Continue to Site

Welcome to our site!

Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

  • Welcome to our site! Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.
Status
Not open for further replies.

Misterbenn

Active Member
Background:
A friend is looking to do some fractal burning (google it) (edit: also called Lichtenberg figure)
and has salvaged a microwave oven transformer (MOT) for this purpose. However he likes to think big and isn't satisfied with the 2kV output so has asked me to look at increasing it.
As a results I've done some research into transformer design and have a few questions that i'm hoping someone here can help with.

At this point i want to save you all the effort of pointing out how dangerous high voltages are (and i'm expecting a flurry of don't do this type comments). I'm aware of the risks of electrocution and the very small amount of current it takes to kill a person. I should mention that I've worked in HV (>20kV) test houses for many years and am fully aware of lab practices for safe testing of HV equipment.

Discussion:
I've no one specific question so this is more of a thought train written down and i'm interested if anyone has anything to add or corrections.

In my initial steps for designing the transformer I wanted to calculate the number of turns required on the primary. To do this i used the EMF equasion
E=4.44 f N_{P} B A
Inputting the mains frequency, voltage and the maximum flux density and area of a E core I have in my bits box:
E = 230 Vrms
f = 60 Hz
B = 1.5 T
A = 4.3 cm^2 = 0.00043 m^2
and re-arranging to get the minimum number of primary turns gives me 1338 turns.
This seems like a huge number of turns and wont fit on my core unless i reduce the wire size (and thus max power). But i know other mains transformers exist in smaller sizes, MOT for example. So how do they do this? the EMF equation seems very clear that you need a sizeable core if you don't want it to saturate at this voltage and frequency.

So it got me thinking ... how do they make a MOT so small without the core saturating.
Answer; An unloaded MOT will saturate and losses skyrocket, but in a microwave application the transformer secondary is always loaded and this reduces the flux density. So you must never use a MOT unloaded. see links **broken link removed**
http://www.qsl.net/kh6grt/page4/xfmr/xfmr.htm

I found this interesting but then thought "why don't wall wart transformers saturate". At the moment I can only think that they must have a large number of turns of small wire on the primary (>1500 turns). And that this is allowable due to the low power of the transformer.

The only example of a high power mains transformer that I can think of is a construction site transformer and these are fairly chunky and expensive.


So back to my friend and the Fractal Burning. If he wants to use a MOT or several in series he must make sure they are always loaded and also limit the arc current.
If we want to build our own transformer using the core that i have we will probably need to increase the frequency via an inverter to at least 1kHz.
If we are dead set on a mains transformer we need a much bigger core, or some way of calculating how much loading is required to keep it out of saturation.

As i said earlier I don't really have a question, this is just my thought process in my attempt to build a mains to several kV transformer. If anyone has any experience or knowledge they would like to share i'm interested

Another question do Lichtenberg figure need AC or can the same be acheived with DC?
Related threads / Links
 
Last edited:
One alterative to using MOTs or trying to re-wind one would be to use a neon sign tranformer. Those can easily output in the tens of kV. Another alternative if you want to work with higher frequencies would be to use a air-gapped transformer core like a flyback transformer. I'm not too familiar in this area, but feel free to look it up. I believe there are some resources online that would do a much better job of explaining it than I could do here.

Also, I think this post might be better moved to the "High voltage" subforum? I'll defer to the mods on that front.

It sounds like you have some experience and have done your homework on this, so I'll spare you the full-length "this is dangerous" spiel that I would normally feel obligated to give. :p However, seeing as this is an open forum and this post may be viewed by others in the future: I recommend that anyone else who reads this post and wants to try it be safe and know the proper safety precautions for working around high voltages.
 
Thanks JLNY, I'll take a look at neon sign transformers. I'm sure i could just buy a suitable transformer but i'm using this problem as a thought experiment so i can increase my understanding of transformers.

At the moment I want to see if there is an analytical way to evaluate the reduction in flux density due to secondary loading. It might be the case that i can use the core I already have provided it is always >50% loaded. It seems this is how MOT transformers achieve size reduction, so i'm interesting in seeing how that calculation is done.
 
Big Clive has a YouTube video on this in which he talks about safety (for those readers who may know less than you do, or value their life) and about alternatives to the microwave oven transformer. I don't recall all the details but there may be some useful information and Clive is very entertaining.
 
Big Clive has a YouTube video on this in which he talks about safety (for those readers who may know less than you do, or value their life) and about alternatives to the microwave oven transformer. I don't recall all the details but there may be some useful information and Clive is very entertaining.

Here is the video for reference. I also had this in mind when I wrote my original post.

 
Last edited:
Thanks both,

I'll add a link to the other thread in my original post for anyone else finding this topic.

Interestingly in the "wood art" thread KMoffett mentioned that he thinks higher voltages give finer detail. It might be an interesting experiment.
 
Hello,

Yes the number of turns can seem high but the frequency is low and that is one of the main factors that makes for a large number of turns. The other is the core size, which i see for this one is small, being less than 1 square inch. In fact, 4.3 square cm isnt even 1 square inch, and that sounds very small for a microwave oven. If you look at a transformer for a higher power rated microwave oven you should find one with a core maybe twice that, which right away reduces the number of turns by a factor of 2 (500 vs 1000 turns for example).

