Analog signal drops after opto-coupler

[ericmar]

Dear all experts,

I need some helps here. I am having this offset problem which causes me getting an incorrect output. I have an analog signal which comes to my circuit and go through an isolcation path before feeding into an ADC and this analog signal drops after the opto-coupler.

The analog signal is first being passed to an op-amp, AD820, then only an opto-coupler, IL300F. After the opto-coupler, the signal passes through another op-amp, AD8601, before it was fed into the ADC, ADS7816. The voltage at each point with reference to each side's ground is as the following:

At pin 3 of AD820, V = 1.933V
At pin 6 of AD820, V = 1.752V
At pin 5 of IL300F, V = 1.818V
At pin 1 of AD8601, V = 1.826V

**broken link removed**

May I know wat is causing this offset? How can I solve this problem?

Thank you very much for those who try to help me in advance.

Best regards,
Eric Mar

 

[IanP]

Try to increase the If current to at least 10mA (or more) by decreasing the value of the 100Ω resistor to, say 62Ω (or less), and then measure these three voltages once more ..
Regards,
IanP

 

[ericmar]

Try to increase the If current to at least 10mA (or more) by decreasing the value of the 100Ω resistor to, say 62Ω (or less), and then measure these three voltages once more ..
Regards,
IanP

I replaced the 100Ω with a 52.3Ω but the voltage still drops!
:|

 

[IanP]

Try different IL300F ..
If the problem persists increase the value of R1 from 49.9kΩ to ≈53kΩ ..
Regards,
IanP

 

[angelote]

Check the transfer gain (K3) of IL300F. This will cause IL300 gain be different to 1. You can try to adjust gain modifying the 49k9 resistors.

 

[artem]

Well this difference :
At pin 5 of IL300F, V = 1.818V
At pin 1 of AD8601, V = 1.826V

could be clarified by high offset votage up to 7 mV (worst case ) (data sheet AD8601) , influence of input bias should not be as it is up to 1 nA (worst case). In addition you can have little bit measurement error , that could somehow explain remnant 1 mV (ot maybe more ).


To explain this difference :
At pin 6 of AD820, V = 1.933V
At pin 5 of IL300F, V = 1.818V

please measure voltage at :
a. R3
b. R1
c. pin 3 opf AD820

What is precision of R1 and R3 ?

Yet, the difference pin 6 of AD820 pin 5 of IL300F
is not much important but pin 3 AD820 to pin 5 of ILF300F.

Although photodiode is raither current source , it could be better to normalize power sources to same voltage level - make them equal AVCC and +V_ISO.

Decreasing of R2 will make sence if AD820 experience saturation when output being set to rail voltage . Till that no decrease is usefull from static point of view, but could change transient behaviour (which is not important to us).

 

[ericmar]

Sorry! I wrote wrongly abt the voltage at pin 6 of AD820. It should be V = 1.752V instead.
However, at pin 3 of AD820, V = 1.933V !

Added after 8 minutes:

Try different IL300F ..
If the problem persists increase the value of R1 from 49.9kΩ to ≈53kΩ ..
Regards,
IanP

I see no point in changing the value of R1 from 49.9kΩ to ≈53kΩ!

In fact, I think a 30KΩ should be used to replace both R1 and R3. This is because according to the page 5 of the datasheet of IL300F, for the K3(transfer gain) to be 1, it falls into group Bin F = 0.945 - 1.061. Based on page 10 of the datasheet, the resistance of the resistor should be 30KΩ.

Am I right?

I did try to replace the resistors from 49.9KΩ with 30KΩ, the voltage drop seems to be reduced, meaning that less voltage drop. However, the calculated output readings based on the input analog signal are still not correct. It is the same as if the resistors have not been changed.

Added after 16 minutes:

Well this difference :
At pin 5 of IL300F, V = 1.818V
At pin 1 of AD8601, V = 1.826V

could be clarified by high offset votage up to 7 mV (worst case ) (data sheet AD8601) , influence of input bias should not be as it is up to 1 nA (worst case). In addition you can have little bit measurement error , that could somehow explain remnant 1 mV (ot maybe more ).


To explain this difference :
At pin 6 of AD820, V = 1.933V
At pin 5 of IL300F, V = 1.818V

please measure voltage at :
a. R3
b. R1
c. pin 3 opf AD820

What is precision of R1 and R3 ?

Yet, the difference pin 6 of AD820 pin 5 of IL300F
is not much important but pin 3 AD820 to pin 5 of ILF300F.

Although photodiode is raither current source , it could be better to normalize power sources to same voltage level - make them equal AVCC and +V_ISO.

Decreasing of R2 will make sence if AD820 experience saturation when output being set to rail voltage . Till that no decrease is usefull from static point of view, but could change transient behaviour (which is not important to us).

The precision of R1 and R3 being used is 1%.
It is unlikely to make AVCC and +V_ISO equal since the circuit is designed in such a way to isolate 2 parts of the circuit. I did notice that there is a voltage difference between the analog ground, AGND, and the ground of isolated circuit, IBP_GND. The voltage difference is about 0.01V.

I would like to know if I use a high resistance resistor to tie both AGND and IBP_GND to balance both grounds, would it void the isolation design?

Once again, thanks for everyone trying to help me with this!
I hope to learn as much as possible with you guys.

Best regards,
Eric Mar

 

[artem]

I dont say that you have to connect them to same voltasge source, but for test purposes connect them to supply with same voltage, also check that these voltages are clean from noise - put HF and LF filtering capasitors per each voltage and check those sources by scope. May be if such noise and due to diode capasitance you can get wrong results .


Could you disconnect AD8601 pin 3 from R1-pin5 IL300F connection and measure voltages:
a. pin 4 IL300F
b. pin 3 of AD820 ?
c. pin 5 IL300F

Can you also tell us voltages of AVCC and +V_ISO.
Yet, what is multimeter input resistance you use to measure voltages ?

 

[ericmar]

I dont say that you have to connect them to same voltasge source, but for test purposes connect them to supply with same voltage, also check that these voltages are clean from noise - put HF and LF filtering capasitors per each voltage and check those sources by scope. May be if such noise and due to diode capasitance you can get wrong results .


Could you disconnect AD8601 pin 3 from R1-pin5 IL300F connection and measure voltages:
a. pin 4 IL300F
b. pin 3 of AD820 ?
c. pin 5 IL300F

Can you also tell us voltages of AVCC and +V_ISO.
Yet, what is multimeter input resistance you use to measure voltages ?

Now that I understand wat u mean by that.
Anyway, I am back to my home already so I'll try it tomorrow.
However, I can tell you that AVCC is +5V and +V_ISO is +8.94V (it should be +9V but I measure it and got +8.94V).