Oscilloscope images of tig welders

[Simclardy]

I would love to compare oscilloscope images from different tig welders.
The only thing i found was a you tube video from this old Tony after i posted some images on tips and tricks forum.

My goal is to take some of the magic out of the decision making when choosing a power source. Forget brand name, i want to see with my eyes what the power source is actually producing. There is alot of advertising about wave shape and the like but they never include the actual image.... just an idealistic image that looks perfect.

So i will start this of with all i have right now, an everlast 250ex. I plan on upgrading to a miller dynasty dx, or a fronius magicwave 230i or htp invertig, in that order.

So here are the images taken so far.
The image with with blue is the amperage. I had the machine set to about 35amps. The overshoot was 65 amps EN.
the other thing i noticed is i can't get below 11 amps. These two facts are the main reason im upgrading. I don't think it matters for bigger material.

Anyway, if anyone can add to the collection of images it could be a great resource!

Sent from my SM-G930V using Tapatalk

[Munkul]

Can you explain the traces for me please?

I.e. "this blue line is volts, on straight DC"
"this yellow line is volts, AC 100hz"

That sort of thing. I'm assuming it's all voltage since you'd need some sort of current clamp and a converter to feed back to the scope...

I can sort of guess, but I'm not entirely sure. e.g. the last two look like a pulsed DC, but I have no real idea.

[Simclardy]

Can you explain the traces for me please?

I.e. "this blue line is volts, on straight DC"
"this yellow line is volts, AC 100hz"

That sort of thing. I'm assuming it's all voltage since you'd need some sort of current clamp and a converter to feed back to the scope...

I can sort of guess, but I'm not entirely sure. e.g. the last two look like a pulsed DC, but I have no real idea.

Sure. Look at image 3&4. The yellow trace is voltage, and the blue trace is amperages. I captured the amperage with a fluke 80i-110s clamp meter. It is set to 10mv/amp. Then i set the scale to 0.1x.
The result is every milli volt equals an amp.
Each grid square is represents 20 amps.
Hope that answered your question.
Let me know if not.
Thanks


Sent from my SM-G930V using Tapatalk

[Munkul]

Thanks for that. It's very helpful.

So what settings correspond to which image? Like i said, I assume pics 3 and 4 are of a pulsed DC, but to be truly helpful you need to state exactly what it is we are looking at.

i.e. some welders light give a rock-steady DC but an AC waveform all over the shop. Or at low amperages it's a bit rough. We don't know.

So I propose if we have a thread like this, the following stats should be included for each image:

- Machine brand and model, including voltage and 1phase/3phase
- Waveform type (i.e. DC, Pulse DC, AC, Pulse AC, squarewave/sine/soft square etc)
- frequency of AC or pulse
- Set amps - peak and background if on pulse


It's interesting, because the small oscillations in voltage must be the feedback mechanisim within the inverter. I'm going to have to get a scope and see the outputs on my Lorch, and post em up too.

[Simclardy]

Thanks for that. It's very helpful.

So what settings correspond to which image? Like i said, I assume pics 3 and 4 are of a pulsed DC, but to be truly helpful you need to state exactly what it is we are looking at.

i.e. some welders light give a rock-steady DC but an AC waveform all over the shop. Or at low amperages it's a bit rough. We don't know.

So I propose if we have a thread like this, the following stats should be included for each image:

- Machine brand and model, including voltage and 1phase/3phase
- Waveform type (i.e. DC, Pulse DC, AC, Pulse AC, squarewave/sine/soft square etc)
- frequency of AC or pulse
- Set amps - peak and background if on pulse


It's interesting, because the small oscillations in voltage must be the feedback mechanisim within the inverter. I'm going to have to get a scope and see the outputs on my Lorch, and post em up too.

Sorry, i did rush this post.
Your right we should try and standardize the test. I think adding the arc gap might be helpful.

- This was ac at about 200hz.
- Balance about 32% cleaning. (You can see the actual% on the scope. Notice they are asymetric + and -. I have no control for that on this machine)
- 35amp setting
- i can't change the wave shape but i think everlast claims square. It's hardly square.


Your right. The fast voltage image is the igbt reaction time and so on

Sorry for the rushed post.

Sent from my SM-G930V using Tapatalk

[Simclardy]

I will redo the test and put all the info in one. It might take me a while, I'm fighting a cold/ mild fever.

Sent from my SM-G930V using Tapatalk

[VPT]

Recently using a AHP I found that the weld came out dull using higher ac frequency. For me the weld got much better when I dropped it down to around 70hz.

[Insaneride]

When you redo the test, can you set the trace so that the current vs volts have the exact starting point? I would expect the current would lag or have a phase shift from volts with a constant current welder. It's probly only a couple degrees but that's why the starting point for each trace needs to be close to exact. I don't have a battery powered scope so I don't dare try it because the short circuit current would fry a mains grounded scope.

