Go Down

Topic: Sensing induction welding machine on/off (Read 769 times) previous topic - next topic

cloxart

Hello everyone,
I was asked to build a device to switch on a solenoid valve for a certain amount of time when an induction welding machine is appyling power to the heating coil. this is to dispense an inert gas on the part being welded to avoid oxidation. that would be a very easy task if i could get some signal from the welding machine itself, but i am not allowed to connect anything there, so I have two solution: the first is a temperature sensor, switching on the valve when the part start heating up, but the temperature there will rise to about 800 C and sensor positioning would be difficult (not much space there).
The second solution I could think of is some sensor to detect the RF field in the vicinity of the heating coil (obviously not near enough to get heated). Anyone has an idea on how I can accomplish that? All I know is that the RF is in the range of 500~600 KHz.

Thanks a lot!


groundFungus

#1
Feb 19, 2018, 01:59 pm Last Edit: Feb 19, 2018, 02:00 pm by groundFungus
I don't know the answer to your question, but I think that more members will see it if you ask a moderator to move it to the General Electronics section.  Use the Report to Moderator link to the lower left of post.

cloxart


PaulS

Quote
when an induction welding machine is appyling power to the heating coil
I would think that that would involve a significant (i.e. measurable) amount of current.

Putting a current sensor on the machine doesn't involve cutting any wires or making any physical connection to the machine. The Arduino is able to read the current sensor, and determine when the current exceeds some value.
The art of getting good answers lies in asking good questions.

cloxart

thank you PaulS,
I thought of that too, but those sensors are quite expensive and they want 25 of those devices, so if i could cut the cost it would be good..., nice suggestion by the way

Thank you

cedarlakeinstruments

thank you PaulS,
I thought of that too, but those sensors are quite expensive and they want 25 of those devices, so if i could cut the cost it would be good..., nice suggestion by the way

Thank you

That raises the obvious question then: what is your budget for the sensing method?
Electronics and firmware/software design and assistance. No project too small

cloxart

they will pay whatever is needed, but obviously the cheaper i can do it, the better it will be

jarnoldbrown

You might consider something like this:
https://learn.adafruit.com/using-melexis-mlx90614-non-contact-sensors/wiring-and-test

Personally, in order of best first, I would go for:

1) Sensing the current draw by the heating coil equipment. Should be fast and reliable. I have no idea what form these induction welding machines take. Are they 240V single phase? What sort of power do they consume?


2) Sensing the RF, although this is a bit of an unknown. You might have to do a bit of experimentaion, and I don't know what equipment you have. Possible problems with RF from nearby units.


3) Sensing the heat/temperature of the work. I think this would be slow and not so reliable.

cloxart

thank you for feedback jarnoldbrown,
current draw sensing is surely feasible, each welding machine draws about 3 KW at full power at 240V...
as i said before i was thinking that an RF sensor could be cheaper... I remember that when i was i boy and i had a CB radio i built a very simple device, taken from an electronic magazine: it had a copper wire wound around the antenna cable and whenever the transmitter was on an LED would lit up... cant remember how it was done but something like that would suffice for my scope

cedarlakeinstruments

@cloxart:
I don't know your customer's requirements, but I try to remind people that cheaper is not necessarily better. What you build needs to be reliable, repairable when you have moved on, and the cost of parts is not the cost of the solution: a few turns of wire around a cable and a rectifier bridge may be ultra cheap but it can have a higher installation cost than a spring loaded current clamp that is built for the specific purpose and may need to be "tweaked" to make it reliable enough.

When I approach something like this, I like to have a target price in mind. A $1 solution is very different from a $100 solution.

e.g., using this: http://www.crmagnetics.com/ground-fault-current-transformers/cr8410 you can build a reliable sensor, but you still have to design/build/document the rest of the circuit.

This: http://www.crmagnetics.com/switches/ac-current/ac-voltage/cr9321 costs a little more, but you can feed its output directly to an Arduino input.

Now, both of these cost more than a wire and a diode, but they are tested and specified for the application and can be purchased off the shelf.

Electronics and firmware/software design and assistance. No project too small

cloxart

@cedarlakeinstruments

obviously I aim for a reliable solution. that said the sensors you suggested are supposed to be used on just ONE power wire. I cannot simply pass the power cable in there, so the request to not modify or tamper the existing machines would not be fulfilled.

jarnoldbrown

I don't believe you will find any sensors that work by clamping on to 3 core mains cable. Does the ban on interfering with the equipment extend to adding an extension lead, and mounting the current sensor there?

Personally, I can't provide any more advice, as I do not have enough information about the apparatus and the way it is set up.

cedarlakeinstruments

Cloxart:
If I were in your position (assuming you are acting in a consultant role), I would point out that the restriction on not modifying the machines will lead to a higher cost solution, including my fee.

That said, not knowing anything about these machines other than what you've mentioned, here are a couple of off the cuff solutions that have been used for similar problems.
- Sense magnetic field of the valve coil.
- Remote IR sensor to detect the coil heating up.
- Local temperature sensor to detect coil heat. A Type K thermocouple is good to over 1,000C and can be made as small as you need it to be, I guarantee that.
Electronics and firmware/software design and assistance. No project too small

Paul_KD7HB

Hello everyone,
I was asked to build a device to switch on a solenoid valve for a certain amount of time when an induction welding machine is appyling power to the heating coil. this is to dispense an inert gas on the part being welded to avoid oxidation. that would be a very easy task if i could get some signal from the welding machine itself, but i am not allowed to connect anything there, so I have two solution: the first is a temperature sensor, switching on the valve when the part start heating up, but the temperature there will rise to about 800 C and sensor positioning would be difficult (not much space there).
The second solution I could think of is some sensor to detect the RF field in the vicinity of the heating coil (obviously not near enough to get heated). Anyone has an idea on how I can accomplish that? All I know is that the RF is in the range of 500~600 KHz.

Thanks a lot!


The answer to your problem lies right in your description. 500-600 KHz is in the standard radio broadcast band. I am really surprised you do not have problems with radio interference.

But, tune a radio to pick up the signal from the welder. Do you hear it go on and off?

Then examine the radio circuit and find the AGC or automatic gain control part of the circuit. This will be a negative voltage that reduces the amplification of the radio circuits, so the sound remains at about the same volume. You can use this voltage to determine when the welder is on and when it is off.

Paul

jarnoldbrown

The OP did say they want 20 of these, so it's possible that there might be more than one unit fairly close by. For that reason I thought the RF route might not be so reliable.
As for 500-600kHz being in the radio broadcast band, yes, if you happen to use Medium Wave.



Go Up