Author Topic: kWeld - "Next level" DIY battery spot welder  (Read 323394 times)

0 Members and 2 Guests are viewing this topic.

Offline tatus1969Topic starter

  • Super Contributor
  • ***
  • Posts: 1273
  • Country: de
  • Resistance is futile - We Are The Watt.
    • keenlab
Re: kWeld - "Next level" DIY battery spot welder
« Reply #475 on: December 06, 2017, 03:26:09 pm »
Hmmm yea. Oooopsy[emoji16] again that wouldn’t be accurate but would give an idea of it.
Nothing that the on-state resistance of the MOSFET's wouldn't give - except better absolute accuracy.
We Are The Watt - Resistance Is Futile!
 

Offline TopQuark

  • Frequent Contributor
  • **
  • Posts: 303
  • Country: hk
Re: kWeld - "Next level" DIY battery spot welder
« Reply #476 on: December 07, 2017, 07:36:41 am »
Frank, really appreciate your answer :D using the mosfets as a shunt is quite a clever idea. How about using the fuse as a shunt? Does the resistance of the fuse change much from one to another, or does the temperature rise of the fuse contribute too much to resistance instability?

If I end up building one for myself, my OCD will probably force me to place a 'real' shunt resistor to get a slightly better current reading. Maybe a 100 microOhm resistor with an amplifier circuit similar to what Dave used for the uCurrent project. But now I am quite occupied with another project I have on hand, so maybe I'll just place Frank's welder on the top of my Christmas wishlist... :P
 

Offline tatus1969Topic starter

  • Super Contributor
  • ***
  • Posts: 1273
  • Country: de
  • Resistance is futile - We Are The Watt.
    • keenlab
Re: kWeld - "Next level" DIY battery spot welder
« Reply #477 on: December 07, 2017, 05:28:05 pm »
How about using the fuse as a shunt? Does the resistance of the fuse change much from one to another, or does the temperature rise of the fuse contribute too much to resistance instability?
Both, as I don't always buy the fuses from the same place. And they heat up significantly, way more than the MOSFETs.
We Are The Watt - Resistance Is Futile!
 

Offline JohnG

  • Frequent Contributor
  • **
  • Posts: 567
  • Country: us
Re: kWeld - "Next level" DIY battery spot welder
« Reply #478 on: December 08, 2017, 03:14:48 pm »
For what it's worth, I have found that Infineon MOSFETs have pretty accurate resistance values, maybe within 10% or better, and the tempco is well specified. My experience has mostly been 100V FETs and above, so it may not be directly applicable here.

I would guess that if you know the FET temp, you can do a pretty good job. For those who have to know, there could be an option to calibrate the current using an external shunt. Not sure if it would be worthwhile, because I suspect the current varies a lot during the weld process.

John

"Reality is that which, when you quit believing in it, doesn't go away." Philip K. Dick (RIP).
 

Offline tatus1969Topic starter

  • Super Contributor
  • ***
  • Posts: 1273
  • Country: de
  • Resistance is futile - We Are The Watt.
    • keenlab
Re: kWeld - "Next level" DIY battery spot welder
« Reply #479 on: December 08, 2017, 11:21:37 pm »
For what it's worth, I have found that Infineon MOSFETs have pretty accurate resistance values, maybe within 10% or better, and the tempco is well specified. My experience has mostly been 100V FETs and above, so it may not be directly applicable here.
Did you find this by measuring, or were these datasheet figures?

I would guess that if you know the FET temp, you can do a pretty good job. For those who have to know, there could be an option to calibrate the current using an external shunt. Not sure if it would be worthwhile, because I suspect the current varies a lot during the weld process.
I was considering calibrating Rdson during production testing, but as mentioned before the absolute accuracy is not that important for this application. And as the MOSFETs only heat up by 25K (accounting for ~10% of error), I didn't go for measuring their temperature either. If I would add temperature monitoring, then to protect the power stage in first place.

In the end, kWeld has proven to make very repeatable welding results, and that's all it has to do  ;)
We Are The Watt - Resistance Is Futile!
 

Offline ovnr

  • Frequent Contributor
  • **
  • Posts: 658
  • Country: no
  • Lurker
Re: kWeld - "Next level" DIY battery spot welder
« Reply #480 on: December 10, 2017, 01:34:08 pm »
In the end, kWeld has proven to make very repeatable welding results, and that's all it has to do  ;)

Let's have a massive round of applause for tatus1969. Congratulations, you've avoided the feature creep hole I invariably spend my days digging even deeper...
 
The following users thanked this post: tatus1969

Offline Fraser

  • Super Contributor
  • ***
  • Posts: 13145
  • Country: gb
Re: kWeld - "Next level" DIY battery spot welder
« Reply #481 on: December 10, 2017, 02:06:46 pm »
Yes, the scourge of many projects. Frank has released a perfectly functional design that he is now enhancing in a controlled manner to provide sensible improvement. The wish list for many projects can become a nightmare as the BoM and design man hours can spiral out of control.

Great work Frank  :-+

Fraser
If I have helped you please consider a donation : https://gofund.me/c86b0a2c
 
The following users thanked this post: tatus1969

Offline rhb

  • Super Contributor
  • ***
  • Posts: 3476
  • Country: us
Re: kWeld - "Next level" DIY battery spot welder
« Reply #482 on: December 10, 2017, 06:12:55 pm »
I went back through the thread quite a bit, but it's a long thread, so please forgive me if this was already addressed.

As I understand it, you're discharging a capacitor bank and you're measuring the energy of the pulse by measuring the current.  The accuracy of the  measurement is limited by the difficulty of measuring a large current under dynamic conditions.

If the charging current is disconnected prior to the pulse and the capacitor bank voltage measured before and after the pulse that should give an accurate measure of the energy from first principles.  There is likely some voltage variation after the pulse is cutoff related to the degree to which the capacitors do not discharge uniformly.  However, this might provide a means for assessing the condition of the capacitor bank.

Is there a problem with using the relationship among capacitance, charge and voltage to determine the energy of a pulse that I have overlooked?  This seems to me a good bit easier and more accurate.  One still has heating losses, but those should not be too hard to measure or estimate using suitably placed thermistors.
 

Offline JohnG

  • Frequent Contributor
  • **
  • Posts: 567
  • Country: us
Re: kWeld - "Next level" DIY battery spot welder
« Reply #483 on: December 10, 2017, 10:17:54 pm »
For what it's worth, I have found that Infineon MOSFETs have pretty accurate resistance values, maybe within 10% or better, and the tempco is well specified. My experience has mostly been 100V FETs and above, so it may not be directly applicable here.
Did you find this by measuring, or were these datasheet figures?



At one point I did (i.e. my tech did) a LOT of voltage drop measurements on Si and SiC MOSFETs and IGBTs at 25, 75, and 125C. The Infineon ones all measured within 10% of that specified in the data sheet. Note that they were easier to measure than high current, low voltage FETs since the Ron was 10s to hundreds of milliohm, but I was surprised at how close they were to the data sheet. Interestingly, so were the measured CV curves.

Not all the FETs were like this... 

John
"Reality is that which, when you quit believing in it, doesn't go away." Philip K. Dick (RIP).
 
The following users thanked this post: tatus1969

Offline ovnr

  • Frequent Contributor
  • **
  • Posts: 658
  • Country: no
  • Lurker
Re: kWeld - "Next level" DIY battery spot welder
« Reply #484 on: December 10, 2017, 11:02:51 pm »
Is there a problem with using the relationship among capacitance, charge and voltage to determine the energy of a pulse that I have overlooked?  This seems to me a good bit easier and more accurate.  One still has heating losses, but those should not be too hard to measure or estimate using suitably placed thermistors.

That only estimates how much energy you drain from the capacitors - not how much gets delivered to the weld site. A majority of the energy consumption is in the capacitors themselves (ESR).

Also, kWeld supports both supercaps and high-energy LiPo cells; the latter are more cost-effective, IMO, and would not work with your proposed measurement method.

Using Rds(on) to estimate current is actually a really common method. In fact, if you're reading this on a computer you're already using a device that does it, namely the CPU core voltage regulator on your motherboard.

