Detect 10V…To.. 260V AC/DC with HCPL-3700?

[frangimeno]

I want to detect voltage between 10V and 260V AC / DC, I have thought to use the HCPL-3700 but if I put the 47k resistors the output is activated, but if I apply 12V the output is not activated.

Does anyone know if it can be done with this chip, or does this one not work and I have to use another type?
Basically I want to activate a 5Vdc PIC with whatever voltage is applied. I don't need zero crossings or anything special, only that it be activated with that input voltage range.
Any ideas?

Thank you very much

 

[rjenkinsgb]

The input side of the HCPL3700 can stand up to 50mA.
As long as the current is below that at the highest voltage in your circuit, you are OK with that device.

47K + 47K means the current at the peak of the AC cycle with 260V input is only around 4mA.

You could reduce the resistor value quite a lot without overdriving the IC; you just need to ensure the resistor power rating is appropriate.

eg. Two 22K 1W or higher resistors should give a rather lower minimum detection voltage.

There are also many other circuits you could use.

 

[frangimeno]

Hi, thanks for your help.

I have two questions:

If I want to detect AC and DC do I have to link the (pin 3 with pin 4) and the (pin 1 with pin 2)? and remove the capacitor ?.

What other solution can you think of?

Thank you

 

[rjenkinsgb]

If I want to detect AC and DC do I have to link the (pin 3 with pin 4) and the (pin 1 with pin 2)? and remove the capacitor ?

No, the device has an internal bridge rectifier so the inputs are not polarity sensitive.

Alternatives are such as using a conventional but high CTR optocoupler with an bridge rectifier, plus the resistors to limit the current.
eg. One with a CTR of 500% will sink fie times more current at the output than the current through the LED; it should work down to near 100uA at the input to switch the 10K load you are using. If the LED can take 20mA, that gives a 200:1 possible input range.

If the electronics are part of the same device as the input voltage (so that is always related to the electronics ground) you can use a differential amplifier with a resistive divider input.

Or an opto type circuit with the input resistors split, so two in series at each input. Add eg. back-to-back zener diodes between the two junctions, so the voltage applied to the "inner" resistors is limited; that also limits the maximum current to whatever those resistors allow; the outer resistors take all the excess voltage.

 

[frangimeno]

Much more complicated.
I will continue with the HCPL3700.

Does the 22uF capacitor shown in the diagram have to be 400V?

 

[ronsimpson]

Does the 22uF capacitor shown in the diagram have to be 400V?

edited: Cap can be low voltage!
The capacitor voltage is 1.414 times higher than the AC voltage. Take the highest the AC could get then find the peak of that then add a little just to be safe. Yes 400V.
Datasheet: HCPL in ac mode. It takes about 5 volts to turn on the part. (4.23/5.0/5.5 min/typ/max) So when you think about low voltage and low voltage you need to substract 5.5V. 12V-5V=7V There will be 7 volts across the resistor.

Input current: You must stay below 50mA. I think it is unsafe to use the part at absolute max current. From experience, the part will fail with time.
It takes 3mA to turn on. (min current)
I do not see how you can reach from 260V to 12V. The part is intended to work from high line 220 to low line 110. I can see maybe getting 10: range.

 

[rjenkinsgb]

The capacitor voltage is 1.414 times higher than the AC voltage.

The cap is across the device internal clamp; a 16V or 25V cap is fine.

The highest voltage between pins 2 & 3 listed in the datasheet is 13.4V, using a direct DC feed rather than the resistors to pins 1 & 4; it should stay below 10V with 2 x 22K series resistors.

 

[AnalogKid]

You can add an external current limit circuit with high voltage transistors, in series with the chip input.

ak

 

[frangimeno]

Hello It seems to me that this is not going to work I am putting 2x4k at 24vdc so that it shoots.
If I put 220v with those resistors, it burns for sure.

 

[frangimeno]

I'll have to think about something else because this doesn't work.
For smaller margins eg 260V / 160V it will work. but this does not look good.

 

[frangimeno]

I have simulated this circuit with 5V zener and it works fine. In the output I will connect the optocoupler.

 

[Diver300]

I have simulated this circuit with 5V zener and it works fine. In the output I will connect the optocoupler.

View attachment 127427

That won't detect the DC.

The problem is that any linear circuit with only resistors that will make sure that the HCPL-3700 operates at 12 V will allow too much current at 240 V.

Assuming that the higher voltages that you need the circuit to detect are ac, then a series capacitor will do exactly the wrong thing, and give you more current with ac. You could put a capacitor in parallel, to bypass current from the HCPL-3700 on ac.

A series inductor would be ideal, but you would need an enormous value of inductor, so it would not be at all practical.

You could limit the voltage with a zener, and then have a second resistor to limit the current to the HCPL-3700 but the first resistor and the zener would produce a lot of heat.

There would be some advantage in having a diode in series, if you are happy that the circuit won't detect negative voltages. You might need a larger capacitor, but the heat generated on ac would be halved.

AnalogKid's circuit is probably the best approach. It would need a diode or a bridge rectifier before it.

 

[ronsimpson]

You can add an external current limit circuit with high voltage transistors, in series with the chip input.

Current limit
I can not find it now but I have used a low current version..... high voltage constant current IC. Can be set to 10mA or any value you want. Up to 900V.
This part is for DC not AC so it needs to have a full wave bridge ahead of it.
I know it cost too much and AnalogKid's circuit works with AC.

 

[frangimeno]

Thank you all for the help and for your time.

 

[Diver300]

I know it cost too much and AnalogKid's circuit works with AC.

No, it doesn't.

AnalogKid's circuit is very neat, but it will be nearly a short with reversed polarity. There has to be a diode in series to prevent reverse current, or the current limit circuit goes after a bridge rectifier.

 

[frangimeno]

Hello, has anyone used the LR8 microchip? It seems like a good solution.

Input voltage + 12V ... 450V

 

[gophert]

You could just get a cheap 5v USB AC adapter to detect if you have 85-230vac present, add a pull down resistor and connect to a logic level OR gate, then use your original circuit/device to detect from 10 to 85v and connect that to the OR gate. The OR gate output tells you if you have power.

 

[Diver300]

You could just get a cheap 5v USB AC adapter to detect if you have 85-230vac present, add a pull down resistor and connect to a logic level OR gate, then use your original circuit/device to detect from 10 to 85v and connect that to the OR gate. The OR gate output tells you if you have power.

I think that the problem would be getting the 10 - 85 V volt sensor to survive the 230 V.

 

[rjenkinsgb]

Try this:

A TLP627 optocoupler, with a 10K load resistor on the output to 5V.

https://docs.rs-online.com/d905/0900766b813d26f9.pdf

Put a diode bridge (4 x 1N4148) in front of the LED.
Connect a 220K resistor across the AC input.

Use two, 33K 0.5W or higher resistors to feed the input voltage to the bridge.

I think that should work from under ten volts, up to 240V; AC or DC.

 

[gophert]

I think that the problem would be getting the 10 - 85 V volt sensor to survive the 230 V.

I have a 12VDC refrigerator that I can either plug into AC outlet or use a 12VDC barrel connector in the side. Two different power sources. The 12VDC barrel connector never sees 230vAC so it is not a problem.

The OP has never clearly stated what kind of situation would cause a single point to see 10vDc and 230vAC. I cannot imagine what kind of a weird solution we might be solving - I think we need to hear the core problem and offer better solutions.