It seems to me the guy is trying to design a traffo for quasi-resonant SMPS where coil capacitance play a crucial role in it.
Of course. Otherwise you'd be measuring the self-resonance between the magnetizing inductance and the stray capacitance. That's not super useful, unless you want to determine the magnetizing inductance.
Tom
Yes, of course. I mistakenly mentioned the self resonance, when I should have said that Ridley recommends measuring the impedance of the primary over a wide frequency range as this tells a great deal about the construction.
Lots of people have been posting, but none have addressed my question: how can I test enamelled wire to discover two important factors - permittivity and thickness. Perhaps there is no easy way, but in my experience in electronics tells me that there usually a way to do almost anything but the trick is to find out what it is.
Some posters say just make a prototype transformer and if it doesn't meet requirements make another a bit different - repeat until good enough. The whole point of my question is about avoiding making multiple prototypes until its right. Because making prototypes that have only a chance of being right is EXPENSIVE - VERY EXPENSIVE in one's time and money spent of materials.
Some say buy from a reputable wire manufacturer and look up the datasheet. Yeah I know I can do that. But try buying a few tens of metres from a reputable US or Euro manufacturer and see how much it costs and how long it takes to ship to you. Then try a Chinese supplier. I guarantee you'll then be looking for ways to use the Chinese wire. China routinely gets things to my door in Australia in a week. A European wire supplier such as Block takes months. American suppliers charge more for shipping than they do for the actual goods.
Some posters point out that calculating self capacitance is complicated and so it is. But I have computer tools to do it in minutes. If I have all the data:-
Wire copper diameter - yes
Air permitivity - yes, 8.854 x 10^-13 F/m
Impreg varnish permittivity- yes, easily measured
winding method - yes, my choice, hex close packed, stacked layers, etc
Interlayer wrap if used thickness & permittivity (mylar or kraft paper)- yes, easily measured
Stacking (winding) efficiency - known from experience with winding machines;
Transformer dimensions - yes, my choice;
but..
wire enamel thickness - that's the question;
wire enamel permittivity - that's the question.
Some posters say just make a prototype transformer and if it doesn't meet requirements make another a bit different - repeat until good enough. The whole point of my question is about avoiding making multiple prototypes until its right. Because making prototypes that have only a chance of being right is EXPENSIVE - VERY EXPENSIVE in one's time and money spent of materials.
Some say buy from a reputable wire manufacturer and look up the datasheet. Yeah I know I can do that. But try buying a few tens of metres from a reputable US or Euro manufacturer and see how much it costs and how long it takes to ship to you. Then try a Chinese supplier. I guarantee you'll then be looking for ways to use the Chinese wire. China routinely gets things to my door in Australia in a week. A European wire supplier such as Block takes months. American suppliers charge more for shipping than they do for the actual goods.
Some posters point out that calculating self capacitance is complicated and so it is. But I have computer tools to do it in minutes. If I have all the data:-
Wire copper diameter - yes
Air permitivity - yes, 8.854 x 10^-13 F/m
Impreg varnish permittivity- yes, easily measured
winding method - yes, my choice, hex close packed, stacked layers, etc
Interlayer wrap if used thickness & permittivity (mylar or kraft paper)- yes, easily measured
Stacking (winding) efficiency - known from experience with winding machines;
Transformer dimensions - yes, my choice;
but..
wire enamel thickness - that's the question;
wire enamel permittivity - that's the question.
Last edited:
For audio applications,
the part of the effective turn to turn winding capacitance, in respect to the total winding capacitance, is so low it can often be neglected.
This is, because the turns are serial connected and turn length and effective voltage between turns is very low.
But if needed, first get the windings turn to turn STATIC capacitance by taking 2 wires
(as long as you can and are still able to twist them, form a loop, and suspend in free air)
measure the exact length, twist tigth with a drill, measure the capacitance between those 2 wires.
Shortcut the loopends of each wire and measure the static capacitance between the 2 wire.
Divide measured capacitance by actual windinglength per turn to get static pF between turns.
To get the usually much higher effective capacitance between layers, first get the STATIC capacitance:
Take the bobbin you want to use, wind end to end tigthly 2 layers, either Z or back forward (in both cases you must bring the wire out at the end to obtain the 4 wire ends ) on top of each other, secure so it does not uncoil (cut open if back forward wound to obtain 2 coils), shortcut the ends of each coil.
