Programming Arduino Pro Mini, TIG Pulse Add-On, missing some code

[Falcon69]

Nominal,

The wires going to the optoisolator for the switch.  I'm not sure of the circuitry inside the welder as to what exactly that switch is doing. So Wouldn't it be a better idea to just have a separate ground for that?  Not tied at all together with the welder pedal inputs. I understand why you have it on an optoisolator though and then connected to the Arduino. It's to shutdown and turn on the Arduino to save battery power.

[rstofer]

So, we would like the pedal and all of the adjustment pots to be isolated from the welder and not even share a common ground.  Total isolation...

Part I: Use a relay as the switch input to the welder and trigger it with the pedal.  Get an accurate measure of switch current and then decide on a relay.  Something with a 5V coil might work well.

Part II:  Isolate the current inputs to the welder...  Here's a completely off-the-wall approach that I really like!  Have the Arduino drive a stepping motor and connect a pot to the shaft.  Total isolation!  In the end, the pot can handle any voltage from any welder or be replaced with one that can.  Universal!

There is a wee bit of an issue and that's the fact that we don't know where 0 is on the pot so, somehow, we need to have some way of backing down from any position to absoute zero when the Arduino starts up.

As part of the coupling between the motor shaft and the pot, add a metal disk with a small hole.  Use a sensor like:
https://www.jameco.com/z/HOA1882-011-Honeywell-Infrared-Transmissive-Sensor_1861380.html

The hole needs to be fairly small because the startup process will back the pot down until it sees the hole and perhaps step it forward until it doesn't see the hole.  That will be the minimum current.  The code will guarantee that the motor is never reversed beyond this point by keeping track of steps.  Obviously, we need to know how many steps until we get to the other end and, usually 1.2 degrees per step.  A 270 degree pot would take 225 steps and I might knock a few off just to prevent mechanical damage.

Maybe the stepper noise will be a concern...  It wouldn't need to be a very big stepper...

[Falcon69]

The problem with the stepper, would require busting open the pedals to connect to the pot inside?

Nominal's recent schematic seems like it will work. Just not connect the switch and pedal grounds together. It's probably fine, but I don't know where those wires are going inside the machine.

I noticed that zero thing you were talking about when I was testing circuit on breadboard. I noticed if i backed off just alittle from the end rotation of the pot, the pulsing seemed fine, but the second i turned it ALL the way to the end, the LED was erratic in behavior. It was acting strange.

Couldn't there be programming that would account for this? Like the Arduino would know to only accept 3%-97% of the output of the potentiometers or something? That way if it goes past, then it doesn't matter. The Arduino wouldn't read the input, or rather kept the input the same as it were at 3% or 97%. Like the Arduino would be programmed to only output or read the voltage from the potentiometer if it was only between .005v to 3.295v. anything less or more than that, arduino would only put out 0.005v or the 3.295v. I'm still new to the Arduino and it's programming, but I remember seeing some coding that is 'IF'. 'I'F voltage < 0.0005 then output 0.0005. 'IF' voltage > 3.295V, then output 3.295V.  Something like that.
 

[rstofer]

The problem with the stepper, would require busting open the pedals to connect to the pot inside?

In my imagination, there would be a jack on the electronics box where the unmodified pedal would plug in and there would be a cord with plug to connect the electronics to the welder.  Or the cord could have connectors as well - just get the sex right.  The input from the pedal would probably be female (like the welder) and the output to the cord would be male.  In effect, the cord is really an extension cord and could be used that way.

I'm not seriously suggesting this approach, I just wanted to come up with a scheme where the pedal voltage and current could be anything and the electronics would just deal with it while not smoking the Arduino.  I could pick ANY potentiometer, for any welder, match the power and resistance ratings and be done with it.  Nothing in the electronics would change at all.

Nominal's recent schematic seems like it will work. Just not connect the switch and pedal grounds together. It's probably fine, but I don't know where those wires are going inside the machine.