 

Offline rhb

  • Super Contributor
  • ***
  • Posts: 3476
  • Country: us
Re: kWeld - "Next level" DIY battery spot welder
« Reply #485 on: December 11, 2017, 01:18:05 am »
ESR is measurable as is the cable resistance.  The relationship between charge and voltage for a LiPo is more complex than the linear relationship of a capacitor,  but I don't see why it is harder than the time variant  Rds(on) of a MOSFET passing a large current for a short period of time. 

A comparison to the measurement of CPU drains is fatuous at best.  That does not involve going from 0 to 1000A and back in a fraction of a heartbeat.
 

Offline ovnr

  • Frequent Contributor
  • **
  • Posts: 658
  • Country: no
  • Lurker
Re: kWeld - "Next level" DIY battery spot welder
« Reply #486 on: December 11, 2017, 01:58:43 am »
ESR is measurable as is the cable resistance.  The relationship between charge and voltage for a LiPo is more complex than the linear relationship of a capacitor,  but I don't see why it is harder than the time variant  Rds(on) of a MOSFET passing a large current for a short period of time. 
Wait. You think that estimating how many joules - with a precision of 0.1J or better - remain in a LiPo based solely on the pack voltage is easier than Rds(on) sensing? Really? Besides, as tatus1969 has said: The weld energy is not calibrated in any way. You're supposed to find the correct energy level by trial and error, at which point the welder will produce consistent results.

Also, the welder does the joule counting during the welding process, cutting off the current the instant the correct energy level has been reached. That means you can't very well measure the voltage before and after.

A comparison to the measurement of CPU drains is fatuous at best.  That does not involve going from 0 to 1000A and back in a fraction of a heartbeat.
True; 1 kA CPUs are still a bit out. Still, we do have 200+ amp CPUs like AMDs Threadripper (with a healthy overclock). And the VRM switches at 500+ kHz.
 

Offline tatus1969Topic starter

  • Super Contributor
  • ***
  • Posts: 1273
  • Country: de
  • Resistance is futile - We Are The Watt.
    • keenlab
Re: kWeld - "Next level" DIY battery spot welder
« Reply #487 on: December 11, 2017, 06:54:30 am »
As I understand it, you're discharging a capacitor bank and you're measuring the energy of the pulse by measuring the current.  The accuracy of the  measurement is limited by the difficulty of measuring a large current under dynamic conditions.
The kWeld firmware uses a measurement cycle of 10 microseconds, that is enough bandwidth to sample the current waveform.

If the charging current is disconnected prior to the pulse and the capacitor bank voltage measured before and after the pulse that should give an accurate measure of the energy from first principles.  There is likely some voltage variation after the pulse is cutoff related to the degree to which the capacitors do not discharge uniformly.  However, this might provide a means for assessing the condition of the capacitor bank.
There are quite a few CD welders that use a similar mechanism. But instead of measuring voltage before and after, they let the user dial in the desired stored energy, calculate the corresponding charging voltage for them, and firing all the energy at once until the capacitors are drained completely. This has the advantage that you can use an SCR as the switch, as it does not have to have the capability to interrupt the flowing current again.

Is there a problem with using the relationship among capacitance, charge and voltage to determine the energy of a pulse that I have overlooked?  This seems to me a good bit easier and more accurate.  One still has heating losses, but those should not be too hard to measure or estimate using suitably placed thermistors.
The idea behind kWeld is to control the energy that goes into the weld. (To achieve that, measuring the flowing current is just one requisite.) With your idea (or the way standard CD welders work) you can only control the total energy consumption. A large part of that goes into the capacitor (or battery) itself, the connecting leads, and the switch.
We Are The Watt - Resistance Is Futile!
 

Offline tatus1969Topic starter

  • Super Contributor
  • ***
  • Posts: 1273
  • Country: de
  • Resistance is futile - We Are The Watt.
    • keenlab
Re: kWeld - "Next level" DIY battery spot welder
« Reply #488 on: December 11, 2017, 06:59:08 am »
ESR is measurable as is the cable resistance.
Yes, but measuring ESR is technically easier, as the MOSFET that you use as the current shunt is typically GND referenced. This way you only need to measure one ground referenced voltage, instead of having to measure a voltage differential. In this way I am following the same rationale that has brought up ESR current sensing in DC/DC converters.
We Are The Watt - Resistance Is Futile!
 

