a toroid winder works.

The large ring has a channel cross-section, and can be split to allow it to be 'threaded' through the core. The correct amount of wire is wound onto the ring, and then unwound from the ring onto the core.
One detail I'm not sure of is exactly how the wire comes off the ring onto the core without losing tension or unravelling - would be interested to see a diagram of this if anyone has one...

ragnar

Sun Oct 28 2007, 03:19AM

Registered Member #63 Joined: Thu Feb 09 2006, 06:18AM
Location:
Posts: 1425

Mike touched on the important point that while you might imagine the carbon brushes as causing shorted turns in your variac, the potential difference between any two turns is only a fraction of a volt. Variacs are pretty clever by design; most people when thinking about how they work don't realise it's not as simple as a tapped transformer.

See, you might imagine the brushes could be only as wide as a single turn of copper, but then you'd realise that you'd still have to space the turns to stop the brushes shorting turn-to-turn. If you look at the interleaved winding styles of big variacs, you can see that they use plenty of turns to minimise the potential difference, turn-to-turn, so that you don't get a shorted turn melting itself off the core =P

A variac simply wouldn't work with copper-to-copper. It would arc and sputter all over the place. Similarly, a variac which had large gaps between the windings (so that two were never connected together through the brushes) would give you a pretty horrendous output, spiking up and down from zero whilever you were adjusting the rotor.

If somebody has a variac handy, try counting the turns on it. Looking at the volts-per-turn, you can calculate what the dissipation will be with e.g. a turn "shorted" to an adjacent turn through brushes with a resistance of e.g. 0.1 ohms. Then you can look at the dissipation when an e.g. 5A, load is drawn between the tap and ground, and the dissipation that would cause in the brushes.

The brushes aren't bolted to huge hunks of steel for nothing

The best way to go really is to just buy one from Variac or second-hand off eBay.

1 2 next

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob