Measuring Wire Feed

[greyshirtguy]

Disclaimer: "Ok, as well as enjoying all things DIY (esp metal work) I'm a also computer nerd and try-hard hardware hacker (eg I love the kind of things you find on www.hackaday.com)"

Having learned how important a stable wire-feed is for successful MIG welding (This is a great site, Malcolm) - I felt curious to measure "how good is my little hobby MIG welders' wire-feed?"

After a little brainstorming I came up with a quick method so my laptop could measure and record it (to an accuracy of approx 4mm):



It's a bit hard to see in the first pic, so take a closer look:



Notice the grey thing in there, with the grey cable coming out of it?



Yep, it's an old USB computer mouse!

After some tinkering and improvising, I found a way to mount a (slightly dis-assembled) mouse so that the wire feed roller movement would be captured by the scroll wheel on the mouse.

I then wrote a (quick and dirty) .NET application to record the mouse wheel movements (18 "pulses" are captured per revolution of the mouse wheel)

I then ran a steel rule carefully over the wheel to record how many pulses in 600mm - which gave a (probably not too) rough calibration.

Now that it was mounted, talking to the PC and calibrated I was ready to record my wire feeder and make some pretty charts in Excel.

Any deviations in wire feed speed would show up most clearly in a chart (instead of a nice straight line - it would have curves/bumps in it).

So I did 10 quick test welds on some 1.2mm mild steel with various power and wire-feed settings and charted the wire-feed over the whole experience. I was glad to see what were acceptably straight lines (indicating a stable feed rate during each weld)



The vertical axis is metres of wire (wow! It didn't take very long to use up 8 metres of MIG wire ). The horizontal axis is time (in seconds). Each of the 10 test welds shows up as an upward sloping line with a (flat-line) pause in between. Weld 9 shows the adjustment in wirespeed made during it.

A close up of one weld shows the stability of the wire feed more clearly:



In some other tests not shown in this chart, I did notice that some really low speeds were not as stable on the higher voltage settings (thankfully they were way too slow to be used at those settings)

I understand most people might not see the "sense" in this project - but it was (for me) a really fun night, and I think I will do some more accurate calibrating and some more recording when welding again.

 

[peterd51]

Hi,

well, I think it's brilliant!

I've seen a number of messages where a user says their wirefeed is so-many feet per second, etc, but on my Clarke I've just got a twirly knob that shows '1' to '9' and I have no idea what speed the wire is actually travelling.

The instructions for the welder just state that once it's set for one level it'll automatically re-adjust for different power settings.

Regards
Peter

 

[rtbcomp]

I like that! Now control the motor from the computer to give a specified and constant wire speed. Pulsed of course to give max torque at low speeds.

 

[rusty69bus]

nice, now there's lateral thinking

 

[Wozzaaah]

Hehe, as we say down here, there's nothing like going all round Bristol to get to Bath.
Nice work all the same though

 

[malcolm]

Excellent stuff there. It would be interesting to see what does vary during welding to mess it us as you go along. The wire at the end has no way of telling what machine it's connected to but ease of welding does vary a lot between machines.

Wonder if it's voltage control - Voltage/current must vary constantly as the wire does it's thing, but it does seem to have a lesser effect on more expensive welders.

Would be fun to hook something up and change the welder (compare all the SIP mods maybe).

One other thing - how on earth did you get attached images in your text? Now that is clever stuff.

 

[Drains]

That's well handy. Could you tweak the app to show metres per min / time, instead of metres used over time? That way you'd get a flat curve instead of the gradient. It would return to zero when you turn off.

And then, can I have one please

 

[rtbcomp]

Wonder if it's voltage control - Voltage/current must vary constantly as the wire does it's thing, but it does seem to have a lesser effect on more expensive welders.

I reckon if the motor supply is taken off the secondary of the big transformer varying loads when welding will vary the motor speed, which will affect the current and the whole thing becoames unstable.

