Manipulating welder output with Arduino

[sportbike]

According to this link: http://arduino-info.wikispaces.com/A...-PWM-Frequency
It appears the PWM output may only be 1khz.

I happen to be working on an IR temperature sensor that outputs a PWM duty cycle variant signal at 1kHz and am looking at analog conversion options.


I found this article with links to a schematic for using a 74HC4053 chip to pretty much eliminate ripple. The output update rate seems to match the input PWM update rate. Would be plenty for a pulse feature. I haven't really looked into it too much but I do have interest in trying to get the PWM output of the sensor I am working with into a usable analog output.

http://www.edn.com/design/analog/432...most-no-ripple

The circuit in the link in the quote should give you a pretty stable output. I bought the chips and then realized it needs a balanced clocking input on one of the channels rather than just the varying duty cycle PWM signal. My sensor application only has the varying output signal and no clock signal running at the same frequency.

[joshuab]

Well, I'm sorry to report that it won't work after all. The problem is that the internal resistance of the switch is about 180 Ohms (measured between the pins with the switch in the "open" state), and that seems to be enough resistance that the welder never gets to max output.

Scratch that. I don't think the measly 180 Ohms open-switch resistance is what's doing it. I think the real problem is that if you're using an SPST switch, you can build a rheostat, but not a potentiometer, and the circuit expects a potentiometer. Simply varying the resistance between any two pins doesn't produce the correct output values. Conceptually, what I think you need is a SPDT switch that throws the circuit back and forth between the "pedal up" and "pedal down" states in response to the PWM signal. Or two SPST switches with one hooked up to a NOT of the other's input. But for now, I'm going to just go back to feeding a straight PWM signal to the pin and get on with my day.

[joshuab]

I have been soldering like mad. Although the hardware design was not 100% complete, I got really tired of stuff falling out all over the place when I was carrying a breadboard around, so I started wiring it up to perf-board. I also ordered a Boarduino so that I could mount the Arduino on headers on the board, instead of having it be a separate component with jumper-wires connecting it to the main-board.

Once I get all the stuff wired up, there will be a round of debugging where I shake out any mistakes I made in the wiring, followed by another round of software revision, shaking out any lingering bugs and refining any small features that I have glossed over in my desire to get to a working product. I am going to a welding workshop this weekend, and it would be really exciting to have this ready to demonstrate, but I'm not holding my breath. The Boarduino comes tomorrow, and there is every chance that I will find a show-stopping problem that I have totally missed. Even if things go smoothly, it's a lot of work to get done, but it's not out of the question.

Here are some pics of the project in its current state.









And a video:

[Rambozo]

Looks like you are getting there.
Did you ever think about using an ardunio perf-shield for a nice compact plug on solution for your wiring points and multiplexer? Something like this.



Since you are back to the PWM you could also add the voltage regulator so that the welder will never get a 5 volt signal even if something happens to stop the PWM. Just dial in the max signal you want then use PWM to lower that as required. I don't see any decoupling caps for those chips, either, something to consider to help prevent noise issues.

[joshuab]

Looks like you are getting there.
Did you ever think about using an ardunio perf-shield for a nice compact plug on solution for your wiring points and multiplexer? Something like this.

I did, but the protoshields that I'm aware of don't have enough space for the components I'm using. For example, the one I have only has space for a total of 10 rows of DIP IC.



The Boarduino sort of turns that on its head: instead of putting the perf-board on the Arduino, it puts the Arduino on the perf-board.

Since you are back to the PWM you could also add the voltage regulator so that the welder will never get a 5 volt signal even if something happens to stop the PWM. Just dial in the max signal you want then use PWM to lower that as required. I don't see any decoupling caps for those chips, either, something to consider to help prevent noise issues.

Great minds think alike. I built a voltage divider to pull the Arduino's 5v down to a max of 2.7v. This makes sure that, no matter what software mistake might happen, the welder will never see over 2.7v from the Arduino. And it simplifies the code, since I can just use the full range of PWM output, instead of tracking max volts and taking output as a percentage of that. Technically, it also about doubles my usable resolution, but that hardly matters.

For the decoupling caps, can you provide a basic schematic? Currently, I've got the chips solder-bridged over to a ground/power bus that runs down the middle of the pins. I'm familiar with the concept of decoupling caps, but I don't really know exactly how they fit in a schematic.

EDIT: Actually, come to think of it, I have seen caps in a schematic, on either side of a component, going from Vcc to Ground. Is that the ones you're talking about?

