Safe Switching Circuit for Capacitor Discharge in 250VDC Jewelry Spot Welder Project

81 32

#1 21660497 30 Dec 2011 21:36

Sarah Harris Sarah Harris

Anonymous
Post #1
21660497 30 Dec 2011 21:36

I have a project in mind.
at the moment, i use a charged capacitor that i short out for the purpose of tack welding in my jewelry making. i would like to make it more user friendly as i have upto 250vdc hanging around at the end of a couple of crocodile clips . i would like to incorperate a way of switching the discharge instantly at the moment of short circuit (see image below ) so that my leads only become live at the point of short circuit. ...make sense ???
what sort of components should i be looking at for the switching ? please excuse my very limited knowlege of electronics
Sarah

Attachments:

forums-discharge-1325209010.jpg Download (8.8 KB)
ADVERTISEMENT
#2 21660498 31 Dec 2011 03:20

Steve Lawson Steve Lawson

Anonymous

Post #2
21660498 31 Dec 2011 03:20

How about something like the attached? SW2 represents the weld point. The P-channel MOSFET only turns on when the welding probes contact the metal (i.e. when a low resistance path is available. Otherwise only a very small current can flow. They still present a high-voltage at the probe tips, but because the current is limited to a few hundred microamps there is no danger.
ADVERTISEMENT
#3 21660499 31 Dec 2011 03:37

Steve Lawson Steve Lawson

Anonymous

Post #3
21660499 31 Dec 2011 03:37

Caveat: I wouldn't bet my life on the circuit that I posted (i.e. NO GUARANTEE!). Probably the best thing to do is wear insulating gloves
#4 21660500 31 Dec 2011 10:05

Bob Loy Bob Loy

Anonymous

Post #4
21660500 31 Dec 2011 10:05

Steve Lawson's idea is way more elegant than mine, but I was wondering if you couldn't just run your live wire through a heavy duty, momentary contact foot switch? Connect the short circuit first, then put the pedal to the metal, so to speak. Cheap, rugged, and ... inelegant.
ADVERTISEMENT
#5 21660501 31 Dec 2011 12:49

Sarah Harris Sarah Harris

Anonymous

Post #5
21660501 31 Dec 2011 12:49

This is my type of idea. ive done some on-line searching and got some sums.
560uf cap x 250v=17.5 joules, discharged in 1ms = 17000W = 68Amps does this sound right ?
could you point me i the right direction for a mosfet with this type of handling capabilities ?
I know nothing about mosfets, and when i read the tech spec i might as well be dyslexic i see letters and numbers but nothing else.
Thanks ever so much for the advice so far, its really apreciated.
Sarah
#6 21660502 31 Dec 2011 12:51

Sarah Harris Sarah Harris

Anonymous

Post #6
21660502 31 Dec 2011 12:51

it could be plan B. i would like to investigate the possibility of plan A first.
thanks
#7 21660503 31 Dec 2011 13:05

Sarah Harris Sarah Harris

Anonymous

Post #7
21660503 31 Dec 2011 13:05

to add, my voltage range is from approx 70v upto 250v max , will a mosfet handle such a range in voltage differences ?
#8 21660504 31 Dec 2011 14:48

Steve Lawson Steve Lawson

Anonymous

Post #8
21660504 31 Dec 2011 14:48

Hi Sarah,

Sometimes "inelegant" is the best solution, such as the hanging tennis ball in the garage to prevent crashing into the back wall ;)

I'm not sure (no experience with this) but the foot switch might have problems with welding shut or the contacts might rapidly degrade. If so, how about a hybrid (see attached). Plus, there's a charge indicator ;)

The question of the MOSFET is addressed in the Notes on the diagram.

Another Caveat: If you have little experience with electronics, then I would caution you not to be messing with high voltages. Also, because I haven't tried either of these circuits, I'm relying heavily on your ability to make adjustments where necessary. There are a lot of unknowns (for instance: that 30 volt protection Zener might divert some of the current needed for a proper weld. Which would mean you would either need to do without it, or use a higher voltage (though the idea is to keep the voltage down below 50 volts which is, in most conditions, the maximum safe voltage). Also, the first time you hit the foot switch with the probes NOT connected to anything, the Zener might actually blow out. In which case you would need to find one that can handle greater energies.

