Arduino Forum

Using Arduino => General Electronics => Topic started by: Deous on May 10, 2018, 03:12 pm

Title: Precise linear optocoupling
Post by: Deous on May 10, 2018, 03:12 pm
Hi, I am trying to find an opto-isilator/opto-coupler that can linearly change voltage.
I have typical 0 to 3-5V input in mind. Output goes to 240V IGBTs
Frequencies would be 50/60Hz and 150-500Khz
What would you recommend? I found IL300

Thanks
Title: Re: Precise linear optocoupling
Post by: MarkT on May 10, 2018, 08:16 pm
500kHz is probably beyond linear optocouplers I'm afraid.  Gain for the IL300 is already dropping at 100kHz.

What are you trying to do?
Title: Re: Precise linear optocoupling
Post by: Deous on May 10, 2018, 08:26 pm
I am trying to reduce voltage on electric cooker stove which I modified.
There is 220VAC that must be reduced with microcontroller - would be best case.
It has also induction cooking elements for higher frequencies - not sure what but I am right now working on normal heating element powered with AC.
Title: Re: Precise linear optocoupling
Post by: DVDdoug on May 10, 2018, 08:28 pm
Congratulations on finding the IL300...  I'm not familiar with it and as you probably know, most opto-isolators are not linear.     The datasheet says "bandwidth > 200kHz, so it may not work at 500kHz.

This is not "easy"...     The LED is still an LED so you'll have to bias your input if you need to get-down near zero volts...  It doesn't turn-on until it gets above 1/2V.   Also the graphs show current  (not voltage)  linearity.   That's not a problem on the output-side where you've presumably got a resistive load, but it's an issue on the input-side.


It's also got two photodiodes.   The "extra " one is intended for feedback (to achieve linearity) on the non-isolated side and you'd have to incorporate that into your drive circuit.


It's photodiode (not a phototransistor) so you'll need some amplification on the output side.



Title: Re: Precise linear optocoupling
Post by: Deous on May 10, 2018, 08:32 pm
Yeah, I guess optocoupler is a bad idea in this case.
What silicon would be good here?
Title: Re: Precise linear optocoupling
Post by: MarkT on May 10, 2018, 08:36 pm
You shouldn't be thinking of sending the induction waveform over the opto-coupler, send its envelope
(use a diode/capacitor detector perhaps?).  You could loosely couple with a small value capacitor for
an HF AC signal, no need for optical isolation.  

I'm not sure the voltage envelope is the right thing really, the output power is probably what you're
really interested in and that's probably easier to measure from the DC supply to the push-pull driver
of the induction circuit.
Title: Re: Precise linear optocoupling
Post by: DVDdoug on May 10, 2018, 08:41 pm
Quote
I am trying to reduce voltage on electric cooker stove which I modified.
There is 220VAC that must be reduced with microcontroller - would be best case.
Heaters normally switch on/off.  I'm sure you've noticed that's how your furnace works.   Same with refrigeration and air conditioning.    You switch the heat on 'till you achieve the target temperature, then you shut it off.  It's super-simple and often done without any microprocessor or software.

It's possible to use phase control (http://www.ilight.co.uk/downloads/iLIGHT%20Binder-HowDimmers.pdf) (like an AC light dimmer) but that's rare for heating/cooling because the heat can't change quickly anyway.     A microprocessor-controlled dimmer requires two opto-isolators...   One for the zero-crossing detector and one for the output.   (Or you can use transformer isolation on your zero-crossing detector).