To do this properly, you need to know everything.  Further, you have to rationalize what you know against reality.  I knew from the start there was no way 300 mA was going through that 10k pot.  Now, there could have been something else in the pedal but from Ohm's Law, I knew that that much current wasn't going through a 10k pot.  I am often reminded:  Ohm's Law is a LAW, not a suggestion.  You can't treat it like a speed limit.

I noticed that zero thing you were talking about when I was testing circuit on breadboard. I noticed if i backed off just alittle from the end rotation of the pot, the pulsing seemed fine, but the second i turned it ALL the way to the end, the LED was erratic in behavior. It was acting strange.

Couldn't there be programming that would account for this? Like the Arduino would know to only accept 3%-97% of the output of the potentiometers or something? That way if it goes past, then it doesn't matter. The Arduino wouldn't read the input, or rather kept the input the same as it were at 3% or 97%. Like the Arduino would be programmed to only output or read the voltage from the potentiometer if it was only between .005v to 3.295v. anything less or more than that, arduino would only put out 0.005v or the 3.295v. I'm still new to the Arduino and it's programming, but I remember seeing some coding that is 'IF'. 'I'F voltage < 0.0005 then output 0.0005. 'IF' voltage > 3.295V, then output 3.295V.  Something like that.

Yes, code is intended to fix up anomalies.  We already know there is going to be some kind of math involved with converting the ADC reading to something we can use to create an output.  If we really look at that, there might be something where the value goes astray.

[Nominal Animal]

Not tied at all together with the welder pedal inputs.

Sure, makes sense.

I understand why you have it on an optoisolator though and then connected to the Arduino. It's to shutdown and turn on the Arduino to save battery power.

Oh no, it doesn't do that.

The pedal has two separate functions: one is the potentiometer (voltage divider), and one is a switch (microswitch).  If we want to use a microcontroller to pretend to be a pedal, we need to provide both.

If we do use a DAC (MCP4911's seem to be more easily available), an TLP240A as the optoisolated switch, and two TLP2168(TP,F) optoisolator pairs to isolate the four signals to the DAC (just to make sure the DAC can be properly controlled), and some resistors and capacitors, the control circuit would look like this, I think:

R10 and R11 are purely optional, just there to stop the I/O pins from burning out if the pins were to be programmed as outputs. Lower right side goes to the welder. You connect the pedal to upper left. Four control pots are drawn in the circuit. I added a tactile button (SW1) just in case; could add a couple more using RX/TX pins (with current-limiting resistors in case they are set as outputs). Pro Micro has selectable internal pullups.

[Falcon69]

In my imagination, there would be a jack on the electronics box where the unmodified pedal would plug in and there would be a cord with plug to connect the electronics to the welder.  Or the cord could have connectors as well - just get the sex right.  The input from the pedal would probably be female (like the welder) and the output to the cord would be male.  In effect, the cord is really an extension cord and could be used that way.

   Yes, that's how I planned to do it. So the pedal is unmodified, and the cable for it is not cut.

I'm not seriously suggesting this approach, I just wanted to come up with a scheme where the pedal voltage and current could be anything and the electronics would just deal with it while not smoking the Arduino.  I could pick ANY potentiometer, for any welder, match the power and resistance ratings and be done with it.  Nothing in the electronics would change at all.

Yes, that is the idea. I can understand this.

To do this properly, you need to know everything.  Further, you have to rationalize what you know against reality.  I knew from the start there was no way 300 mA was going through that 10k pot.  Now, there could have been something else in the pedal but from Ohm's Law, I knew that that much current wasn't going through a 10k pot.  I am often reminded:  Ohm's Law is a LAW, not a suggestion.  You can't treat it like a speed limit.

Yes, I was off on my original measurement, you guys were right.  I should have done the calculation for ohm's law, I just was not thinking.

Nominal,

So SW1 would be used to turn the Circuit on/off, since it has it's own power source, the battery bank?

