6pence Resistance Spot Welder

Place for dealers to post items for sale.
User avatar
pguk   100 W

100 W
Posts: 107
Joined: Mar 16 2012 4:30pm
Location: North England, birthplace of the Industrial Revolution

Re: 6pence Resistance Spot Welder

Post by pguk » Jan 06 2017 9:56am

I finally have a layout for the power board which I believe I'm happy with:
6pencePwrShot.png
6pencePwrShot.png (157.43 KiB) Viewed 913 times
I have posted the kicad project files in the first post - for a proper look you can load into kicad to view each layer more clearly.

So, I have the Schottky flyback diode surface mounted, and 3 TVS diodes sat on top of the FETs. I have increased the number of FETs to 8 for those with ambition to weld brass. Gate resistors are on the underside right underneath the FET legs, and pops up through a via next to the gate leg. I made some via stitching so that both sides of the board carry weld current. The blue stripes are where there is no solder mask on front side. The idea being is to run copper wire (2.5mm sq times 3) from the input lug of LHS directly to the FET drains and then from the FET sources to the output lug on RHS. I was thinking to use 2 oz copper, board prices above 2 oz go up dramatically.

cheapest quote from allpcb for a batch of 10 shipped to me:
1oz = $35
2oz = $50
3oz = $89
4oz = $191

I reckon 2 oz should be ok, what do you all think? I have gerber files ready to submit.
If there's enough interest I'm happy to supply these at cost to those buying a control module.

edit: 1oz copper should be ok as we are not relying in any way on the board foil to carry weld current.
Last edited by pguk on Feb 05 2017 4:02pm, edited 2 times in total.
6pence spot welder https://endless-sphere.com/forums/viewt ... 31&t=81274
"It is all nice and proper, but only an idiot would think that the father is sixpence to the good on the transaction." C.S. Lewis' Mere christianity

User avatar
fechter   100 GW

100 GW
Posts: 14637
Joined: Dec 31 2006 3:23pm
Location: California Bay Area, USA

Re: 6pence Resistance Spot Welder

Post by fechter » Jan 06 2017 10:44am

What is the range for the weld time?
Can the software be tweaked for longer times?

I'm thinking this might work as a trigger for a microwave oven transformer style welder. Using a large solid state relay, the trigger just needs to be a 5-12v low current pulse.
"One test is worth a thousand opinions"

User avatar
pguk   100 W

100 W
Posts: 107
Joined: Mar 16 2012 4:30pm
Location: North England, birthplace of the Industrial Revolution

Re: 6pence Resistance Spot Welder

Post by pguk » Jan 06 2017 11:38am

tbh i haven't looked into compatibility with MOT style welder, but:

both pulses and delay between them are set to whatever values req'd. that's why I needed to use the lcd display.
a push of a button increases p2 in 1ms steps, p1 and delay increase in 0.1 ms steps.
it uses an mcp1407 fet driver chip which puts out ~ 11v (=12v less a diode drop if using a 12v adapter), at 6 amps.

So - fits your specs..? You could reprogram to make a button press cause different value of step changes, if these are too cumbersome for driving MOT.
6pence spot welder https://endless-sphere.com/forums/viewt ... 31&t=81274
"It is all nice and proper, but only an idiot would think that the father is sixpence to the good on the transaction." C.S. Lewis' Mere christianity

User avatar
fechter   100 GW

100 GW
Posts: 14637
Joined: Dec 31 2006 3:23pm
Location: California Bay Area, USA

Re: 6pence Resistance Spot Welder

Post by fechter » Jan 06 2017 1:06pm

Output is good. What is the range for the time (max ms)?