[joshuab]

Okay, now I have done a lot of reading about bypass caps, and I think I have it figured out, at least a little. By good fortune, I ordered a handful of 0.1 uF caps a while back, planned for a different project that fell by the wayside. I've stuck one of them on the Vcc pins of my two IC's, bridging between pin and ground.

Learning something new every day!

[Rambozo]

Okay, now I have done a lot of reading about bypass caps, and I think I have it figured out, at least a little. By good fortune, I ordered a handful of 0.1 uF caps a while back, planned for a different project that fell by the wayside. I've stuck one of them on the Vcc pins of my two IC's, bridging between pin and ground.

Learning something new every day!

You got it. For best results they are installed as close to the chip as possible. The value is determined by the frequency of the noise you are trying to block. Often times to get a wider filter band on logic chips you will run a 0.1µF and 1µF in parallel or a 0.01µF and 0.1µF. It is always good practice to use decoupling caps on circuits with fast transitions or pulses. The data sheets will often include recommended values based on the internal characteristics of the chip.
Here is a good article to get you started.

http://www.edn.com/electronics-blogs...s-best/4328787

[sportbike]

In my (limted) experience with digital signals, it seems 0.1µF is pretty universal to at least get you in the ballpark for this. I do things on my motorcycle for data collection and sometimes use them to eliminate noise, etc.

[joshuab]

You got it. For best results they are installed as close to the chip as possible.

This was no problem. The pc-board I'm working with has interleaved power and ground rails, and on either side is a pad of three holes for the pins of the IC that straddles the rails. So each IC pin, when wired up, has one hole free for an additional connection. This puts the capacitor pretty much as close to the pin as it can get, without actually soldering it to the pin itself.

Now there's an image! Put the capacitor on the pin like a backpack, straddling the Vcc and ground pins.

[Rambozo]

Now there's an image! Put the capacitor on the pin like a backpack, straddling the Vcc and ground pins.

I've actually done that more than a few times. Usually because I forgot to include it in the board layout (Doh!) or I needed to add another in parallel to capture a different frequency of noise.

[joshuab]

I've got some more pictures for y'all. The Boarduino came tonight and I spent about an hour and a half soldering it all up. If you saw how many components were involved, you'd laugh. I'm a slow solderer. The other thing that came in the mail was the parts to pass the power connector through to the inside of the box. There is a panel-mount 2.1 mm barrel jack on the box that passes through to a plug on the inside. This, combined with mounting the Boarduino on headers instead of soldering directly to its PCB, allows me to remove the Boarduino if I want to, for any reason. It's a general-purpose component, so who knows? Maybe I'll want to pull it out and use it for something else. Why limit myself?

The main thing left, on the hardware front, anyway, is to finish wiring up the connections to the Arduino. There are about 20 pins to hook up. It's the home stretch, but I'm going to put it off for tomorrow, when I'm fresh.



What the board might look like when mounted in the box.



Close-up of the mainboard. This, with the exception of about twenty trace-wires, is complete.



With the Boarduino removed.



My awesome cable. I needed a seven-connector, small-gauge, stranded cable. I spent about a week trying to figure out where I could purchase such a thing locally, before I realized that an Ethernet patch cable would do fine.

[Jason]

Nice..
Did you cut the power supply out of the arduino or was it already seperate?

[everlastsupport]

The decoupling cap will normally smooth a certain frequency, like 60Hz. So you might want to drop a scope on there and see what you have. I know I had a PIC problem many years about on a street bike and .1uF did solve it, but it was a shot in the dark. PICs are a little more picky (no pun) than Atmel.

On the rheostat vs potentiometer, rheo will handle more power both do the same. reho will be wire wound and a little larger, but last longer and is tougher.

[joshuab]

Nice..
Did you cut the power supply out of the arduino or was it already seperate?

Not sure if I understand your question. The Boarduino comes as a kit, with PCB and all parts ready to solder. It includes a basic power supply, with 2.1mm barrel jack, voltage regulator, electrolytic capacitor, and so forth. It's basically totally self-contained, including the microprocessor, power supply, input/output pins, and FTDI interface for programming. All I did was install an additional panel-mount 2.1mm barrel jack on the wall of the enclosure, and wire it to a 2.1mm plug inside the box, so that when I plug a power supply into the side of the box, it is passed through to the Arduino's jack. This approach is very versatile, since the Arduino can handle up to 20v input (although at above 12v, heat dissipation will become an issue). I can plug the Arduino into a wall, or I can plug it up to a battery pack. An option for future development would be to put a rechargeable battery inside the box and wire it up with a charge controller so when it's plugged in, it charges the battery, and when it's not plugged in, it runs off the battery.

