A little success

--
Congratulations! :-)

...and, in the process, you learned a few things electronic.

Here is something to implement at the later date in your project. My Lincoln

175 which I use primarily for tig welding has a fixed time for Argon post weld flow. This can waste a lot of gas if you are stitching and touching up the welds the do not stay very hot for long time after the arc has extinguished. The post flow time should be based on the duration and current used at the time of the last weld. This is a job for a micro but it could be done with discrete circuitry. Since I still feel very much like a newb at this, I'd appreciate any comments.

Regards,

Boris Mohar

Got Knock? - see: Viatrack Printed Circuit Designs (among other things)

void _-void-_ in the obvious place

Which is A Good Thing. Nobody is born knowing about this stuff and pretty much everybody has let the Magic Smoke out of a few things along the way. Hang in there, Ignoramus27362!

--
Rich Webb   Norfolk, VA

yep... very many things...

i

Yep... It is somewhat like computer programming, using logic to debug things and verify assumptions etc.

i

You want a circuit that models the temperature rise of the tip and also the cooling time for that temperature rise. An RC time constant is probably a model that is good enough. Charge a capacitor through a resistor with a fixed voltage that represents the steady state tip temperature (this should be higher when the current is higher). When the current goes to zero, this voltage will also be zxero, and will discharge he capacitor through the same resistor. When the cap voltage falles below some fixed threshold, representing the safe tip temperature without gas, switch the gas off.

In parallel, I would have a single shot that switches the gas on immediately when current starts (unless you can figure out how to have it anticipate the current) and holds it on for a minimum time, regardless of what the cap is doing.

I think that what you say makes sense. I would say, go for it, but have a lower minimum on postflow time to avoid programming bugs ruining your tungstens.

On my welder, postflow time is settable. I can say that, for example, postflow should be 8.4 seconds and preflow should be 5.3 seconds and starting current would be 44.8 amps and the welding current would be

78 amps and the time to reach welding current should be 3.2 seconds and s0o on and so forth. It is a programmable welding machine, which is kind of nice.

Later on, when I am done with my current inverter project (whether I succeed or blow up the IGBTs), I may try to do an automatic welding setup where the tig torch is moved along a straight line at a predetermined speed and the welder is controlled from a linux laptop, let's say. But I should not scatter myself too much, right now I am focused on the inverter.

i

but be *VERY* careful when messing with the big stuff - its the difference between "oops, there goes a 555" and "Oh f*ck, where are my fingers"

Cheers Terry

But there's a significant difference - if you crash a program, it doesn't usually let the smoke out - you just edit it and run another edit/compile/run/debug cycle.

With real parts, when you let the smoke out, you actually _learn_ stuff. :-)

Good Luck! Rich

Well, obviously, the place to start looking is at the current post flow timer/regulator. It might even have an adjustment already. But if not, then find out just exactly what the mechanism does, to figure out what you need to do to control it. (timer relay? dashpot?)

Please report back here with the results of your investigations. :-)

Good Luck! Rich

wrote: ...

Well, I've already posted my thoughts, which are, in a nutshell, the polite way, start small. Get an XR2206 working on a breadboard reliably, and _then_ start playing with jacking up the output voltage and stuff.

And read.

Good Luck!

Hey Iggy, if you want PWM, just grab a UC3842 from a dead monitor board, or at worst, send a buck to a mail order catalog. If you're headed for a half or full bridge circuit you'll want dead time too, in which case a TL494 might be better. (3524 / TL494 / KA7500 type chips are push pull output, not overall variable duty cycle, so I don't really know what you could get that's a cross between them..)

Tim

-- Deep Fryer: a very philosophical monk. Website:

May I suggest an experiment? Remove the pull up resistor from the output and see what the output does. Also, measure the pull up resistor with an ohm meter (while it is out of the circuit).

It depends on what kind of software you are talking about. Seriously. Around where I work, software bugs == lost money.

You do have a point though, but these chips are not terribly expensive. I ruined a $0.29 555 timer, and a $3.59 XR2206, so far, no big deal.

The amount of time that I spent on this project, already, is enormous and dwarfs the cost of chips by at least 2 orders of magnitude.

I have piles of used 1960's circuit boards with plenty of free components too.

i

By the way, Rich, do you have any thoughts regarding my XR2206 issues (see my separate thread on that). The square wave output works, but has too low voltage.

i

1 down, unknown to go.

As I read it, pin 10 connects only to a bypass capacitor whose other end is grounded. Pin 12 is grounded.

Getting an XR2206 working on a breadboard reliably is exactly what I am working on.

It kind of half works.

It produces pulses. I can see them on my scope.

The trouble is that the pulse amplitude is too small. The variable duty cycle application that makes the chip change the timing resistor for top vs. bottom of the cycle (which changes their relative duration), involves sending over 2V to pin 9, from the output on pin

1. My output is about 0.67 volts in amplitude. That's my current issue.

i