Run capacitors, guage of wire

Hello All,

I am in the final phases (sorry, bad pun) of upgrading my RPC from a homemade 7.5hp to a 20hp (with 1hp pony motor). So I needed to rebalance the phases and alter the run capacitors. The old unit was based on an old, old Phasmatic cabinet that looks like it dates back the Korean War. At least judging by the capacitors that came with it, which are slightly larger than lantern batteries yet rated only for 20uF. Still work fine, but I only had six of them (120uF, the cabinet was marked as a 5hp unit). So I bought a half dozen new 50uF motor run capacitors, and in conjunction with an assortment of 2nd hand capacitors of various values I had on the shelf, I tuned the RPC per the W.Fitch instructions. Got it tuned quite well, I think.

But it raised a question, how much CURRENT do the run capacitors have to handle? Do they need to be able to handle the full potential current of the RPC, or just some portion of that based on the amount that they are balancing the voltage? The question is important to me, as I want to know what gauge wire I ought to be using to connect the capacitors.

When I upgraded the RPC from a 50amp circuit to a 75 amp circuit, I upgraded the wiring as well. From a probably inadequate 10ga that I inherited from my father, to a new 6ga line that I installed. I just added a 1988 Mazak QuickTurn 10N, and needed more power. The old capacitors had threaded terminals to which I was able to route #6 conductor (left over from the new wiring job, waste-not). See first attached image. The new capacitors one buys today seem to all have 1/4 inch spade terminals. Since there are four spade connectors per terminal, I used two of them to double up the #12ga wire. See second attached image. This was taken while I was still 'tuning' the amount of capacitance, so it is a bit jury-rigged and prior to soldering the flag connectors and insulting the bare conductors. But it is easier to see what I did before it was all covered up. Now there are four spades per terminal and I could use all of them, but the whole thing is just spot welded to a post so I am sure the current capacity is not nearly as high as the older capacitors. Also I am sure each spade connector is only good for a few amps anyway.

I am having an issue where on maximum acceleration/deceleration of the spindle I am having voltage fluctuations that sometimes trip an e-stop condition. So I am investigating various areas of my RPC where I think the current capacity might be lacking. The wiring for the capacitors is the lightest gauge in the entire system. From the Mazak to the RPC the wiring is #2 (what it came with), everything else is #6. Voltages are AB-228v, AC-243v, BC-248v where AC is the incoming line and AC, BC are the generated phases. Under a steady load the generated phases are pulled down (which is why it is recommended they be set high) to AB-225v, BC-231v, AC-228v; which I thought was reasonably uniform. I have a bunch of different motor sizes in the shop and you cannot please them all, so I balanced for the largest load. It might be that #6 cable to the RPC is just not enough. But I was also curious as to the role the capacitors played relating to current loads. By the way, the steady state draw on the three phases (spindle running 2k) is 27, 24, and 26 amps. Again, I thought reasonably balanced.

Pictures of the RPC included for those interested. All salvage material except for a Lovejoy coupling bought from an industrial surplus supplier on eBay for $30.

Thanks,
Doug