I just looked, the TLP2168 is not available from any of the big suppliers (Arrow, Mouser, Digikey). It is factory pre-order, and therefore may require a HUGE quantity to buy. There are a few on eBay, but I don't like buying IC Chips off eBay.

EDIT: Looks like FindChips has located some from Verical, chip1stop, and Arrow, but it is unsure if that is in America or not.

[Falcon69]

Nominal,

I looked at data sheet, and am curious, though, it's probably not that big of a deal since the values can be adjust (or multiplier) in the code to make it work.  But I have to ask...

The Vf of that tlp2168 for the LED inside is 1.4-1.8volts.  Does that mean that the LED inside will not turn on until 1.4 volts?  If that's the case, how is the machine going to see the back ground current then? If that isn't on during pulsing, the molten metal pool will cool too quickly during pulses.

[Falcon69]

oh, wait...

You mentioned earlier that the DAC is 16bit, and if i understand correctly, that works like a highly advanced shift register. Meaning it reads all the bits (zeros and ones) before it executes the string of bits.  Meaning, it can receive the flashes of light from the LED in the optocoupler (on/off state), and then it calculates the voltage from that and adjusts the output.

Did I understand that right?  I have never worked with DAC's before.

[Nominal Animal]

So SW1 would be used to turn the Circuit on/off, since it has it's own power source, the battery bank?

No, just in case a button is needed for the UI. It can be used to put the microcontroller to deep sleep (and wake up from it), but it'd be better to just disconnect the power bank when not in use.

For a real on/off switch, I'd just take an USB micro B cable, cut it in half, and put a 2xon-off toggle switch in between. (Pin 1 is VCC, and you want to disconnect that when off. Only pin 1 and pin 5, GND, need to be connected for a pure charging cable.)

I just looked, the TLP2168 is not available from any of the big suppliers (Arrow, Mouser, Digikey).

There are other drop-in equivalents, like TLP2161, TLP2662, and so on.

You mentioned earlier that the DAC is 16bit, and if i understand correctly, that works like a highly advanced shift register.

10-bit, with 16-bit commands, yes.  Think of it as a serial-to-parallel latch, followed by a resistor ladder.

If we look at the DAC end: Normally, CS and LDAC are high. To send data, CS is pulled low. Then, on each rising SCK edge, the SDI line determines each bit. If we send 0011xxxxxxxxxx00, and finally pull LDAC low and CS high, the output of the DAC will be xxxxxxxxxx₂ / 1024 of the maximum voltage within 5 microseconds, and stay there. (0000000000₂ = 0, 1000000000₂ = 512, 1111111111₂ = 1023, and so on.)

(If LDAC is tied to the ground, then rising edge of CS will cause the data to latch. That probably makes sense, and would free pin 7 on the Pro Micro.)

The signal inversion due to optoisolators do mean that we must use SPI mode 1,1 instead of mode 0,0; but fortunately that is not a problem on the Pro Micro/ATmega32u4.

Pin 10 is internally connected to the LED, and can be connected to an external LED (with a suitable current-limiting resistor), and PWM'd to control its intensity.

[rstofer]

The MCP4911 DAC will operate at Vdd=2.7V..5.0V which is nice since it includes both 3.3V and 5.0V logic and the isolation in the schematic should work fine.

However, should there be a welder whose pedal voltage is outside this range, the DAC is not going to work/survive.  I don't know if there are welders using, say, 12V for pedal voltage.

[Falcon69]

there is this one, excepts up to 36v but only puts out max 11v.  So, probably won't work.

http://www.ti.com/lit/ds/symlink/dac8760.pdf

I know some machines are 12 volt and another that operate on 42 volts. I'm not sure which machines those are.  I just remember reading about it when I was looking to buy compatible foot pedals for friends machine.  So, I'm not sure yet, what the voltage is coming out of my friends' machines.

