# What is Current Transfer Ratio In a Linear Optocoupler?

Joined Oct 26, 2012
434
Current Transfer Ratio (CTR) in a linear optocoupler:
• Is it always temperature based? Ie, amount of drift relative to change in ambient temperature? Or does it sometimes just mean linearity, regardless of temperature?
• Does a smaller CTR mean less drift, more linear, or...?
• What does a CTR of "100% to 200%" mean?

#### dl324

Joined Mar 30, 2015
16,896
Is it always temperature based? Ie, amount of drift relative to change in ambient temperature? Or does it sometimes just mean linearity, regardless of temperature?
It has to be dependent on temperature because the beta of the output amplifier changes with temperature, current, ...
Does a smaller CTR mean less drift, more linear, or...?
No. Plus CTR degrades over time.
What does a CTR of "100% to 200%" mean?
It's the ratio of the output current to the input current.

#### Papabravo

Joined Feb 24, 2006
21,217
CTR has nothing to do with drift or linearity. In optocouplers generally it represents how much of the diode current is reflected in the transistor current. It is a function of many things, and it usually degrades over time. That is why there is compensation circuitry in isolated DC-DC converters when they are used to feedback the output voltage to the controller. A smaller CTR means less of the diode current is present in the output transistor. A CTR of 100% means that if the forward current in the diode is 10 mA, then the collector(emitter) current will also be 10 mA.

I'm not familiar with ANY optocouplers that are specifically intended for linear operation. Could you give me a part number so I can check the datasheet.

#### crutschow

Joined Mar 14, 2008
34,386
I'm not familiar with ANY optocouplers that are specifically intended for linear operation.
They commonly have dual outputs so one output can be used for negative feedback at the input to give a linear input to output signal.
The IL300 is one example.

Joined Oct 26, 2012
434
• Is it always temperature based? Ie, amount of drift relative to change in ambient temperature? Or does it sometimes just mean linearity, regardless of temperature?
It has to be dependent on temperature because the beta of the output amplifier changes with temperature, current, ...
Meaning, i think, that the amount of feedback changes with load and temp, which in turn alters the amount of amplification. I'm guessing that would only apply to an opto with transistor output, and not an opto with a photosensor/photodiode output?
• Does a smaller CTR mean less drift, more linear, or...?
No. Plus CTR degrades over time.
That wasn't a yes/no question. Also, don't opto's with dual sensors compensate for degradation.

• What does a CTR of "100% to 200%" mean?
It's the ratio of the output current to the input current.
I was asking specifically about the numbers "100% to 200%". Does it mean that, with changes in temperature and load, ratio can vary by anywhere from 100% to 200%? Or...?

Last edited:

Joined Oct 26, 2012
434

#### Papabravo

Joined Feb 24, 2006
21,217
That sounds like a linear relationship of some sort.

LOC110
https://www.mouser.com/datasheet/2/205/LOC110-1391998.pdf
The datasheet will quote an average value or a value at a particular current and it is not constant.
The following simulation exposes the actual behavior where the middle trace show CTR as a function of diode current.

The diode current is swept in a linear fashion from 0.1 mA to 60 mA in increments of 0.1 mA. The blue trace of that sweep is linear, but the current through the transistor is NOT. The output window gives actual values at the two points 1 mA and 30 mA. You can do things in simulation that might break an actual part.

I'll have a look at the datasheet you referenced.

#### Papabravo

Joined Feb 24, 2006
21,217
I see what they are doing. What you have is not a traditional optocoupler and term CTR was never mentioned in the datasheet. What you have is two matched photodiodes that are electrically isolated but expected to behave and respond in identical fashion. In that way there can be circuitry on the LED side that has a picture of what is happening on the isolated side. Since this is a diode instead of a transistor alternate circuitry must be utilized.

#### dl324

Joined Mar 30, 2015
16,896
I'm guessing that would only apply to an opto with transistor output, and not an opto with a photosensor/photodiode output?
You're not going to get any gain without a transistor.
That wasn't a yes/no question.
Does a smaller CTR mean less drift
No
Does a smaller CTR mean more linear
No

Happy now?
Also, don't opto's with dual sensors compensate for degradation.
No. That's done in the circuit using the optoisolator.
I was asking specifically about the numbers "100% to 200%". Does it mean that, with changes in temperature and load, ratio can vary by anywhere from 100% to 200%?
100% means the output current is the same as the input current. 200% means the output current is twice as large as the input current.

#### panic mode

Joined Oct 10, 2011
2,731
suppose you have a product like this:

the two sensing diodes are matched to some degree but... they are never going to be exact.
measuring value on one of them, one can get an idea of what the other one is seeing (a proportional value)
so you need to build a circuit that compensate for that - and this will need to be calibrated.
not sure about your objectives but this is something i try to avoid.

one simple way to avoid opto problems like temperature drift, calibration, etc is to not use opto as a precision analog device (which is delicate, sensitive and costly). when talking about cost, i am not even considering the hardware cost but the time needed to manually calibrate each and every unit. so why not just use an opto to serially transmit binary code? then choice of opto and all opto related issues are gone! granted, both input and output side will need digital circuitry (MCUs). one of them will need an ADC and the other will need a DAC. but in general this is low cost (typically built into MCU) and a much more stable solution due to well defined resolution and performance of needed references.

