Help with a one shot device

I could assume you have experience here...but just in case you don't...

There are a BUNCH of problems.

I used to think that to spot weld, you just clamp the material and stuff a lot of current thru it for several seconds until it gets red hot. That works fine for things that can stand the heat. If you're welding something like battery tabs, thermal fuses, things that can't stand the temperature rise, you need LOT more current for much shorter time. So the level of difficulty depends on what you're trying to weld.

You might find that the inrush current is higher than 15A when you short the secondary. My spot welder made from a microwave transformer draws 40A. You want to switch the circuit on at a zero crossing. If you don't you'll have all manner of grief with blown relay contacts etc. You want an integral number of input cycles. If you don't, depending on the hysteresis in the core of your transformer, you'll get random weld power. In the extreme case, you leave the transformer in a state that causes it to saturate with the next shot. This blows up whatever you're using to switch the primary. Much of this gets solved by using a solid state relay with internal circuitry to "close" on zero crossing. The SCR will turn off at a zero crossing too. I use a 240V 40A SSR.

It's pretty simple to build the counter and control circuitry into a single chip. Mine just uses an analog ramp to synchronously drive the solid state relay. The control is analog, but it adds a complete cycle at a time at each threshold point. IIRC, I got best results on battery tabs using about 6 cycles of 60 Hz. But it can vary a lot with secondary resistance. Work HARD on minimizing secondary path resistance.

I posted a circuit at the link below to show how simple the whole thing can be. I can send you the pal equations if you're interested.

I'm a fan of minimilist designs. The design ain't pretty, but it works.

If I were to do it again, I'd probably use a PIC processor just cause they're easier for most people to find and program. Once you bite the bullet and build the PIC/AVR/Whatever you like programmer, you'll find that most anything you can do with a 555 is quicker/easier/better done with a microcontroller.

here's a link to the whole shebang.

So, I strung it all together. Tried to weld battery tabs. Electronics works perfectly, but the welds were tricky and unrepeatable. What you really need is a LOT more current for a much shorter time. Got MUCH more consistent tab welds with a commercial CD welder. I got lucky. The CD welder cost me less than the wire I bought to wind the transformer secondary.

mike

--
Return address is VALID but some sites block emails
with links.  Delete this sig when replying.
.
Wanted, PCMCIA SCSI Card for HP m820 CDRW.
FS 500MHz Tek DSOscilloscope TDS540 Make Offer
Wanted, 12.1" LCD for Gateway Solo 5300. Samsung LT121SU-121
Bunch of stuff For Sale and Wanted at the link below.
MAKE THE OBVIOUS CHANGES TO THE LINK
http://www.geocities.com/SiliconValley/Monitor/4710/

I was going to reply with the typical 555 timer circuit, but I think you can get away with a simpler circuit, which requires no separate power supply. Here is one:

Timing is controlled by Ct and Rt. Before a button press, the gate of the mosfet is at 12V. When you press the button, it gets dragged down to gnd, and slowly comes back up by recharging the cap through the timing resistor. Since I don't know how long you want it to stay on for, I don't know how big to make the cap and resistor. In general, the thing will turn off about when t = Rt * Ct seconds have elapsed, but that's a guess based on a turn on voltage of 8V for the mosfet. That may be high, but it's in the ball park.

One thing is that because of the configuration, the action of turning on the coil has the side effect of pulling down the 12V rail slightly, thus speeding up the turn-off of the mosfet, and likewise the turning off of the coil causes the +12V to rise a bit, causing the mosfet to turn on faster, so it shouldn't chatter when turning on and off. If you find it's chattering, you can use a smaller smoothing cap (C3), which will allow more movement of the 12V rail, and thus a quicker turn on and off.

470uF is bigger than you probably need, but if it's too small, the relay may cycle on and off at 60Hz. Not a good thing.

Standard precautions and disclaimer applies. You MUST ENSURE that the pushbutton is isolated from the user, probably using some kind of rubber cover. Also, make sure the the neutral and ground aren't swapped by testing the socket you plug it into.

This assumes you are in the US. If you are in europe, don't use this circuit without modification by somebody who knows what's up with the local power.

--
Regards,
   Robert Monsen

"Your Highness, I have no need of this hypothesis."
     - Pierre Laplace (1749-1827), to Napoleon,
        on why his works on celestial mechanics make no mention of God.

there are ready made timers designed to wire in or plug into a socket that operate from 120 volts and gives you relay contacts.

the key words are ONE SHOT relays. DELAY off Relays.