The PCB layout looks nice. I wouldn't depend on the board copper to carry the output current at all. A separate chunk of copper or aluminum would be needed. People blew the traces off on some of the riba welders. Hefty copper wire soldered to the board is the easiest approach.
"One test is worth a thousand opinions"

User avatar
pguk   100 W

100 W
Posts: 107
Joined: Mar 16 2012 4:30pm
Location: North England, birthplace of the Industrial Revolution

Re: 6pence Resistance Spot Welder

Post by pguk » Jan 06 2017 3:00pm

sorry, to be clear - there is no ceiling on the value for 2nd pulse. 1st pulse and delay have maximum value of 9.9ms, due to the larger variable type used in the software. I find I need very small values (under 1ms) for p1 & delay when welding with lead acid.

I do recall people having to fix traces on the JP welder - I hope I learned the lessons with this layout. There is room for 3 lengths of 2.5mm solid copper laid flat. This is plentiful in UK as we use it to wire our houses. This could be stacked to a number of layers. Or as you say put aluminium or copper bar there instead.
thank you for your feedback fechter :)

I have a commitment for 2 boards so far. I will go ahead and submit gerbers to the fab house :)
Last edited by pguk on Jan 07 2017 7:18am, edited 1 time in total.
6pence spot welder https://endless-sphere.com/forums/viewt ... 31&t=81274
"It is all nice and proper, but only an idiot would think that the father is sixpence to the good on the transaction." C.S. Lewis' Mere christianity

Valld   10 mW

10 mW
Posts: 23
Joined: Dec 06 2016 8:03am

Re: 6pence Resistance Spot Welder

Post by Valld » Jan 07 2017 12:46am

Long vacation... Please count me for 1 power board as well.

User avatar
fechter   100 GW

100 GW
Posts: 14637
Joined: Dec 31 2006 3:23pm
Location: California Bay Area, USA

Re: 6pence Resistance Spot Welder

Post by fechter » Jan 07 2017 2:13am

pguk wrote:sorry, to be clear - there is no ceiling on the value for 2nd pulse. 1st pulse and delay have maximum value of 16ms iirc, due to the larger variable type used in the software. I find I need very small values (under 1ms) for p1 & delay when welding with lead acid.

I do recall people having to fix traces on the JP welder - I hope I learned the lessons with this layout. There is room for 3 lengths of 2.5mm solid copper laid flat. This is plentiful in UK as we use it to wire our houses. This could be stacked to a number of layers. Or as you say put aluminium or copper bar there instead.
thank you for your feedback fechter :)
So could the second pulse be 100ms?

Good on the copper. That's how I would do it. Having room for it on the board helps a lot. Even a single strand of 2.5 will not be likely to blow off like the foil on the board.
"One test is worth a thousand opinions"

User avatar
pguk   100 W

100 W
Posts: 107
Joined: Mar 16 2012 4:30pm
Location: North England, birthplace of the Industrial Revolution

Re: 6pence Resistance Spot Welder

Post by pguk » Jan 07 2017 7:17am

Valld wrote:Long vacation... Please count me for 1 power board as well.
okay, one is reserved for you. Thanks for your valued input Valentin :)
fechter wrote: So could the second pulse be 100ms?

Good on the copper. That's how I would do it. Having room for it on the board helps a lot. Even a single strand of 2.5 will not be likely to blow off like the foil on the board.
Yes. I just tried it - actually I've never been all the way up there before you asked. It will go a whole lot higher than that as there is plenty of room for a large number on the 2nd row of the lcd. It is rather cumbersome to get high numbers (lots of button pushes), it really needs coding for a fast scroll when up or down button is held. There is also the possibility of swapping the buttons for a rotary encoder. This would be for a later version tho.

I have put four M3 mounting holes on the power board having the same spacing as those on the control module and the LCD. Control module then sits on top of the power board on M3 standoffs.
Last edited by pguk on Jan 07 2017 1:18pm, edited 1 time in total.
6pence spot welder https://endless-sphere.com/forums/viewt ... 31&t=81274
"It is all nice and proper, but only an idiot would think that the father is sixpence to the good on the transaction." C.S. Lewis' Mere christianity

User avatar
spinningmagnets   100 GW

100 GW
Posts: 11528
Joined: Dec 21 2007 10:27pm
Location: Ft Riley, NE Kansas

Re: 6pence Resistance Spot Welder

Post by spinningmagnets » Jan 07 2017 10:28am

I'd like to go over my understanding of the functioning of this unit for just a minute. The length of the timing of the single or dual pulses is easy to grasp. By going to 8 FETs, I can see how each FET would run cooler, and it would allow a higher amperage to flow through the board at its max current. How to adjust the current?

