Help with Wire size for capacitor bank

Thread Starter

Christopher Wolford

Joined May 30, 2018
5
Skip to needed info and questions if you want.

Background:

I started a project, converting DIY ac welder to DC. I dont like to follow along blindly, I like to learn and understand what i'm doing. Poof it eats up madding amount of time. I'm a dad, student and part time worker, so i'm sort on that atm. Currently my attempt at makeing bus bars with out a proper Machine shop came out less than perfect leading me down a new path, needing new information.

I am at my wits ends atm and I need an answer to my question(s) but would also like to be pointed at the books or websites that would help me figure this stuff out on my own in the future.

Needed info:

I have 10 Malloory 10,000 50 VDC CGS103U050R4C, That I am putting in parallel to get the ripple current and high energy storage for this welding application. The capacitor bank will be feed by a single phase 60hz 120v ac stepped down, to about 30v ac feed to a 150 amp bridge rectifier.working with 4 awg wire. duty cycle of about 20% @ 90 amps and max out put is about 125amps but the duty cycle drops to like 10%. we are talking 2 min weld time max out of 10 min.

I tried to make 2 banks, with 4 bus bars attaching the bus bars to the top of the capacitors. Had issues keeping bus bars straight, drilling holes prefect, and not a lot of room to keep them from short circuiting.
4 and 6 awg lug connectors will not fit the capacitors, too big and would be hard to keep them from shorting. lead me to the bus bar idea.

questions

Then it hit me:
New idea: Each capacitor only supply's part of the current, therefore I could use smaller wire to feed each screw top capacitor to a thick copper bus bar.
Does this seem like correct circuit thinking? Will this work safely?

How to gauge wire? +20% of ripple current of each capacitors 8.5 amps @ 120hz. ?

Also my understanding is that in welding the capacitor bank ripple current only needs to supply 30-40% of total welding current. I went well above this just to be sure.

Thanks for your time.
 

#12

Joined Nov 30, 2010
18,224
This drawing demonstrates the math and shows that using 12 AWG wire can result in a peak loss in the capacitor wiring of about 40% what you lose in the six foot cables. I came up with a very different ripple current than you did, even accounting for the idea that I used p-p voltage. An obvious error in this model is the fact that your capacitors probably have a lot more internal resistance than an inch of 12 AWG wire. Anyway, I think 12 AWG wire is a good choice and 14 AWG wouldn't be much worse, comparing everything to the 4 AWG cables.

This is a sloppy estimate but it shows feasibility.

ps. I have found that copper pipe makes a cheap and handy buss bar.
 

Attachments

Thread Starter

Christopher Wolford

Joined May 30, 2018
5
I have not looked at connectors but was thinking 10 to 8 awg wire, but if some one can show smaller working, it saves money.

The length from cap to bus bar would be like 6 in to a 1 foot at most. I can shorten if need be.

4 awg from rectifier to bus bars, 4 awg from bus bar to welding leads.

Looking at esr of 17 mOhms. each cap, it drops as you put them in parallel if I do recall.

Thanks for the info.
 

ebp

Joined Feb 8, 2018
2,332
If you use wires to the caps, try to keep the length from the common point to each capacitor reasonably close to the same. This helps to assure that the capacitors share ripple current equally. A little resistance in the wire is not a bad thing, as it helps a bit in balancing. 12 AWG is probably a good size, in part because it is the maximum size that will fit in some crimp-on ring lugs.10 paralleled 12 AWG conductors have the same cross sectional area as 2 AWG
 

#12

Joined Nov 30, 2010
18,224
try to keep the length from the common point to each capacitor reasonably close to the same.
That seems like a good idea at first glance, but...
A foot of 12 AWG is 1.6 milliohms. These capacitors are 17 milliohms. Therefore, the caps have more than 10X the resistance of the connecting wires. In theory, the resistance of the wire is insignificant. Still, there is no law that says you can't give each capacitor a foot of wire and just zig-zag the extra length.
ps. I calculated for 2 feet of 12 AWG wire as if 20 jumpers could be made from 24 inches of wire. Now that we know the resistance of the capacitors, that specification is a lot tighter than it needs to be. Using a foot for each capacitor still leaves the wire resistance at, "insignificant".
 

Thread Starter

Christopher Wolford

Joined May 30, 2018
5
i'm not a 100%, but does the ESR value not drop cause I will be putting this capacitors in parallel. Its been a while since I looked at the circuit book, just now getting back to it. would the 1/Rt work here, the caps are all the same, in parallel and seen as a resistance load. I have to admit, I have not learned that much on caps, at this point. I not sure how equivalent series resistance works when I pop them in parallel. Also I was reading about an enormous resistance at the arc point during welding operation. There is also the fact that I will be putting 50-100 ohm bleed resistor(s) on the cap bank.

You all raise good points about voltage drop, losses, and etc.. my first concern was related to the high amp load and making sure that the load was split on each branch of the parallel caps. So I could use smaller wire, and not have a fire/ melting / break in the circuit.