[Simclardy]

When you redo the test, can you set the trace so that the current vs volts have the exact starting point? I would expect the current would lag or have a phase shift from volts with a constant current welder. It's probly only a couple degrees but that's why the starting point for each trace needs to be close to exact. I don't have a battery powered scope so I don't dare try it because the short circuit current would fry a mains grounded scope.

Not quite sure what your asking. The trace is on the voltage and the current (blue) is locked in actual time. Any lag or lead you see is real. I did not shift them (i did shift them vertical for clarity)
I am not sure what to expect on a welder but with a capacitive load the current leads voltage and on inductive loads voltage leads. Resistive loads would be in phase with each other.
Yes, connecting a scope to a welder with high frequency start is dangerous. The rigol probe im using gives me the protection i need.


Sent from my SM-G930V using Tapatalk

[Simclardy]

Recently using a AHP I found that the weld came out dull using higher ac frequency. For me the weld got much better when I dropped it down to around 70hz.

It has been a while since i experimented but i don't think i saw the improvement you got. And because my unit is not the best on the low amp range i think the high hz helps me control the arc.
I am truly stunned by the amount of current overshoot with this everlast.


Sent from my SM-G930V using Tapatalk

[Munkul]

Some inverters like high freq, some like lower. That's my experience, anyway.

My old ESAB liked 120hz, no other frequency kept the tungsten in as good a shape. Just a little peculiarity.

My Lorch doesn't care about the frequency as far as tungsten goes, but it wets out the puddle much nicer at lower frequency say 60hz, as well as a lot stiffer and directional at 200hz.

Inverters are weird about balance, as well. The ESAB liked 65% EN, the Lorch likes 75%. I think it's individual to the hardware design for each manufacturer.

[Simclardy]

Some inverters like high freq, some like lower. That's my experience, anyway.

My old ESAB liked 120hz, no other frequency kept the tungsten in as good a shape. Just a little peculiarity.

My Lorch doesn't care about the frequency as far as tungsten goes, but it wets out the puddle much nicer at lower frequency say 60hz, as well as a lot stiffer and directional at 200hz.

Inverters are weird about balance, as well. The ESAB liked 65% EN, the Lorch likes 75%. I think it's individual to the hardware design for each manufacturer.

Maybe these types of images could help answer these questions? If we could see what the current looks like at different hz and compare that to different machines it could shed some light.

Sent from my SM-G930V using Tapatalk

[thegary]

What year was your EX250 made. The reason I ask is they went from analog to all digital I believe 2016 or 2017 on the EX250. I am going from memory so those dates might not be correct. I know the one I got is the all digital one . When I purchased it they still had some of the older version left but I wanted the newer version . So far it has bean able to do anything I have needed to do with it. I will say I do not like its for stick welding but I did not buy it for stick welding. I do need to stick weld some with it so I wish it worked better for that.

[Insaneride]

Not quite sure what your asking. The trace is on the voltage and the current (blue) is locked in actual time. Any lag or lead you see is real. I did not shift them (i did shift them vertical for clarity)
I am not sure what to expect on a welder but with a capacitive load the current leads voltage and on inductive loads voltage leads. Resistive loads would be in phase with each other.
Yes, connecting a scope to a welder with high frequency start is dangerous. The rigol probe im using gives me the protection i need.


Sent from my SM-G930V using Tapatalk

What you said, voltage leads in an inductor, current leads in a capacitor. I would expect a noticeable current lag in a transformer machine. Maybe there's no lag in an inverter that's why I ask if the horizontal was synced between channels. I would still expect a lag in an inverter.

[MCTig]

I would love to compare oscilloscope images from different tig welders.
... i want to see with my eyes what the power source is actually producing. There is alot of advertising about wave shape and the like but they never include the actual image.... just an idealistic image that looks perfect..

This is a very interesting project to pursue, but if you want to know what is actually being produced you need to use an analog oscilloscope. Digital oscilloscopes are great for certain tasks, their multi-color outputs can make work easier, and they certainly make capturing the waveforms easier as well, but because of filtering what they display on their screens typically can be a lot better than what the actual signals look like when those signals are noisy.

A digital oscilloscope's bandwidth specification tells you you the sampling rate of its A/D converter, it does not tell you how that signal is subsequently processed prior to being displayed. This isn't a hypothetical issue. I have four digital oscilloscopes, but when I need to see transients or noise on a signal I use an analog oscilloscope. Using a digital oscilloscope for your project defeats the purpose because the oscilloscope produces "an idealistic image that looks perfect" even when the input signal isn't.