It's also possible  to cycle it on & off (without feedback) with a short duty-cycle (maybe around 1 second) to get 'high", "medium", and "low" settings, etc...  If you turn it for 1 second and then off for 1 second in a repeating cycle, that's "half temperature"

Quote
Yeah, I guess optocoupler is a bad idea in this case.
Yes, you should isolate your Arduino from the power line!!!!    ...For your safety, the safety of the Arduino, and the safety of your computer if you happen to connect USB at the same time.
Title: Re: Precise linear optocoupling
Post by: TomGeorge on May 11, 2018, 01:27 pm
Hi, I am trying to find an opto-isilator/opto-coupler that can linearly change voltage.
I have typical 0 to 3-5V input in mind. Output goes to 240V IGBTs
Frequencies would be 50/60Hz and 150-500Khz
What would you recommend? I found IL300

Thanks
I am trying to reduce voltage on electric cooker stove which I modified.
There is 220VAC that must be reduced with microcontroller - would be best case.
It has also induction cooking elements for higher frequencies - not sure what but I am right now working on normal heating element powered with AC.
Trying to control an IGBT in a linear mode will produce all sorts of problems, the main one being the voltage drop across the IGBT and the need for a method of dealing the energy  dissipated in it.
IGBTs work best a ON/OFF devices, so PWM control or AC Phase control will be needed, so you do not need a linear response type opto-coupler.
HOWEVER

Inductive cooking uses a power oscillator to drive the inductive element, it might not appreciate a lower input supply voltage.
Does the cooker already have an energy level control for each element?

Tom.... :)
Title: Re: Precise linear optocoupling
Post by: Deous on May 11, 2018, 03:42 pm
There are no energy level controls - they are gone and circuits as well )
Title: Re: Precise linear optocoupling
Post by: TomGeorge on May 11, 2018, 04:00 pm
There are no energy level controls - they are gone and circuits as well )

Okay, so you are going to build a new power driver circuit.
Can you post a proposed circuit?
Tom.. :)
Title: Re: Precise linear optocoupling
Post by: allanhurst on May 11, 2018, 11:57 pm
I'm looking forward to this!

tee hee.......

Why?

 because that's stuff for an experienced professional in a specialist  field , and involves, among other things, the design of some clever magnetics.

I couldn't do it first time off.  No chance. Lots of magic smoke.

Allan
Title: Re: Precise linear optocoupling
Post by: Deous on May 12, 2018, 03:31 am
Okay, so you are going to build a new power driver circuit.
Can you post a proposed circuit?
Tom.. :)
I was thinking about something like this:

(http://i65.tinypic.com/bf1kew.jpg)

However optocoupling won't do the proper job here.
I need second subcircuit that would 'smooth out' the pwm ratio into linear relationship
There must be a simpler circuit


Title: Re: Precise linear optocoupling
Post by: Deous on May 12, 2018, 04:03 am
Also found some materials pertaining low-pass filtering PWMs from Arduinos


https://www.allaboutcircuits.com/technical-articles/low-pass-filter-a-pwm-signal-into-an-analog-voltage/ (https://www.allaboutcircuits.com/technical-articles/low-pass-filter-a-pwm-signal-into-an-analog-voltage/)


Comes down to this:

(https://i.stack.imgur.com/eDUFY.png)

...  :smiley-neutral:
Title: Re: Precise linear optocoupling
Post by: allanhurst on May 12, 2018, 06:00 am
If you're going to post a Sallen and Key filter, at least get the values right.


C1  |=  C2.

And of what use would it be to you anyway in this application?


Allan
Title: Re: Precise linear optocoupling
Post by: TomGeorge on May 12, 2018, 06:10 am
Hi,
Why do you want to go to analog power control?
What will the analog voltage control and how?

If you are replacing a U/S (UnServicable) unit, do you know the frequency it operated at?
Do you know the inductance on the coil?
Does the coil have a capacitor connected across it?

The inductive cook tops I have come across appear to be PWM/HF for efficiency and induction performance.

Tom... :)

Title: Re: Precise linear optocoupling
Post by: allanhurst on May 12, 2018, 06:12 am
Quote
The inductive cook tops I have come across appear to be PWM/HF for efficiency and induction performance.
Quite.

Not a trivial design task.