Once dialed in for a certain job, no further adjustment would be necessary. However...is the current determined by the max amp capability of the battery used? That's fine if it is, I am just curious if there is some easy way to use a very high-amp battery, and then dial it down in-between the battery and spot-welder (without altering the performance of the unit).

I accept that I may end up needing to use nickel for the parallel connections to the cell tips. however, I am still certain that thick copper is best to connect those nickel strips to each other for the series connection. Also, I have not completely given up on the possibility of brass working for the parallel connections, since I have seen a video of it being done on a video of an expensive machine.

The final form of a machine that is dedicated to one job can be very simple, but the experimentation process demands some adjustability.

User avatar
fechter   100 GW

100 GW
Posts: 14637
Joined: Dec 31 2006 3:23pm
Location: California Bay Area, USA

Re: 6pence Resistance Spot Welder

Post by fechter » Jan 07 2017 4:46pm

Yes, the current is a function of the battery and total circuit resistance. Generally, you want as much current as you can get. Weld energy will be current x time. Varying the time gives you total control over the weld energy.
With a small battery, it's hard to get enough current and the electrodes have more tendency to stick.
"One test is worth a thousand opinions"

User avatar
pguk   100 W

100 W
Posts: 107
Joined: Mar 16 2012 4:30pm
Location: North England, birthplace of the Industrial Revolution

Re: 6pence Resistance Spot Welder

Post by pguk » Jan 09 2017 7:10am

Concisely put by fechter, I agree with all of that.

I have a book by Horowitz & Hill bought back when I was a student entitled 'The Art of Electronics'. This kind of expresses the idea that there isn't an exact right or wrong way to achieve an outcome with electronics. You can get there by experiment, trying this and that and getting a feel for what works, just as well as by heavy maths gymnastics. From crossbreak's thread, my solution ended up a little different to riba's, although both work.

From Okashira's posts on the other thread, he was reporting great results with his beefed up JP welder using some kind of lithium battery for weld current running around 9 volts iirc. The JP unit uses T0-220 version of the FET which can't handle as much current as the smd version which I chose to use. I have increased qty from 6 to 8. My power board has incorporated a lot of ideas from the other thread especially the use of flyback and TVS diodes, which is why I made it open source :)

I believe you have to shoot for the lowest system resistance as possible. As fechter says you need as much current available as possible, weld energy is regulated by time. This means fat weld cables (25mm sq - not too long), and a weld source battery having low internal resistance. My initial testing last year was using a small leisure battery. I hardly got any metal to melt, so in the video I have put my car battery in parrallel. Someone reported good results with a single gel type lead acid battery iirc.

I had enough sales over the weekend (thank you) to fund a good batch of power boards so I'll look over the layout one last time before I send gerbers off.

Edit 10.1.17 Gerbers submitted to fab house for audit
11.1.17 audit passed. Lead time 3-4 days, then 11 days or so shipping time to get to me. QTY 20 on 2oz copper.
Last edited by pguk on Jan 11 2017 6:21am, edited 1 time in total.
6pence spot welder https://endless-sphere.com/forums/viewt ... 31&t=81274
"It is all nice and proper, but only an idiot would think that the father is sixpence to the good on the transaction." C.S. Lewis' Mere christianity

zwillx   1 µW

1 µW
Posts: 3
Joined: Jan 07 2017 11:48pm

Re: 6pence Resistance Spot Welder

Post by zwillx » Jan 10 2017 11:19pm

exciting! subbed and interested.

quick (and prob obvious) question, would an inline capacitor (bank) add anything to the output? I've seen ultracaps mostly replace car batteries anyway, so it seems like some combination could provide a benefit?

then, would it be OK to trickle charge battery while working so voltage wouldn't drop? I guess I mean something similar to what alternator provides when the battery is in its native environment. not sure how fast they drain in use, but I'd rather a combo of cap/charger than an extra lead-acid. plus, already have the charger :)

thanks for all the hard work and talent!