[Simclardy]

This is a very interesting project to pursue, but if you want to know what is actually being produced you need to use an analog oscilloscope. Digital oscilloscopes are great for certain tasks, their multi-color outputs can make work easier, and they certainly make capturing the waveforms easier as well, but because of filtering what they display on their screens typically can be a lot better than what the actual signals look like when those signals are noisy.

A digital oscilloscope's bandwidth specification tells you you the sampling rate of its A/D converter, it does not tell you how that signal is subsequently processed prior to being displayed. This isn't a hypothetical issue. I have four digital oscilloscopes, but when I need to see transients or noise on a signal I use an analog oscilloscope. Using a digital oscilloscope for your project defeats the purpose because the oscilloscope produces "an idealistic image that looks perfect" even when the input signal isn't.

I am no expert, but as i understand it, as long as your memory depth, bandwidth, and sample rate are adequate you can properly represent the waveform. My scope has 300mhz bandwidth,
200.0Msa/s, and 5.60M pts.
This should be more than enough for this slow signal.
Am i missing something?

Sent from my SM-G930V using Tapatalk

[Simclardy]

What you said, voltage leads in an inductor, current leads in a capacitor. I would expect a noticeable current lag in a transformer machine. Maybe there's no lag in an inverter that's why I ask if the horizontal was synced between channels. I would still expect a lag in an inverter.

I could be wrong but i think the important factor is what type of load. So the load would be the arc gap. This should be fairly linear.
If i connected my probe to the welder input then i would see the lead or lag caused by the coil or cap. But on the output i have a potential (voltage) once i connect a load the race is on, if it was a capacitor the electrons would find no resistance and rush in to fill the cap (current), as it fills the pressure (voltage) would then start to climb.
I want to stress, i am not an authority on this subject and welcome correction.




Sent from my SM-G930V using Tapatalk

[Simclardy]

What year was your EX250 made. The reason I ask is they went from analog to all digital I believe 2016 or 2017 on the EX250. I am going from memory so those dates might not be correct. I know the one I got is the all digital one . When I purchased it they still had some of the older version left but I wanted the newer version . So far it has bean able to do anything I have needed to do with it. I will say I do not like its for stick welding but I did not buy it for stick welding. I do need to stick weld some with it so I wish it worked better for that.

Mine was purchased in 2014ish? It has rocker switches in the middle instead of the flat touch pad.

Sent from my SM-G930V using Tapatalk

[MCTig]

Am i missing something?

This is a much more specialized topic than appropriate here. But, as an example, your Rigol displays all signals at the same intensity, whereas an analog oscilloscope displays transients at an intensity appropriate for the time they occupy on each cycle. That is, if a glitch appears roughly only once every 10 sweeps, it would be 10x dimmer on an analog oscilloscope and hence would be a much more accurate representation of the true output.

I have two Rigol oscilloscopes and they're great for many things. But, if I wanted to achieve the goals you stated for your project, I definitely would use my Tektronix 2465B.

[Simclardy]

This is a much more specialized topic than appropriate here. But, as an example, your Rigol displays all signals at the same intensity, whereas an analog oscilloscope displays transients at an intensity appropriate for the time they occupy on each cycle. That is, if a glitch appears roughly only once every 10 sweeps, it would be 10x dimmer on an analog oscilloscope and hence would be a much more accurate representation of the true output.

I have two Rigol oscilloscopes and they're great for many things. But, if I wanted to achieve the goals you stated for your project, I definitely would use my Tektronix 2465B.

Even though it might be a specialized topic if the results are not accurate there's no point in doing the test so I think it's reasonable to make sure the results are accurate. You are right when you say that the trace does not vary in intensity proportional to the duration of the signal. However this I don't believe is a problem. For example if my current trace is showing a peak of 65 amps but in fact that was only occurring one out of every ten Cycles I would see Jitter in the trace and I would see the other state of the signal. What I am seeing is a very stable signal. As I said before this is also a very slow signal. the rule of thumb is your bandwidth should be five times the signal speed. The other factor is the refresh rate of the screen my rigol refreshes 50,000 times per second. To be certain I called rigol and explained my experiment and they agreed that everything should work fine. Of course they might be biased but they are a reputable company. Now if I had the choice between an analog and digital scope maybe I would go with the analog scope. For my experiment and comparison of different welders I feel confident this is adequate.


Linked is an article discussing analog vs digital if anyone's interested.

https://www.electronicdesign.com/com...g-oscilloscope

Sent from my SM-G930V using Tapatalk

[Simclardy]

Ok the first photo is just my warm up pad, with the torch fixed with a very small gap.
Second photo shows peak high frequency voltage around 2700 volts. Third photo shows my machines minimum amp setting. You can see EN is still up at about 17amps. The rms is 13amps.

