I need help to detect an AC burned heater element

[igeorge]

Hello,
I need help to achieve a very difficult task for me.
We have in the plant few machines for welding plastic, based in heating plastic and pressing the parts together.
The company which made them is out of business so no luck there.
Each machine run at 110 VAC, and has almost 100 heaters.
The heaters are from 250W to 450W, and are controlled by a SSR and a heat controller which fire them at different phase angle to control the amount of heat.
The issue is that they burn out, one, or more and is hard to detect it without manual visual inspection of each part.
My task is to find a solution to detect if any heater in the machine is burned out.
I started by looking at ACS712 from allegro micro. It seems to be good , but i have a problem - ANALOG
The chip give out 0 to 5 volts with middle point at 2.5 volts for zero amps. Amps in the positive side goes 2.5 and up, and the negative side 2.5 and down.
So the output, at any amps reading is a sinusoid at 60 HZ.
For my point of view, all i need is to get a digital 0 or 1 for the result, so i can input it to a PIC.
0 if the current is under 1 Amp, and 1 if it is over 1 AMP
As i mentioned before i do not have a clue about analog, op amp to condition , or amplify the signal, or the rest. I just know plain basic of analog.
Please help , and if you have a solution which you believe it will works, a small drawing, even by hand but with component value will be great.
Thank you

[Gevo]

Hello,

Maybe someting with a magnetic (Murata AS-M15TAN-R) sensor ? This give you a 5-volt low/high signal when detect a magenetic field.

Regards,
/ Gevo

[Archangel]

Hello igeorge,
Could you elaborate a bit on the conditions you want to test the device? It seems you want / need to test during the welding operation. Can you install a tiny reed switch along the power wires? Honestly I do not know if a reed will activate on an A/C load wire, a coil like a Tape Head or a guitar pickup will. If you do not need to test during operation, then a simple continuity test should work, neon lights used to work very well for that.

[igeorge]

Thank you Gevo for suggestion. I have to test it first because i do not know the value of magnetic field at 10 amps.
I will update you when i get it

[igeorge]

Hello igeorge,
Could you elaborate a bit on the conditions you want to test the device? It seems you want / need to test during the welding operation. Can you install a tiny reed switch along the power wires? Honestly I do not know if a reed will activate on an A/C load wire, a coil like a Tape Head or a guitar pickup will. If you do not need to test during operation, then a simple continuity test should work, neon lights used to work very well for that.
Thanks Archangel for reading my post and offer to help.
Here are the conditions:
I must check during the weld process, and i have to make sure that the current going through the wire is over 1 amp.
I do not think the other solution is feasible, due to the number of wires involved.
What i have in mind is to have a PCB for 16 heaters and bring the wire heater to it to one terminal, run a trace through some device to get the current, then get out on the next terminal and go to heater.
Current transformers is not a solution, as they are bulky, not to many models PCB mount and i still need to interpret the reading.
My option was for ACS712 , but if you look at the attached graph you will see that the output is + and -
I need somehow to convert it, or do something that i will get only the + portion, but not pulsating, i need it as a continuous line.
Also, i need that any voltage over 2.6 volts to be a logic 1 and any voltage from 0 to 2.6 to be a logic ZERO. In this case, i will have my inputs to PIC set to read 1 for a good heater, and 0 for a bad heater.
Thank you

[Archangel]

Thanks Archangel for reading my post and offer to help.
Here are the conditions:
I must check during the weld process, and i have to make sure that the current going through the wire is over 1 amp.
I do not think the other solution is feasible, due to the number of wires involved.
What i have in mind is to have a PCB for 16 heaters and bring the wire heater to it to one terminal, run a trace through some device to get the current, then get out on the next terminal and go to heater.
Current transformers is not a solution, as they are bulky, not to many models PCB mount and i still need to interpret the reading.
My option was for ACS712 , but if you look at the attached graph you will see that the output is + and -
I need somehow to convert it, or do something that i will get only the + portion, but not pulsating, i need it as a continuous line.
Also, i need that any voltage over 2.6 volts to be a logic 1 and any voltage from 0 to 2.6 to be a logic ZERO. In this case, i will have my inputs to PIC set to read 1 for a good heater, and 0 for a bad heater.
Thank you
Ok so I am thinking, use 2 diodes in parallel with their polarities reversed so there is a path for both phases, then only run the forward phase through the device.
0 amps = 2.5v 1 amp or more, is above 2.7v.
What do you think, will this work?

