Modifing a Battery Spot Welder because of Controller Issues

[sam_sam_sam]

Re: Modifing a Battery Spot Welder because of Controller Issues

Update

Well things did not go very well not everything would fit in this case so I had to go with plan “B” which is to cut down the the heat sink and put the triac back on the controller board again —>

this actually went better than I thought it would in that I cut the heat sink the width as the triac and a piece of non conducive paper that is used for making battery packs and put it on the circuit board so it would not short out

Now I have figure out how I am going to mount the small controller transformer and where

I also have to figure out how to mount the amp and volt meters on the face plate of the welder which will be a challenge because of how the welder arms are

There is also another challenge with welder arms because I change the configuration of the transformer secondary winding wires and where they go so I have to buy some welding wire that fit the welding arms

I will post some pictures when I get done with this project

[sam_sam_sam]

Re: Modifing a Battery Spot Welder because of Controller Issues

Well I had time to work on this battery spot welder exterior controller

The test results went well and the concept will work now I have mount everything in the controller case and fix the transformer and welder fixture and hook up amp meter and volt meter

[sam_sam_sam]

Re: Modifing a Battery Spot Welder because of Controller Issues

The parts came in and I will see if I have time over weekend and see if I get something working

[Curious.George]

Re: Modifing a Battery Spot Welder because of Controller Issues

The one question I have what type SCR do I use and how do I hook it up

I do not have much experience with high power SCRs so example would be nice

That depends on the amount of "energy" that it will be called upon to pass.

Note that the essential aspect of this approach is that you have to "pre-buffer" that energy packet (e.g., in a cap) and let the SCR just "dump" it into the load (the weld via the Xformer).

You can't, for example, have the SCR fed by a DC supply in any way because it won't shut off until the current flowing through it drops to zero -- which won't happen if there is a persistent "supply" present.

E.g., you want to have a charging circuit that you can control (turn on/off) that dumps charge into a storage device (capacitor). You need a way of monitoring how much charge has thus far been accumulated. And, you need a way of turning OFF that charge path once you've reached the proper amount of "energy". Then, let the SCR "dump" ALL of that charge into the load.

The load impedance (and voltage on the cap -- which reflects the amount of stored energy) will determine the maximum current through the switch. That, in turn, will determine how long the current will flow (until the charge is depleted). The current and time will determine the thermal shock that the switch will experience. The recovery time (time it takes you to recharge for the next shot) plus the thermal resistance of the switch will determine how quickly it can cool the junction to stay in the SOA.

Note that "SCR" is just a shorthand for a "4 layer device that shuts itself off". You can make such a device using other (e.g., 3 layer) devices. The point is that you want to rely on virtually nothing to turn the switch off else you risk something screwing up and leaving it on "too long" (and melting the switch!). And, in making that determination, you have to consider other devices in the circuit may screw up due to the large current spikes present, ground bounce, the phase of the moon, etc.

[sam_sam_sam]

Re: Modifing a Battery Spot Welder because of Controller Issues

I prefer the SCR approach -- just trigger it and let it "self-reset". No possibility of it "staying on" and melting the circuit. Move the design challenge into the precharge of the cap.

The one question I have what type SCR do I use and how do I hook it up

I do not have much experience with high power SCRs so example would be nice

Thanks

[Curious.George]

Re: Modifing a Battery Spot Welder because of Controller Issues

You could make a design that tolerates reboots mid-weld. Something has to command turn-on (and sustain) the switch, so a mosfet would need a latch. Another example uses 1,200A (10msec) SCR.

I prefer the SCR approach -- just trigger it and let it "self-reset". No possibility of it "staying on" and melting the circuit. Move the design challenge into the precharge of the cap.

[redwire]

Re: Modifing a Battery Spot Welder because of Controller Issues

You could make a design that tolerates reboots mid-weld. Something has to command turn-on (and sustain) the switch, so a mosfet would need a latch. Another example uses 1,200A (10msec) SCR.

This example uses 705,000uF 16V cap bank https://www.instructables.com/id/Cap...r-battery-tabs
Popular now is using a car battery and mosfets.

I like MOT design, one guy measured 500A 16mm^2 wire 2T; not sure if this is good or bad. It' s just cheaper compared to lots of money on capacitors.

[Curious.George]

Re: Modifing a Battery Spot Welder because of Controller Issues

"Dual pulse spot welding: allows the use of two pulses to make a weld. The first pulse is generally used to displace surface oxides (contaminants) and plating, so the second pulse welds the base materials. This feature also reduces spitting."

The first pulse is around 1/8 of the main pulse (time), pause between pulses is variable - long around 500msec.

