electronics help

[dinges]

I'm faced with an electronics problem that has little to do with RE, but I'm sure that the knowledge is on this board so I will ask it here anyway. Well, technically, it -may- be a little RE related as the finished tool will be used to spotweld battery tabs to NiCd & NiMH batteries...

It's a capacitor discharge spotwelder. For those who don't know what it is, these links explain best:

http://www.philpem.me.uk/elec/welder/

http://www.philpem.me.uk/elec/welder/images/schematic.png

http://sunstoneengineering.com/site/pages/products  (these products make me drool)

My design is mostly based on the schematic as found in the first link, with some modifications. First of all, it uses more capacity, about 2 F at 25 V (30 pcs of 68.000 uF/25 V elcos, total ~600 Ws). I have modified the trigger by adding a NE555 to debounce the switch and to make sure that continuous pressing of the trigger gives just one (~ .5 s) fired pulse to the thyristor. The .5 s time is pretty long (the thyristor will discharge the capacitor in milliseconds, if not microseconds) to make sure that the relay will properly disconnect the PSU as otherwise the thyristor (SCR) would remain conductive due to the large supply current (~13 A) of the PSU.

(hi-res schematic can be found here: http://www.anotherpower.com/gallery/dinges/spotwelder_2?full=1)

The thyristor that I intend to use is a large beefy P202CH12, rated at 325 A continuous and 3500 A pk. It should be more than fine for this application. However, the thing that I can't figure out is how to properly drive the gate.

In the datasheet (http://www.westcode.com/p0327wc.pdf  (warning, it's an 8 MB .pdf file)) on page 3 it says, under item nr. 4 'gate drive': 'the recommended gate drive is 20 V, 20 ohm'. Sounds pretty straightforward. But, when I look at the graph on page 7, 20 V at 1 A seems to lie way outside the maximum-rated boundaries. Also, on page 2, it says that maximum forward peak gate voltage is 12 V. Obviously I'm missing something here... but what... ?

To the point, the questions I have:

1. using 12V, what resistor value R would you recommend so that the thyristors turns fully on in the shortest amount of time, yet doesn't exceed its maximum ratings ?
   
2. would a snubber network over the K-A be needed for this application ? The datasheet recommends 22 ohm & .22 uF but I don't think one would be needed in this application. But I may be wrong.
   
   

Any advice from those having experience driving thyristors will be greatly appreciated.

Peter.

[Opera House]

I own a C-H spot welder for strain gauge protective covers.  All these devices I have seen have used a SCR dumping into a transformer to develop the current.  If you have access to a toroid transformer this would be easy to build.  The gate current will require a surprising amount of amps in order to prevent propagation delays in the die and overheating.  Consider a pulse transformer or a cap on the gate drive resistor.

[Flux]

I agree. The units I used had transformers usually with a single turn secondary of 3" wide copper foil. I see no reason why a toroidal wouldn't' be ok with a thick tape secondary.

I am sure you will get enough current without the transformer but I thing you will bet a lot of spectacular bangs and a lot of electrode damage. Ratio about 20;1 from memory.

I also agree about pulse driving the gate, I prefer a pulse transformer connected as a one shot blocking oscillator with positive feedback, but a capacitor discharge from 10v with a 5ohm gate limit should be ok. I wouldn't let a 555 near the gate circuit you will most likely get a chain reaction with the main discharge firing the 555 from the inevitable spikes.

I seem to remember the best natured welder we had was very old with a big relay discharging the capacitor into the transformer. With a thyristor drive I would also add a freewheeling diode across the transformer primary although it may not matter.

[Norm]

Just wondering how large of a battery do you

plan on making? any particular size?

           

[commanda]

What mainly concerns me with your circuit as drawn is the difference between relay pull-in time vs firing the scr.

Maybe a 2 pole relay, the second NO contacts supplying power to the PNP.

Amanda

[dinges]

Hello Amanda,

I've thought about that too but I don't think it's an issue, as the PSU is an 'Oztules' conversion with current limit at 13A. So the PSU shouldn't mind. But there should be enough of a timegap (no power being supplied from the PSU) to make sure that, as the capacitor bank is discharged, no current is being supplied by the PSU as the 13A would keep the thyristor conductive (the P202CH12 needs a maximum of 600 mA to remain conductive after being fired)

After the thyristor stops conducting, the PSU may start delivering energy to the capacitor bank again. I expected the 0.5 s delay to be doable with relais.

So, even though it's not an elegant solution, I think it does the trick in this case. If it doesn't, I'd love to be corrected

My main worry for not using a relay in the way you describe (to trigger the PNP) is contact bounce. Or isn't this an issue ?

Peter.

[dinges]

Norm,

There is no battery in this apparatus. It is used for welding tabs to batteries.

Peter.

[dinges]

I figure this method works, as there are commercial applications and homebuilt contraptions that work by discharging a large capacitor bank through a thyristor. They fire the thyristor in even cruder ways than I intend to do. But maybe using a switch or even relay to pull the gate of the thyristor 'high' is the fastest, simplest and best way after all ? And contact bounce is not an issue ?

I hope there are simpler ways than pulse transformers to fire this thing reliably.

Peter.