User avatar
pguk   100 W

100 W
Posts: 107
Joined: Mar 16 2012 4:30pm
Location: North England, birthplace of the Industrial Revolution

Re: 6pence Resistance Spot Welder

Post by pguk » Jan 11 2017 11:16am

You mean capacitor bank as weld current source right? Sure, some do use capacitors, you just need rather a lot of them, at least a farad of good quality caps = more pricey than a lead acid setup. Beware car audio capacitors - they mostly are not what they look like on the outside.
When setting up for charging as you weld, the TVS Diode's reverse bias conducting characteristics are what you need to look at. In the BOM I have specified using 3 of 12v rated. Your charger will push the batt voltage above 12v, so if those TVSs begin to conduct, you should choose a higher voltage TVS. Trade off is that as they begin to conduct later, thy may be less effective for diverting current from those MOSFETs.
6pence spot welder https://endless-sphere.com/forums/viewt ... 31&t=81274
"It is all nice and proper, but only an idiot would think that the father is sixpence to the good on the transaction." C.S. Lewis' Mere christianity

Roncos   1 mW

1 mW
Posts: 19
Joined: Oct 08 2016 5:03am

Re: 6pence Resistance Spot Welder

Post by Roncos » Jan 22 2017 5:15am

Hi PGUK

This looks brilliant. Is there any possibility that you could post a description or a photo of the whole system so we can see how it all goes together, especially the lead acid battery.

Saw your stuff on ebay. Do you plan to sell the power boards also.

User avatar
pguk   100 W

100 W
Posts: 107
Joined: Mar 16 2012 4:30pm
Location: North England, birthplace of the Industrial Revolution

Re: 6pence Resistance Spot Welder

Post by pguk » Jan 22 2017 2:23pm

Roncos wrote:Hi PGUK

This looks brilliant. Is there any possibility that you could post a description or a photo of the whole system so we can see how it all goes together, especially the lead acid battery.

Saw your stuff on ebay. Do you plan to sell the power boards also.
Hi Roncos

Thanks for the kind words. This has been a slow burn for me, but we're getting to the end game at last I hope.
As I said at first, I don't have time to get into full production of these welders for folks, but I can continue to offer a built and tested control module with small parts. I have with me now 22 of the power boards which I worked on over christmas. Three of these are being built up as a special, when they're done I'll be able to post a 'how to' series of photos on here soon. Power board is open source - kicad files are in the first post. These can be had with gate resistors fitted for £6 from me for those buying a control module.

The exciting news is this.
One of the power boards I'm building up is for Boobyjohn in Denmark. He has a crazy number of hours of coding experience, including embedded systems, and has kindly offered to refactor the code with me. It will take a little time yet till we have something we are happy to release, but when that happens I will put it up on here also. The code was developed on Arduino, so its really accessible. I'm learning new stuff with what Boobyjohn is sending through.
Edit: this never quite got off the ground - no matter, no need for otp coding for a simple switch !
Last edited by pguk on Aug 08 2017 11:05am, edited 1 time in total.
6pence spot welder https://endless-sphere.com/forums/viewt ... 31&t=81274
"It is all nice and proper, but only an idiot would think that the father is sixpence to the good on the transaction." C.S. Lewis' Mere christianity

Roncos   1 mW

1 mW
Posts: 19
Joined: Oct 08 2016 5:03am

Re: 6pence Resistance Spot Welder

Post by Roncos » Jan 22 2017 3:43pm

Follow up question. Has anyone commented on dangers of using FETs in parallel?

