ok, so I am trying to select some capacitors for a welder i am building,
I found a stockpile of 64000μF caps for $3 each, however they do not have too many markings on them besdes that, other than to say they are 20V and SURGE 25VDC. I was thinking of using two banks of these, for a two step discharge, however, I do not know if they are capable of handleing the current discharge. they are a scew terminal can, i'll get a pic to post when i get home from work.
Is there a way i can safely test if one of them would handle being shorted?

the other question is about SCR/thyristors/relays.

I'm plannign on using one of these three as a switch to allow current to flow to the discharge electrodes. I have it set up so that the time the switch is open can be varied, however i do not know enough about the SCR/Thyristor to tell if it has a response time quick enough. the Relays probably would, though i dont know if i could find one to handle the current load. (at least 2kA)

I've read several posts, and articles on the SCR/Thyristor, but can't seem to find the answers I am looking for (probably just didnt recognize them)


An help and advice (on these or the project as a whole) would be greatly appreciated.
Thanks in advance

Mike

 

1. The capacitors are going to be fine with the discharge current. But they willl need to be reformed before use (hopefully, the voltage on them won't be right at their limit). Use a DC supply and a 10 - 20K resistor to allow each one to charge up to its rated voltage over a period of time. That lets the dielectric film reestablish itself.

2. SCR's and Triacs are AC devices. If triggered on DC, they will latch on until current through them falls below some low level. A power FET or two is more the thing to use.

I don't have my catalogs here, so I can't come up with numbers for FET's, but someone will probably make a recommendaion.

 

ok, so I am trying to select some capacitors for a welder i am building,
I found a stockpile of 64000μF caps for $3 each, however they do not have too many markings on them besdes that, other than to say they are 20V and SURGE 25VDC. I was thinking of using two banks of these, for a two step discharge, however, I do not know if they are capable of handleing the current discharge. they are a scew terminal can, i'll get a pic to post when i get home from work.
Is there a way i can safely test if one of them would handle being shorted?

the other question is about SCR/thyristors/relays.

I'm plannign on using one of these three as a switch to allow current to flow to the discharge electrodes. I have it set up so that the time the switch is open can be varied, however i do not know enough about the SCR/Thyristor to tell if it has a response time quick enough. the Relays probably would, though i dont know if i could find one to handle the current load. (at least 2kA)

I've read several posts, and articles on the SCR/Thyristor, but can't seem to find the answers I am looking for (probably just didnt recognize them)


An help and advice (on these or the project as a whole) would be greatly appreciated.
Thanks in advance

Mike

If your current load is indeed 2,000 amps then you are in no position to come up with a DIY solution. No disparity meant by that, but that kind of current flow would challenge even an experienced electrical engineer and is going to cost some serious cash for any kind of switching components rated for that much current. You need a different plan

 

I'll agree the 2kA number is a bit high.
I've seen this design done before (battery tab welder using a .6F cap bank)
looking at the math i did again, i think i assumed too low of a load resistance for welding at max charge (20V on a .64F cap bank)
most of the welding i will be doing will be on 14Ga down to 22ga stainless.

i've had to make alot of assumptions based on some material properties i havent found yet, hence the 2kA. if it really is that high...lol then yeah i am in a bit of a pickle.

though here is another question.... is there a wway to measure the ESR of a cap? is it just as simple as sticking an ohmmeter across the terminals?

 

I'll agree the 2kA number is a bit high.
I've seen this design done before (battery tab welder using a .6F cap bank)
looking at the math i did again, i think i assumed too low of a load resistance for welding at max charge (20V on a .64F cap bank)
most of the welding i will be doing will be on 14Ga down to 22ga stainless.

i've had to make alot of assumptions based on some material properties i havent found yet, hence the 2kA. if it really is that high...lol then yeah i am in a bit of a pickle.

though here is another question.... is there a wway to measure the ESR of a cap? is it just as simple as sticking an ohmmeter across the terminals?

Well sometime a cap manufacture list publish the specification. There are capacitor testers that will measure it directly. Here are a couple of quick google hits:

http://www.emcesd.com/tt020100.htm

http://octopus.freeyellow.com/esr.html

 

A couple of comments:

First, 14 ga stainless may be a bit much for a 0.6 to 1F capacitor discharge welder at no more than 20 V. I can get tack welds on 0.010 wall SS tubing. They are certainly not structural.