It looks like however, that all of them that drop down low enough (3.3v min), for supply voltage, have max input of 5.5v.  For my machine, that would work just fine, but for my friends' machines, if their's is more then the 5v, maybe have to design the circuit board to except either of the DAC's.  Maybe can find one with same foot print and same pin out, would make it easy.  Still, it doesn't seem like it can be universal.

[Nominal Animal]

The MCP4911 DAC will operate at Vdd=2.7V..5.0V which is nice since it includes both 3.3V and 5.0V logic and the isolation in the schematic should work fine.

Quite true, but since we don't have any specs for the pedal voltage, I do like the PWM option better. Plus it should be simpler to do; fewer things that could go wrong. I finessed it a bit:

The idea is that the two switches and five potentiometers form a 3×7 pin block, with the VCC pins omitted from the two switches. This allows different configuration in software, with up to seven switches, or up to five potentiometers, as long as their total number is at most seven. In my opinion, three buttons for the OLED UI and four pots sounds best, but what do I know? Not much.

I'm thinking the capacitors and connections to the welder would be on a separate board, connected with 5 wires to the Pro Micro carrier board (which would not need to be much larger than the Pro Micro; just enough additional room for the connectors). It looks like 450 V tolerant electrolytics should be easy to find; hopefully that'd be enough.  If not, they'll go bang, but should not harm either the Pro Micro or the welder (since they'll just go either short or open).

For a power switch, I'd just cannibalize an USB cable, and splice in an 1xON-OFF or 2xON-OFF switch (see e.g. the Wikipedia article on USB connectors for details). Since the current needed is well under 500mA, you can use any old cable, even the el-cheapo thin ones.

rstofer, do you see any issues with this one?

Because I do have a Pro Micro at hand, at least one 10k potentiometer, and a collection of random capacitors, I think I could write an example firmware during the holidays. Make it CC0-1.0, so anyone could build on top of it, too.  (Because the inputs do not have current-limiting resistors, one does need to be careful which pins are set as outputs, but that's about it.  Having one of the input pins configured as output and high yields a short, which usually burns that pin out.  Then again, the Pro Micro clones seem to work just as well as the originals (except they use the Arduino Leonardo bootloader), and cost $5 or so, so it's not that big of a loss, as long as one is aware of the issue.)

[Falcon69]

Thank you Nominal,

Okay, those DAC's are readily available. Mouser shows over 1800 in stock, bit pricey at over $3 each, but I don't care of cost too much, if this works. It's still alot less then spending $4000 on a new machine.

So, I'm not really understanding what R5 and the switches do.  I guess the switch is to select different code saved into the Arduino, in case the DAC needs to be switched out for a 12v version for other welders? R4 is still the potentiometer inside the pedal? Or does that correspond now to the Pedal_A, B, and C inputs now, leaving two potentiometers on the board for something else? R1 and R2 are still for Background Time and Main Time.

I'm guessing the resistor still needs to be on Pedal_B P3 and on the switches as well, like in previous schematic. (see, I'm learning)

So the capacitor bank for machine input, those are electrolytic, or all 1uF and below ceramic?

What If when I design the board, It's all one board, but I separate them with a slit routed in it, to separate the welder inputs from rest of circuit, and have the TLP5701 straddle over the slit? I think I've seen that done before on old Television Boards to isolate power.

[Falcon69]

Nominal,

I was planning on using a TP4056 tied to a 18650 3.7v battery, with a step-up converter to get to 5v for the arduino (or just use the tp4056, battery, and step down module to 3.3v arduino), then just have the USB connector sticking out of the project box for recharging the battery. That battery should last months on a single charge on this simple circuit, i would think.

[Nominal Animal]

So, I'm not really understanding what R5 and the switches do.

You do not *need* to use them; they're just there to make sure that if you do need them, they're available.

The first firmware I'd write would use R1, R2, and R3 potentiometers (and the pedal).  R1 would control the pulse rate (between say 0.2 and 2 seconds), R2 the duty cycle (between say 10% and 90%), and R3 the off-time current (as a percentage of maximum).