So, you've considerable time to, e.g., recharge a cap in preparation for the next firing.

1. acquire parameters from user
2. move to non-volatile memory
3. accumulate charge for first pulse
4. dump charge into coil
5. allow processor to reset (e.g., watchdog or EMI)
6. acquire parameters for next pulse from NVRAM (delay, precharge level)
7. lather, rinse, repeat

I.e., there's no need for the MCU to keep running THROUGH the "disturbance"; just plan on (re)starting up AFTER it's over!

[sam_sam_sam]

Re: Modifing a Battery Spot Welder because of Controller Issues

I have a plan

I modified this controller to be able to have the heat sink and the transformer all on one board

https://www.ebay.com/itm/Spot-welder...53.m2749.l2649

One note about this controller it nice device but it has to many steps to program it
The other controller is easier to step up and use so we will see how well this controller works

And waiting for this to come in

https://www.ebay.com/itm/NY-D04-40A-...53.m2749.l2649

I have this part already

https://www.ebay.com/itm/NY-D04-NY-D...53.m2749.l2649

I was just going to take the heat sink off the other controller that in the photo but it would interfere with the power in and power out pins

I can not have this way so I started with another plan

I will be able to fit all the parts in it I will also cut a hole for a fan to keep the heat sink cool if I have room in the case to do this

Yeah I know that a 100 amp unit is over kill but the 30 amp ones do not last

I thought about using the optic sensor parts of this board but it setup is different so I could not do it that way so I remove the optic sensor circuits and use the other optic sensor circuit

I just have three wires to hook for this part and the power supply for the controller board and hook up the foot switch pedal I should be rocking and rolling soon I hope

I have quite a few batteries to weld

[sam_sam_sam]

Re: Modifing a Battery Spot Welder because of Controller Issues

those are good, but any type will do.
i think you only need one at the end with the microcontroller - as close to the board as possible.

another option woiuld be to screen the cable with something.
maybe aluminium or copper tape - and earth it.

Thanks for information

[stj]

Re: Modifing a Battery Spot Welder because of Controller Issues

Are taking about this type

Should I have one on each end of ribbon cable and hot glue them in place

those are good, but any type will do.
i think you only need one at the end with the microcontroller - as close to the board as possible.

another option woiuld be to screen the cable with something.
maybe aluminium or copper tape - and earth it.

[redwire]

Re: Modifing a Battery Spot Welder because of Controller Issues

But what is the timing of them? I.e., how much "recovery time" between each of the first two pulses and then the "main event"?
And, what is the relative energy dumped in each case (I assume the two initial pulses are "small" compared to the followup)?

"Dual pulse spot welding: allows the use of two pulses to make a weld. The first pulse is generally used to displace surface oxides (contaminants) and plating, so the second pulse welds the base materials. This feature also reduces spitting."

The first pulse is around 1/8 of the main pulse (time), pause between pulses is variable - long around 500msec.
You're right though, a really good welder measures current to determine the energy being delivered.

Those flat ferrite ring or clamps, you can use one with a loop (if two ribbon cable thicknesses fit). A round donut works if you can fit the ribbon connector through.

[sam_sam_sam]

Re: Modifing a Battery Spot Welder because of Controller Issues

ferrite ring?

Are taking about this type

Should I have one on each end of ribbon cable and hot glue them in place

[stj]

Re: Modifing a Battery Spot Welder because of Controller Issues

This might be possible but what about the ribbon cable this is a problem by it self

ferrite ring?

[Curious.George]

Re: Modifing a Battery Spot Welder because of Controller Issues

I believe there is an initial double pulse as a pre-heat, followed by the main weld long pulse. Somewhere they found it does better with batteries, fewer outright punch through holes.

But what is the timing of them? I.e., how much "recovery time" between each of the first two pulses and then the "main event"?

And, what is the relative energy dumped in each case (I assume the two initial pulses are "small" compared to the followup)?

[redwire]

Re: Modifing a Battery Spot Welder because of Controller Issues

I believe there is an initial double pulse as a pre-heat, followed by the main weld long pulse. Somewhere they found it does better with batteries, fewer outright punch through holes.

My guess is the controller crashes whenever a burst of EMI from an arc happens, either during pre-heat or the main pulse. It might be from the transformer too. EMI would explain it acting unpredictably.

First thing to try would be to ground the controller MCU/DC side to earth-ground. This can be a short piece of wire, or a cap+resistor if there is a chance the MCU-side has some oddball grounding. The 1MEG resistor bleeds off leakage currents from the power transformer so no charge build up and the 0.02uF cap is a short at RF frequencies.
Using RC grounding lets it float so at DC or 50/60Hz AC frequencies there is no ground loop possible, but at RF frequencies it is grounded. Note one display board trace is very close to shorting to the mounting spacer.
Attach the ground (wire or RC) to the main board's filter cap C1 (-) or Vreg U1 (tab?). The footswitch common might be OK if one end goes to MCU ground. I would ohm it out.