Allan
Title: Re: Precise linear optocoupling
Post by: Deous on May 12, 2018, 03:39 pm
Right now I am not touching induction heater/cooker but regular resistance based heater because it needs more power around 1-2.5KW

I just want simple isolated microcontroler regulation that is't some insane pulsing or ugly-wave modulating but oldschool linear regulation of AC just as if it was a resistor on the knob.

I see industry has many problems with isolating smd from power circuits. And it becomes so expensive at some cases because people don't understand how simple it is. I was once thinking that a nice servo motor turning power resistor would be cheaper and better than some expensive fancy silicon controller for few hundred bucks....
Title: Re: Precise linear optocoupling
Post by: allanhurst on May 12, 2018, 05:20 pm
The normal way of controlling a resistive load with an ac source is to use a triac. There have been several threads on this on this site covering this problem.

Isolation from the controller is generally by means of an optotriac.

The basic idea is similar to that used in domestic incandescent light dimmer switches.

try googling ac power control.

Allan
Title: Re: Precise linear optocoupling
Post by: Deous on May 12, 2018, 05:38 pm
Why not just use a coil transformer with pwm low pass filter converting it to linear output?
You can achieve really nice frequencies - I am checking it now and looks promising.
Title: Re: Precise linear optocoupling
Post by: allanhurst on May 12, 2018, 07:11 pm
Quote
Right now I am not touching induction heater/cooker but regular resistance based heater because it needs more power around 1-2.5KW
So how are you going to control the high power?

Allan
Title: Re: Precise linear optocoupling
Post by: Deous on May 12, 2018, 07:48 pm
Here is some very basic sketch:

(http://i65.tinypic.com/2s7uflx.jpg)
Title: Re: Precise linear optocoupling
Post by: allanhurst on May 13, 2018, 05:52 am
You can't use 2 N-channel fets - you need an N and  a P.

And your transforrner will have to generate at least the p-p value of the AC plus a few volts to work.

And there will have to be a whole load of protection stuff to prevent overvolting the gates - both + and -ve.

Try a triac. It's been done many times. And it works.

Allan

PS   there IS a way to do this..... but try the triac first. Much easier.
Title: Re: Precise linear optocoupling
Post by: Deous on May 13, 2018, 06:27 am
But triac is non-linear. It's just not stable. You have any model in mind?
Also how you can overvolt gates if max secondary output is fixed?
Besides few volts over threshold usually will just put fet in saturation mode

Easier way is of course good cooker IGBT )
Title: Re: Precise linear optocoupling
Post by: allanhurst on May 13, 2018, 11:10 am
The p-p value of 240vac is about 340 volts. So your transformer needs to deliver 360v p-p.
 Fixed, you say. Interesting transformer. Bet you can't buy one.

But suppose the mains varies +/- 10% ?

The 'on' device's gate  needs about 10v higher than the supply - but it's ac, so how will this 10v track the ac waveform? And what about the other device? that'll have -360 volts on it's gate - it certainly won't survive that.

And you're right ,  IGBT's would be a better choice.

Triacs non-linear? - of course. But you want to switch the current, not control it linearly, so this doesn't matter.

Allan
.

Title: Re: Precise linear optocoupling
Post by: Deous on May 13, 2018, 04:29 pm
What are you talking about??
What transformer? gates are fully isolated and only 5V or less is going to them.
There is another simple solution of two pnp and npn power BJTs connected parallel.
I am surprised there are no circuits for linear voltage control of mains AC.
Title: Re: Precise linear optocoupling
Post by: ted on May 13, 2018, 05:12 pm
I am surprised there are no circuits for linear voltage control of mains AC.
I am not, everyone is using diac/triac, if they need a linear respond they to that by software.

On the other hands, linear response is sometimes needed for instruments not for a stove.
Just simple thermostat do the job with accuracy 1%
Title: Re: Precise linear optocoupling
Post by: Deous on May 13, 2018, 07:16 pm
I am not, everyone is using diac/triac, if they need a linear respond they to that by software.