However, if you find that you want to add those other modes, or add a three-button menu interface (up, down, and select), you need those three buttons.  And you still have one potentiometer free. You might want to use it for decay rate (so that when you release the pedal, the welder always sees a smooth easing at that rate).

It is horrible to find out you created a board that you cannot use, and must redesign.  I like to have a bit of future-proofing.

R4 is still the potentiometer inside the pedal? Or does that correspond now to the Pedal_A, B, and C inputs now, leaving two potentiometers on the board for something else?

The pedal potentiometer is connected to Pedal_A, Pedal_B, and Pedal_C, and the pedal switch to Pedal_D and Pedal_E.  There are up to five control potentiometers (knobs) you can connect to the Pro Micro in addition to the pedal.

I'm guessing the resistor still needs to be on Pedal_B P3 and on the switches as well, like in previous schematic.

The Pro Micro does have internal programmable pullups, so they're not really needed. You can use standard pullup wiring (switch between ground and a 10k resistor to VCC, and input connected to the point the switch connects to the resistor), which protects against programming bugs. Debouncing capacitors can be used in series with the switches (because at microsecond scales, the switch does not change states cleanly, but bounces a few times between open and closed), but that too can be done in software.

So the capacitor bank for machine input, those are electrolytic, or all 1uF and below ceramic?

I am not certain; I'm not even certain on how much capacitance is needed to get comfortable results.  Remember, the existing pedals use capacitors with very high voltage ratings.  I do not know if that is for compatibility, or actually necessary with all machines.  I would personally do a compromise, and find capacitors rated at say 450 V.

As an RC filter, the break frequency when using the capacitors shown, would be around 15 Hz.  At 50% duty cycle, the peak-to-peak ripple would be around 0.1%.  Is that enough? Is that overkill?  I do not have enough practical experience to know; as I said, I mostly use digital sensors myself.

What If when I design the board, It's all one board, but I separate them with a slit routed in it

That helps with isolation, sure.  I was thinking of using a separate board more for ease of putting into an enclosure (you could put the capacitors on both sides of the board.  I like Hammond diecast enclosures, and you could put a conductive separator (plate, or just circuit board material) to help with the EM noise.  I don't know if that matters, but that's what I'd do.

I was planning on using a TP4056 tied to a 18650 3.7v battery, with a step-up converter to get to 5v for the arduino (or just use the tp4056, battery, and step down module to 3.3v arduino), then just have the USB connector sticking out of the project box for recharging the battery. That battery should last months on a single charge on this simple circuit, i would think.

OLED and Pro Micro do consume quite a bit of power. Haven't measured it, but if it was say 150 mA on average, I would not be surprised. A 2000 mAh USB power bank (a single 18650 cell, from a reputable local seller) costs about 5€ here, so would provide about a day of continuous running. I think it would be easier if you had a short USB cable with a male connector sticking out from the enclosure, and a couple of those cheapo power banks, and simply switch them when they run out; keeping the other in a charger.

That way you could have a power switch on the enclosure, too between the two ends of the USB cable; one end sticking out, the other end inside connected to the Pro Micro. Ferrite bead or a choke where it exits the enclosure. KISS.

[Falcon69]

ya, good idea on the Battery Bank and switching them out on the fly.

The capacitors in the pedal with high voltage rating, i'm guessing that is to eat up the voltage spikes that could happen with the plugging and unplugging of the cable maybe?

Ya, I'm down for that, adding extra switches/potentiometers for later features, if they ever get implemented.

Okay, so the internal pull-ups and stuff you are talking about is what you meant by being careful on how you program and pin selection in previous posts.  That makes sense.  Otherwise, we would have to put in resistors so we don't fry the pin outputs/inputs.

ya, I'm well familiar with the debounce thing. I had that problem with another circuit I was working on for Hall Effects Sensors. Is a pain to experience that, and then have to redesign and make the boards for it.

