kWeld - "Next level" DIY battery spot welder

[Fraser]

Having watched the videos of the probes in use, I think there is some room for improvement in that area as great care is needed to not touch the probes to each other. My Chinese transformer based welder uses approx 1.5mm hard drawn copper probes that are the required diameter for the contact surface with the tabs. To refinish the ends you just draw a file across them  Seem to work fine in my use.

Longer handles on the probes provides better directional control than you presently have over the probe placing. Your current probe design looks a bit like trying to write with a really short pencil.... its hard to control it well. In my experience you will likely get away with a sleeker and longer probe design. Maybe insulated brass bar around the same diameter as a ball point pen and grub screws holding the copper probe tips in place ?  Commercial systems I have seen use such an arrangement and there does not seem to be a problem with too high a resistance in the probes. My Chinese welder uses brass 'arms' to hold the probe tips in an arrangement similar to that which you have used. There are many possibilities as you know. With a lathe it is possible to create all manner of exotic probe designs including copper tips with a stepped offset for close parallel working 

The joy of your design is that the builder of the unit may modify it to suit their needs and desires. You are still developing the design of the complete system so please do not take my comments as any form of criticism.

I will add some pictures of other probe designs in a minute

Best Wishes

Fraser

[Fraser]

Other probe designs, some good, some needing refinement !

Fraser

Links to original pages:

http://zeva.com.au/Projects/SpotWelder/
https://endless-sphere.com/forums/viewtopic.php?f=2&t=2633&start=400
https://endless-sphere.com/forums/viewtopic.php?t=83411
http://frikkieg.blogspot.co.uk/2010/01/

[Kjelt]

Having watched the videos of the probes in use, I think there is some room for improvement in that area as great care is needed to not touch the probes to each other.

Watch the latest video again, he already solved the issue in software 

[tatus1969]

Having watched the videos of the probes in use, I think there is some room for improvement in that area as great care is needed to not touch the probes to each other.

Watch the latest video again, he already solved the issue in software 

I did what I can, but I think the automatic mode is still to be used with caution. That was maybe the "educational" part of my second last video. Having said that, this caution is advisable when using it in general, as you are dealing with power levels of 10kW+ here. And, if you decide to run it from Lipos (they don't really like this abuse), make sure to store them in a fire safe place. I have an old ammunition box for that.

But it's safe to say that nothing dangerous can happen from an accidental electrode short, apart from you getting a heart attack, and possibly damaged electrodes or battery tabs. The system has a hardware pulse duration limiter (250ms), and a fuse to prevent permanent short circuit current. I don't like the battery connection yet, that should definitely receive some heat shrink tubing as well. Imagine what happens when one of the bolts gets loose...

[tatus1969]

@Fraser: thanks, that is very helpful. And I certainly don't take it as an offense. I am not quite happy with the electrode holders, there is way too much machining involved them and they are too bulky. It's just the resistance that is perfect, probably in the order of < 100uOhm. I think I will make some experiments with a coaxial design, a hard drawn copper rod, a brass tube, and grub screws.

I don't think that I will reduce the copper rod diameter, the robustness testing showed that they generate lots of heat. Although quite some of it will be transferred heat from the weld spot. But I think a minimum thermal mass is important.

I think I will change the tip shape to something like this:

[SvanGool]

I added an aux DC input that can be fed with 10~12V when the system is used as a CD welder. The problem here is that the internal holdup capacitor is only good for 250ms. The power caps will probably not be recharged quick enough to keep it from power cycling otherwise.

You have a max setting of 500 Joule (which is already huge). At 500F 15V (6x2.5V, to stay safe) the start capacity (CV²/2) is 56 kJ and the end capacity @12V is 36 kJ, so the budget is 20 kJ which is 40 pulses, before charging, that means that keeping your capacitor/microprocessor won't be a problem. This will be linearly less when the capacitors are smaller (besides that their internal resistance is normally higher also). The 3000F pack has a specced internal resistance of 1.74 mOhm and will probably need some larger gauge cable to come in the neighborhood of the energy-budget that you currently have.

[Fraser]

@Kjelt,

Many thanks. Sorry, I totally missed the significance of the change shown in the latest video. There is a warning sound plus delay before the welding voltage is presented to the probes. An excellent refinement for the auto mode . I was concentrating on the additional data the display presents after the weld. Thanks again for getting me to re-watch the video.