The arc is very unstable. I used a 3/32" tungsten 2% lanthonated. A thinner tip would be better.

Let's just say it's not fun welding small stuff with this machine.
Cheers
Sandy

Sent from my SM-G930V using Tapatalk

[MCTig]

To be certain I called rigol and explained my experiment and they agreed that everything should work fine. Of course they might be biased but they are a reputable company.

If Rigol made both analog and digital oscilloscopes asking them for technical information to help choose between the two types might have been reasonable. But, they only make digital oscilloscopes so of course they said digital was fine.

The issue isn't refresh rate, memory depth, bandwidth, etc. Yes, the repetition rate of the pulses is low so your digital oscilloscope has no trouble displaying the big features. But, does a given welder provide clean pulses, or could there be noisy, intermittent transients on those pulses that affect the arc but that your digital oscilloscope effectively filters out and doesn't display? Again, your stated purpose is to see "what the power source is actually producing." With a digital oscilloscope you're seeing a sanitized version of what the power source is actually producing so you won't see certain kinds of possibly-relevant features that you would see if they were there had you been using an analog oscilloscope.

[Simclardy]

If Rigol made both analog and digital oscilloscopes asking them for technical information to help choose between the two types might have been reasonable. But, they only make digital oscilloscopes so of course they said digital was fine.

The issue isn't refresh rate, memory depth, bandwidth, etc. Yes, the repetition rate of the pulses is low so your digital oscilloscope has no trouble displaying the big features. But, does a given welder provide clean pulses, or could there be noisy, intermittent transients on those pulses that affect the arc but that your digital oscilloscope effectively filters out and doesn't display? Again, your stated purpose is to see "what the power source is actually producing." With a digital oscilloscope you're seeing a sanitized version of what the power source is actually producing so you won't see certain kinds of possibly-relevant features that you would see if they were there had you been using an analog oscilloscope.

If Rigol made both analog and digital oscilloscopes asking them for technical information to help choose between the two types might have been reasonable. But, they only make digital oscilloscopes so of course they said digital was fine.

The issue isn't refresh rate, memory depth, bandwidth, etc. Yes, the repetition rate of the pulses is low so your digital oscilloscope has no trouble displaying the big features. But, does a given welder provide clean pulses, or could there be noisy, intermittent transients on those pulses that affect the arc but that your digital oscilloscope effectively filters out and doesn't display? Again, your stated purpose is to see "what the power source is actually producing." With a digital oscilloscope you're seeing a sanitized version of what the power source is actually producing so you won't see certain kinds of possibly-relevant features that you would see if they were there had you been using an analog oscilloscope.

Rigol does not sell analog but they do sell scopes that can show the contrasting intensity. Their mso units.
Again he said it was not necessary.


If the noise is within my speed capture ratings it will pick it up. If it's outside it won't matter to a mains type power source.

Im tempted to buy a 100mhz analog scope because they are so cheap, and they look cool, but i see no practical need.

Im curious what your background is? How did you come to own 4 scopes?

Sent from my SM-G930V using Tapatalk

[MCTig]

Im curious what your background is? How did you come to own 4 scopes?

I actually have six counting two credit card-size ones, five of which are digital. The reason I have five is each has some capability or capabilities I needed that aren't duplicated in the others so it's not like I collect them for fun. For example, the Fast Fourier Transform FFT function of Rigol oscilloscopes is a useful substitute for a spectrum analyzer for the type of measurements I need from it.

I've used oscilloscopes in my profession for over 40 years but my interest in them in recent years is more for studying obscure types of automotive electronics. Digital oscilloscopes definitely have their advantages (otherwise, I wouldn't have five of them), but analog oscilloscopes have other advantages (otherwise, I wouldn't have one of them).

By the way, a significant reason why digital oscilloscopes have taken over isn't that they're "better," it's that they do everything most people need an oscilloscope to do at a much lower price than an analog oscilloscope would cost to manufacture.

[Simclardy]

I've used oscilloscopes in my profession for over 40 years but my interest in them in recent years is more for studying obscure types of automotive electronics. .

Your teasing me. What's your profession? Lol, you don't have to say.
Do you follow scanner danner?

I purchase my scope about 6 years ago to help me learn electronics. As an electrician i never found a strong need for it. Now i go to it more and more. When troubleshooting arc fault breakers i have the siemens intelliarc tester but if that fails out comes the scope.
The first image was an anomaly that was causing the trip.
The other use i have found is to capture generator waves that i install. Second photo is a Kohler gen. Not the prettiest.
Then i came across scanner danner and i am very interested in troubleshooting all kinds of engine issues by analyzing spark, injector, battery current etc.
Cheers

Sent from my SM-G930V using Tapatalk