Building a Hybrid Toroid Transformer

[Matt Watts]

Been a tough battle so far, but I did manage to make it to my first milestone--I have molds.

Going to order some DevCon 10120 today and try my first pour next week.

Thinking about what may be a suitable lubricant to dust the molds with prior to pouring.  Any suggestions?   Pam maybe?

[TinMan]

Been a tough battle so far, but I did manage to make it to my first milestone--I have molds.

Going to order some DevCon 10120 today and try my first pour next week.

Thinking about what may be a suitable lubricant to dust the molds with prior to pouring.  Any suggestions?   Pam maybe?

Nice work on the molds Matt.

I would use PVA release agent. It comes in liquid form,paints on to the molds with a brush,dries like cling wrap,and dissolves with water.

Brad

[Grumage]

Been a tough battle so far, but I did manage to make it to my first milestone--I have molds.

Going to order some DevCon 10120 today and try my first pour next week.

Thinking about what may be a suitable lubricant to dust the molds with prior to pouring.  Any suggestions?   Pam maybe?

Dear Matt.

Making moulds has been almost a life long career!  .  I would suggest a proper barrier as the Devcon is more of a paste than a liquid. I would suggest using cling film between the mould and the castable just to make sure it doesn't adhere to it.

Cheers Grum.

[Matt Watts]

Thank you guys for pitching in.  Watching this rookie make and pour molds ought to be a hoot for you guys.   

So I ordered some SprayOn MR311 mold release agent along with the Devcon.

Do you think I should still press-in a layer of cling wrap Grum?  Spray both the mold and the wrap prior to pouring?

I'd much rather have a bit of distortion in the casting than have the casting stuck in the mold.

This HDPE is pretty solid and could probably take a hard bop, but I sure don't want to have to dig the resin out or cut a new mold.


Future looking, I may choose to mix a ferrite (or other magnetic) powder with resin depending upon how well/poorly the Devcon performs.  What is the viscosity limit for such a mold?  I'm sure if it's too thick it will never fully fill-out the mold, then the only way to get the stuff in there would be to inject it.  My hope is that gravity and a simple air bleeder hole in the top cavity will be sufficient.

And speaking of magnetic powder, is there any place you know of that sells small quantities of this material?  I bumped into a fellow by the name of Daniel Donatelli that can get me whatever I want as long as it comes in a shipping container.  I suppose if this gizmo does what we want I can consider a 50 ton purchase, but for now this is still strictly experimental.

[Grumage]

Dear Matt.

It's been many a year since I last used Devcon but I'm sure it resembles putty, you might find you will have to press the inner toroid pretty firmly. May be Brad could fill in the blanks ?

And yes I would employ both " belt and braces " although HDPE is pretty slippy the last thing you want is a damaged mould.

Cheers Grum. 

[Matt Watts]

I discovered there are two types of steel impregnated Devcon, 10110 which is a putty and 10120 which is called "liquid steel".  Based on the mixing ratio, I tend to agree even the liquid form probably won't be even as thin as syrup.  I really didn't want to use putty like Brad because once it's set, the inner core cannot be reused for anything else.

My hope was to make a clamshell like outer core where I can try different inner cores, different winding ratios, etc.  Once a fairly good combination is found, then I can try different outer core material.  Surely a drop-the-inner-core-in, place the top-half of the outer core on, wind the primary and test, is the fastest way to see what the capabilities are.

Actually getting pretty excited about this device.  Mainly because I worked on a similar concept months ago and got stuck on the geometry.  What Brad came up with reopens the door.  Getting this to a point where some serious power numbers can be evaluated is the kind of stuff I live for.

[TinMan]

Dear Matt.

It's been many a year since I last used Devcon but I'm sure it resembles putty, you might find you will have to press the inner toroid pretty firmly. May be Brad could fill in the blanks ?

And yes I would employ both " belt and braces " although HDPE is pretty slippy the last thing you want is a damaged mould.

Cheers Grum. 

Yes,devcon is a thick putty-much like pottery clay.
Matt. I hope you put a small positive tapper on the wall of your mold's?,so as once the bond between mold and product is broken,the product will fall out easily.


Brad

[Matt Watts]

Matt. I hope you put a small positive tapper on the wall of your mold's?,so as once the bond between mold and product is broken,the product will fall out easily.

