optocoupler(817C) calculation required

[Mithun_K_Das]

I want to use 817C optocoupler to isolate a source and need to measure the voltage. I need the calculation for voltage change in pin3 for change of voltage in pin1.

 

[alexan_e]

There is no way you will be able to do what you want and preserve accuracy.
I have seen this done with devices called linear optocouplers
they are essentially an optocoupler with two identical output receivers and you use one of the outputs for feedback to make sure that the output has the level that it should according to the input/output relation

You can find typical application circuit s in the datasheet.

Alex

 

[Raza]

Hi,
Here are two images from the data sheet showing the Collector current of Photo transistor versus to CE voltage of 817C and the transfer ratio versus forward current. The transfer ratio of 817C is 200-400%. You can calculate your required parameters based on the datasheet provided information.

 

[Mithun_K_Das]

The whole calculation is becoming so much complex to me for just a little measurement. Is there have any simple and linear way?

 

[alexan_e]

Do you think they would have created the linear optocouplers if it was so easy to use a simple optocoupler and get accurate results?
Even if you calculate it the result will be temperature and device dependent.

Alex

 

[poorchava]

I've done some experiments with linear optocouplers. Although Perkin-Elmer, not Vishay, but results weren't actually that good. The signal coupled drifted heavily with temperature. Also reaching steady state took quite a long time.

If you are looking for something easier, i'd suggest getting some standalone ADC or a small microcontroller which has one (like TI MSP430) and then interfacing it with the rest of the ciurcuit using SPI or UART which are very easy to couple optically with ordinary couplers.

 

[Raza]

HI,
Perkin-Elmer produces the Optocouplers for their instruments which are mostly Laboratory equipment. They need a precision measurements for the Test Works and obviously they are expensive.

 

[poorchava]

I don't really think they are expensive. In Poland (where, by the way Vishay linear optocouplers are unavailable) Perkin-Elmer couplers cost about PLN 20-25 which is equivalent to about $8. As this is kinda speciality component, I don't think they are expensive.

On the other hand for that kind of money u can buy an MSP430F2013 and a dual optocoupler to send ADC results to central processing unit.

 

[Mithun_K_Das]

I want to design a small solar charge controller (10A-20A: /12V/24V/48V).In this case I need to measure the solar voltage but as it is not possible to keep a common GND so i need to isolate the solar GND from Battery GND and also need to measure the solar panel voltage for generating a dynamic PWM signal to control the charging current. The whole system should be within 5-8$ only. So I need to design in cost-effective way.

 

[Raza]

Hi poorchava,
I was just comparing the market for both the manufacturers (as I have worked with Perkin-Elmer, so I knew a bit about them). Here in my country the vishay is about $0.1 per piece and PE is about $0.8 per piece.

 

[poorchava]

Man... I'm starting to seriously think about moving to Pakistan... From what you say and from the other topic some time ago (about PIC MCU prices in Pakistan) it seems that converting national currencies to USD, electronic components are 5-10 times cheaper in Pakistan than in Poland. That is kinda outrageous....

But on-topic:
If your application is the solar charger, I would definitely go with MSP430 + ordinary optocpupler. MSP430 are designed for ridiculously low power consumption (i think as low as 100 nA in sleep mode, and 1 uA per MHz in active mode). On top of that programming is very simple. A debugger you can buy for $4.30 with free shipping (Launchpad). Moreover: they are very cheap, so using an ordinary garden variety dual optocoupler you can easily fit into your $8 budget.

 

[doraemon]

Hello!

(i think as low as 100 nA in sleep mode, and 1 uA per MHz in active mode)

I hope it was a typo. 1µA/MHZ is still to be developed.
The recent devices (F5x) consume about 250 µA/MHZ. If you use a
F5X at its maximal frequency (25MHz), then you will consume about
6 mA.
Now the FRAM devices are still a little bit better and can drop below
100µA/MHz. Please check the docs since I am not 100% sure about the
figures, but I think 250µA/MHZ and 100µA/MHZ is pretty close.

Dora.

 

[poorchava]

You are right, I was writing straight from my memory without checking. The MSP430F2013 with which I'm working right now has power consumption of 220uA/MHz while active, 0.5 uA in standby mode, and 0.1 uA in RAM retention mode. Sorry for disinformation