> How deep can you stack the laminations?
Magnetically: as deep as you like.
But for a good transformer, you want the most iron area with the least length of copper.
Assume you want a 1 square inch core and consider several core shapes and the length of turn in the first layer:
Round: 1.128" diameter, 3.54" turn
Square: 1" square, 4" turn
2:1 Rectangle: 0.707"x1.414", 4.24" turn
4:1 Rectangle: 0.5"x2", 5" turn
I have only computed the "first turn", not the average MLT. I'm lazy. But the trend will be the same all the way up the winding.
There are ways to make round cores, but they cost a lot.
With flat-stack laminations, square is best but a 2:1 stack is only 6% more copper loss, and even a 4:1 stack is only 25% worse than square-stack (assuming all the same wire gauge). Aim for square, but don't let it get in the way if you need a core area that is in-between standard EI sizes, or if you have to fit a certain space.
> is it necessary to isolate with paper (or what have you) the winding layers?
The voltage between adjacent turns, on 50/60Hz windings, is usually about 0.1V Volts. The varnish is good for over 100 volts. There is no reason to further insulate adjacent turns (obviously, or else magnet wire would have beter insulation).
Where you get in trouble is when you wind a layer then come back with a second layer. The voltage between start and finish is the sum of the per-turn voltages. If you are winding 20 turns per layer, it is 20*2*0.1= 4 Volts, no problem. But if you are winding fine wire on a wide core it could be a thousand turns, 1000*2*0.1= 200 volts, and you might want to slap some fishpaper between windings. Check your actual voltage per turn values from the design; "0.1V" is just my rough rule of thumb. It does not vary a huge amount for wall-power or audio transformers or chokes; it will often be less. But if you have a super-hot core iron and run it in heavy flux you may have more. And if you design for 400Hz or higher (aircraft systems, inverters) you better do the math even for very few turns per layer.
You also need insulation between windings if there is any large voltage (chokes don't matter unless they are dual-winding common-mode blockers). Wall-power should always be heavily insulated from the output winding. Tube plate output transformers have a large DC voltage added to the audio swing and normally want extra insulation so the B+ does not land on the guy touching the speaker terminal.
> paper (or what have you)
I am sure you know that "paper", ordinary paper, is only a little better than nothing, and worse if it wicks moisture into the winding. I don't know if you can still buy good-old "fishpaper", the classic 1930s inter-winding stuff, but I would use a heat resisting plastic film today.
Magnetically: as deep as you like.
But for a good transformer, you want the most iron area with the least length of copper.
Assume you want a 1 square inch core and consider several core shapes and the length of turn in the first layer:
Round: 1.128" diameter, 3.54" turn
Square: 1" square, 4" turn
2:1 Rectangle: 0.707"x1.414", 4.24" turn
4:1 Rectangle: 0.5"x2", 5" turn
I have only computed the "first turn", not the average MLT. I'm lazy. But the trend will be the same all the way up the winding.
There are ways to make round cores, but they cost a lot.
With flat-stack laminations, square is best but a 2:1 stack is only 6% more copper loss, and even a 4:1 stack is only 25% worse than square-stack (assuming all the same wire gauge). Aim for square, but don't let it get in the way if you need a core area that is in-between standard EI sizes, or if you have to fit a certain space.
> is it necessary to isolate with paper (or what have you) the winding layers?
The voltage between adjacent turns, on 50/60Hz windings, is usually about 0.1V Volts. The varnish is good for over 100 volts. There is no reason to further insulate adjacent turns (obviously, or else magnet wire would have beter insulation).
Where you get in trouble is when you wind a layer then come back with a second layer. The voltage between start and finish is the sum of the per-turn voltages. If you are winding 20 turns per layer, it is 20*2*0.1= 4 Volts, no problem. But if you are winding fine wire on a wide core it could be a thousand turns, 1000*2*0.1= 200 volts, and you might want to slap some fishpaper between windings. Check your actual voltage per turn values from the design; "0.1V" is just my rough rule of thumb. It does not vary a huge amount for wall-power or audio transformers or chokes; it will often be less. But if you have a super-hot core iron and run it in heavy flux you may have more. And if you design for 400Hz or higher (aircraft systems, inverters) you better do the math even for very few turns per layer.