Looking at alot of boards just now on google images that have optoisolators, it looks like the majority of them just stop the copper pour underneath the optoisolators, in a line that goes across the whole board.  So what I was seeing before on old TV Boards, must have been for something else? Because they actually had a slit routed across the board.

[Nominal Animal]

The capacitors in the pedal with high voltage rating, i'm guessing that is to eat up the voltage spikes that could happen with the plugging and unplugging of the cable maybe?

Anything that sparks tends to generate all sorts of EM, that could get coupled to the long wires.

Okay, so the internal pull-ups and stuff you are talking about is what you meant by being careful on how you program and pin selection in previous posts.  That makes sense.  Otherwise, we would have to put in resistors so we don't fry the pin outputs/inputs.

Yes, exactly.  If the firmware accidentally makes one of those pot or switch pins outputs, and sets them high, the next time the user presses the button or turns the pot all the way, will cause a short that is likely to burn out the pin.  It is a common thing to happen when playing with microcontrollers in the Arduino environment.

If one wants to, one can solder the pull-up resistors to the push buttons; no need to have them on the board at all. 1k or above will work fine; I like 10k. (At 5V, each 10k pull-up consumes 0.5 mA. I think the internal ones are equivalent to 33k or so.)
I really like the idea of having each button or potentiometer connected to a 3-pin header; keeps it neat but modular.

I'm well familiar with the debounce thing.

In software, I like the approach where you act on the change in state as soon as it is detected, but ignore further changes for a "debounce duration", say 20ms (0.020 seconds) or so. 

In the welder case, the most common software debounce (like implemented in several Arduino libraries) that only accepts a change in state after the new state has been observed for long enough, might work better: it ignores spurious spikes (say caused by radiated EM).

Because they actually had a slit routed across the board.

It is for isolation. If you consider a high voltage spike, a routed slot is better than a bare board, and a plastic or mica baffle inserted into the slot is even better than an empty slot.

I really like how Big Clive describes these things. In this video, he tears down an UK USB power supply that has a baffle inside a routed slot for better isolation (it's quite a dense little thing), and later on even does a high-voltage test on it.

With respect to the welder and pedals, I do believe the spikes, if any, would be caused by radiated EM due to sparking. Don't forget: the very first radio waves were generated with spark gaps, and relatively long wires close by will act as antennas.

[rstofer]

It's a TIG welder, we know there will be high frequency and high voltage generated to start the arc.  We can also suspect that the torch cable will be laying somewhere near the pedal cable.

[Falcon69]

Sorry all, I been sick the last couple days. I did receive those 3.3v Arduino Mini's.  I'll test those when I feel better.

rstofer. 

Yes, most TIG machines have high frequency start, however, mine is a DC machine. All in one, does TIG, Stick, and MIG.  It does not do HF (I wish it did though), it does Lift Arc. I wish it was HF, cause then I could TIG Aluminum. As it is, I have to buy a spool gun to be able to MIG aluminum.  The machine is a Lincoln MP 210

[Falcon69]

I'm still ill.  Sorry no updates.

As soon as I feel better (think I have the flu), I'll be hooking up this 3.3v Arduino and seeing if I can get the machine to fire this time. I think the 5volt arduino just didn't see enough voltage from the machine to fire it up.

Hopefully I'll be feeling better next few days and can get this done.

[Pm31416]

Keep us update, i would like to build this for my Everlast St140. Did you follow the initial wiring diagram from the Youtube video ?

[Falcon69]

yes, I will try. I've been pretty ill the last week and a half. I'll keep ya posted though when I can.

[Pm31416]

Did you finally tried it ? Wonder if it's working well !

[Falcon69]

Sorry, I've been ill, and with the holidays and everything I haven't had time to try it yet.  I did get the 3v arduino's in, so I can try when I have time.  I'll update this once I get it figured out.

[Mackn918]

I have been trying to bud this for a year. Did you get it working? Can you help others now?