Fraser

[tatus1969]

You have a max setting of 500 Joule (which is already huge). At 500F 15V (6x2.5V, to stay safe) the start capacity (CV²/2) is 56 kJ and the end capacity @12V is 36 kJ, so the budget is 20 kJ which is 40 pulses, before charging, that means that keeping your capacitor/microprocessor won't be a problem. This will be linearly less when the capacitors are smaller (besides that their internal resistance is normally higher also). The 3000F pack has a specced internal resistance of 1.74 mOhm and will probably need some larger gauge cable to come in the neighborhood of the energy-budget that you currently have.

I think that you have a mistake in your calculations (or maybe I misunderstood something). When you put six capacitors of 500F/2.5V each into series, then this makes an 83.3F / 15V capacitor. When going 15V -> 5V, it can deliver 8333J. Which is still more than enough, and you still don't need the aux DC input. Anyway, that's only 50ct extra 

Edit: I think I got it: you are using 6x BCAP0650 PBCAP3000 270 K04 in series. That means that you will need to add something that limits current to an acceptable level. At (1.74 + 0.14) mOhm, the starting current would be way above the 2kA limit of my welder. Hence to use larger gauge and maybe longer probe leads.

[SvanGool]

Sorry, I was not clear: I am talking about 6x3000F/2.7V (Maxwell BCAP3000 P270 K04, see also here) in series. Which makes a 500F/16.2V capacitor pack (if charged to the edge, I prefer 6x2.5V=15V).

You are correct, if the max starting current is 2 kA (the advised max is 1.9 kA for the BCAP3000) and the cap pack resistance = minim. 1.74 mOhm (6 x 0.29) then at least 6.2 mOhm is needed for cabling and tips. The advantage is, that your pulse length for a 200 J pulse would only need to be app. 7 mS and everything would be within the spec of the capacitors (no ammunition box needed anymore).
 

[Kjelt]

But you need a mortgage on the house to buy these supercaps    I looked them up on Farnell and they are €200 each 
Very nice though if you find them in a dumpster.

[SvanGool]

The price at Farnell is, for this case, not very representative 

If you follow the ebay-like sites for a period, you can find 6 Maxwell new/used caps in the range of 650F-3400F for a price much lower than the single Farnell price, but do watch out for "shady" sellers.
I'm trying to establish with tatus1969 what the lowest workable capacitor type would be, but we first need to have our "math" correct.
At the end, it is also preferable to have balancing or over-voltage protection for the caps.

[tatus1969]

16.2V capacitor pack (if charged to the edge, I prefer 6x2.5V=15V).

I've implemented a 5.8F/140V pack in the past, and we had the manufacturer in house for technical discussions. They stated that, in order to maintain their lifetime, you do not need to derate voltage, and it is much more important to keep the cell temperatures as low as possible. It's like for any other electrolytic capacitor: every 10°C less doubles the lifetime.

cap pack resistance = minim. 1.74 mOhm (6 x 0.29)

Be careful, the datasheet states this as the maximum initial ESR. There will certainly be some spread to lower values, so you need some headroom. The software based overcurrent detection will immediately trip at >= 2kA, and the measurement accuracy is not very high. It depends on production spread of the MOSFETs, the gate drive voltage (which is decreasing during the pulse), and the brass bus bar construction repeatability. Things like mounting torque...

I'm trying to establish with tatus1969 what the lowest workable capacitor type would be, but we first need to have our "math" correct.

At the moment I am busy building up price structure and stock, but I've already set one of the next tasks to investigate this: how should a power capacitor look like that
- delivers enough energy to the load in a short enough time
- stays within given voltage limits during this
- has the lowest possible price to achieve the above

It is not only a matter of ESR, you need to juggle between that and the voltage span. At higher voltages, you can allow more ESR.

Some numbers are already there for my welder:
- max input voltage: 30 V
- min input voltage for welding: 12 V
- min input voltage to keep microcontroller from power cycling: 4 V
- resistance of power switch, cabling and electrodes: approx 2.5 mOhm (can be reduced to <<1mOhm by using more material)
- resistance of weld spot: approx 1 - 4 mOhm

At the end, it is also preferable to have balancing or over-voltage protection for the caps.

As we don't need to be energy efficient, I am thinking of a passive solution with beefy equalization resistors. Their current must exceed the maximum required equalization current, and can be calculated from capacitance spread and leakage current. Of course that doesn't work as overvoltage protection.

[tatus1969]

A first lot of 25 PCBs is ordered, and I have the system part list and pricing complete. I will add articles to my web shop today, stay tuned!

And I have changed the electrode design like this now. The cable will be soldered to the electrode holder. The tip will be held into place with one screw/nut. The (brass) nut will be soldered to the tube to ease electrode replacement.

Please tell me what you think about it? The construction is much simpler now and can be machined quicker.