[joshuab]

... as expected, I ran into some hardware-related issues. I have passed pins 6 and 7 of the welder's connection into the box so that I can selectively indicate to the welder whether the box is attached or not. This way, I can power down the box if I want to, and return it to panel control, without having to disconnect the plug. These leads need to be Normally-Open, and then be Closed whenever the box is powered up. I tested an analog switch IC that I have, and with no power, the pins showed open on my multimeter. With power, they showed about 190 Ohms (which I believe to be the internal resistance of the IC, since that's present between all of its four switch-pin-pairs when they are closed). So I figured they would work perfectly, but they didn't. For some reason, even when the board is powered down, the welder detects that the device is plugged in. I am going to troubleshoot some more, just to make sure I haven't accidentally shorted a wire somewhere (I don't think so--multimeter says the pins are open), but in the mean time, I have ordered some Normally-Open optically-isolated solid-state relays. This will provide total isolation between the welder's electrical circuit and the Arduino's, and should be as close to just flipping a manually-operated switch as it comes.

[Rambozo]

A grounded metal case might be better for keeping the HF noise at bay.

[sportbike]

... as expected, I ran into some hardware-related issues. I have passed pins 6 and 7 of the welder's connection into the box so that I can selectively indicate to the welder whether the box is attached or not. This way, I can power down the box if I want to, and return it to panel control, without having to disconnect the plug. These leads need to be Normally-Open, and then be Closed whenever the box is powered up. I tested an analog switch IC that I have, and with no power, the pins showed open on my multimeter. With power, they showed about 190 Ohms (which I believe to be the internal resistance of the IC, since that's present between all of its four switch-pin-pairs when they are closed). So I figured they would work perfectly, but they didn't. For some reason, even when the board is powered down, the welder detects that the device is plugged in. I am going to troubleshoot some more, just to make sure I haven't accidentally shorted a wire somewhere (I don't think so--multimeter says the pins are open), but in the mean time, I have ordered some Normally-Open optically-isolated solid-state relays. This will provide total isolation between the welder's electrical circuit and the Arduino's, and should be as close to just flipping a manually-operated switch as it comes.

Do you have a manual toggle for turning the box on? Could just make it a double pole switch and connect one pole to the power for the box, and the other to make the connection to let the welder know it is active.

[joshuab]

Do you have a manual toggle for turning the box on? Could just make it a double pole switch and connect one pole to the power for the box, and the other to make the connection to let the welder know it is active.

That solution is head-smackingly obvious, and yet I totally missed it. Never would have thought of handling the situation that way. Good suggestion. I'm still going to try solving it with the equipment I have on hand (vs. buying another power switch), but it'll be my fallback. The current suggestion that I'm going to test is whether a pulldown resistor will make the switch work properly. I didn't realize that analog switch ICs were leaky (vs. a relay, for example), and that may be the cause of my issue.

A grounded metal case might be better for keeping the HF noise at bay.

Absolutely. For the time being, I'm using the plastic case, and once everything is worked out with lift-start, I'm going to try to solve the issue of protecting against HF. One would expect that a grounded metal case would form a Faraday cage and protect against HF. But then again, the box is going to be connected to the welder by the cable, and it's hard to believe there wouldn't be some HF leakage down the cable. And where are you going to get a ground source from either? It'd be silly to have to drive a rod in the ground when you wanted to weld, but I've also read that grounding HF down household electrical wiring isn't a good idea either. Anyway, the HF question is big enough that I'm putting it off entirely for now.

[sportbike]

That solution is head-smackingly obvious, and yet I totally missed it. Never would have thought of handling the situation that way. Good suggestion. I'm still going to try solving it with the equipment I have on hand (vs. buying another power switch), but it'll be my fallback. The current suggestion that I'm going to test is whether a pulldown resistor will make the switch work properly. I didn't realize that analog switch ICs were leaky (vs. a relay, for example), and that may be the cause of my issue.

Yes, semi conductors often do have some leakage across them. I've run into issues when using LED's on the outputs of ECU's for motorcycles in that they maintain a dim glow unless some resistance is added to pull the "open" circuit to actual "open"

[Rambozo]

A lot of digital logic chips are tri-state designs, so good practice is to use appropriate pull-up or pull-down resistors to control the floated state. Also check the data sheets on your devices as many require unused inputs and outputs to be pulled up or down to prevent unwanted operation. Sometimes what happens will be clear from the block diagram, but other times you can get some pretty unpredictable issues due to the way the internal circuitry shares junctions. When in doubt follow the datasheet. Of course there are the rare occasions where you can use the high impedance output to your advantage such as Charlieplexing.