Unfortunately, this is going to take some sanding which I've already started.  To my discovery, when HDPE is cut, the material left behind tries to warp into the missing space.  In effect, the current taper is actually backwards from what I need.  A quick runout clearly shows it.  So I'll hand sand and polish for a while until I have a few thousandths of taper.

I knew this wasn't going to be easy, but it's still cheaper than having something professionally made, plus I get to learn why we call them professionals.

[TinselKoala]

Ugh, HDPE is not my favourite material to work with. Not as bad as UHMW but still.... ugh.  Try Delrin instead: it will be dimensionally stable, and you can machine it to a very smooth surface easily. Probably won't even need any mold release compound.

[Matt Watts]

Ugh, HDPE is not my favourite material to work with. Not as bad as UHMW but still.... ugh.  Try Delrin instead: it will be dimensionally stable, and you can machine it to a very smooth surface easily. Probably won't even need any mold release compound.

Delrin would have been my first choice, but it's a bit pricey and I didn't see the thickness I needed.

I have never tried UHMW before, but the specs appear quite similar to HDPE--isn't too awful expensive and I can get all kinds of thicknesses.  Says it's self-lubricating too.

For CNC routing of plastic, I found an Amana Tool single flute spiral bit that works magic even on my rather sloppy home-built CNC machine.  This bit is so hard and so sharp it cuts plastic like butter.  I only wish they made a variant that had some taper to it.  If they did, I could match the taper in my CAM program and do pocket cuts for molds with ease.  Some day maybe I can get a "real" mill and make this kind of work a hobby that pays for itself.  'Til then, use what I have or can afford.  Russ keeps hounding me to get or build a 3D printer, but I just can't see myself tinkering with something like that.  I'd need to have 28 hour days if I were to jump into that fray.

[JimBoot]

Starting a rewind tonight. Got an iron outer core a metglass inner. Unfortunately my SG only goes as high as 1mhz but we'll see what happens

[Grumage]

I discovered there are two types of steel impregnated Devcon, 10110 which is a putty and 10120 which is called "liquid steel".  Based on the mixing ratio, I tend to agree even the liquid form probably won't be even as thin as syrup.  I really didn't want to use putty like Brad because once it's set, the inner core cannot be reused for anything else.

My hope was to make a clamshell like outer core where I can try different inner cores, different winding ratios, etc.  Once a fairly good combination is found, then I can try different outer core material.  Surely a drop-the-inner-core-in, place the top-half of the outer core on, wind the primary and test, is the fastest way to see what the capabilities are.

Actually getting pretty excited about this device.  Mainly because I worked on a similar concept months ago and got stuck on the geometry.  What Brad came up with reopens the door.  Getting this to a point where some serious power numbers can be evaluated is the kind of stuff I live for.

Dear Matt.

That sounds like a plan, cast two halves of the Devcon on a really flat surface so you can drop in your inner Torroid.

With regards to your cutter, Tungsten Carbide? You can put a taper on quite simply if you have a Diamond cutting disc to hand, those used for ceramic tile cutting work really well. With my background I have a dedicated bench grinder fitted with a rather expensive Diamond in Bronze lapping wheel, the TC just doesn't stand a chance! 

Cheers Grum. 

[ion]

As I mentioned before, it would be a good idea (if we want to get all scientific and all that  ) to measure a cast toroid for AL value and plot high frequency loss of the cast material to get some kind of a baseline.

Maybe I'll order some of the putty and see what I can find.

Regards

[Matt Watts]

That's a good idea ION.  You may have to walk me through the procedure.  Thinking I should run a validation against my inner core that already has factory specs:
http://www.cwsbytemark.com/images/OD%20740.pdf

CK740060 is stamped on the core; chart says AL of 206.
If I can produce a number within 5% then the same procedure can be used for:
  1.  the outer core alone
  2.  outer core with embedded core testing primary
  3.  outer core with embedded core testing secondary

Those four numbers should give us a hint what might be happening inside there.

[TinMan]

As I mentioned before, it would be a good idea (if we want to get all scientific and all that  ) to measure a cast toroid for AL value and plot high frequency loss of the cast material to get some kind of a baseline.

Maybe I'll order some of the putty and see what I can find.

Regards

Ion

Below is a pic of the new inner secondary-along with info.
The inductance is quite low on this one,although i did not measure the inductance on either of the other two before i cast the outer core around them,so i have no reference to go against ATM.