[Fraser]

Nice new probe design

Looking forward to being able to purchase this unit.

Fraser

[Kjelt]

The cable will be soldered to the electrode holder.
Please tell me what you think about it? The construction is much simpler now and can be machined quicker.

I do like the new tip connection. The cable to holder I am not so sure.
I am no mechanical expert so could be wrong but thinking out loud, a soldered connection should be inferior to a well crimped connection where the metals will become one without much of a transition resistance. Esp. with huge currents like in this application I would choose a direct crimp connection or hard welding connection.

[tatus1969]

The cable will be soldered to the electrode holder.
Please tell me what you think about it? The construction is much simpler now and can be machined quicker.

I do like the new tip connection. The cable to holder I am not so sure.
I am no mechanical expert so could be wrong but thinking out loud, a soldered connection should be inferior to a well crimped connection where the metals will become one without much of a transition resistance. Esp. with huge currents like in this application I would choose a direct crimp connection or hard welding connection.

Good point. I ordered parts to make a build and see how it performs. The soldering will be done with solid solder / flux / blow torch, in order to achieve a good fill. I always have in mind that everybody should be able to do the machining/assembly work, and not everybody has a suitable crimp tool that can handle 6mm diameter. Me neither... I mainly want to see kit parts.

[Fraser]

I see the main module is now in your shop. Are you taking pre-orders ?
If so, I would like to pre-order a unit plus some of the cables with pre-crimped ends.

Fraser

[tatus1969]

Just a quick update: the first items are online on https://www.kicksurfer.de/index.php/shop/
I will add more tomorrow, as it is late here.

The first batch of electronic sets is ordered and I estimate that I can add them to the stock by mid August. It will not be possible to place an order for that before, but as I appreciate your help here I want to respect the order in that you registered interest. If you want to buy an electronics set, then please register as a customer on my site and give me a PM here that you actuall plan to purchase an electronics set. I will reserve one set for you for the duration of one week after stock arrival.

All the other items will not be that problematic, as I can source them quickly. However, I will not build up a large stock of them because I think that most of you will want to source as much as possible locally and do the machining/soldering/crimping/assembling work by yourself.

[tatus1969]

I see the main module is now in your shop. Are you taking pre-orders ?
If so, I would like to pre-order a unit plus some of the cables with pre-crimped ends.

Fraser

You are quicker than I can type 
My shop system (Wordpress+Woocommerce) unfortuately does not support pre orders 

[Fraser]

Reservation request PM sent. I am a committed customer of your kit     

Fraser

[tatus1969]

It will not be possible to place an order for that before

I'll have to take that back, I just wasn't familiar enough with that system.

Sorry for the confusion!!!

WOW, that's an endless world of possibilities there! If someone needs a guide to setting up a Wordpress + Woocommerce server, here is another expert   

Okay, no more products added there yet, but I managed to set up the full chain for the electronics module, including
- accepting pre-orders now - I'll serve them in the order that they come in
- the option whether you want me to complete it (solder LCD and dial pot), or you want to do that
- world wide automatic shipping rate calculator
- safe credit card payment via stripe

I'll add Paypal as well, but my business account there needs verification and they told me that this would take 2-3 weeks  :shock:

I've set the limit to 20 copies now, to accomodate the possibility that not all of the 25 ordered are good.

Please order only one per person for now, in order to allow others receiving one. I need to refuse larger orders for the first batch. No problem of course later, after successful ramp up!

If you plan to purchase other options that are not in the store yet, please wait and you'll be offered a combined shipping rate. So you don't need to pay shipping twice.

I've activated the website again and do accept orders 

[Fraser]

I just went to pre-order. The price has gone up from 89 Euro to 99 Euro.

Is this correct please ?

Fraser

[tatus1969]

I just went to pre-order. The price has gone up from 89 Euro to 99 Euro.

Is this correct please ?

Fraser

No, I separated it differently before. It was
- 89€ for the partially soldered board
- 10€ for the additional components LCD / dial pot / pin header / knob

Now it is everything included in one package. I thought that it didn't make much sense to sell these separately. I initially did that because, if you want to design a housing for it, then you will want to use a different potentiometer type. This one does not have a mounting thread. But that part is worth less than 1€, so I canceled that now.

[Fraser]

 No worries

I see from your letter you will be adding additional items. I have one controller board in my basket. I will add to that later.

Best Wishes

Fraser

[tatus1969]

The remaining items are online as well. As I will be doing all the soldering, machining, and assembly work by myself for the time being, I need to set the prices for that relatively high. When this product is ramping up nicely, then this may change in the future.