Measure the static capacitance between the 2 coils.
(The difference between Z and back forward is not only the voltage distribution, but is also by how the wire falls into the grooves, so wind accordingly).
(To get the static capacitance when winding as above, but with layer isolation, wind with the intended isolation and tigthness!)
The EFFECTIVE capacitance depends on the effective voltage distribution between the layers.
With serial connected Z winding the voltage between the 2 layers is half, so the effective capacitance is 1/4 of the static capacitance.
With serial connected back/forward wound layers the voltage between the 2 layers is 1 at one end and 0 at the other, the effective capacitance is 1/3 of the static capacitance.
Biggest part of the appearing effective primary capacitance is the primary to secondary capacitance because of the usually high voltage difference between winding layers, the more you interleave the higher it will be.
Add also effective layer and layer end strays to core core, effective turn to turn capacities as needed.
the part of the effective turn to turn winding capacitance, in respect to the total winding capacitance, is so low it can often be neglected.
This is, because the turns are serial connected and turn length and effective voltage between turns is very low.
But if needed, first get the windings turn to turn STATIC capacitance by taking 2 wires
(as long as you can and are still able to twist them, form a loop, and suspend in free air)
measure the exact length, twist tigth with a drill, measure the capacitance between those 2 wires.
Shortcut the loopends of each wire and measure the static capacitance between the 2 wire.
Divide measured capacitance by actual windinglength per turn to get static pF between turns.
To get the usually much higher effective capacitance between layers, first get the STATIC capacitance:
Take the bobbin you want to use, wind end to end tigthly 2 layers, either Z or back forward (in both cases you must bring the wire out at the end to obtain the 4 wire ends ) on top of each other, secure so it does not uncoil (cut open if back forward wound to obtain 2 coils), shortcut the ends of each coil.
Measure the static capacitance between the 2 coils.
(The difference between Z and back forward is not only the voltage distribution, but is also by how the wire falls into the grooves, so wind accordingly).
(To get the static capacitance when winding as above, but with layer isolation, wind with the intended isolation and tigthness!)
The EFFECTIVE capacitance depends on the effective voltage distribution between the layers.
With serial connected Z winding the voltage between the 2 layers is half, so the effective capacitance is 1/4 of the static capacitance.
With serial connected back/forward wound layers the voltage between the 2 layers is 1 at one end and 0 at the other, the effective capacitance is 1/3 of the static capacitance.
Biggest part of the appearing effective primary capacitance is the primary to secondary capacitance because of the usually high voltage difference between winding layers, the more you interleave the higher it will be.
Add also effective layer and layer end strays to core core, effective turn to turn capacities as needed.
Last edited:
If you have the ability to model your design but lack only the thickness and permittivity of the insulation, then you must have an idea of the sensitivity of the design to those two variables. It should be safe to assume that the permittivity is about 3 (for polyimide) and unlikely to be as high as 6. The thickness should be less than 100um.
https://www.professionalplastics.com/professionalplastics/ElectricalPropertiesofPlastics.pdf
As to magnet wire, for example, valid for IEC 317-8 Datherm, IEC 317-13 Damid, IEC 317-20 Dasol
Conductor diam 0.3mm, min 0.296, max 0.304
Overall diam, Grad 1, min 0.319, max 0.334
Overall diam, Grad 2, min 0.335, max 0.352
As you can see, even if you would know the exact permittivity the large tolerances still limit the accuracy of the calculation.
With low dk isolation material used between layers, things get even worse, there is thicknes variation not only do to manufacturing tolerances, but also depending on wire pressure.
The more you sqeeze, the smaller the dimension and the higher the dk, especially low dk Nomex because the low dk is only in uncompressed state do to the air pockets of the material. Then there are plastics, again manufacturer tolerances, combined that with the coldflow under pressure and you are left with materials of comparable high dk such as polyester.
In my view, there are too many unknows and variables for calculating the capacitances to any degree of accuracy without basing it at least on the measured results of a minimum of 2 layers wound on the bobbin of the actual transformer the way i described in #27.
Conductor diam 0.3mm, min 0.296, max 0.304
Overall diam, Grad 1, min 0.319, max 0.334
Overall diam, Grad 2, min 0.335, max 0.352
As you can see, even if you would know the exact permittivity the large tolerances still limit the accuracy of the calculation.
