Tolerances with Higher AL Cores vary... Is it Temperature?

[thermionic]

edit: see follow up posts

Hi,

I've just received a batch of 5 different inductor values, wound on AL400 and AL3800 cores. The values (<150 mH) on the 400 cores are within 2%. However, the 440mH, 500mH and 1H values wound on AL3,800 can only achieve +/- 10%. The winding firm tells me the cores specify +/- 10% and that cores with adjustable gaps (the ones with the screws) won't achieve any better (he claims to have tried them). The winder said he'd done better than this, but the values weren't consistent...

For the project I have in mind, +/-10% is unusable. Can anyone give me an idea as to how I can improve on this? Please don't tell me to buy a load and put them in series - that's not funny!

Thanks in advance.

Justin

 

[PRR]

The obvious thing is to wind them high, then take turns off. If DCR is critical, add a resistor.

 

[abbey road d enfer]

Narrow-gapped half-cores need to be pressed and rubbed against each other before QC'ing them. You'll find probably 5% change between before and after adjustment. If the inductance is too high, old-timers would put a piece of cigarette-paper between one of the legs.

 

[thermionic]

Well, this is rather embarrassing...

[Sidenote]: I've been with the winder making inductors before. He's been in the game decades and has an illustrious list of clients in the military sector etc. He always overwinds (via CNC machines) and gradually reduces until measurement hits tolerance (as PRR suggests). He always 'grinds' the cores prior to lacquering.

I just got onto the phone with him, complaining that his 3,800 cores were anything up to 25% off. To illustrate my point, I went down to the workshop and started re-measuring with him on the phone... Guess what? They were within 2%. Inductors which measured 480mH last night, now measure 440mH (I had this written on them).

Upon telling him this, he stated that he'd seen the same problem. When he came into work and the workshop was cold, they'd measure differently.

I really don't know what to think. I sent some to a friend who has a sophisticated HP analyser and he measured different values.

The AL400 cores seem stable, regardless. My only thought is to find a lower AL core for the larger inductors, and go for more turns. Instead of 3,800, maybe go for AL1200 and increase the turns?

I'm now wondering about the ramifications of this with regards to the client who the item utilising the inductors is destined for... Maybe he'll like it; "hey, whenever I use this unit I notice different details...it's really revealing"

Thanks.

Justin

These are RM10 cores btw.

 

[thermionic]

Anyone? Have you seen ferrite cores with higher AL specs change with temperature? Is it purely temperature? Could there be other factors? I've re-checked the values with my LCR meter and it repeats the same values when measured within minutes - it's when you leave them for a while the value changes.

The AL400 cores measure rock solid, regardless of temperature.

Cheers,
J

 

[abbey road d enfer]

It seems to be difficult to get proper info on the subject; most articles talk about temperature drift without actually giving numbers.
I eventually found this:
http://www.magnet-tech.com/core/MnZn/characteristic_graphic1.htm
As you can see, some materials have almost +100% drift over 40°C temp variation.

 

[thermionic]

It seems to be difficult to get proper info on the subject; most articles talk about temperature drift without actually giving numbers.
I eventually found this:
http://www.magnet-tech.com/core/MnZn/characteristic_graphic1.htm
As you can see, some materials have almost +100% drift over 40°C temp variation.

Jeepers! We seem to talk about inductor winding on a daily basis here, how come this hasn't been discussed?

Many thanks, Abbey - much thinking needs to be done.

Justin

 

[Dan Kennedy]

With RM6 size and AL400 material I also see temperature
variations.

I figure the internal temperature of the gear once it's reached
equilibrium will tend to stabilize things.

A cop-out I know, but an oven seems a bit overkill.

 

[thermionic]

With RM6 size and AL400 material I also see temperature
variations.

I figure the internal temperature of the gear once it's reached
equilibrium will tend to stabilize things.

A cop-out I know, but an oven seems a bit overkill.

The RM10 / AL400s I have are perfectly stable. It's the RM10 / AL3800 that are the problem. They seem to measure best in the morning, before the heating's warmed the workshop up.

An oven could be a good investment - you could prepare your lunch at the same time. That's far more important than any inductor tolerance.

Justin

 

[abbey road d enfer]

To Dan Kennedy:
An RM6 AL400 is basically an AL2200 with a 0.05mm gap. Dividing AL by a factor of 5.5 roughly reduces the drift by the same coefficient.
By using an RM10 AL400 (or maybe AL250, since the larger size allows more turns) with a 0.21mm gap, the drift will be reduced by a factor 10-12. Clearly one case of the bigger the better.

To thermionic:
I don't have data for RM10 AL3800 but I suspect it's ungapped or very small gapped; I think you'll have to go for an RM14 AL1000, where the larger bobbin may compensate for the lower AL. A P30 core AL2000 with 0.05mm may be right for you.

 

[Dan Kennedy]

In discussions with my coil winder we came to the same basic conclusions.

It's a six year old board at this point, no changes in the product lifetime now,
and as I said, the variations in actual use are minimal.

But I won't push the capabilities of the core next time. All of the parameters
get fiddlier. Winding tensions, assembly, fragility of wire, all have been issues.

 

[thermionic]

Abbey - I've given your suggestion to the winding firm. Let's see if they can get a handful of samples made and how they work out. RM14 sounds more expensive though!

Many thanks - this is really specialist stuff - much appreciated.

Justin