It would be interesting to build an interface to measure voltage & current and plot that on a graph.

 

[rtbcomp]

That's well handy. Could you tweak the app to show metres per min / time, instead of metres used over time? That way you'd get a flat curve instead of the gradient. It would return to zero when you turn off.

And then, can I have one please

Yeah, it would show fluctuations nicely.

 

[shenion]

Shows how good the Clarke's are. They have a regulated wire speed controller.

Would be interesting with a SIP; they use a cheap voltage drop circuit.

 

[malcolm]

Tried my SIP with the wire feed powered from a car battery. It's annoying I can't find the photos, but it was quite amusing with jump leads connected to the little wire motor terminals and at one point an unsuccessful attempt at stepping down the voltage with a car ignition coil as a resistor (3 ohms not enough).

Anyway - didn't make a lot of difference to weld stability.

 

[greyshirtguy]

Shenion: I haven't seen the Clarkes around here in Aust' - My Welder is a "CIGWeld, TransMIG 165" (A Thermadyne company)....and yes, I would find it very interesting to measure and compare models known to have problematic wire-feeds - like the SIPs.

Simondrains: I like your suggestion - I'm sure that will end up in version 2 of the app (which is currently pretty "rough" due to rushing it to get something working in a single evening)

Malcolm: I have already started thinking about ways to record voltage and current (having visions of recording all parameters that influence the welders performance).

Regarding inserting attached images in text - is this what you mean.....?


1. Attach some images....


2. Go back to your message in the "Post Editor" and once your text cursor (aka carat) is positioned where you would like to insert an image....click the little triangle next to the attachments button (paper-clip) to get a menu of your attached images. Select (click on) an image name from the drop-down list to insert that image inline...


BTW: You dont get to see the images in the "Post Editor" when editing - instead, you see a "code" for each inline image:

 

[peterd51]

Hi,

nice one...I bet no-one here knew we could do that!

Regards
Peter

 

[greyshirtguy]

Hi,

nice one...I bet no-one here knew we could do that!

Regards
Peter

Hmmm, I remember seeing a number of posts on this forum with inline images...which is why I hunted around for the how to do it in the first place.

I'm still not so sure that's exactly what Malcolm as asking about....

 

[peterd51]

Hi,

in previous messages when someone asked how to include photos teh only options offered were to put it on a web-site or add it to the gallery and link it in.

Either option isn't too usefull for the odd snap to illustrate a point in a message, but your method is.

Regards
Peter

 

[malcolm]

That's brilliant! I'd no idea it was possible to do that. It's even possible to attach images in the first place using the paper clip instead of scrolling down. Love the highlighting on your screenshots too - go on - how did you do that?



For the voltage measurements I noted that someone broke a voltmeter measuring a TIG - might be a risk for a sound card (is that what you are using?). Wondered if a tape recorder would be safer - I have an old 8 channel data recorder knocking about.

 

[shenion]

Shenion: I haven't seen the Clarkes around here in Aust' - My Welder is a "CIGWeld, TransMIG 165" (A Thermadyne company)....and yes, I would find it very interesting to measure and compare models known to have problematic wire-feeds - like the SIPs.

Hmm, that is exactly what my Clarke looks like inside. Same wire feeder.

 

[Drains]

That is just smart. Hang on while I 'ava a go...



Yes!!! That rocks!

(thanx for the pic snowy )

 

[zardoz]

One other thing - how on earth did you get attached images in your text? Now that is clever stuff.

It seems there are a few ways of doing it.
One is just use the option to use photobucket etc, instead of uploading to the server from your computer. You can use images from the web. You can use the 'insert image' thingy (in between 'insert email' and 'wrap text')

like this

Yes that hackaday site is great. I have always wished i knew more about electrics so i could do more. I had a go at college, but couldnt get me head around most of it. too many diodes and capacitors, transistors and er stuff.

 

[steelnett]

brilliant, can i try?