Measuring weld cable resistance with Ohms Law

Thread Starter

anishkgt

Joined Mar 21, 2017
549
Hey all,

I have spot welder project in mind. In this project i am hoping to add a feature to measure the weld current on every weld. The resistance of the weld cable is only measured once during assembly. Using that and the weld voltage, measured at the Positive (+) electrode together can calculate the weld current using ohm's Law.

Here is what i hope to accomplish or could be wrong completely. Once the electrodes are shorted the base of Q1(iSet) is set high by the microcontroller. This would set a constant current of 1A by the op-amp via the feedback pin on U1. A voltage drop at the cable would be read at pin Vw which would be the total resistance from the positive electrode to gnd. The MOSFETs have a 9mOhm Rds(ON) resistance so minus it from the total resistance would give the cable resistance. So during a weld Vw would read the voltage, with ohms law calculating current would be possible.

The voltage from the supercap is 8.1v, R1 is set 8Ohms to set the current at 1A.

Am i correct with this assumption ?
 

Attachments

Last edited:

jpanhalt

Joined Jan 18, 2008
11,087
Hey all,

I have spot welder project in mind. In this project i am hoping to add a feature to measure the weld current on every weld. The resistance of the weld cable is only measured once during assembly. Using that and the weld voltage, measured at the Positive (+) electrode together can calculate the weld current using ohm's Law.
No. Inductance plays is big part at such high currents. I use #4 AWG welding cables with my CD battery tab welder. Its short cables (ca. 30") quite noticeably jump with each weld.

I do know the beginning voltage, so I simply made a table of voltage versus other weld variables, like nickle strip thickness and subjective quality of weld.
 

dendad

Joined Feb 20, 2016
4,452
Why not just measure the current directly?
Your circuit seems quite complicated to me. Or am I missing something?
For one idea, a Hall effect current sensor can do it I think. The sensor can be mounted a little away from the cable if the current pulse is too high.
 
Last edited:

Thread Starter

anishkgt

Joined Mar 21, 2017
549
Why not just measure the current directly?
Your circuit seems quite complicated to me. Or am I missing something?
For one idea, a Hall effect current sensor can do it I think. The sensor cam be mounted a little away from the cable if the current pulse is too high.
Cost, BOM. Don't want to add to the existing.
 

Thread Starter

anishkgt

Joined Mar 21, 2017
549
No. Inductance plays is big part at such high currents. I use #4 AWG welding cables with my CD battery tab welder. Its short cables (ca. 30") quite noticeably jump with each weld.

I do know the beginning voltage, so I simply made a table of voltage versus other weld variables, like nickle strip thickness and subjective quality of weld.
Is that a DIY or a brand ?

Well i am not expecting accurate readings just a ballpark. Something to learn as i make it.
 

Thread Starter

anishkgt

Joined Mar 21, 2017
549
Your circuit seems quite complicated to me. Or am I missing something?
The design and function is fairly simple. MOSFETs Q3-Q8 function as a switch and are driven by their corresponding drivers U2-U7. They switch the stored charge from the supercapacitors C2-C7. These capacitors are in 3s2p configuration with a max voltage of 8.1v. The trigger pulse is send by U8, the Atmega328 from pin 10.

The constant current source is comprised of the opamp U1, U8 send a pulse to Q1 which in turn siwtches on the power to U1, Q2 base current is controlled by U1's feedback.

Hopes this helps to understand the schematic.
 
Last edited:

sparky 1

Joined Nov 3, 2018
756
The power supply has an amperage and voltage characteristic for each spot welding setting. The cable must be rated high enough to meet manufacturers tested output of the controller and transformer together. That series of load tests is fairly accurate. Scaling a sample down may be necessary.The data can then be adjusted for a given material thickness and further refined to pinpoint accuracy.
Going further, from a working system a sense feedback loop could be designed. Likely the amount current stored and the rate of discharge cannot exceed the load test. Transformer rating of 50% duty cycle. The number of cycles until the spot welder needs a rest and is also thermal related.
Going beyond that would require cooling. For example picking an arbitrary number. after 12 spot welds the transformer needs cooling time or the transformer exceeds 50% that would entail water circulation around all the hot parts,
 
Last edited:

Thread Starter

anishkgt

Joined Mar 21, 2017
549
Yea for every weld the voltage and current varies which is what i am trying to achieve here. I could measure the resistance of the weld cable prior to hooking it up with welder but i'd like to have it within the welder it self.

@jpanhalt why would it not work ? The U1 part of the circuit would be idle during welds. It would be just there to measure the resistance initially during assembly. Otherwise its the other set MOSFETs that do the switching. I've completed a spot welder project using a MOT earlier and yea there is a lot of inductance. Longer the cable more the inductance.
 