I do know the duty cycle play to my advantage, it lets smaller wire be used for bigger current, not a lot i could find on it. Other than welding leads 8, 6, 4, 2 gauge wire and there loads at duty cycle x%.

the circuit will be mounted in a wind tunnel, a custom made box, with 4 120mm pc fans. 2 for the intake, 2 for the exhaust. The rectifier is mounted to an old heat pipe pc cooler, plan on having the heat cooler with a fan mounted outside the box, but the lower part of the cooler in the box for wiring easy and extra cooling.

This has been a fun project so far, thanks for your help guys. I'm good with math and other stuff, not spelling and grammar. Sorry if my post is hard to follow.
 

jpanhalt

Joined Jan 18, 2008
11,087
If you reconsider bus bar,here is what I did:
upload_2018-5-31_14-48-6.png

Bought the copper strip of the right width and thickness. Then it was only a matter of drilling and a few cuts (jig saw). As you found out, electronic copper can be hard to machine because it is so stringy. I just use an oil-based cutting fluid. Ordinary kerosene or diesel will work. Fastened like that, the capacitors are self-supporting and do not roll around.
 

Thread Starter

Christopher Wolford

Joined May 30, 2018
5
If you reconsider bus bar,here is what I did:
View attachment 153450

Bought the copper strip of the right width and thickness. Then it was only a matter of drilling and a few cuts (jig saw). As you found out, electronic copper can be hard to machine because it is so stringy. I just use an oil-based cutting fluid. Ordinary kerosene or diesel will work. Fastened like that, the capacitors are self-supporting and do not roll around.
I used anvil and hammer with cooper pipe, was good and straight, but i don't have a drill press, and my hand held drill caused issues. wrong bits too. I came up with many better ways to do the bus bar, but at this point i'm in need of a better welder and need to get this done. I was going for neat and cool looking factor, now i'm going for practical and get the job done. Love the photo btw. I may have over done my cooling system XD. How that running with out 1 inch spacing on the caps for cooling?
 

Thread Starter

Christopher Wolford

Joined May 30, 2018
5
unless someone has some more to say, will be running this with with 10 awg wire from each cap to the bus bar (found 10-12 awg ring connectors). It may be over kill, 10 awg is rated for 30-40 amps based of normal tables, for 0-4ft crutchfield has it listed for 85 amp (max) draw for a car amplifier.

I don't feel each wire will see more than 10 amps, as the ripple current for the caps are about 9 amps. some theory somewhere says they really only put out about 40% of the welding current anyways. 125*.40= 50amps.

anyone got a comment on if the esr drops as you put the caps in parallel?
anyone think the logic on the caps splitting the current is sound. even if they split all 125 thats still only 12.5 amps each cap( even tho max ripple is 8.5 amps). there will be air blown over the single conductor wires, and they will only be used 2 mins out of 10, based on duty cycle.

I dont have a lot of law backing me up that i know of, looking for pros inputs.
Thank all of you for your time and input.
 
Skip to needed info and questions if you want.

Background:

I started a project, converting DIY ac welder to DC. I dont like to follow along blindly, I like to learn and understand what i'm doing. Poof it eats up madding amount of time. I'm a dad, student and part time worker, so i'm sort on that atm. Currently my attempt at makeing bus bars with out a proper Machine shop came out less than perfect leading me down a new path, needing new information.

I am at my wits ends atm and I need an answer to my question(s) but would also like to be pointed at the books or websites that would help me figure this stuff out on my own in the future.

Needed info:

I have 10 Malloory 10,000 50 VDC CGS103U050R4C, That I am putting in parallel to get the ripple current and high energy storage for this welding application. The capacitor bank will be feed by a single phase 60hz 120v ac stepped down, to about 30v ac feed to a 150 amp bridge rectifier.working with 4 awg wire. duty cycle of about 20% @ 90 amps and max out put is about 125amps but the duty cycle drops to like 10%. we are talking 2 min weld time max out of 10 min.

I tried to make 2 banks, with 4 bus bars attaching the bus bars to the top of the capacitors. Had issues keeping bus bars straight, drilling holes prefect, and not a lot of room to keep them from short circuiting.
4 and 6 awg lug connectors will not fit the capacitors, too big and would be hard to keep them from shorting. lead me to the bus bar idea.

questions

Then it hit me:
New idea: Each capacitor only supply's part of the current, therefore I could use smaller wire to feed each screw top capacitor to a thick copper bus bar.
Does this seem like correct circuit thinking? Will this work safely?

How to gauge wire? +20% of ripple current of each capacitors 8.5 amps @ 120hz. ?

Also my understanding is that in welding the capacitor bank ripple current only needs to supply 30-40% of total welding current. I went well above this just to be sure.

Thanks for your time.
What a small world. I am doing the exact same thing with a cheap Zeny Ebay 130A flux core welder and I have the same number of 10000mf capacitors with the SAME QUESTION! I decided to wire capacitors with individual wires to a thicker bus wire (like size 8) because I don't have a bus bar or a copper pipe handy. Also Bus bars are great but I have space issue in the case and I don't want to modify the case or mount the fan outside as other people did. I can fit 6 of them at the back of the transformer and 4 of them at the front. We'll see how it works. How did yours turn out?
 
Top