So, basically, all I'm providing, here, are ideas. It's up to you to tweak them into something that actually works
#9 21660505 31 Dec 2011 15:51

Sarah Harris Sarah Harris

Anonymous

Post #9
21660505 31 Dec 2011 15:51

Thanks tor your advice and concern. so far i have managed to work out how to make the inverter with active flyback for the charging. its just there are certain things i have never had any dealings with, mosfets (infact any FET ) being one of them.
its quite intimidating reading responses to some posts (on other forums) by knowledgeable members answering novices questions and bombarding them with technical advice way over their heads and belittling they ignorance so i have tried to work this out on my own, but on this one i was totally stumped.
with my project, i have been successfully using it for some time, gloves are not practical, neither really is a foot switch (I'm disabled with acute spinal injuries) but it has on the odd occasion bit me when ive dropped the grounded component on the floor and am scrambling about trying to retrieve it hence starting my quest to remedy the problem. the mosfet looks a promising solution.
the zenner is a nice touch but as you say the break down may be enough to trigger the mosfet and take the full blow of the discharge.. ive re-draw your first circuit into my software, and in principle it looks like it will work so i will give it a go and post how i get on.
Thanks again and Happy new year.
Sarah
#10 21660506 31 Dec 2011 16:10

Sarah Harris Sarah Harris

Anonymous

Post #10
21660506 31 Dec 2011 16:10

to add (again)
reading through the second circuit description again, the voltage divider is set to supply 210v to the foot switch with the input voltage at 250v which drops to 66v when the input voltage is 70v.????
i guess your drawn them round the wrong way ? or have i miss read something ?
and on second glance neither of these circuits will work on a variable input voltage ? or will they ?
#11 21660507 31 Dec 2011 16:35

Sarah Harris Sarah Harris

Anonymous

Post #11
21660507 31 Dec 2011 16:35

spotted (cringes with embarrassment ) you,ve reversed the ground polarity..
not an option as every thing (inc LV and power supply ) share a common grounding to earth
ADVERTISEMENT
#12 21660508 31 Dec 2011 18:15

Steve Lawson Steve Lawson

Anonymous

Post #12
21660508 31 Dec 2011 18:15

Yes, as long as it can handle the maximum voltage. I would go with about 150% of the maximum voltage (depending on how noisy the environment is and how likely surges and spikes are. But, with my 'rule of thumb' or 150%, 250VDC max would be a minimum Max Vds of 375 Volts.
#13 21660509 31 Dec 2011 19:02

Steve Lawson Steve Lawson

Anonymous

Post #13
21660509 31 Dec 2011 19:02

Probably the best thing to do regarding MOSFETs is read up on them. Here are a few suggestions:
* http://www.youtube.com/watch?v=Te5YYVZiOKs
* http://www.electronics-tutorials.ws/transistor/tran_6.html
* http://www.electronics-tutorials.ws/transistor/tran_7.html
* http://www.fairchildsemi.com/an/AN/AN-9010.pdf
* http://www.ecse.rpi.edu/~schubert/Course-ECSE-6290 SDM-2/1 MOSFET-1 Basics.pdf

Consider that most MOSFETs in general use are Enhancement mode (normally off), and the N-channel is more common, diverse, and usually cheaper than the P-channel. The MOSFET in my solution(s), here, are all Enhancement mode.

Regarding the polarity reversal--it's pretty simple to change the polarity back to Neg ground--just turn everything around and replace the N-channel MOSFET with a P-channel MOSFET. The Gate, Source, and Drain go in the exact same places. The trick will be finding an adequate P-Channel MOSFET for this application.
#14 21660510 31 Dec 2011 19:20

Steve Lawson Steve Lawson

Anonymous

Post #14
21660510 31 Dec 2011 19:20

You're right, I designed this for a voltage of 250VDC. To get it to work over the range of 70 to 250 is a bit more challenging. Maybe the attached will work.
#15 21660511 01 Jan 2012 09:42

Sarah Harris Sarah Harris

Anonymous

Post #15
21660511 01 Jan 2012 09:42

Thank you so much for your patience and input., ive read through the attached links, and really surprised by how useful they are why did i not discovered these before !!!!!!
But as you said previously, P type are pretty much impossible to find. ive been up all night searching and reading, serious brain overload happening LOL.
I guess ive got some serious re-thinking to do and try to establish how i can incorporate N types instead. heres a pic of what ive got (with a added switching latch circuit) I was controling flyback with IC1, but adding a latch that resets when the flyback voltage is reached controlled by VR1 seemed a more sensible option rather than the circuit charging constantly . the voltage indicator (referred to at the bottom) is a 10 led bargraph.
ive added your mosfet idea to the drawing .
i guess it shows that reversing the polarity would be tricky, i havent got it grounded to earth as i use a 12v wall wart power supply
#16 21660512 01 Jan 2012 14:26

Sarah Harris Sarah Harris

Anonymous

Post #16
21660512 01 Jan 2012 14:26

today ive been doing some more head scratching.
Would a thyristor work instead if a mosfet ? .
#17 21660513 01 Jan 2012 14:28

Sarah Harris Sarah Harris

Anonymous

Post #17
21660513 01 Jan 2012 14:28

for got to add pic
#18 21660514 01 Jan 2012 14:38

Steve Lawson Steve Lawson

Anonymous

Post #18
21660514 01 Jan 2012 14:38

How about this: use an optoisolator: Have Q3 turn on the LED in the opto isolator and use the opto's transistor in place of Q3. That way you can isolate the secondary and then turn the diode around and make it a positive ground system.