The MCU connects directly to the ribbon cable going to LED display/keypad, and the ribbon cable makes a great antenna.
Just pay attention to lead dress, loose mains wires that go close or touch the ribbon cable, or PC boards- will radiate EMI which gets into the MCU.

I don't know the values of the triac's snubber, C5+R17. R17 should be 10-39 ohms 3W and the cap at least 0.1uF or 0.22uF I think they might have cheaped out on C5. 0.01uF is useless and the triac might even get damaged.

I think a diode fell off the board for this guy: Aliexpress poor quality chinese NY-D04 spot welder. Very sad.

[Curious.George]

Re: Modifing a Battery Spot Welder because of Controller Issues

The controller has some malfunction which means that it does not cycle the pluses more than once or none at all and should do it any where from 1 pluse to 20 pluses because you can not change the setting does not respond to its input micro controller failure

and

But the problem with that version was that the transformer was not big enough to get it hot enough to make a good weld because it only does two pluses instead of a maximum of 20 pluses

and

The normally setting might be 15 pulses at 300 to 500 milliseconds this normally gives you the best results depending on how good the metal that is used on the cell and the tabbing metal is

What is the advantage of many (short?) pulses over a single longer pulse?

And, need those pulses be closely spaced in time (to take advantage of the metal still being "hot" from the earlier pulse(s))?

How does the circuit control the current into the short? And, does it ENSURE that the specified current passes? Or, just LIMIT the current to that value?

It seems a simpler design would be to control the ENERGY delivered to the weld. Use the MCU to control the accumulation of charge on a capacitor and then just shunt that into the transformer/coil (via 4-layer device). Note when the current stops flowing and start the next charging cycle.

Depending on how quickly the pulses must follow each other (if, indeed, that is a requirement), you'd move all of the "smarts" to the charging side of the event -- which is EMI-clean -- and just let the MCU fire the triac and then *CRASH* (i.e., design so the MCU expects to be reset after the trigger event)

[sam_sam_sam]

Re: Modifing a Battery Spot Welder because of Controller Issues

maybe put a tin shield over the board and earth it.

This might be possible but what about the ribbon cable this is a problem by it self

[sam_sam_sam]

Re: Modifing a Battery Spot Welder because of Controller Issues

Probably

Likely the MCU side is not earth-grounded and the EMI from a spot weld causes it to crash.

Why the resistor at 1meg what will that do your only using a 0.02uf capacitor what voltage spikes is going to hold down with that high of resistance / capacitor smooth the spikes I understand this and I understand that if you have the right resistance it will discharge the capacitor over time

I would add a 0.02uF cap parallel with 1MEG, from DC(-) (at the filter cap or 5V reg) to the enclosure chassis/earth ground so the MCU does not experience high level EMI.

This can be done up to a point what can I really do for this not be a problem

The display/keypad board may also pick that up. The ribbon cable is also vulnerable. Mains wiring needs to be away from the MCU and keypad and ribbon.

Please explain further what parts would I use and how would I hook it up I understand what it suppose do but I have never had to add one

One guy got his board to work better by beefing up the snubber and adding a cap for noise filtering.

What part of the circuit do you need to see I can do a very close close up I just need to what part

Have to see the board's values for those parts. Is your 9-12VAC power transformer decent, not sagging.

The pins are not tied to the metal part of tirac

The big 100A triac BTA100-800B I think has a live tab so you need an insulated heatsink or insulated mounting kit.

It does not if you hit it once in a while but if you do a big battery pack it will get quite warm
if you are pulsing the tirac many times in a short period of time to get a better weld

The normally setting might be 15 pulses at 300 to 500 milliseconds this normally gives you the best results depending on how good the metal that is used on the cell and the tabbing metal is

But it apparently doesn't heat up much, as it is only on for a few 100 msec and many seconds rest.

[sam_sam_sam]

Re: Modifing a Battery Spot Welder because of Controller Issues

I was able to cut circuit board where the optic sensor is and separate them it will be easier to mount the tirac to a heat sink that I have

on the outside of the enclosure the enclosure has a cover that covers the tirac from another setup that used the same tirac

Yeah I know stealing parts from one setup and use on another setup really but other setup would have had me use a separate timer which I did not want to do

The optic sensor is now mount on a small double sided circuit board that I have to make small circuits for testing

I just have to use some hook up wire to make all the circuits come together and test it out and see how well it will work