-- in any case, if you want to use a 555 timer, get your self a simple wallwart @ apox 12 vdc,300 Ma or more supply, a 555 timer ic, and a

12 v coil relay with dual poles that are 10 amps ech. use both poles together to double up the current and some caps/snubbers across the contacts to absorb the flyback. or use a smaller relay to drive a large mercury relay which handles those kind of loads just fine.

the 555 timer is to run in one-shot mode, this mode forces you to release the trigger before you can put the timer into a short on state again.

there you will find info on how to construct it.

P.S. the relay coil should have a diode across it to absorb fly back energy. also use something like a 2N2222 transistor to drive the relay coil from the output of the timer. make sure you drive the transistor hard enough to saturate it, this will keep the heat down. i am sure others here will offer you drawings, i don't have that ability at this time.

Jamie,

Thanks, I will look for more info on one of these ready made units to see what the cost would be. Very nice website link too.

Mike,

Thanks, I am making mine from a MOT also. Had not really considered the larger fast amp draw. Great info on the whole process.

Did you find a 220V transformer (I have had no luck finding anything other than 110V units from the scrap piles) from a microwave or was it

110V originally and you rewound the primary too?

Robert,

I like the no other power supply idea, will have to give this a good looking over and put something together soon.

THanks again for all the replies and advice.

Jeff

I'm in the US. They're all 120V. The SS Relay is 240V simply 'cause I had it in the junk box. A microwave transformer is a lousy choice for a spot welder. Designed to be low coupling for the microwave application. That's not what you want here. But they are cheap and easily available. If you're gonna wind the primary, start with a different core.`. mike

--
Return address is VALID but some sites block emails
with links.  Delete this sig when replying.
.
Wanted, PCMCIA SCSI Card for HP m820 CDRW.
FS 500MHz Tek DSOscilloscope TDS540 Make Offer
Wanted, 12.1" LCD for Gateway Solo 5300. Samsung LT121SU-121
Bunch of stuff For Sale and Wanted at the link below.
MAKE THE OBVIOUS CHANGES TO THE LINK
http://www.geocities.com/SiliconValley/Monitor/4710/

MIke,

I am in the US also. I don't really have the ambition to rewind the primary, and yes, the MOTs are readily available at least.

Thanks aga> Jeff wrote:

considered

anything

it

I

for a

application.

available.

I agree, toss the shunts, but I'm not sure that's the whole story. Is the core made of the same stuff as a real transformer? YOu also lose some coupling with the side-by-side windings. I expect there are some rules of thumb about this?? mike

--
Return address is VALID but some sites block emails
with links.  Delete this sig when replying.
.
Wanted, PCMCIA SCSI Card for HP m820 CDRW.
FS 500MHz Tek DSOscilloscope TDS540 Make Offer
Wanted, 12.1" LCD for Gateway Solo 5300. Samsung LT121SU-121
Bunch of stuff For Sale and Wanted at the link below.
MAKE THE OBVIOUS CHANGES TO THE LINK
http://www.geocities.com/SiliconValley/Monitor/4710/

HI Clarence

It's always

Clarence_A wrote:

nice to hear from people

who have a contribution to the thread.

See, you can make everybody happy if you post creatively... Suggest you have another beer and watch more rasslin' on TV. mike

-- Return address is VALID but some sites block emails with links. Delete this sig when replying. . Wanted, PCMCIA SCSI Card for HP m820 CDRW. FS 500MHz Tek DSOscilloscope TDS540 Make Offer Wanted, 12.1" LCD for Gateway Solo 5300. Samsung LT121SU-121 Bunch of stuff For Sale and Wanted at the link below. MAKE THE OBVIOUS CHANGES TO THE LINK

Top poster -- PLONK!

Yea, those have already been knocked out, all I have left is to figure out the timing method and get some rod stock for the arms, which is no big deal.

Thanks for the reminder though.

I think I am going to use 2 transformers to ensure adequate amps since I am doing slightly thicker metal than alot of what I see these created for. Need to do 2 sheets of 24 or 22 gauge sheet steel together.

Jeff

Remember that there's no free lunch. If you expect twice the power out, you'll need twice the power in. Put the primaries in series and run it off 220 if you expect any real gain.

If you're only doing one job, put an ad on a local newsgroup and get someone with a thirst for beer and a welder to help you. mike

--
Return address is VALID but some sites block emails
with links.  Delete this sig when replying.
.
Wanted, PCMCIA SCSI Card for HP m820 CDRW.
FS 500MHz Tek DSOscilloscope TDS540 Make Offer
Wanted, 12.1" LCD for Gateway Solo 5300. Samsung LT121SU-121
Bunch of stuff For Sale and Wanted at the link below.
MAKE THE OBVIOUS CHANGES TO THE LINK
http://www.geocities.com/SiliconValley/Monitor/4710/