At 120vac with a 1 inch core (6.45 square cm) the min number of turns without any safety factor is 349, so at 240vac that would go up to around 700 turns and just a little less for 230vac, and with the safety margin and stacking factor that would bring it up even higher.

Straight from an old Magnetics Inc. handbook for a 1 inch square EI lamination stack:
A=1 in^2=6.45 cm^2
F=60 Hertz
Bmax=(58.1 * 10^3)/(K*N) Gauss per volt (K is the stacking factor roughly 0.8 to 0.95)

The only way around this high turns count dilemma is to move to a higher frequency where we first rectify the mains, filter it, then chop it up into 10kHz or higher. That's when we see a significant reduction in the number of turns required. Even if we only went up to 600Hz we'd see a 10 to 1 reduction of turns, so instead of 350 turns we'd only need 35 turns! Unfortunately the metal used for EI cores that run offline probably wont work well at 600Hz so we'd have to get a different core material or only increase the frequency a little.

If you can get a hold of a bigger used microwave oven you will find a better transformer core to start with, and it will already have primary turns.
 
Last edited:
Thanks MrAl useful info. So do wall warts just use very small wire on the primary in order to get the compact size? They must have a very high number of turns as well.

I work in aircraft electrical design and that is usually 400-800Hz 3 phase, which explains the reduced size of the magnetics. I'd mistakenly thought a 8kVA 3phase transformer designed for 115Vrms 400Hz would easily be re-purposed to a ~100W mains transformer. I understand the issues much more now, but its all good learning!

I've attached a photo of the core i'm working with, salvaged from my works scrap bin. The cross section is actually 1.1in by 0.8 in (2cm by 2.8cm) but with a stacking factor of 0.76 which gives a core area of 4.315 cm^2
 

Attachments

  • IMG_20170306_105049_small.jpg
    IMG_20170306_105049_small.jpg
    23.1 KB · Views: 335
Hi,

Very interesting, i had worked on a few aircraft converter designs which were 400Hz. They were a little noisier :)
Yeah for aircraft a big factor is the weight.

It's unfortunate that three phase cores have an effective magnetic area equal to the center leg area, even though there is a lot more steel involved. That means they dont make really good single phase cores unelss you dont mind the extra weight. I suppose if you wind on the outer two legs that lets you fit more turns though, for what it's worth (two coils instead of just one allow 2x turns).

Yes the wall warts will have much thinner wire on the primary because they seldom have to handle very much current on the secondary. If the secondary is 1 amp and the output is 12v, with a 120vac input that would mean around 100ma on the primary which means light gauge wire can be used. So it's the combination of the low output current and low output voltage that allows a lighter gauge wire on the primary. You might note that the higher current wall warts (like 4 amps output at 12v) are made with a different technology, a special kind of converter that uses the rectified and filtered line voltage to power a very small high frequency transformer. The rectified DC is chopped up into a high frequency 10kHz or higher. That's the secret to better wall warts. Of course if you need an AC wall wart i guess you are stuck with the old technology with just a transformer inside.
 
I'm glad that you work with 400 Hz magnetics and fully understand the difference in core material and construction with respect to a 50/60 Hz one.

The problem will be the winding insulation techniques. You may have to use triple-insulated magnetic wire, perhaps kapton film between layers, and then enclose everything in RTV.............. maybe an oil-filled enclosure.

Why don't you try a Tesla coil? It is easier to generate very high voltages with them.
 
Thanks MrAl, it's good to have someone confirm my suspicion. I've wracked my brain to see if there is a better way to use the geometry of the core I have but no joy so far.

Schmitt trigger, I'd thought of a Tesla coil but my friend wanted a solution with minimal effort. When I first found the core in scrap I thought I'd hit the jackpot. I've also considered voltage multipliers but they give DC outputs and it looks like fractal burning needs AC. That's my next line of enquiry 'can you perform fractal burns with DC?'
For insulation I was going to use kapton tape with a 50% overlap on the core and between layers and submerged in oil. Anyway that's academic at the moment.

Actually in my job I design 3ph inverters for power conversion and aircraft motor applications so am quite familiar with +10khz switching inverters. We rarely use transformers with the exception of the core I found in scrap which was a 9phase autotransformer.
 
Hello again,

Lastly, if you happen to have a second core of the same size you can stack them side by side and create a single core with area two times the area of one core. Maybe with some means to keep them from vibrating.
 
I tried it today with the one high voltage diode that was on the MOT when I removed it. Half wave, not DC. It worked but not very well.
 
Somewhere on you tube (I cant find it at the moment!), they use a arc welder as the transformer. Might be a fair bit safer? the current should be high enough to burn the wood if you use pretty salty water to get a decent current path. The welder gets hot quickly though, so you have to take it slow. BTW dont use welding rods, use a nail and the earth clip.
 
Status
Not open for further replies.

Latest threads

New Articles From Microcontroller Tips

Back
Top