High Voltage Planar Ferrite transformers. ( Intial Experiments ) / High Voltage / Forums

Forums

4hv.org :: Forums :: High Voltage

« Previous topic | Next topic »

High Voltage Planar Ferrite transformers. ( Intial Experiments )

first previous 5 6 7 8 next last

Move Thread

LAN_403

Ash Small

Wed Sept 24 2014, 11:04AM

Registered Member #3414 Joined: Sun Nov 14 2010, 05:05PM
Location: UK
Posts: 4245

Patrick wrote ...

the whole idea from the beginning of this thread involved fractional turns, so I need all the comments I can get. Test board is nearing completion.

I'd forget the fractional turns thing for now, and run some tests with integer turns, in order to get some idea of what's practical and achievable

Uspring

Wed Sept 24 2014, 11:12AM

Registered Member #3988 Joined: Thu Jul 07 2011, 03:25PM
Location:
Posts: 711

The two outer limb coils used independently will have finite coupling, but when they are forced, by terminal connection, to operate together, the core central limb can be split by a magnetic wall and the coupling vanishes.

Coupling doesn't depend on how you drive the coils. It depends solely on the geometry.

Similarly, my first and second diagrams have the same topology, so in the perfect limit of infinite core/vacuum permeability ratio they must have the same inductance. Period.

I believe that the topological transformation you have in mind for the middle diagram is cutting through the central leg and joining the limbs so that the turns are adjacent. That looks indeed equivalent to the top diagram.
If you accept, that there is coupling, it will have changed by moving the turns, making the top and middle circuits different. That depends on my assumption, that the coupling of the outer limbs in the middle diagram is less than 1. Its topological variant has k=1.

Dr. Slack

Wed Sept 24 2014, 12:22PM

Registered Member #72 Joined: Thu Feb 09 2006, 08:29AM
Location: UK St. Albans
Posts: 1659

No, the transformation I had in mind I said is specifically not cutting through the middle leg, but lifting the wire over the entire core. Imagine you are holding a cored inductor in one hand, and the pair of wires going to it in the other. Turn the core over so a different face is towards the wires. That's the extent of the transformation needed.

Anyhows, a much better way of looking at magnetically the same solution. This is much closer to a practical winding scheme for Patrick, as I've just remembered that he is in fact using multiple cores. Instead of splitting the central limb magnetically in half by considering each outer limb separately, let's split the centre limb in half by making it from two cores side by side. Magnetically, this is exactly the same situation, but it is probably easier for unbelievers to see that it works.

Now unfortunately, Patrick's OP shows 3 cores. This means that he could implement 1.33 or 1.67 turns by this scheme with trifilar wire. But for simplicity, let's stay with half turns as the diagram is easier. The extension to third turns with 3 cores or Nth turns with N cores is trivial.

The diagram below is for 1.5 turns effective wound on two cores, or an even number of cores, side by side. It is wound from bifilar wire, which avoids joints inside the transformer, or excess wire to get the joints outside, probably a good thing. To make it easier to see what's going on, I have made the two strands different colours.

Consider 2A entering the inductor. It splits as 1A into each wire. Look carefully at each core and count the magnetising ampere turns round it. 3 turns round each core, that's 3At for 2A input, 1.5 turns effective. The only place you may have trouble counting is on the left, where the wires go off to the supply. Don't forget that they join through the supply, so constitute two more Amps, making a total of 3At on the left as well.

Now consider the voltage generated by changing B field in the cores. The parts of the winding that go round one half core generate only half the voltage of the turns that go round both halves. So they act like half a turn. They are in parallel, so the two parts stay balanced.

Notice in the gap between the two cores, there are two more bits of wire, but they carry current in opposite directions, so generate no net H field. This is the only excess wire penalty that this arrangement suffers. In trying to make an efficient transformer, you really want as much of your wire doing electromagnetic work, and as little as possible just getting from A to B, as it's all causing IR drops. These two wires in the middle are the only cost of doing half turns this way. Using the outer limbs for half turns would be much more expensive in terms of IR drops, and leakage inductance for that matter, than this multiple core approach.