On the other hands, linear response is sometimes needed for instruments not for a stove.
Just simple thermostat do the job with accuracy 1%
Well obviously they have no choice ) Me including
Why there is such a problem controlling amplitude of AC ??
This is crazy.
Most of power circuits I came across basically rectify main high voltage and produce the desired waveform out of it back again.
This is ridiculous in my opinion. We have technology to make our lives easier then why the won't make simple linear AC amplitude IC...
Title: Re: Precise linear optocoupling
Post by: ted on May 13, 2018, 07:32 pm
Well obviously they have no choice ) Me including
Why there is such a problem controlling amplitude of AC ??

Induction or not thermostat do not care about that
Amplitude control is very simple, high voltage audio amplifier 2000W and volume control, if you want to kept sine shape.
The same results of heating as SCR  will give.
Title: Re: Precise linear optocoupling
Post by: Deous on May 13, 2018, 07:40 pm
Yes I want to keep sine shape and linearly change the AC mains voltage and use all mains power from that voltage.
Heating SCR? What's that? Please share the solutions to that problem
Title: Re: Precise linear optocoupling
Post by: ted on May 13, 2018, 07:56 pm
The solution is simple as in traditional stove - turn on and of the power supply to heating element, plug in plug of the stove , thermostat decides for how long, no one is reducing amplitude of power supply.
Induction stove has a generatotr about 100kHz, which will stop working at reducet voltage .....
SCR = Silicon controlled rectifier
Title: Re: Precise linear optocoupling
Post by: MorganS on May 14, 2018, 02:09 am
I just want simple isolated microcontroler regulation that is't some insane pulsing or ugly-wave modulating but oldschool linear regulation of AC just as if it was a resistor on the knob.
Have you ever looked at an "old school" resistive power controller? Like the controls on a W-class Melbourne tram? Or any DC motor controller built before the 1950s?

Yes, they use a resistor and a knob. However the resistor weighs about the same as the motor and has about the same cooling requirements. When you're at 50% power, the 'other' 50% is waste power at the resistor. For something the size of a tram or train, the knob is a giant lever that the driver operates. If you look at a modern train control panel, they still have the giant lever because that's what the older drivers are used to. You also don't want to control a 1000-tonne train with a knob that might snap off if your jacket pocket catches it wrong.

The "old school" light dimmers you might have seen in houses are actually really clever triac circuits which don't require vast amounts of cooling for any resistor.
Title: Re: Precise linear optocoupling
Post by: Deous on May 14, 2018, 04:12 am
Have you ever looked at ...
Thank you for preaching and justifying today'd electronic engineering for not having what is exactly needed or desired.
Title: Re: Precise linear optocoupling
Post by: allanhurst on May 14, 2018, 05:08 am
Various posters have pointed out that what you exactly  require is impossible without great power loss.

A variac with motor control night achieve your needs.

But far more expensively than a triac approach for a heating element.

Allan
Title: Re: Precise linear optocoupling
Post by: Deous on May 14, 2018, 02:11 pm
Well ok, but unfortunately the project I am working on apparently needs a linear change of the AC in time.
I am just designing that for some complex heating process of ceramic and other materials that may take even few days and involves some complex profiles. This looks very difficult and I see there is no easy solution. Any fast switching with such power causes all kinds of EM interference. I will look at SCR and maybe some triacs that may be somehow adjusted to linear dimmers. Maybe there are some other silicons suitable for that.
Title: Re: Precise linear optocoupling
Post by: MorganS on May 14, 2018, 03:47 pm
Why do your ceramics care about E-M radiation?
Title: Re: Precise linear optocoupling
Post by: ted on May 14, 2018, 04:03 pm
He is lost in the jungle , he is expecting that TV set will work at 10V AC - similar electronics is inside of induction stove....
He is crying no simple solution - look at my posts
Title: Re: Precise linear optocoupling
Post by: Deous on May 14, 2018, 05:23 pm
Why do your ceramics care about E-M radiation?
There is some various em sensitive measuring equipment around.
Title: Re: Precise linear optocoupling
Post by: TomGeorge on May 14, 2018, 11:40 pm
There is some various em sensitive measuring equipment around.