Let me know if you would like help with that. And then, if it is CircuitWizard you are using to capture your schematic, then can you post the .cwz file--then I'll make the changes and show them to you.
#19 21660515 01 Jan 2012 14:41

Steve Lawson Steve Lawson

Anonymous

Post #19
21660515 01 Jan 2012 14:41

I don't have much experience with Thyristers. But maybe an SCR.
#20 21660516 01 Jan 2012 15:26

Sarah Harris Sarah Harris

Anonymous

Post #20
21660516 01 Jan 2012 15:26

Sorry,i thought thyristors and scr's were same thing
I have worked with opto isolators in the past so i kinda know how they work.
Heres what ive got upto now
#21 21660517 01 Jan 2012 19:22

Steve Lawson Steve Lawson

Anonymous

Post #21
21660517 01 Jan 2012 19:22

OK, here's my version, using an opto-isolator, so the polarity can be reversed on the Secondary so an N-channel MOSFET can be used.

BUT, it occurred to me that perhaps you don't need such a high voltage. What you need is current. So, what if you paralleled several large low voltage capacitors, such as 3 or 4 300uf 50V caps. You could also switch caps in or out for more or less weld heat. Then you wouldn't have the shock danger if you ran it at, say 35Volts. Paralleling caps is a good way to overcome ESR (Effective Series Resistance). With a lower ESR, you should get more current on the discharge.
#22 21660518 01 Jan 2012 19:26

Steve Lawson Steve Lawson

Anonymous

Post #22
21660518 01 Jan 2012 19:26

In regards to the low voltage supply on the Secondary side, if you move the bar-graph circuit to the primary, then if you can find a second opto-isolator that can handle the high voltage, you would only need a resistor to limit the current to the opto's LED.
#23 21660519 01 Jan 2012 19:44

Sarah Harris Sarah Harris

Anonymous

Post #23
21660519 01 Jan 2012 19:44

the problem i found is that standard capacitors dont last very long when shorted out on a regular basis. photoflash caps dont tend to come in the 14000uf size that would be required at 50v and as the are damed expensive (and difficult very to source )and bulky to parallel them up would be cost prohibiting. i built this with bits i had lying around and made them work. so it was all based around the capacitor that i had.
if you know of a capacitor range that can with stand being shorted out apart from dedicated photoflash one please let me know .
thank you again for your help, i will have a closer look at it in the morning, i ve had a little too much wine tonight Happy new year and thank you again.
Sarah
#24 21660520 01 Jan 2012 21:39

Sarah Harris Sarah Harris

Anonymous

Post #24
21660520 01 Jan 2012 21:39

Ok, i couldnt wait
i've still got in mind whether a SCR could be used. they are cheap compared to 70a mosfets, and what i have drawn also significantly simplifies your ideas. But i dont know enough about scr's to determine whether this will work. circuit wizard seems to suggest it does, but it does have trouble seeing the inverter voltage. but with a virtual power source it seems to work. ive lft out the voltage indicator to simplify it (but we know how that works ) sorry to be a pain.
#25 21660521 01 Jan 2012 21:40

Sarah Harris Sarah Harris

Anonymous

Post #25
21660521 01 Jan 2012 21:40

forgot again
#26 21660522 01 Jan 2012 22:02

Sarah Harris Sarah Harris

Anonymous

Post #26
21660522 01 Jan 2012 22:02

trying to find out about scrs, i came across this http://uk.farnell.com/on-semiconductor/2n6397g/thyristor-12a-400v-to-220/dp/9556753
it says in its description "Peak Non Rep Surge Current Itsm 50Hz:
100A" does this mean that it will handle a short blast of 100A ??
#27 21660523 02 Jan 2012 15:38

