Hi,
When assembling a transformer core there will always remain some airgap even when you don't want it (toroids excepted).
And it's not insignificant either: I found out that I can raise the primairy inductance of a coil on a double c-core by over 30 % just by squeezing the cores together with even more force
The gap will not only lower inductance but also increase leakage inductance. Bad thing...
So I was wondering if I can put some "magnetic field conducting" stuff between the two core halves to fill the gap
I've been thinking about ferro fluid such as used in tweeters but I am afraid it will probably leak out sooner or later.
Better would be some glue/kit with such properties? Does it exist? Or do I have to mix up something myself by adding some iron particles to it?
When assembling a transformer core there will always remain some airgap even when you don't want it (toroids excepted).
And it's not insignificant either: I found out that I can raise the primairy inductance of a coil on a double c-core by over 30 % just by squeezing the cores together with even more force
The gap will not only lower inductance but also increase leakage inductance. Bad thing...
So I was wondering if I can put some "magnetic field conducting" stuff between the two core halves to fill the gap
I've been thinking about ferro fluid such as used in tweeters but I am afraid it will probably leak out sooner or later.
Better would be some glue/kit with such properties? Does it exist? Or do I have to mix up something myself by adding some iron particles to it?
maudio said:And it's not insignificant either: I found out that I can raise the primary inductance of a coil on a double c-core by over 30 % just by squeezing the cores together with even more force
Yes, but ask yourself what will happen when some DC flows in the transformer. The gap is a magnetic resistance that helps make the primary inductance more nearly constant with (AC) level and DC.
It depends on what your motivation for doing so is. Yes, you increase the inductance but this also bring about core saturation at lower current levels. It will NOT increase the basic power handling capability of the transformer... that tends to come on a pounds per watt per LF response basis.
In OPTs and even power transformers, that gap is your friend.
Can you tell us what you are trying to achieve?
In OPTs and even power transformers, that gap is your friend.
Can you tell us what you are trying to achieve?
Can you tell us what you are trying to achieve?
Sure, I am winding an esl-transformer and my aim with this idea was to keep the number of turns needed as low as possible. The basic maths being:
Less turns = less capacitance + less leakage = more bandwidth
So therefore I wanted to maximize the inductance/turn. I also think reducing the gap will decrease leakage. I measured an increase in resonance frequency when reducing the gap.
I know less gap will lower the core saturation limit but that is not my primairy concern as I plan on using rather large cores. DC currents are also non-existent in the application.
If you are certain that there will NO DC... then by all means. But listen to the cat... the gap tends to make your design prediction of L better. In other words don't go to extremes.
The bulk of your reduction in leakage inductance will be achieved through interleaving anyway.
See if he beats me again...
The bulk of your reduction in leakage inductance will be achieved through interleaving anyway.
See if he beats me again...
poobah said:
Ok, so I understand that too little gap might raise a few new problems. But am I wrong in thinking that a toroid core is wound from one long strip of iron and will not have any gap at all? Are they using material with higher losses for toroids to compensate?
Well I guess it's all a matter of taking things into account in the design stage.
With only 1 layer of primairy wire and 4 layers of secundairy there isn't much to interleave... I'll probably just add another layer of prim on top of the secundairy to sandwich it in between.
Yes... a tape wound core can be considered gapless.
There are lossy toriods usually made of dusts and ceramics. They are popular in power inductor designs.
You may consider winding your primary trilfilar or quadfilar to improve interleaving... capacitance will be your deciding, or limiting, factor then.
What type of core are planning on using?
There are lossy toriods usually made of dusts and ceramics. They are popular in power inductor designs.
You may consider winding your primary trilfilar or quadfilar to improve interleaving... capacitance will be your deciding, or limiting, factor then.
What type of core are planning on using?
lapping and clamping high perm ferrite core faces is necessay to get the best core coupling/least gap
to a 1st order aproximation the leakage inductance is not changed by the core - the leakage inductance is a coil geometry issue, it is the coil inductance from flux not linking the 2 windings - the core is usefully placed where the flux does link the windings
to a 1st order aproximation the leakage inductance is not changed by the core - the leakage inductance is a coil geometry issue, it is the coil inductance from flux not linking the 2 windings - the core is usefully placed where the flux does link the windings
makes a lot of sense, Jcx. I came up with this because I saw the frequency of the resonance peak (the one due to Lleak and Cpar) of my unloaded transformer shifting slightly upwards when reducing the gap and reduced leakage seemed a logical explanation to me at the time. But it might as well have something to do with the changing parameters of the primairy.
Poobah, right now I use some c-cores that I salvaged from old mains transformers. So I don't have much data on them, they are proably quite standard. Made with 0.3mm (0.01") laminations and core area is around 13 cm^2 (just over 2 square inch). I plan to stack two of them doubling core area. Maybe further down on the road I'll look for some special high-mu cores but for now they will do I hope.
I don't think I fully understand your suggestion about trilfilar or quadfilar primairy winding. Do you mean parallelling several wires for the primairy? I was planning on using 5 or 6 parallel wires for the primairy on my next try-out, but my objective was just to fill the entire winding space (62 mm / 20 turns) while maintaining low winding height. I don't see how that would reduce leakage. But I guess you mean something different... Maybe you can explain this a little further?
thanks to you all,
Martin
Simplest is bifilar (two wires, trifilar is three wires, quadrifilar, four). You wind a pair of wires at the same time so that they lay adjacent to one another. One is the primary, one is the secondary. You get very low leakage inductance because the primary and secondary are so tightly coupled. You also get high capacitance between primary and secondary because of the adjacent wires. That doesn't matter if the transformer is 1:1, but I expect you're thinking of 1:50, so it does matter, and disqualifies the idea.