Offline rhb

  • Super Contributor
  • ***
  • Posts: 3476
  • Country: us
Re: kWeld - "Next level" DIY battery spot welder
« Reply #489 on: December 12, 2017, 12:44:33 am »
Interesting.  Thanks for the explanation.
 

Offline carracing111

  • Contributor
  • Posts: 21
  • Country: ch
Re: kWeld - "Next level" DIY battery spot welder
« Reply #490 on: December 31, 2017, 05:39:01 pm »
Hello everyone.

I am waiting to buy a KWeld spot welder. It has an incredible appearance!

Meanwhile I'm doing some capacitor testing with my Arduino Spot Welder.

I have purchased two Maxwell 58F units to build two different welders and power my Arduino Spot Welder, and I also hope to use them with my future  Kweld !

At the moment I have not obtained any acceptable welding with Maxwell 58F. It seems that the capacitors do not have enough power.
I have charged the capacitors with a small CC CV power supply (12V, 5A).
I have connected the Arduino Spot Welder directly to the capacitors to perform the tests. The set works, but the welds are very loose, the solders hardly leave a mark on the nickel.

The arduino spot welder is working perfectly with a car lead battery.
Of course I have checked that the capacitors have reached their working voltage, 12.15v. I do not understand why even the first weld is not acceptable.

I just made some tests with the same configuration but with 5 car audio capacitors, they are Chinese and of poor quality. In theory it is assumed that 2 parallel capacitors are 1 farad.

With the Chinese capacitors, the Arduino Spot Welder works very well!





I can not understand why the Maxwell capacitors do not work! I bought them on Mouser and they are not cheap! Can the capacitors be defective?

I have tried both units and the result is just as bad.
Do they need more than 12V to work properly? I can not think of anything else.

I would greatly appreciate any ideas or advice.

Happy New Year !
« Last Edit: December 31, 2017, 05:40:40 pm by carracing111 »
 

Offline tatus1969Topic starter

  • Super Contributor
  • ***
  • Posts: 1273
  • Country: de
  • Resistance is futile - We Are The Watt.
    • keenlab
Re: kWeld - "Next level" DIY battery spot welder
« Reply #491 on: January 01, 2018, 01:54:28 pm »
I have purchased two Maxwell 58F units to build two different welders and power my Arduino Spot Welder, and I also hope to use them with my future  Kweld !
Happy New Year to you too!

I assume that it is this module:  BMOD0058 E016 B02

A quick look in the datasheet tells that they have 22mOhm ESR. This is way too much for welding. The modules seem to be optimized towards energy density, not power density. If you still want to use them, I would suggest ripping both modules apart and reconfigure them in a 3S4P configuration. That way you'd achieve 2.75mOhm ESR at 5.3V, which should result in 870A weld current with kWeld. That should be good enough for 0.1mm nickel, maybe 0.2mm.
We Are The Watt - Resistance Is Futile!
 

Offline carracing111

  • Contributor
  • Posts: 21
  • Country: ch
Re: kWeld - "Next level" DIY battery spot welder
« Reply #492 on: January 01, 2018, 05:10:51 pm »
Thank you very much for the reply.

That's right, it's the Maxwell BMOD0058 E016 B02.

I chose these modules inspired by this design: https://www.thingiverse.com/thing:2588371

In the article he says that they work well and that's why I chose them.
These modules are not cheap, If I can not get the spot welder to work with just one module, it might be worth it to return them and order different capacitors ...

In the electrical diagram of the article where these capacitors are used, I have seen that there are 2 resistors of 10ohm and 100W in parallel with the capacitors.

I thought that the purpose of the resistors was only to discharge the capacitors after using the soldering iron. Maybe they are necessary for it to work properly?

I'm sorry, but my knowledge of electronics is still very basic.

Regards.
 