Last edited:

#### crutschow

Joined Mar 14, 2008
34,386
the two sensing diodes are matched to some degree but... they are never going to be exact.
True, but the match may be good enough to transmit AC signals, such as audio, with sufficient fidelity, if connected in a proper feedback circuit.
A digital approach can obviously be more accurate, but may be overkill, depending upon the application.

#### panic mode

Joined Oct 10, 2011
2,731
true though we still do not know what the application is

Joined Oct 26, 2012
434
I'll try to ask more clearly.
Does a smaller CTR mean:
• less drift
• or, more linear
• or something else?
Happy now?
You seem angry.
No. That's done in the circuit using the optoisolator.
I'll try to ask more clearly. Doesn't the compensation circuit built using linear optoisolator with dual sensors compensate for degradation?
100% means the output current is the same as the input current. 200% means the output current is twice as large as the input current.
I'll try to ask more clearly. If 100% means the output current is the same as the input current, and 200% means the output current is twice as large as the input current, then what does a CTR of "100% to 200%" mean? That is varies with load, temperature, or other factors?

we still do not know what the application is
DC-DC power converter with secondary-side regulation. Need 0.5% linearity or better, @ Vout of 0V to 5V. Hope that helps!

#### crutschow

Joined Mar 14, 2008
34,386
DC-DC power converter with secondary-side regulation. Need 0.5% linearity or better, @ Vout of 0V to 5V.
Why such tight linearity?
Isn't the voltage reference, such as a TL431, for the controlling feedback in the secondary side, as is usually done?

#### ronsimpson

Joined Oct 7, 2019
3,034
DC-DC power converter with secondary-side regulation. Need 0.5% linearity or better, @ Vout of 0V to 5V. Hope that helps!
Power supplies I build with opto isolators do not require linearity. When the supply is in regulation the isolator current is at a constant current. Changing load does not change current in the isolator. With reason, the CTR does not effect the current in the output of the isolator.

#### panic mode

Joined Oct 10, 2011
2,731
what does a CTR of "100% to 200%" mean? That is varies with load, temperature, or other factors?

it means when you buy optocouplers, you may get one with CTR as low as 100% or as high as 200% or anything in-between. and that is at same test conditions (same temperature, input current, output voltage etc.). things get worse when temperature, aging, variation of input and output parameters etc. are factored in. so you are getting parts with very wide range for tolerance and it is insane to try to rely on it.

this is why you normally look for CTR that is high enough that it can do the job, then design your circuitry so it does not depend on the exact CTR value.

as was already pointed out, normal way is to pass value in digital form (on/off) rather than analog. and that is exactly what takes place in SMPS.
the TL431 is sensing if the Vout is above or below setpoint. this turns LED in optocoupler on or off (nothing in-between since this is a digital information, state is simply either on or off).

then the output side of the optocoupler is tied into feedback of your SMPS controller. the provided feedback works much like thermostat - it is not an analog value but constant switching on/off provides the regulation. unlike thermostat this happens much faster.

Last edited:

#### crutschow

Joined Mar 14, 2008
34,386
the TL431 is sensing if the Vout is above or below setpoint. this turns LED in optocoupler on or off (nothing in-between since this is a digital information, state is simply either on or off).
No, the output from the TL431 to the opto is a DC current as determined by the output voltage, and the output of opto is the DC feedback voltage to the converter chip..

Since it's part of a negative feedback loop, a small change in transfer gain of the opto doesn't affect the regulated voltage significantly.

#### BobTPH

Joined Jun 5, 2013
8,922
I'll try to ask more clearly.
Does a smaller CTR mean:
• less drift
• or, more linear
• or something else?

Here, I will make it easy for you.
CTR has nothing to do with drift or linearity.
So, no, it does not mean less drift or more linearity. It is simply the ratio of the output current to the input current.

That is like asking, about resistors, "does a higher resistance mean a lower temperature coefficient?" The two are not correlated. One does not say anything about the other.

#### Papabravo

Joined Feb 24, 2006
21,217

Here, I will make it easy for you.

So, no, it does not mean less drift or more linearity. It is simply the ratio of the output current to the input current.

That is like asking, about resistors, "does a higher resistance mean a lower temperature coefficient?" The two are not correlated. One does not say anything about the other.
I think that was my lead in post#3. Apparently, the TS suffers from selective comprehension, or some such affliction.

#### ronsimpson

Joined Oct 7, 2019
3,034
as was already pointed out, normal way is to pass value in digital form (on/off) rather than analog. and that is exactly what takes place in SMPS.
the TL431 is sensing if the Vout is above or below setpoint. this turns LED in optocoupler on or off (nothing in-between since this is a digital information, state is simply either on or off).
I do not know what type of power supply you have, but TL431 is run in the linear region. There is no ON/OFF. The TL431 is slow and will not switch on/off fast. It is a wonderful amplifier and voltage reference.