I seem to remember someone who I trusted about these things telling me that FETS do not share the load equally when in parallel. A thermal runaway like: the one that initially heats up slightly faster than the others looses resistance and gets more current and heats up more.... so very quickly one FET is taking all the current red hot and the others are doing nothing. Maybe it doesn't matter for short pulses? Maybe they have fixed this with modern designs? Any info?

User avatar
fechter   100 GW

100 GW
Posts: 14637
Joined: Dec 31 2006 3:23pm
Location: California Bay Area, USA

Re: 6pence Resistance Spot Welder

Post by fechter » Jan 22 2017 3:55pm

FET resistance increases with temperature so they tend to share nicely. The duty cycle of the welder is so low that they won't heat much anyway. Equal path resistance in the connections will be more important.
"One test is worth a thousand opinions"

User avatar
spinningmagnets   100 GW

100 GW
Posts: 11528
Joined: Dec 21 2007 10:27pm
Location: Ft Riley, NE Kansas

Re: 6pence Resistance Spot Welder

Post by spinningmagnets » Jan 22 2017 4:30pm

A controller powering a motor will have the FETs cycling very rapidly with no rest, also housed inside a case that holds-in heat. You raise a good question, but I am certain these will not have any heat issues.

Roncos   1 mW

1 mW
Posts: 19
Joined: Oct 08 2016 5:03am

Re: 6pence Resistance Spot Welder

Post by Roncos » Jan 22 2017 5:17pm

fechter wrote:FET resistance increases with temperature so they tend to share nicely. The duty cycle of the welder is so low that they won't heat much anyway.
I did a quick search to check the point about resistance increasing or decreasing with temperature. These papers seem to say that resistance goes down as temperature goes up.
http://web.eecs.utk.edu/~tolbert/public ... n_2004.pdf
http://web.eecs.utk.edu/~tolbert/public ... t_2004.pdf
What do you think?

Agree about the duty cycle. The FETs will not have enough on time to heat up much anyway so it is just a point of curiosity.

User avatar
pguk   100 W

100 W
Posts: 107
Joined: Mar 16 2012 4:30pm
Location: North England, birthplace of the Industrial Revolution

Re: 6pence Resistance Spot Welder

Post by pguk » Jan 23 2017 5:12am

I just took a look a the data sheet for these FETs. There is a graph of RDS_on against Temp which shows a steady increase of on resistance from 0.8 mohm to 1.6 mohm as temp rises from -60 to +-80 degrees.

The FETs stay close to ambient temp in practice. More important I think regarding good current sharing of paralleled FETs is the provision of individual gate resistors. Initially I was convinced that a single gate resistor was the way to go - following the advice of one particular builder. However I since found that industry best practice is to use a dedicated gate resistor for each device.
6pence spot welder https://endless-sphere.com/forums/viewt ... 31&t=81274
"It is all nice and proper, but only an idiot would think that the father is sixpence to the good on the transaction." C.S. Lewis' Mere christianity

User avatar
fechter   100 GW

100 GW
Posts: 14637
Joined: Dec 31 2006 3:23pm
Location: California Bay Area, USA

Re: 6pence Resistance Spot Welder

Post by fechter » Jan 23 2017 12:21pm

Roncos wrote: I did a quick search to check the point about resistance increasing or decreasing with temperature. These papers seem to say that resistance goes down as temperature goes up.
http://web.eecs.utk.edu/~tolbert/public ... n_2004.pdf
http://web.eecs.utk.edu/~tolbert/public ... t_2004.pdf
What do you think?
Those are both Silicon Carbide transistors. Different animals. Old-school bipolar junction transistors conduct more with temperature, so were prone to current hogging and thermal runaway. The FETs we are using have a positive temperature coefficient.
"One test is worth a thousand opinions"

Boobyjohn   1 mW

1 mW
Posts: 18
Joined: Jun 13 2015 1:03pm

Re: 6pence Resistance Spot Welder

Post by Boobyjohn » Jan 23 2017 2:27pm

The exciting news is this.
One of the power boards I'm building up is for Boobyjohn in Denmark. He has a crazy number of hours of coding experience, including embedded systems, and has kindly offered to refactor the code with me. It will take a little time yet till we have something we are happy to release, but when that happens I will put it up on here also. The code was developed on Arduino, so its really accessible. I'm learning new stuff with what Boobyjohn is sending through.
I do consider this news exciting as well. We (i.e., the community) have already been through a few different spot welder designs that have been left unsupported. With this design fully open source I really hope that we can work out a highly robust and useful platform for spot welding. I also hope that most people will continue sourcing their controller and power boards through pguk for as long as he wishes to provide these boards.

That being said, what SW defined feature would you appreciate the most?
I see a few opportunities (from top of head):
1. Save/reload of configuration. The controller has an EEPROM good for 100k write cycles. No need to reconfigure every time it boots!
2. Pre-configured profiles (made up profiles below):
2.1 0.3 mm Nickel strips, 50 Ah battery -> P1: 2 ms, delay 5, P2 12 ms
2.2 0.3 mm Nickel strips, 70 Ah battery -> P1: 1 ms, delay 5, P2 9 ms
2.3 0.5 mm Nickel strips, 70 Ah battery -> P1: 3 ms, delay 5, P2 15 ms
3. Usability improvements, e.g., long press on inc/decrease buttons -> Rapid increase/decrease of selected parameter.
4. Safety features? (Dunno about this one)
5. Calibration features?

Ad 2) This would have to be crowd-sourced somehow, by people submitting the parameters that worked well for them. I could then incorporate the database into the programming...

Ad 5) Maybe a, e.g., binary search would be helpful in quickly finding the best value for a parameter. That way, the user just needs to input boundary values and telling the controller if the weld fire "got too little" or "too much" current.

zwillx   1 µW

1 µW
Posts: 3
Joined: Jan 07 2017 11:48pm

Re: 6pence Resistance Spot Welder

Post by zwillx » Jan 24 2017 9:09am

Boobyjohn wrote: 3. Usability improvements, e.g., long press on inc/decrease buttons -> Rapid increase/decrease of selected parameter.
Definitely this for me. in TIG welding I've been annoyed w button operations like, for example, pulses per second. imagine clicking tenths to get from one to 30.0. My opinion is its worth learning the actual settings you want, and dialing them in, rather than depend on auto-settings. this allows you to adjust for something like a cold workshop or variations like using brass strips. And users can share their settings and results to establish norms, online rather than in the welder. By the way, has anybody had success w copper or brass? More difficult for the welder but much better economy it would seem!

User avatar
spinningmagnets   100 GW

100 GW
Posts: 11528
Joined: Dec 21 2007 10:27pm
Location: Ft Riley, NE Kansas

Re: 6pence Resistance Spot Welder

Post by spinningmagnets » Jan 24 2017 9:23am

No success yet on copper or brass. It's close, but more fails than good welds. They need much higher current, but shorter pulses. The slits help, so maybe using two strips per parallel string might work better? (infinite slit). For low resistance busses, magnets and compression look better right now, so go straight to copper if that interests you.

The only reason brass was even considered is because in theory it should be "easier" to spot-weld onto the cell tips, then spot welding copper series bars onto the brass would be very easy.

joby1   10 mW

10 mW
Posts: 29
Joined: Dec 15 2016 11:21pm

Re: 6pence Resistance Spot Welder

Post by joby1 » Jan 24 2017 7:26pm

mudflap5 wrote:Recently finished a 20S12P battery using the 6pence welder. The welder worked very well, although it took a few attempts to get the pulse timing right. I think timing has a lot to do with the size and capacity 12 volt battery, using a very large deep cycle marine battery for this build. The nickel strips were .25mm thick. The power board was built with the FET's bolted to an old heat sink from a burned up controller, and using the flyback and avalanche diodes. Made the welding tips from 4 guage solid copper wire. Here is a picture of the pack and the welder is in the background.
trying to wrap my head around pack design,
in the post it says 20s10p, is that battery not a 24s10p?

I am counting batteries and getting confused.. :)

thanks
Joby

Post Reply