Offline tatus1969Topic starter

  • Super Contributor
  • ***
  • Posts: 1273
  • Country: de
  • Resistance is futile - We Are The Watt.
    • keenlab
Re: kWeld - "Next level" DIY battery spot welder
« Reply #493 on: January 01, 2018, 05:26:05 pm »
this design: https://www.thingiverse.com/thing:2588371
Maybe you contact the author of that unit, he is talking about 0.22mOhm ESR but the datasheet clearly states 22mR. That is a factor 100 off.

In the electrical diagram of the article where these capacitors are used, I have seen that there are 2 resistors of 10ohm and 100W in parallel with the capacitors.
I thought that the purpose of the resistors was only to discharge the capacitors after using the soldering iron. Maybe they are necessary for it to work properly?
Thats not relevant, they are for discharge. Just make sure to use the thickest wires that you can afford. Use AWG8 or better.

I'm sorry, but my knowledge of electronics is still very basic.
Pay attention to what you do, you are dealing with high energies. You can easily hurt yourself or burn down your house if you do not fully understand what you are doing. Also take great care of handling Lithium batteries, they are potentially dangerous.
We Are The Watt - Resistance Is Futile!
 

Offline carracing111

  • Contributor
  • Posts: 21
  • Country: ch
Re: kWeld - "Next level" DIY battery spot welder
« Reply #494 on: January 01, 2018, 06:13:43 pm »
Thanks, I have contacted the author of the article.

Although I believe that I will return the Maxwell modules and I will obtain the ones you recommend or if you sell them directly.

Regards.
 

Offline tatus1969Topic starter

  • Super Contributor
  • ***
  • Posts: 1273
  • Country: de
  • Resistance is futile - We Are The Watt.
    • keenlab
Re: kWeld - "Next level" DIY battery spot welder
« Reply #495 on: January 02, 2018, 10:32:34 am »
if you sell them directly.
I am currently working on the next batch kWeld, and on a customer project (a motor controller), so I didn't have time to continue the stationery kWeld power supply. I think that I can continue this in three weeks, as it is already almost mature. I plan to sell the capacitor balancer board with and without capacitors.
We Are The Watt - Resistance Is Futile!
 

Offline tatus1969Topic starter

  • Super Contributor
  • ***
  • Posts: 1273
  • Country: de
  • Resistance is futile - We Are The Watt.
    • keenlab
Re: kWeld - "Next level" DIY battery spot welder
« Reply #496 on: January 02, 2018, 10:36:32 am »
A quick update: I had time to test the Turnigy graphene 3S / 6000mAh / 65C. It delivers about the same current as the smaller and cheaper nanotech 3S / 5000mAh / 65-130C battery. The only weakness is that the leads are only AWG10. Does anyone know of affordable Lipos that have AWG8 and XT150 connectors?
We Are The Watt - Resistance Is Futile!
 

Offline Derriell

  • Newbie
  • Posts: 4
  • Country: ch
Re: kWeld - "Next level" DIY battery spot welder
« Reply #497 on: January 04, 2018, 06:53:59 pm »
It has arrived today   :-+

How many mOhms (approximately) will it lose when using thicker cables and solder the cables (partly) directly to the board?
 

Offline tatus1969Topic starter

  • Super Contributor
  • ***
  • Posts: 1273
  • Country: de
  • Resistance is futile - We Are The Watt.
    • keenlab
Re: kWeld - "Next level" DIY battery spot welder
« Reply #498 on: January 05, 2018, 10:45:41 am »
It has arrived today   :-+

How many mOhms (approximately) will it lose when using thicker cables and solder the cables (partly) directly to the board?
Great!

You can estimate this directly from the AWG relationship, I use to use Wikipedia. The table is for solid cores, you can add maybe 15% to the numbers for fine stranded cores. https://en.wikipedia.org/wiki/American_wire_gauge

The standard setup is AWG8, total length 0.8m.

I don't recommend soldering the excess strands to the PCB, just cut them. The cable lugs have negligible resistance.
We Are The Watt - Resistance Is Futile!
 

Offline LuboN

  • Newbie
  • Posts: 1
  • Country: sk
Re: kWeld - "Next level" DIY battery spot welder
« Reply #499 on: January 05, 2018, 10:08:18 pm »
Try to use two or three smaller capacity LiPos parallel, that way the resistance of leads and connectors divide  ;)
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf