Hi Diy-ers,
I want to built my own PotCore inductors around 1 Henry.
Does anyone know how to calculate it - is there any calculator for Pot Core inductors?
Many thanks
I want to built my own PotCore inductors around 1 Henry.
Does anyone know how to calculate it - is there any calculator for Pot Core inductors?
Many thanks
You divide the required inductance by the specified Al value of the potcore, take the square root, and then that's the required number of turns.
For example, for 1 H required inductance and Al = 400 nH/turn2:
√(1/400E-9) = √(2500 000) ~= 1581 turns.
You will need a really big core and coil former and very thin wire for that many turns. You could also go for an ungapped core, those are available with higher Al values, but they are inaccurate and quite nonlinear.
For example, for 1 H required inductance and Al = 400 nH/turn2:
√(1/400E-9) = √(2500 000) ~= 1581 turns.
You will need a really big core and coil former and very thin wire for that many turns. You could also go for an ungapped core, those are available with higher Al values, but they are inaccurate and quite nonlinear.
Unfornually I need accuracy and lineraity. I can get a potcore with an Al of 4500 I saw
Then you need only 471.5 turns.
What is the tolerance on that 4500 nH/turn2? If it is large, there is probably no airgap.
What is the tolerance on that 4500 nH/turn2? If it is large, there is probably no airgap.
The most linear pot cores are the ones with lowest AL. Typically that's 160nH/rtTurn. A P18 (18mm diameter) inductor can just about approach 1H on such a core, using 0.06mm dia wire. It needs 2500 turns.
I want to build these inductor for a passive line XO.
I saw that the Merchant guys use pot cores with Al = 4000 nH/turn.
So I guess 0,06mm wire is to thin right?
I was looking these cores https://de.aliexpress.com/item/1005006162241283.html?mp=1&gatewayAdapt=glo2deu
I saw that the Merchant guys use pot cores with Al = 4000 nH/turn.
So I guess 0,06mm wire is to thin right?
I was looking these cores https://de.aliexpress.com/item/1005006162241283.html?mp=1&gatewayAdapt=glo2deu
Last edited:
If the AL really is 4000nH then that would indicate an ungapped core with 25% or greater tolerance for its AL number. So seems unlikely that Marchand is using that AL, perhaps its a typo for AL=400 which is a practical value with a very small gap and likely around 5% tolerance.
0.06mm is quite practical with some experience and gentle handling. Assuming that is that the DCR (or Q) isn't too much of an issue. If it is, a bigger core is called for.
The cores in your link are ungapped - how to tell is the AL numbers are in the thousands and have 25% tolerance which isn't too practical for creating filters for audio crossovers.
0.06mm is quite practical with some experience and gentle handling. Assuming that is that the DCR (or Q) isn't too much of an issue. If it is, a bigger core is called for.
The cores in your link are ungapped - how to tell is the AL numbers are in the thousands and have 25% tolerance which isn't too practical for creating filters for audio crossovers.
+/- 25 % tolerance on Al according to your link. Those must be ungapped, otherwise it would be 5 % or less.
Edit: abraxalito already posted this while I was still typing.
Edit: abraxalito already posted this while I was still typing.
In their calculator the mention AL=4000, but 25% tolerance seems to be to much for an XO I guess.
Last edited:
I downloaded the Marchand 'calculator.exe' program, it does default to AL=4000 when the XM46 is selected. However that number is user editable. Perhaps some users don't care much about the tolerance of their PLLXO so are happy to design with ungapped cores?
abraxalito, what will be the effect of these tolarances. My impression is 25% is much, but how will it effect the filters
Last edited:
I recall the corner frequency of a 2nd order LC is inversely proportional to SQRT(L). So I'd expect variation of +/-12% of the corner freq.
If the working impedance of the filter isn't critical that would work yes. But ungapped inductors don't have such good stability over temperature as gapped ones. Perhaps the value will shift over time too.
Mouser stocked them last time I looked but I just checked and now they only have ungapped ones. Send me a PM, I might be able to help out.
Looking again, I can see gapped cores on Mouser now, for example : https://www.mouser.com/ProductDetail/EPCOS-TDK/B65671T0400A048?qs=/sLciWRBLmCe%2BJUr9kG6gg==. I've never been able to work out whether the item is a pair or just a half core and also what other accessories I would need to order to get a finished inductor.
Looking again, I can see gapped cores on Mouser now, for example : https://www.mouser.com/ProductDetail/EPCOS-TDK/B65671T0400A048?qs=/sLciWRBLmCe%2BJUr9kG6gg==. I've never been able to work out whether the item is a pair or just a half core and also what other accessories I would need to order to get a finished inductor.
They will certainly work but earlier on you cited linearity as a desirable characteristic. The higher the AL number, the poorer the linearity. But perhaps its good enough?
Something to consider is the use: if it's for signal applications (~line level) you don't need to worry about the maximum induction, but if it is used in a crossover filter or other large signal/current application, you need to make sure that the induction doesn't reach the saturation level