Doesn't answer the question.. :o


"Around" where?
In your "laboratory"/"workshop"?


Tom.... :)
Title: Re: Precise linear optocoupling
Post by: ted on May 15, 2018, 12:00 am
10m - or more, this is how far interferences can go from 1000W, 100 kHz square wave transmitter
Title: Re: Precise linear optocoupling
Post by: allanhurst on May 15, 2018, 12:14 am
10m - or more, this is how far interferences can go from 1000W, 100 kHz square wave transmitter

''

Couple it to a decent antenna and  it'll be detectable continents away.

It's called longwave radio.

Chopped ac -ie with a triac and reasonable suppression ( common mode chokes, X rated caps etc) , not very far.  and eg Murata make excellent mains noise suppression gadgets based on such techniques..

You can buy solid state relays with which you can play such tricks and contain such suppression devices for reasonable prices.

Allan
Title: Re: Precise linear optocoupling
Post by: Deous on May 15, 2018, 01:05 am
Good idea thanks. Switching those power elements on and off many times really affects some work on rf stuff in that small lab ) Anyways I ordered couple of power triacs and will try to experiment with them.
Title: Re: Precise linear optocoupling
Post by: ted on May 15, 2018, 02:08 am
induction stove power module is not a resistive load for triac
Title: Re: Precise linear optocoupling
Post by: allanhurst on May 15, 2018, 02:18 am
No, but the OP says he's going to use a resistive heating element - or do I misunderstand you,Deous?

And if you're going to use a triac with suitable suppression devices, check out this forum for threads on variable phase ac control. It's been done many times before and gives you effectively near analog control with good resolution.

Allan
Title: Re: Precise linear optocoupling
Post by: ted on May 15, 2018, 02:25 am
So many times he changed his mind so I don't know what he is doing, and misleading name of the thread ......
Title: Re: Precise linear optocoupling
Post by: TomGeorge on May 15, 2018, 02:33 am
Hi,
@Deous, have you finished any of the projects from your previous threads?

Tom... :)
Title: Re: Precise linear optocoupling
Post by: allanhurst on May 15, 2018, 03:02 am
And ( a thought )

Deous : how do you expect to heat a ceramic with induction heating? - they're insulators.......

Allan
Title: Re: Precise linear optocoupling
Post by: Deous on Jun 10, 2018, 02:48 am
I just came back with update:
Replacing that mosfets in series with triac works perfect.
Circuit needs ac cap though - oven also flashes sometimes at different frequencies
All works perfect because triac requires really small voltage - 0.6-0.9V to switch on
Title: Re: Precise linear optocoupling
Post by: TomGeorge on Jun 10, 2018, 03:54 am
Hi,
Good to hear.

Can you please post a copy of your new circuit, in CAD or a picture of a hand drawn circuit in jpg, png?

Thanks.. Tom.. :)
Title: Re: Precise linear optocoupling
Post by: Deous on Jun 11, 2018, 05:23 am
It is the same picture as posted one in this thread. Additionally I found out an easy way how to generate a PWM signal using ESP32 that can be adjusted from 1Hz to 10MHz ))
It is perfect to control all kinds of devices. Posted code on other forum
Title: Re: Precise linear optocoupling
Post by: TomGeorge on Jun 11, 2018, 09:07 am
It is the same picture as posted one in this thread. Additionally I found out an easy way how to generate a PWM signal using ESP32 that can be adjusted from 1Hz to 10MHz ))
It is perfect to control all kinds of devices. Posted code on other forum
Sorry this is what you posted? ? ?
(http://forum.arduino.cc/index.php?action=dlattach;topic=546661.0;attach=261728)
PLEASE a hand drawn complete circuit diagram please, to help complete this thead for anybody looking for solution to similar problem they may have.
Thanks.. Tom.. :)