Second, while I can see the logic in using a charging resistor, 20K resistor seems a bit large. The TC for that charging a 0.6F capacitor is several hours.

Third: The currents are probably exaggerated, but they are significant. My set up with 0.6F and 18 to 20 volts using #4 AWG welding cables (about 36" each) causes a very noticeable jerk in the cables.

I am approaching capacitor charging with current limiting to 4 to 6 A. Discharge is with an SKKT 71/72 Semikron SCR rated at 125A continuous and 1600 A pulse on the low side, as described in the previously given references to the Poor Man's Battery Welder.

John

 

That 20K value was for a one-time use to insure the plate dielectric reformed under low stress conditions.

 

Measuring ESR can't be done with a ohmeter. It can be done with a VNA (Vector Network Analyzer). I'm guessing that instrument might be out of your price range. There is a cheaper instument called a gain/phase meter that you might be able to use, cheaper yet would be a signal generator and a scope. The problem is that ESR for most capacitors is pretty small, and it is a function of frequency, which is important if you try to charge the capacitor with a step function like you would in a switching regulator or in your application. I'd lend you my VNA, but I'd need your house for collateral. (LOL). You probably could rent one for say $4661/mo

http://www.metrictest.com/product_info.jsp?mfgmdl=HP%20E8364B-014-080

 

Thanks for all the advice, i really appreciate it.

makes me wonder, though, if the answer is right here in front of my face.
here is the printing ont he capacitor

23D85681N01
64000UF
20WVDC
Surge 25VDC
8952L 4702 (2)
36DE1673
7M138

my guess was the top line was the part number, then its Farad rating, followed by voltage ratings. the last three lines i think may be manufacturer markings, though I really dont know.


I had a feeling that going for the 14Ga with this was a bit more than it would be able to handle, though my main intent was for much smaller thicknesses.
the current plan is to charge the cap banks with a SEPIC DC/DC converter fed by a PC power supply. still working through the equations

I had seen a link to a 'dual pulse' capcitor discharge welder that this one company had built, and was going to try and mimic the same function, though one of the stumbling blocks i am hitting is with the SCR switch. If the signal to the gate on the SCR drops to ground, will it cut the current flow through it? or will the SCR continue to allow current flow till it drops below a certain level? ( this is one area i've been confused about with SCR opperation). and if the second condition is true, is there a way to "force" it off?

thanks again everyone,

 

Thanks for all the advice, i really appreciate it.

makes me wonder, though, if the answer is right here in front of my face.
here is the printing ont he capacitor

23D85681N01
64000UF
20WVDC
Surge 25VDC
8952L 4702 (2)
36DE1673
7M138

my guess was the top line was the part number, then its Farad rating, followed by voltage ratings. the last three lines i think may be manufacturer markings, though I really dont know.


I had a feeling that going for the 14Ga with this was a bit more than it would be able to handle, though my main intent was for much smaller thicknesses.
the current plan is to charge the cap banks with a SEPIC DC/DC converter fed by a PC power supply. still working through the equations

I had seen a link to a 'dual pulse' capcitor discharge welder that this one company had built, and was going to try and mimic the same function, though one of the stumbling blocks i am hitting is with the SCR switch. If the signal to the gate on the SCR drops to ground, will it cut the current flow through it? or will the SCR continue to allow current flow till it drops below a certain level? ( this is one area i've been confused about with SCR opperation). and if the second condition is true, is there a way to "force" it off?

thanks again everyone,

"If the signal to the gate on the SCR drops to ground, will it cut the current flow through it? or will the SCR continue to allow current flow till it drops below a certain level? ( this is one area i've been confused about with SCR opperation). and if the second condition is true, is there a way to "force" it off?"

No, and that is one of the main difficulties with working with SCRs in DC circuits. The triggering is easy and very sensitive, however once a SCR starts to conduct the only way to turn it off is the remove the trigger current AND (the hard part) remove the forward current flow to zero. Hence at DC a SCR is a turn on once and stay on until power is interrupted externally. In AC service the reversal of current flow automatically turns off the SCR at the end of one conduction cycle.