Anyway,what kind of testing should i do,or you would like me to do before i cast the outer core around this one.
I wound this one as neat as i could,but still have to try and space the outer loops a little better yet.


Brad.

[Chet K]

TinMan
Re: Testing cores
To Note....ION is presently redoing his internet service due to Clear channel dropping all subscribers .
may be spotty responses for a few days yet.

just passing it along..

Chet

[ion]

TinMan
Re: Testing cores
To Note....ION is presently redoing his internet service due to Clear channel dropping all subscribers .
may be spotty responses for a few days yet.

just passing it along..

Chet

Yes my internet is dialup right now and not working that well, so I will let Smudge and or Verpies take over for the initial testing. If you have a data sheet on the core, no need to calculate the AL value as it should be on the data sheet.

http://www.dextermag.com/resource-center/magnetic-core-calculators/AL-inductance-calculator.

For high frequency loss factor, it is a bit more tricky, but you can get a rough idea by sweeping frequency and noting the amplitude across a current shunt resistor, but this is only good at high frequencies, and then capacitance will wreak havoc with the reading. There are netter ways, perhaps others can suggest. Another crude method would be to measure the heating value of the core vs. frequency over a slow sweep.

Gotta run right now.

Cheers, later
ION

[Matt Watts]

Hopefully this is a good reference:
http://meettechniek.info/passive/magnetic-permeability.html

[wattsup]

@tinman

Nice rebuild but you are basically making the same mistakes.

wattsup

[TinMan]

@tinman

Nice rebuild but you are basically making the same mistakes.

wattsup

As this is my design,and we are still working on what is going on with it,how is it that !im! making mistake's ?

[gyula]

Hopefully this is a good reference:
http://meettechniek.info/passive/magnetic-permeability.html

Hi Matt,

From Brad's measurement (200 turns=120uH on his given sized core) it is possible to use software to arrive at the magnetic permeability and the A_L inductance factor of the core he made.
I would refer to the free Mini Ring core calculator (I mentioned this yesterday at the OU forum, see  http://overunity.com/15395/partnered-output-coils-free-energy/msg465294/#msg465294 ) and TinselKoala also drew attention to another free program, the Coil32 (see http://coil32.net/ and for this latter program Windows, Linux and Android versions are available).

From these programs the magnetic permeability comes out as u=2.2 and A_L is 3 nH/N². (Of course, this latter A_L value comes also from the link ION referred to in his above post.)

Naturally, Brad's inductance meter uses a certain (and fixed) test frequency which is normally anything between 80 to 120 Hz, maybe some hundred Hz, this may be included in the specifications. This test frequency cannot be changed for a given L meter, unless it falls in the high-priced category where the test frequency choices can be many. At the OU forum a good suggestion was to check the test frequency with an oscilloscope while measuring an inductance (in case it is not included in the user manual).

The link you gave can surely be useful to measure the permeability of a core, the advantage is that the test frequency can be chosen at will. The test does need a good, non inductive resistor though.

Gyula

[ion]

Dear Matt and gyula

Thanks for your added links to nail down the permeability of the core with and without the putty.

I also believe it would be a good idea to check for high frequency losses to get an idea of the  of the core upper limit or knee of the core. This should not be that difficult and a sweep is not needed as just a few test frequencies should tell the story.

I would put some ratio (to the primary turns)  on the core as a secondary, add a low value load resistor that reflects the generator impedance transformation and plot power drop off with frequency.

The method outlined in Matts link also covers HF losses.

regards

[Matt Watts]

As this is my design,and we are still working on what is going on with it,how is it that !im! making mistake's ?

I had the same question Brad.

I suppose one could say we are all making the same mistakes until we hit on the perfect combination, in which case it's mind over matter--that being, if you don't mind, it don't matter.  

One thing I can say for sure about the construction:  If your casting of the inner core works well, this is clearly the way to go.  Making a complete HTT using only magnetic powder, resin and copper wire is a slick idea.  Basically disposable devices--they either work well or go in the bin.  No big loss either way.

My approach using off-the-shelf inner cores and a casted two-piece outer core is focused on being able to change-out the inner core.  My feeling is there may be an optimal ratio of core material and geometry at play here.  If with some testing it becomes apparent the outer core must be made of something more substantial then powder and resin, I'll look towards a manufacturer building me a custom hollowed out core.  Hoping this isn't necessary, because I'm certain it will lead to a cost prohibitive solution.

