Electric Automation Forum
Forum » General Discussion » Design high current air cored inductor
Topics: Design high current air cored inductor on General Discussion
#1
Start by
Charlie
09-08-2013 11:42 PM

Design high current air cored inductor

Anybody ever build their own LISN? Need advice for design of high current air cored inductor!!

Mainly due to lead time issues I am considering building a couple of CISPR 25 compliant LISNs for automotive EMC work. These have a really simple network so the only real challenge is an air cored inductor capable of doing (ideally) 100A continuous. Any advice folks? Frequency range is 150kHz to 108 MHz. Is it really as simple as some turns of "Litz" round a former? I have never designed an air core inductor in my life other than for high current switch pulse testing when the "design" was a couple of turns of equipment wire that was kicking around with cable ties to stop it leaping apart!!
09-08-2013 11:43 PM
Top #2
Charlie
09-08-2013 11:43 PM
Update - sleeping on it am not convinced I have to use "Litz" wire. The HF amplitude of the current is going to be very low so ...... Sure it will add some parasitic resistance in series with the inductor and reduce the Q but I suspect that it wont matter that much as the allowable impedance characteristic is quite generous in the standard. On the other hand I need to be able to form the coil without a large hammer and I think I need approx 5mm (0.1" to our amercian friends) diameter wire to carry 100A so .....
09-08-2013 11:43 PM
Top #3
Terry
09-08-2013 11:43 PM
I used sleeved multiple strand varnished wire to eliminate eddy current heating. 27khz 30A peak sine. Flat induction heating coil.
09-08-2013 11:44 PM
Top #4
Stephen
09-08-2013 11:44 PM
Don't forget the stray field. A solenoid will have a huge stray field, so nearby things will affect the rf performance. Orientate the coils orthogonally and make sure there is nothing ferrous or low resistance nearby. Consider toroidal coils.
09-08-2013 11:44 PM
Top #5
William
09-08-2013 11:44 PM
I built one a few years ago. I based it loosely on the DIY EMI series from Keith Armstrong. I used 5 inch PVC pipe as a form. For 5 ┬ÁH it was something like 6 or 7 turns with a very specific pitch. I used 6 AWG (0.184") spaced with cotton clothesline to maintain the pitch. I had to tweak the coil after it was placed in the housing. Even though it wasn't very close the proximity to the metal box definitely had an impact on the inductance. I compared it to our 16A commercial LISN and and it was pretty close below 20 MHz. Looked at it with a network analyzer and the impedance was far from flat between 20 and 30 MHz. Lead dress and component selection cleared most of it up. Getting flat response to 108 MHz will be tricky.

The coil gets surprizingly hot. I didn't have enough ventilation and the PVC jacket on the wire started smoking at 65 Amps. If I need to rebuild it to use again I'll go to 4AWG and add forced cooling. All in all a pretty good learning experience.
09-08-2013 11:45 PM
Top #6
Charlie
09-08-2013 11:45 PM
Thanks for your helpful comments folks. I am going to construct the LISN as two completely separate boxes (one for each line) which is actually the norm for these CISPR25 types. Then I only have one coil per box so I don;t need to worry about stray fields coupling. However William has confirmed my fear that the box will affect the inductance value but I guess I could oversize the box. I had already thought about a toroidal form (thanks for suggesting it Stephen) but if my calcs are correct it would get a fair bit bigger. The only other things in the box will be a 1uF film cap on the input and a 100nF + 1kOhm for the measurement port. For the immediate little EMC challenge we have I only need the LISN to be accurate up to about 5 MHz which should hopefully make my life easier.
Reply to Thread