[Pic_User]

Hi igeorge,

If you just need an un-calibrated “Go - No Go” measurement, a simple “peak detector” could work.
If a silicon diode has too much junction drop, maybe a germanium or Schottky diode would work.
Just a diode to a capacitor circuit on the output of the ACS712.

5645-Adam-

[igeorge]

Thanks Archangel,
I do not know if it will works, but i will try it. I am some how concerned about the power dissipation as they are 5 amps going through the diodes, and they will be 32 on one PCB for 16 heaters. I will test and let you know.
Thanks again
Ion

[igeorge]

Hello Adam,
Thanks for reply but i am afraid, like i mentioned before about the power dissipation on the diodes.
I have to do some tests to see if it will work.
Thanks
Ion

[Pic_User]

Hi Ion,
You are right to be concerned with the several watts per diode dissipation with diodes passing Amperage..
But I am proposing using them in the 5V output circuit of the ACS712.
They would have almost no current. and would give your digital pin a 2.5V to 5V swing.
If the peak Voltage was a little too high for your logic levels then add a “Voltage divider (= two resistors).
-Adam-

[Ioannis]

Maybe a simple diode bridge with a optocoupler can give a low cost solution.

Put the ac pins of the bridge in series with the element. Short the dc pins together.

Put an optocoupler with a small series resistor in parallel to the ac pins of the bridge. Use a small integration of the pulses of the output optocoupler.

You are done!

Ioannis

[igeorge]

Adam and Ioanis, both of the solutions will works if i combine them. Optocoupler in series with heater is not feasible due to the amount of heat dissipation on the limiting resistor for the diode on the optocoupler. Also , as current varies from 0 to 5 amps, will be difficult to get a resistor value to light the led in the opto at everything over one amp, but do not fry the LED at 5 amps. Like Adam suggested, i will try to use the output from the ACS712. Here, i am going back to Ioannis advice to use an integrator to get a proper DC voltage out of the sine wave. From the DC voltage i need to use a comparator to get an output fro any voltage over 2.6 volts.
Ioannis, do you have any schematic for an integrator which will works on the voltage diagram posted above.
Let's say the Amps going through the heater are 5 A, so i will have a sine wave going up and down between +2vdc to +3vdc, where the 2.5 volts is the zero crossing point for zero amps. I need some kind of schematic to convert that sine wave to a dc line at 3 volts
Best regards
Ion

[Ioannis]

You did not followed exactly my thought.

Look at the attached schematic. You have to calculate the small resistor (I use SMD of 470 Ohms at 0805 size) according to bridge and Opto characteristics.

No need for power parts.

Light and dirty cheap.

The integrator or a simple delay will be done in software. When you have powered your load, current will flow through the bridge. A small voltage drop is then developed across the AC terminals of about 1,4-2 volts. The reverse voltage is small enough not to destroy the opto led.

So, when you have powered the load, normally the opto transistor will give a square waveform of line frequency. If the load is faulty, no current will flow and no pulses will be generated.

Ioannis

[Nicmus]

One other approach will be to do a search in this forum for “transformerless” you will find ideas of simple circuits of powering up small loads (a LED part of the optocoupler in your case) .
Use one of the circuits in parallel with your switching device (relay, triac,…etc) and monitor its control signal and the optocoupler output.
If your heater is OK you will have a signal from your optpcoupler when the heater is not energzed. In case the heater is open you get no signal when the controlling device is OFF.

HTH

Nick

[igeorge]

Thank you Nick
I will look on this too. It might be a good solution to get rid of the power components problem
Regards
Ion

[Nicmus]

Using one of the modules, suggested above, parallel with the switching device and one parallel with your load and ORing the outputs of the optocouplers (preferably open collector type) you can have all your parts inside your machine and the only connections to the monitoring device will be two isolated wires that will act as a “switch”. The only time the “switch” shows open will be when your heater is open.

Nick