An engineer is making a high current transformer. He has Litz wire from previous projects but he needs about three times the current and is thinking that he can wind three parallel bundles of Litz wire rather than having to wait while the new size is manufactured. Listenng to this it occured to me that he would probably get more even induction among the separate strands in all those insulated wires if he twisted the thee Litz bundles together.

Since this is expensive custom-made wire he would like to have an opinion as to whether twisting is likely to lower losses rather than just running an experiment.

Does anybody here have any relevant experience or insight?

Thank you in advance for any helpful replies.

 

No personal experience, but the question was interesting enough to do a search. There are eight types of litz wire, including bundles of litz wire. It appears that when litz wire is bundled, it is usually weaved/braided without a core or twisted around a fiber core. The rationale seems consistent with your assumption (source #1 reference):


Source: 3rd reference:
There are eight standard types of litz wire available, including:

• Type 1: A single twist with elective external insulation.
• Type 2: A twisted bundle of Type 1 weaved together with an option for external insulation.
• Type 3: Pre-insulated bundles of Type 2 braided/weaved together with an option for external insulation around the full bundle.
• Type 4: Twisted bundles of non-insulated Type 2 with a fiber core.
• Type 5: Bundles of insulated Type 2 with a fiber core.
• Type 6: Bundles of Type 4 with a fiber core.
• Type 7: Braided insulated wire then formed in a rectangular shape.
• Type 8: Single twisted insulated strands compressed into a rectangular shape.

I did not find nor look specifically for what properties each type of bundling has. It might be that the braiding is for added flexibility rather than an electronic reason.

References:
1) https://usatransformers.com/transformer-winding-conductor/
2) https://info.triadmagnetics.com/blog/litz-wire-vs.-copper-foil
3) https://www.agilemagco.com/blog/5-benefits-of-using-litz-wire/ (Litz wire types)

Some or all of the referenced companies offer free consultation. Maybe a call?

 

IMO far too many unknown variables here to say much informative, other than, running an experiment would be a wise course of action.

http://edge.rit.edu/edge/P10022/public/team_docs/Wires_Casing/NEWENGLANDlitzSpecs.pdf

 

I have this rebus to solve many times every year. Just every lovely 100-300 kHz SMPS have this sacramental question at winding of ferrite transformer. If say having the 10...50 or more Amperes in bobbin and one may know the Focault layer thickness is ca 30 microns or even 100 micron, any, the wire being twice-as-much thick is just senseless waste of copper and money. Therefore unwritten (may be somewhere written) law is to divide the wire into smaller insulated strands, and practice recommends a diameter 0,5...0,8 for most cases except the MHz scale where 0,2-0,3 is more handful; and put it accordingly much identical together to get those desired 3,5 A/mm2 or even 2,5 A/mm2 for higher frequencies.

So, to push it free or pre-wind it in the braids (pigtails) that sure is question more of the work-culture than belief of which is better by quality. For me, I always braid it.

Tech is rather simple, take a nail with hook in drilling machine or screwdriver machine, push the two nails somewhere on the shelves or walls or trees; let it be even in worst case at the street, but far enough let the braid length would be enough for winding that bobbin until calculus, and when one have run those 10...30...100 times from nail to nail (according calculus) with wire producer bobbin in hands, then cut the wire, pull one nail content toward screwdriver machine and push the gun-button. Wait a bit until it looks nice and release. Your litz is brand ready.

Take in account it may become some 5...10% shorter than original distance between the nails. Theoretically (but none do like that) You may put into formula the work frequency and pigtail winkels per centimetre, then the wire impedance PROBABLY may be optimized. But it is more a thing of belief for calming the conscience nor the actual need per se.

 

You don't give any scale to the problem, for example frequency, size & style of bobbin/core and size of existing bundles. It could be the trifiler winding will have a lower leakage inductance than the triple litz bundle on the basis of poor fill factor of the latter. I have made many Lit bundles using the method mentioned above but certain constructions are more compact than others so my final choice of wire size/strands is influenced by that. A bundle of only 3 is a small number and often results in a lumpy result with poor fill factor and consequentially higher leakage inductance.