Steve Lawson Steve Lawson

Anonymous

Post #27
21660523 02 Jan 2012 15:38

An SCR might work but here are the differences between an SCR and a MOSFET:
* MOSFETs are faster [50-100ns]. The 2N6397 takes a few microseconds to turn on. But, for your application that may be OK.
* MOSFETs are resistive and can have very low channel resistances (on the order of tenths of an ohm or less). Thus, in switch mode they tend to deliver more power to the load and can run very cool. The 2N6397 drops 5.6volts at 100A, that's a 560W power loss. Even at 10Amps the voltage drop is around 1.8V for an 18W power loss. Again, this may not matter in your case--only experimentation will tell. But, the implication is that the SCR will likely get hot (again, might not matter since this is a VERY low frequency application Also, as you pointed out, the 2N6397 has a 100A Peak Non Rep Surge Current, and according to the datasheet, this baby can take 100A for a few milliseconds, so it might just work!
* MOSFETs require, essentially, no current to drive them (save the current needed to charge/discharge the parasitic capacitance). The 2N6397 takes anywhere from 10ma to 30ma to trigger. But, there's plenty of current to steal, so probably not a problem.

In the circuit you uploaded, I'm not sure, but it looks like the SCR will be stressed by the large negative voltage applied through R9. I attached a circuit that is a lot simpler and doesn't stress the SCR. I'm a little nervous about R9, though. I would feel better if that was at 1 Meg or more. If you agree, perhaps you can add another MPSA92 to Q9 in a darlington arrangement, and bump R9 up to a Meg or more.
#28 21660524 02 Jan 2012 17:22

Sarah Harris Sarah Harris

Anonymous

Post #28
21660524 02 Jan 2012 17:22

Your new circuit does make it all a lot simpler, but doesnt allow for variable voltage.
What do you mean when you say that the scr was "Stressed" ??
I did put a diode between Q1 and the gate of D1, but uploaded the wrong file. This diode did make simulation in circuit wizard run better, but i dont know how accurate the program is, and its component library is poor so i cannot run the circuit at working voltages
#29 21660525 03 Jan 2012 04:41

Steve Lawson Steve Lawson

Anonymous

Post #29
21660525 03 Jan 2012 04:41

Hmm... I thought my new circuit _did_ allow for variable voltage -- I took into account voltages between 70 and 250. All that is needed is enough current to trigger the SCR as long as it isn't more current than the gate can take (2A).

And I took another look at your circuit and it appears that I'm wrong. Everything is fine.
#30 21660526 16 Jan 2012 21:43

Sarah Harris Sarah Harris

Anonymous

Post #30
21660526 16 Jan 2012 21:43

sorry to ressurect this post again but i have hit another hurdle that i cannot quite get my head around. i tried both circuits on breadboard ,albeit at lower voltage (maplins component list is rubbish) and success.
the issue now is the refresh time, aprox 30 secs. i deliberately kept it slow to allow the capacitor to recover. which has got me thinking (dangerous i know ), if i made a bank of the output capacitors and associated switching ,using my circuit (sorry steve ) and then used a 4017 to switch them on incrementally, this would give me up to 10 ready to fire caps which would mean by the time i fired off the last one, the first on would be ready again. i have drawn up an idea in the pic below, there will obviously 10 of everything, does this look like it would work with the opto-isolator grounding out the scr until its turn comes ? i have the 4017 circuit worked out so omitted all of that.the 4017 is in moving hole mode
Thank you for your help so far
Sarah
Create an account, log in here. You will receive points by participating in discussions.
Join this discussion.

Install Elektroda application

Topic summary

The discussion centers on designing a safe and user-friendly switching circuit for discharging a charged capacitor at up to 250VDC in a jewelry spot welder application. The main challenge is to enable instantaneous switching at the weld point to ensure the leads become live only during the short circuit, minimizing shock risk. Proposed solutions include using P-channel and N-channel MOSFETs as switches triggered by contact resistance, with considerations for voltage range (70V to 250V) and current handling capabilities. The difficulty in sourcing suitable P-channel MOSFETs led to exploring N-channel alternatives and polarity reversal via opto-isolators for isolation and control. SCRs (thyristors) were also considered due to their cost and surge current capacity, with analysis of their slower switching speed and higher power dissipation compared to MOSFETs. Practical issues such as capacitor longevity under repeated shorting, the use of photoflash capacitors, and the possibility of capacitor banks switched sequentially by a 4017 decade counter were discussed to improve refresh time and welding frequency. Additional circuit elements like voltage indicators (10-LED bargraph), latching circuits for inverter control, and footswitch alternatives were evaluated. The importance of proper gate drive, polarity, and current limiting components (diodes, resistors) was emphasized to ensure reliable operation and safety. References to educational resources on MOSFET operation and circuit simulation tools (CircuitWizard) were provided to aid understanding. Overall, the conversation highlights the trade-offs between component availability, electrical characteristics, safety, and user ergonomics in developing a capacitor discharge switch for high-voltage spot welding.
Summary generated by the language model.