Thread Starter

anishkgt

Joined Mar 21, 2017
549
Seems like the U1 is gonna be a CV rather than CC. my bad. Seems like getting CC to work would be harder than CV. Anyone have any advise on how to get CC and not CV.
 

sparky 1

Joined Nov 3, 2018
756
you said "In this project i am hoping to add a feature to measure the weld current on every weld. "
on page 65 of the pdf below shows a good approach.
The hall effect sensor in the cut out section of the toroid. The ferrite toroid is placed over the cable.
This method deals with the magnetic content that will be colliding in the lattice of the weldment, very precise.
The very basic microwave oven transformer RSW does not address the circuits or controls with dynamic resistance.
So the improvements using MOT RSW will likely improve. thanks I appreciate the effort to make us aware of dynamic resistance.
If the hall effect toroid runs into proprietary issues there are other current sensors you can make.


Most of the basic circuits used to make the dentist spot welder are shown in this extensive paper.
A modification for higher output might be upgrading some of the output components but the current and timing look good.
Going further, from a working system a sense feedback loop was designed and published.
https://ir.lib.uwo.ca/cgi/viewconte...ng.com/&httpsredir=1&article=1129&context=etd
 
Last edited:

Thread Starter

anishkgt

Joined Mar 21, 2017
549
@sparky 1 That's a bit of work and cost involved. Not planning on to spend much on this project. I wouldn't want much accuracy to the current measured. Just an approximate but thanks for the link.

@jpanhalt I've seen that the schematic nor the code is open-source. He had earlier shared the schematic but that was not updated in anyway moreover i'am not very much interested in copying his work. I prefer learning things and doing them along the way. In one of his videos he says he measure the cable resistance the same way but not sure how he does it.
 

jpanhalt

Joined Jan 18, 2008
11,087
@jpanhalt I've seen that the schematic nor the code is open-source. He had earlier shared the schematic but that was not updated in anyway moreover i'am not very much interested in copying his work. I prefer learning things and doing them along the way. In one of his videos he says he measure the cable resistance the same way but not sure how he does it.
I have not followed his work. It's a very log thread, so these comments are based on a few "soundbites."

1) The one thing he does that I found attractive was to add a "cleaning" pulse. Some commercial units do that too. It might be nice in a commercial environment, but for me, it's not essential. You can tell a lot by the "feel" when welding a tab. It's the sound, look, whole gemisch. Since each weld with 2 electrodes gives 2 welds, I generally use 4 welds per connection. If one looks bad, I add another one or two (a pair) more (maybe putting one electrode in a previous pit). That doesn't happen very often.

2) In order to provide #1, he turns mosfets on and off. I use an large SCR with a high single pulse rating.

3) Some of his early versions used lipo batteries directly as the current source, rather than capacitors. I do not need the portability provided by lipo's, and that approach seemed to me it would be problematic. Eventually, he got to specifying which lipo's worked and which didn't. Some recent posts seem to be using capacitors more.

So, basically, the decision is how much control you need (i.e., commercial/non-commercial) and whether you want a cleaning pulse. I am comfortable as a hobbyist with my toy:

1595257489210.png1595257539206.png

The right image is set to 15 V OCV, The multi-turn pot numbers are offset about +10 V, which I didn't bother trying to adjust..

It is based simply on this design: https://www.philpem.me.uk/elec/welder
 

Thread Starter

anishkgt

Joined Mar 21, 2017
549
1) The one thing he does that I found attractive was to add a "cleaning" pulse.
His design does not use a dual pulse approach. He claims his welds are consistent because he measure the energy put into the strips so a single pulse works all the time. When i had my MOT version it was build with dual pulse. Always gave consistent welds never failed.

I tried to incorporate a way to measure the resistance without removing the cable but can't figure it would otherwise.
 

Attachments

jpanhalt

Joined Jan 18, 2008
11,087
Sorry, early on there was some discussion of a cleaning pulse, I guess it never got incorporated. I certainly didn't read most of the posts.
 

Thread Starter

anishkgt

Joined Mar 21, 2017
549
So to measure the voltage based on the schematic in post#16, generally with a DM its at the 1 & 2 of J1 but how is it measured using and arduino. Is it referencing each pin to ground ? or reading both pins as in the schematic which would give two values. How can i convert them to a value ?
 

sparky 1

Joined Nov 3, 2018
756
A PC uses software and USB connection to communicate with the micro-controller.
to auto calibrate, to set output power, to adjust weld timing, and so on is done more smoothly on a PC
and when all the sub-routines work the program is uploaded then the PC is disconnected.

The 10 year old does "hello world" because no one told him he can't
 
Last edited:
Top