What I am certain of is we will have some answers in the coming weeks.  Whether they be good answers or not...

Time will tell.

[TinMan]

I had the same question Brad.

I suppose one could say we are all making the same mistakes until we hit on the perfect combination, in which case it's mind over matter--that being, if you don't mind, it don't matter.  

One thing I can say for sure about the construction:  If your casting of the inner core works well, this is clearly the way to go.  Making a complete HTT using only magnetic powder, resin and copper wire is a slick idea.  Basically disposable devices--they either work well or go in the bin.  No big loss either way.

My approach using off-the-shelf inner cores and a casted two-piece outer core is focused on being able to change-out the inner core.  My feeling is there may be an optimal ratio of core material and geometry at play here.  If with some testing it becomes apparent the outer core must be made of something more substantial then powder and resin, I'll look towards a manufacturer building me a custom hollowed out core.  Hoping this isn't necessary, because I'm certain it will lead to a cost prohibitive solution.

What I am certain of is we will have some answers in the coming weeks.  Whether they be good answers or not...

Time will tell.

I have a feeling this one will not work well with the low perm core,but i will continue on just so we can see what direction is best.

My gut feeling ATM is that both inner and outer cores should be of high permability . It would be nice if we could have a ferrite inner core wound,and then cast a high perm ferrite core around that as our outer core as well. Can i buy ferrite powder from some where ?.

Im off to do some research to see how they make ferrite toroid cores.

[wattsup]

@tinman

Yep - what I mean by making the same mistake is pretty simple. Arbitrarily winding that secondary coil full around is one. This is the most critical point in the build because once it is encapsulated with an outer core, there is no turning back. This would require that this first stage be taken slowly and investigated as much as possible. If you are really interested I could spend the time required to give you some simple tests that would help hone in on the best method but it takes time and requires commitment. This means you cannot just blow it off if the first results don't pan out like you envisioned like I really feel you did with your zero force build and like so many do after they go to all the trouble of building something to test and throw in the towel after a few token experiments, knowing full well that while the device is on the bench you could learn so many more side effects with very little extra effort. That's why I keep on pushing guys left and right (off tangent as @Smudge put so well).

It's like inspecting a device that is in the center of a dark room. One flashlight shines on it and you get a pretty good idea of what it is but with two or three flashlights placed around the device , you will get a much better idea.

This would be a pretty crazy and difficult task because there are so many variables to consider and the intended topology closes certain doors of effect while it opens others but to understand this would take some time, if you have the stomach for it and realize the results may take time to mature into a solid understanding (good or bad) (if such exists), I would be willing to take up the challenge as a part time project that would still leave you and me open for our other shenanigans - hehehe.

I had taken six months off OU and regular work to finish the outside of my garage that I built two years ago plus I completely redid my home while super insulating it to match the garage. It's all done and I will be making a new time lapse video (garage video already on youtube at https://www.youtube.com/watch?v=kbLyewMeBt4 ) of the total works but I am now free to return to OU research, but this time, I will be integrating my Spin Conveyance construct in all my works and I am more then sure you guys will start to see the effects on our bench in a new light because if you guys continue like this, you will be turning in circles for more years and years.

Some of the things you may require are as follows;

1) Variac to be able to experiment with low voltage 50 or 60HZ AC straight from the mains.
2) Computer sound card connected to a stereo amplifier*. (*Something that you don't mind blowing up - hahaha)
3) Frequency generator software to push AC sinewaves into the coiling via the sound amplifier outputs and permit adjustment of frequency up to 20khz while playing with left/right phase and amplitude.
Get the software from my ftp site here and enjoy this new flexibility: http://purco.qc.ca/ftp/Software/frequency-generator/
4) Try to find a tape head used in old 4 track cassette players.
5) 3-4 elcheapo compasses.
6) Try to build yourself a Diode Carousel (search Google or here). https://www.youtube.com/watch?v=nj0ShZV6bhY

The first thing I would ask about your core in core build is why choose a toroid form? Whatever happens in the toroid can happen on a straight core without the polarity cancellation of the coil ends. I would migrate to a toroid form only after the straight form was fully understood. Straight cores would enable slide in flexibility to make quick inner coil changes. hehehe

All the best.

wattsup