With low dk isolation material used between layers, things get even worse, there is thicknes variation not only do to manufacturing tolerances, but also depending on wire pressure.
The more you sqeeze, the smaller the dimension and the higher the dk, especially low dk Nomex because the low dk is only in uncompressed state do to the air pockets of the material. Then there are plastics, again manufacturer tolerances, combined that with the coldflow under pressure and you are left with materials of comparable high dk such as polyester.
In my view, there are too many unknows and variables for calculating the capacitances to any degree of accuracy without basing it at least on the measured results of a minimum of 2 layers wound on the bobbin of the actual transformer the way i described in #27.
Last edited:
NOBODY said that.
Many experienced builders said: "make a test winding and measure it" which you steadily refuse.
Once you measured it, you have the necessary constants to scale up and down as needed.
No need to build a full transformer, junk it and repeat at random as you imply.
* You can even build just the primary.
* You can wind "100 turns" on the bobbin you will use later to mimic reality even better, with and/or without core material inserted.
* You can even wind "X" turns or layers. on a piece of plastic tube as suggested by Conrad Hoffman although geometry will vary: average turn length is higher on a rectangular core than on a round one but no big deal, rDC and capacitance per meter will roughly stay the same.
Test/experiment cost is near zero because that wire can be later reused.
All experiments are more accurate than any simulation.
.
JMFahey:-
1. RE "nobody said that" - Essentially, they did. A single test winding is not sufficient, The number of layers and the number of primaries, secondaries, and terteries all have a big influence on effective self capacitance. But, proposed final transformer or a simplified test winding, it comes down to the same thing. Say I design a relavent test winding, and order the wire - a minimum quantity. Some time later it will turn up and I will spend time winding and testing, only to find it worn't do. So, titivate the desing, heck, need a differet former size and different wire gauge. Order more wire and wait, Test again, better but still no good. Order more wire, rinse and repeat. One could see months go by doing this, even with the rapid shipping from China. Go through $200 worth of wire too.
2. You said "Once you measure, you have the constants, you can scale up." Well, yes, if you can measure the two parameters enamel thickness AND enamel permittivity. That's the big question - how do I do that? It is not sufficient to just measure capacitance, or enamel capacitance - it does not scale.
3. You said "test cost is zero because the wire can be re-used." Not true. Handling degrades reliability. Copper rapidly work hardens. Successive winding and unwinding leaves you with wire having bends and kinks that you can never fully get rid off. In any case the transformer will be impregnated for reliability and stability. The impregnant of course has its own permittivity. Once a winding has been made, if it can't be used as a winding, it is scrap.
4. You said "All experiments are more accurate then any simulation". Yes, if the experiment conditions are very close to the final reality, and the simulation is dodgy. Calculation methods for transformer/inductor self capacitance has been well established for many years - but need the enamel properties as inputs. Not a problem for a major manufacturer who is going to buy millions of metres of wire - the wire factory will bend over backwards to get the sale. I'm only going to buy 10s of metres.
People please understand that if I didn't need to know a way of measuring enamel thickness and permittivity, I would not have asked. I you know a way, I'll be in your depth. Telling me I don't need to know doesn't help.
Calculation always takes less time than experimenting.
1. RE "nobody said that" - Essentially, they did. A single test winding is not sufficient, The number of layers and the number of primaries, secondaries, and terteries all have a big influence on effective self capacitance. But, proposed final transformer or a simplified test winding, it comes down to the same thing. Say I design a relavent test winding, and order the wire - a minimum quantity. Some time later it will turn up and I will spend time winding and testing, only to find it worn't do. So, titivate the desing, heck, need a differet former size and different wire gauge. Order more wire and wait, Test again, better but still no good. Order more wire, rinse and repeat. One could see months go by doing this, even with the rapid shipping from China. Go through $200 worth of wire too.
2. You said "Once you measure, you have the constants, you can scale up." Well, yes, if you can measure the two parameters enamel thickness AND enamel permittivity. That's the big question - how do I do that? It is not sufficient to just measure capacitance, or enamel capacitance - it does not scale.
3. You said "test cost is zero because the wire can be re-used." Not true. Handling degrades reliability. Copper rapidly work hardens. Successive winding and unwinding leaves you with wire having bends and kinks that you can never fully get rid off. In any case the transformer will be impregnated for reliability and stability. The impregnant of course has its own permittivity. Once a winding has been made, if it can't be used as a winding, it is scrap.