You also need insulation between windings if there is any large voltage (chokes don't matter unless they are dual-winding common-mode blockers). Wall-power should always be heavily insulated from the output winding. Tube plate output transformers have a large DC voltage added to the audio swing and normally want extra insulation so the B+ does not land on the guy touching the speaker terminal.
> paper (or what have you)
I am sure you know that "paper", ordinary paper, is only a little better than nothing, and worse if it wicks moisture into the winding. I don't know if you can still buy good-old "fishpaper", the classic 1930s inter-winding stuff, but I would use a heat resisting plastic film today.
Well, carpenter, I've got a question for you. Where did you get your cores from? When I found out how much large value/large amperage inductors cost I decided that I'll try my own at winding chokes. I need 4 100mH 5A chokes (for current sources) and 4 10mH 10A chokes (for the PSU), the prices of these are far from the prices of the 100mH 10A chokes you needed, but still are pretty high. My only problems are finding the cores, and calculating the core sizes and air gap. If you know of any good sources of both the parts (cores) or information (how to calculate such things) I'd greatly apreciate it...
Mark A. Gulbrandsen said:
I've done work before for other members of the forum but you know this is just for the love of DIY, I don't do a living with this. It will depend on the project and cost since my core source is in Mexico city so I have to pay for freight and buy the minimun quantity of lamination, plus phone, faxes, packing and other expences, is not much though.
Now, for the calculation of gap and size I use Hanna's method which is by continuous aproximation since you have to play with the number of wire turns, gap and inductance. This method uses a graph. You can find this information on the Radiotron Designers Handbook, page 248 of the fourth edition and of course in other books. For doing this work you must have an inductance bridge, which I do.
Feel free to contact me by email.
Mark A. Gulbrandsen said:
Old transformer lamination, which I dont use, is a tricky thing since you will have to burn the trafo in order to calcinate the varnish then separate and clean the lamination. This usually requieres phisical force plus the deformation you may have from heating you will end with distorted parts, dificult to work with.
Phisical work over the lamination diminishes the permeability. So it is much easier to work with new lamination with known characteristics, Oh boy, and cleaner!!!
Fortunately here I can get most E/I trafo materials such as M19, M15 and M6. The lower the number the better for bandwidth. So for PSU trafos and chokes I use M19 or M15 and for audio M6.
there are ways to dismantle and old tranny without damaging them, i do one or combinations of these:
1. soak the tranny into a solvent to loosen up the varnish, i use denatured alcohol and variuos types of paint thinners...
2. cut acrross the copper wires using hacksaws, or cutting disks.
with the copper out of the way, stripping the core into single laminations can be done... i never apply heat as i am aware it may change the core's characteristics..i favour grain oriented steel of ga.29 variety and use then when i can find them.....
1. soak the tranny into a solvent to loosen up the varnish, i use denatured alcohol and variuos types of paint thinners...
2. cut acrross the copper wires using hacksaws, or cutting disks.
with the copper out of the way, stripping the core into single laminations can be done... i never apply heat as i am aware it may change the core's characteristics..i favour grain oriented steel of ga.29 variety and use then when i can find them.....
joan2 said:
Never done this befor, but I would try to spray or dip the laminations in water and some diluted acid on it, say water +vinager. but you will have to experiment the reaction and then wash the lamination thoroughly.
Laminations, being mostly iron, I would expect quite a bit of oxidation reaction. on the other hand I would not go as far as seeing oxide over the surface as on rusted steel but I would look only for a dull greyish color, indication of a very thin oxide layer
Hi Tony,
How are you doing with those trafos?
I'll be buying EI 150 to do some 500 and 750VA power trafos for a couple of amp projects.
Since I'm bringing in some material I can make gapped coils for CLC supplies if there is enough interest. What I can do right out of my hand are 5 mH 8 Amps and 15 mH 4 Amps or there abouts on EI-100, hand wound, perfect layering, varnished and super low Re. Other values are possible. Just for the record, this is not a business (well, so far).
You wont find this type of compact coils ready made on the market as far as I know.
How are you doing with those trafos?
I'll be buying EI 150 to do some 500 and 750VA power trafos for a couple of amp projects.
Since I'm bringing in some material I can make gapped coils for CLC supplies if there is enough interest. What I can do right out of my hand are 5 mH 8 Amps and 15 mH 4 Amps or there abouts on EI-100, hand wound, perfect layering, varnished and super low Re. Other values are possible. Just for the record, this is not a business (well, so far).
You wont find this type of compact coils ready made on the market as far as I know.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.