What you need is lots of interleaved sections. Perhaps split the primary into four sections and the secondary into three. You wind a section of primary, then secondary, then primary, etc etc. The Radio Designer's Handbook has details on this sort of thing.
What you need is lots of interleaved sections. Perhaps split the primary into four sections and the secondary into three. You wind a section of primary, then secondary, then primary, etc etc. The Radio Designer's Handbook has details on this sort of thing.
Damn that cat is quick...
Yes, I mean winding the primary (same turns) in four layers and paralleling them. RDH4 does not provide charts for this... but I think I can link you some info that does. The trick is that the current gets divided by four, but you have 4 coils, so the amp-turns and the inductance stays the same.
This might work better since your turns ratio will be so high...
Yes, I mean winding the primary (same turns) in four layers and paralleling them. RDH4 does not provide charts for this... but I think I can link you some info that does. The trick is that the current gets divided by four, but you have 4 coils, so the amp-turns and the inductance stays the same.
This might work better since your turns ratio will be so high...
Winding the primairy and sec together (bifilair) as the speedy cat pointed out is indeed not possible I am afraid, not only because of capacitance (I am aiming at much more than 1:50) but also because I don't think the wire isolation will withstand the 5-6kV between prim and sec 
But winding several primairies together will indeed help improve things?
pointed out is indeed not possible I am afraid, not only because of capacitance (I am aiming at much more than 1:50) but also because I don't think the wire isolation will withstand the 5-6kV between prim and sec But winding several primairies together will indeed help improve things?
Well I'm very interested, so far I failed to find any information on this. My reason for doing it was to reduce winding height, as it allows for use of much thinner wire. The drawback that I can think of is that it increases prim to sec capacitance. But I'll be putting thick insulation layers between all windings to reduce that.
OK... I might be all hosed up here but, my guts tell me that interleaving parallel windings should have the same beneficial effect as it does in series windings.
I can't find the references I was looking for.
In either event, you should be winding for a full window... that is key to minimizing leakage inductance.
I can't find the references I was looking for.
In either event, you should be winding for a full window... that is key to minimizing leakage inductance.
OK, since we seem to lack theoretical evidence on the subject I'll just go ahead and try it out in practice 
I'll wind one with a full window with 5 primary wires in parallel and see how it measures with only one of the wires connected, compared to all wires connected in parallel. That should reveal a thing or two.
About the way I should be interleaving I'm not sure yet, I am studying the RDH4 on that subject right now. I think I will use the P-S-P scheme for this trial, let's not complicate things
Thanks for helping to you all, I'll let you know when I found out more...
I'll wind one with a full window with 5 primary wires in parallel and see how it measures with only one of the wires connected, compared to all wires connected in parallel. That should reveal a thing or two.
About the way I should be interleaving I'm not sure yet, I am studying the RDH4 on that subject right now. I think I will use the P-S-P scheme for this trial, let's not complicate things
Thanks for helping to you all, I'll let you know when I found out more...
I have done plenty of paralleling of primaries for the purpose of improving cross-regulation in setups with multiple secondaries. Now, cross regulation is all about: all the flux linking all the windings. So while I wasn't measuring leakage inductacne directly... I was measuring its effects. I had to ponder this all for a moment, but I am sure the RDH4 equations for leakage will hold... or very nearly so.
When you get ready to wind the real one... we'll discuss different wire gauges in the paralleled windings to ensure current balance... it gets better!
When you get ready to wind the real one... we'll discuss different wire gauges in the paralleled windings to ensure current balance... it gets better!
getting rid of the air gap in xformer cores
Excellent !! Squezzing the plates works!
As for making them using "ferro fluid" you might check out powdered metalergy = ferro alloy metal powder or even NiCAD powder (used in making healthy magnets like the "pole piece" in larger speakers.) ... stuff the powder in a ferro box, squeze it down etc ...
I would bet there is more going on with squezzing the core plate than just reducing the distance between the plate, removing the air. Might be some sort of piezo electric effect as well.
Excellent !! Squezzing the plates works!
As for making them using "ferro fluid" you might check out powdered metalergy = ferro alloy metal powder or even NiCAD powder (used in making healthy magnets like the "pole piece" in larger speakers.) ... stuff the powder in a ferro box, squeze it down etc ...
I would bet there is more going on with squezzing the core plate than just reducing the distance between the plate, removing the air. Might be some sort of piezo electric effect as well.
Well Poobah, it seems it worked out. Wound this coil:
- primary 1: 1 layer 5 wires parallel, 20 turns each
- secondary: 4 layers of 260 turns, total 1040 turns
- primary 2: 1 layer 5 wires parallel, 20 turns each
a quick and crude measurement of the unloaded resonance gives:
- only one of the 5 wires in the first prim layer connected: 11.5 khz
- all wires in first prim layer connected: 24 khz
- all wires in first and second prim layer connected: 26 khz
So there a lot to be gained by paralleling wires, and surprisingly little by interleaving.
Poor interleaving results might be explained by the low total winding heigth, the coil is around 60 mm/ 2.4" long and 3.5 mm/0.14" high. It seems rdh4 isn't that far off indeed.
- primary 1: 1 layer 5 wires parallel, 20 turns each
- secondary: 4 layers of 260 turns, total 1040 turns
- primary 2: 1 layer 5 wires parallel, 20 turns each
a quick and crude measurement of the unloaded resonance gives:
- only one of the 5 wires in the first prim layer connected: 11.5 khz
- all wires in first prim layer connected: 24 khz
- all wires in first and second prim layer connected: 26 khz
So there a lot to be gained by paralleling wires, and surprisingly little by interleaving.
Poor interleaving results might be explained by the low total winding heigth, the coil is around 60 mm/ 2.4" long and 3.5 mm/0.14" high. It seems rdh4 isn't that far off indeed.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.