ONET4201PA: Optical fiber receiving circuit

Greetings,

ONET4201PA can be used for 1Gbps applications as you noted. However, i am not sure about OPA855. I am assigning this case to Amplifier group for their comment.

Regards,,Nasser

Nasser san

Thank you immediately answer.

your message:
"I am assigning this case to Amplifier group for their comment.

About the above
Should I wait for an answer in this thread?

Regards, takahashi

Hi takahashi,

The OPA855 can support a transimpedance bandwidth >1GHz, but it will depend on the input capacitance of the photo diode and the gain resistor value required. How much current will the photodiode produce for a logic "1"? And what is the required output voltage for a logic "1"?

Sean Cashin san,

Thank you for your reply.

> The OPA855 can support a transimpedance bandwidth> 1GHz,

⇒Because the target is a communication rate of 1 Gbps
At a minimum, the transimpedance bandwidth of 625MHz or less

Must be supported.

> but it will depend on the input capacitance of the photo diode and the gain resistor value required.

⇒ What is "input capacitance of the photo diode"?
Is it the capacitance between the terminals of the photo diode?
This time, APD is used for the photo diode.
We are considering several types of APD, and the capacitance between terminals is 0.68 to 2pF.

> How much current will the photodiode produce for a logic "1"?

⇒The product to be developed has a linear motion between the transmitter and the receiver.
Therefore, the current generated by the photodiode depends on the distance between the transmitter and the receiver.

There is a target maximum communication distance, but there are various uses, so
We plan to evaluate it with a prototype.

> And what is the required output voltage for a logic "1"?

⇒ 3mVp-p.
The output of OPA855 is input to the limit amplifier ONET4201PA.
This is because the data sheet states that the ONET4201PA guarantees the level of differential CML output with an input of at least 3 mVp-p.

The OPA855 can support a transimpedance bandwidth> 1GHz,
However,
For this application
To increase the maximum flight distance between the transmitter and receiver,
Is it okay to imagine that if you try to transimpedance the weak current from the APD with high amplification, the frequency characteristics will deteriorate and the bandwidth that can be supported will be low?

Regards, takahashi

Hi takahashi, The OPA855 sounds like a good fit for this transimpedance amplifier application.

Yes, the capacitance across the diode will influence the amplifier gain.

0.68p to 2pF is reasonable for >1GHz bandwidth.

3mVpp sounds like a minimum value; it will be difficult to have a good signal to noise ratio with an output that small.

You may want to consider the LMH32401 if you need to switch the amplification for small signals and large signals, but it will not reach 1GHz.

Sean Cashin san,

thank you for your reply.

> The OPA855 sounds like a good fit for this transimpedance amplifier application.

⇒Adopt OPA855.

> 3mVpp sounds like a minimum value; it will be difficult to have a good signal to noise ratio with an output that small.

⇒ With the set distance between the transmitter and receiver, The frequency characteristic of OPA855 meets 1GHz,

From the noise signals, the distance at which RSSI at a level that can be recognized

as a correct signal can be obtained is likely to be the ability of this testing machine.

> You may want to consider the LMH32401 if you need to switch the amplification for small signals and large signals,

but it will not reach 1GHz.

⇒ Thank you for the information. This time, we will test with OPA855.

When configuring a circuit that inputs the output of OPA855 to the limit amplifier ONET4201PA,

Are there any other points to be aware of regarding OPA855 peripheral parts other than the following?

・ Connect between FB pin and IN- pin, Rf, Cf

・ Bias voltage supplied to IN +

data sheet: 10.2 Typical Application section The contents described in are different because the application is different.

Please teach me. Because TIA recognizes that the current input is voltage-converted and amplified.

We recognize that the APD "anode" can be directly connected to the IN-pin.

Do you need external parts on the OUT pin side as well?

Please confirm.

Regards, takahashi

Hi takahashi, 

The main concern is to keep the APD very close to IN-, since you need the APD capacitance and Cf to set the amplifier's high frequency gain above 7V/V. Any isolation inductance on the way to the APD will make the amplifier oscillate.

You can use a +Vbias and connect the anode, or a -Vbias and connect the cathode (this is how the LMH32401 is configured).

I think you will only need part of the circuit from figure 55. The circuit on page 1 might be more useful to you. There are no external parts needed on the output for the amplifier, but maybe you could use a comparator for the system-level design. 

Sean Cashin san,

thank you for your reply.

> The main concern is to keep the APD very close to IN-,

since you need the APD capacitance and Cf to set the amplifier's high frequency gain above 7V / V.

Any isolation inductance on the way to the APD will make the amplifier oscillate.

⇒ Is it okay to recognize that it is necessary to make the layout between the APD cathode and IN- very close to prevent the amplifier from oscillating?

> You can use a + Vbias and connect the anode, or a -Vbias and connect the cathode (this is how the LMH32401 is configured).

⇒ "(this is how the LMH32401 is configured)" Is OPA855 wrong about the above? Connect the anode using + Vbias.

> I think you will only need part of the circuit from figure 55.

The circuit on page 1 might be more useful to you.

There are no external parts needed on the output for the amplifier, but maybe you could use a comparator for the system -level design.

⇒ Thank you.

However, what is worrisome is impedance matching.

output:

The output of OPA855 is input to ONET4201PA.

The input of ONET4201PA has a 50Ω terminating resistor inside.

It is pulled up to VCC. Therefore, it is necessary to set the transmission line between OPA855 and ONET4201PA with 50Ω.

So what about the output of the OPA855?

It seems that it does not have a terminating resistor inside, so Do I need to terminate in parallel or in series with 50Ω?

input:

Is impedance matching between the APD cathodes --IN-necessary?

Regards, takahashi

Hi takahashi, Yes it is okay to recognize that they need to be very close. No, it is not wrong. Either direction can be used. 

You can put a 50Ω resistor in series with the OPA855 output for impedance matching. Impedance matching is not necessary on the input side because it is going to be a very short distance.