Now I've got the two half turns on physically distinct cores, explain where this coupling between the two half turns is coming from. That's why I said it's an interesting question. In the outer limb arrangement, when the 'half turns' are used independently, there is coupling. When *used strictly in parallel* there isn't, the magnetic wall is maintained in the core central leg by the electrical parallelism, so there is not the effect of coupling, whether or not it looks like there is coupling. This is magnetically and electrically equaivalent when used in the same balanced way, but now there is obviously no coupling if used unbalanced. An interesting conundrum, I don't understand why I don't understand it yet, but then I don't understand relativity or QM either.

But it needn't matter to Patrick, as this is a practical scheme that will work, even if we don't agree on why the equivalent scheme works.

<edit>

Whoa, stop press, it's even easier than that with multiple cores. Put three turns round one core with one wire. Put three turns round the second core with another wire. Put these two primaries in parallel. Stack the two cores side by side and wind the secondaries round both cores. I'll let you do the math.

This results in slightly more wasted wire than the diagram above, as there are now three cancelling wires going up/down the middle gap. But it affords a much nicer mechanical winding arrangement, and a simpler less head-scratching way of seeing that it also works.

</edit>

Patrick

Thu Sept 25 2014, 09:57PM

Registered Member #2431 Joined: Tue Oct 13 2009, 09:47PM
Location: Chico, CA. USA
Posts: 5639

ive finalized everything and am ready for the toner transfer to board.
but a problem has popped up. previously this printer worked fine, about 8 months ago. and its only used for PCB making.

Having gotten it all plugged in and ready to go it prints like this:

This smearing wont go away...

it should print like this:

And this is the printer, brother 2170W.

i just cant figure out why it wont work after ive done everything i can think of to clean it. I dont know if the photo-drum is bad.

Shrad

Fri Sept 26 2014, 08:12AM

Registered Member #3215 Joined: Sun Sept 19 2010, 08:42PM
Location:
Posts: 780

charge leftover on the drum? have you cleaned it with something like methanol or isopropyl?

Patrick

Fri Sept 26 2014, 10:07AM

Registered Member #2431 Joined: Tue Oct 13 2009, 09:47PM
Location: Chico, CA. USA
Posts: 5639

Shrad wrote ...

charge leftover on the drum? have you cleaned it with something like methanol or isopropyl?

yep the pattern changed but it still smears the whole page. Having cleaned it, it was rolled with a uniform layer of powder, then after another bad print i looked, and the spots and sploches of toner where random on the photo-drum, and looked like the invert of the garbage page.

so, at 11$ ill try a new photo-drum, or buy a whole new printer just for PCBs at 110$.

Patrick

Sun Sept 28 2014, 10:40PM

Registered Member #2431 Joined: Tue Oct 13 2009, 09:47PM
Location: Chico, CA. USA
Posts: 5639

Sorry, for the lack of updates. But not being able to print a usuable transfer on Brother printers. With out a circuit bard i cant test differnt configurations.

Patrick

Tue Sept 30 2014, 05:06AM

Registered Member #2431 Joined: Tue Oct 13 2009, 09:47PM
Location: Chico, CA. USA
Posts: 5639

trying for a good transfer, double sided insertion mount board.

Crap transfer

i dont think im getting it hot enough.

At this point im wondering if i should just print with a 3D printer, a single layer mask?

Patrick

Wed Oct 01 2014, 01:44AM

Registered Member #2431 Joined: Tue Oct 13 2009, 09:47PM
Location: Chico, CA. USA
Posts: 5639

Circuit board ready to be drilled and populated.

Start of etching.

it got real dark quickly, then I think the etch slowed greatly.
I decided to clamp it to the table, as it seemed prudent.

I added H2O2, enough to get it light again.

2 oZ copper, double sided board.

Electra

Wed Oct 01 2014, 10:21PM

Registered Member #816 Joined: Sun Jun 03 2007, 07:29PM
Location:
Posts: 156

Patrick good to see it progressing, don't think I was expecting to see so small mosfets. Out of curiosity what devices are they?

Trouble is I'm too set in my ways, to me I still think lots of power needs to be huge heatsinks with rows of big mosfets, and a few gigantic capacitors for reassurance. I can see now if you need it very light and small, conventional won't cut it.

first previous 5 6 7 8 next last

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