4. You said "All experiments are more accurate then any simulation". Yes, if the experiment conditions are very close to the final reality, and the simulation is dodgy. Calculation methods for transformer/inductor self capacitance has been well established for many years - but need the enamel properties as inputs. Not a problem for a major manufacturer who is going to buy millions of metres of wire - the wire factory will bend over backwards to get the sale. I'm only going to buy 10s of metres.
People please understand that if I didn't need to know a way of measuring enamel thickness and permittivity, I would not have asked. I you know a way, I'll be in your depth. Telling me I don't need to know doesn't help.
Calculation always takes less time than experimenting.
Last edited:
True. But calculation can also lead you down the wrong path in a hurry if you don't correlate calculations with reality.
Tom
Bucks, it has indeed been stated that the wire enamel is a minor contributor. A statement is not necessarily true, and in this case it certainly is not.
And in fact it is indeed not hard to see that it IS a major contributor. Imagine two good conductors, both round, next to each other, just touching. Most of the space between the 2 conductors is air, an insulator, but in fact there will be good conduction between the two, because electric current finds the PATH OF LEAST RESISTANCE and concentrates itself in the small region where the round wires touch.
Just as current finds and concentrates itself in the path of least resistance, an electric field finds and concentrates itself in the path of least resistance, which is where the wires touch, since the permittivity of enamel is much greater than the permittivity of air.
Conductivity is to current what permittivity is to electric fields - the math is the same.
Since capacitance between 2 conductors is proportional to the electric field quantity between them, the wire enamel has a major influence on capacitance..
I'm not searching for an easy formula. I already have the formula, which is complicated, but with PC's that is not a problem. It's known as Massarini's formula. What I am searching for is a way to test enamel thickness and permittivity so I can use the formula.
And in fact it is indeed not hard to see that it IS a major contributor. Imagine two good conductors, both round, next to each other, just touching. Most of the space between the 2 conductors is air, an insulator, but in fact there will be good conduction between the two, because electric current finds the PATH OF LEAST RESISTANCE and concentrates itself in the small region where the round wires touch.
Just as current finds and concentrates itself in the path of least resistance, an electric field finds and concentrates itself in the path of least resistance, which is where the wires touch, since the permittivity of enamel is much greater than the permittivity of air.
Conductivity is to current what permittivity is to electric fields - the math is the same.
Since capacitance between 2 conductors is proportional to the electric field quantity between them, the wire enamel has a major influence on capacitance..
I'm not searching for an easy formula. I already have the formula, which is complicated, but with PC's that is not a problem. It's known as Massarini's formula. What I am searching for is a way to test enamel thickness and permittivity so I can use the formula.
Bucks, if you search on postings by me on this and another forum, you will find that I am in my late 70's. I have been designing transformers and building transformers, mainly power and audio output, for nearly 50 years. I was for several years the chief engineer for a transformer manufacturer.
I didn't bake the theory - it comes from books and professional journals.
As I said earlier in this thread, if you are a manufacturer and buy wire in quantities of millions of metres, getting trustworthy data from the wire supplier is no problem. As is keeping a few thousand dollars worth of wire in stock for immediate use. But working at home and buying in tens of metres, its a different story. One has to buy on eBay or at best electronic retailers. Either way data is not available and wire takes time to be shipped.
The world has changed too. when i worked for the transformer company, we used to buy our wire from Philips and Rola - both very good quality manufacturers with application engineers ready to help. But Philips is no longer in the wire business and Rola got taken over by Plessey and went broke. Now everything comes from China.
I didn't bake the theory - it comes from books and professional journals.
As I said earlier in this thread, if you are a manufacturer and buy wire in quantities of millions of metres, getting trustworthy data from the wire supplier is no problem. As is keeping a few thousand dollars worth of wire in stock for immediate use. But working at home and buying in tens of metres, its a different story. One has to buy on eBay or at best electronic retailers. Either way data is not available and wire takes time to be shipped.
The world has changed too. when i worked for the transformer company, we used to buy our wire from Philips and Rola - both very good quality manufacturers with application engineers ready to help. But Philips is no longer in the wire business and Rola got taken over by Plessey and went broke. Now everything comes from China.
Last edited: