EQ Inductor tolerance

[dirty1_1garry]

Hi duy'ers!
Need your opinions.

Emulating and prototyping inductor based eq.
Calculated frequencies in a range of 10kHz - 20kHz with Q=1 bell gave me pretty low inductance range from 20mH up to 1mH. For HF band I used RM-6 cores with AL = 100nH, N48 core material with pretty tight 5% tolerance.

But I ran into the problem that even with 5% core for low inductance I need just a couple decades of turns and just a slight turn number change gave me pretty drastic inductance change and as a result not correct frequency point (for example instead of 14kHz I’ve got 13kHz).

As I want to get high band as precise as possible will try to minimize a such inductance tolerance. Here is my ideas how:
idea #1 - use cores with even less permeability, I’ve found on mouser RM-5 cores with AL = 40nH, but it’s 4 times more expensive (K1 core material)
idea #2 -  trim capacitance for given inductance but with change of Q a little bit
idea #3 - try to wind air gap inductor?

Any suggestions for best result?)

Cheers,
Igor

 

[dirty1_1garry]

Or use trim-able inductors? Hmm...

 

[warpie]

Have you read this?

http://www.diygallery.de/DIYsites/inductor.html

 

[dirty1_1garry]

Have you read this?

http://www.diygallery.de/DIYsites/inductor.html

once read it, but it's worth repeating. Thanks!

 

[gyraf]

..make a handful of them and select them in pairs?

 

[gyraf]

measure, select and pair AL of core..?

 

[abbey road d enfer]

idea #3 - try to wind air gap inductor?

RM6 Al100 is already air-gapped.
Do you know you can trim the inductors by wiping the two halves against each other?

 

[ruffrecords]

Have you read this?

http://www.diygallery.de/DIYsites/inductor.html

Lot's of good information there. One thing I would add is that the smallest dc resistance of the coil is not usually necessary. The Q of the circuit depends on its characteristic impedance which is typically a few hundred ohms. As long as the dc resistance of the coil is small compared to this then the Q of the circuit will be largely independent of the coil dcr. It only becomes an issue when you need really large inductance like 5H or so for bass bands.

Cheers

Ian

 

[dirty1_1garry]

RM6 Al100 is already air-gapped.

I mean winding air core inductors.

Do you know you can trim the inductors by wiping the two halves against each other?

Good idea! As typically I got higher inductance I can increase gap distance by using layer of tape for example and as result - lower inductance - profit!)

 

[dirty1_1garry]

Lot's of goof information there. One thing I would add is that the smallest dc resistance of the coil is not usually necessary. The Q of the circuit depends on its characteristic impedance which is typically a few hundred ohms. As long as the dc resistance of the coil is small compared to this then the Q of the circuit will be largely independent of the coil dcr. It only becomes an issue when you need really large inductance like 5H or so for bass bands.

Hi Ian, thanks for your opinion!

 

[abbey road d enfer]

I mean winding air core inductors.

Air core implies winding techniques that are not DIY, like honeycomb. Standard winding results in excessive parasitic capacitance.

Good idea! As typically I got higher inductance I can increase gap distance by using layer of tape for example and as result - lower inductance - profit!)

I was actually recommending wiping as a way to increase inductance. If the inductance is already excessive you should decrease the turns count.
Adding spacers (like rolling paper) results in quite large variations.

 

[warpie]

Adding spacers (like rolling paper) results in quite large variations.

Can you please expand a bit on this? Do you mean that by separating each layer of turns, the tolerance changes a lot?

 

[dirty1_1garry]

Can you please expand a bit on this? Do you mean that by separating each layer of turns, the tolerance changes a lot?

Abbey mean adding spacer between two cores halves, that increase gap distance and as a result lowering inductance

 

[warpie]

Abbey mean adding spacer between two cores halves, that increase gap distance and as a result lowering inductance

Ah OK, thanks

 

[dirty1_1garry]

I've reemulated eq points for real inductance that's I've got after inductor winding.  Adjusting capacitors ( make them lower for a little bit sharper bell) gave pretty satisfying result. See attached.

This is a way to solve the problem this time, but a more stable way of having a more accurate inductance has to be tried.

 

[abbey road d enfer]

This is a way to solve the problem this time, but a more stable way of having a more accurate inductance has to be tried.

When I wound prototypes, I had a tolerance of about +/-5%, that I could somewhat adjust by rubbing the core halves. Then when I had them subcontracted for production I asked for a better tolerance. The subcontactor had to select them, because there was still about the same variations, so they had to rewind those that were too far.
IIRC, the Al of small-gap cores has a tolerance of +/-3%. The ungapped are much worse, I think +/-20%.

 

[dirty1_1garry]

IIRC, the Al of small-gap cores has a tolerance of +/-3%. The ungapped are much worse, I think +/-20%.

I use gapped of course. Including because ungapped have too high AL.
Also just faced with a fact that Epcos RM-6/5/4 (regard to manual) have core type with center hole where it's possible to install adjusting screw for changing gap value. That can be a solution too.

 

[abbey road d enfer]

I use gapped of course. Including because ungapped have too high AL.
Also just faced with a fact that Epcos RM-6/5/4 (regard to manual) have core type with center hole where it's possible to install adjusting screw for changing gap value. That can be a solution too.

It's a possiblity, but it increases the cost.
I think adjusting the capacitor value is also a good solution. If you need to match response, you want to pair inductors.