ONET4211LD: Collapse of floodlight waveform

Greetings,

1). Based on these waveforms, it seems one level does not reach all the way to the top - as far as i can see.

2). Do you see an issue if using PRBS type of data pattern such as PRBS10?

Regards,,Nasser

Nasser san

thank you for your reply.

> 1). Based on these waveforms, it seems one level does not reach all the way to the top --as far as i can see.

⇒Yes. I am thinking the same.
In the circuit of data sheet P17 to check the modulation current
The voltage waveform of MOD- and MOD + was confirmed with a differential probe.
There was no problem.

> 2). Do you see an issue if using PRBS type of data pattern such as PRBS10?

⇒The data signal used to check the operation of ONET4211LD is
It is K code K28.5 of 8b10b.
(Repeat of 0011111010 and 1100000101)

There is one thing to worry about.
If the threshold set in R BIAS MAX is exceeded
I think the shutdown flag will be '1'.
If I make a mistake in setting R BIASMAX, will the shutdown flag be '1' before it looks like this waveform?

Nasser san

thank you for your reply.

Changed R BIAS MAX to make IBIAS MAX slightly larger than I TH MAX
I did, but the waveform did not change.
No failure detection has been issued.

I think it's natural,
Is it okay to recognize that the MODSET resistor needs to be set even when using the APC function?

Nasser san

thank you for your reply.

① Check the voltage between MOD + and MOD-pin with a differential probe.
You can check the well-organized signal waveform.
(2) When checking the voltage between MOD + and GND pin with a probe,
You can check the broken signal waveform.

The circuit I have confirmed to work is
This is a reference circuit for the data sheet p17 figure 25.
This circuit has a single-ended laser interface
I think it's a driving circuit.

Based on (1) above, it was judged that there was no problem with the laser drive current, but
It means that it is actually driven by the current waveform of ② by single-ended drive?

And I found the following in a past forum article:

e2e.ti.com/.../onet4211ld-onet4211ld-spice-model-for-laser-modulation-current-output ONET4211LD #

The draft documentation referenced in this article includes
A reference circuit for differentially driving the laser interface
It is described.

That this circuit can drive the laser with the well-organized signal waveform of ①
I'm looking forward to it.

Please give us your opinion on the above.

Regards,Takahashi

Nasser san

thank you for your reply.

> Internal to the part, device is driving the signal differentially. To get the optimum performance, output impedance on MOD + or MOD- must be balanced or the same.
⇒ In other words, even in the circuit of data sheet p17 figure 25
If the output impedance of MOD + and MOD- is the same,
Can I recognize that it can be driven by the differential waveform obtained by the measurement method (1) above?

> So we should be able to modulate the laser in single ended or differential mode.
⇒I can drive with DIN + and DIN- differential signals even in single mode, right?

> If we drive the laser single ended then the other end has to have a correct or matched impedance.
⇒ MOD- is connected to VCC via a resistor 20Ω.
MOD + is connected to VCC via a resistance of 15Ω + LD (internal resistance of 5Ω).
In other words, the impedances of MOD- and MOD + are almost the same.

> If you use your scope probe across MOD + / MOD-then you should see twice amplitude versus MOD + to GND.
⇒Yes. I understand that this is because two positive and negative voltages are added.

> Please note when using scope probe single ended i assume you are AC coupled.
⇒ When observing between MOD + and GND with a normal probe,
Is it okay to pay attention to observing in AC coupling mode?

I've been trying different methods over the last few days, but the problem hasn't been resolved.

Problem: When observing the waveform between MOD + and MOD- with a differential probe, a shaped waveform can be observed, whereas it can be observed.
When observing the floodlight waveform of LD via the APD module,
A broken waveform is observed.

The optical output (mW) of the LD was observed with an optical power meter.
There is a fluctuating offset as shown below.

I estimated that APC is not working properly and confirmed the following.

(1) Check the capacity of CAPC
The data sheet recommends 200nF.
However, with 625MHz modulation, I thought it was a bit small.
When I changed the capacity to 440nF, it improved as follows.

(2) Confirmation of BIA Spin
The inductor between the BIAS pin and the LD cathode is
I understand that it is for separating the modulation current and the BIAS current.
The inductor was previously recommended by Nasser san in another thread.
It implements Murata BLM03HG102SN1D.
220nF between the inductor temporary side (LD cathode side) and VCC A capacitor is connected.
Also, 1nF between BIAS pin and GND recommended in the data sheet.
A capacitor is mounted.

If you check the voltage waveform between BIAS pin and GND in the above connection,
As shown below, it does not oscillate, so I think there is no problem.

Due to the CAPC change in (1), the optical output became a little stable.
In (2), it was confirmed that the voltage of BIAS pin was stable.

In addition, the optical output ave (mW) measured by the optical power meter and The theoretical values ​​set for the R APCSET resistance are almost the same.

Optical factors such as the observation performance of the APD module and the disturbance of the projection waveform due to the return light of the laser are unthinkable.
Also, if MOD +, MOD- impedance mismatch is a problem,
I think that the differential current between MOD + and MOD- is disturbed.
But this is not disturbed.
I think that the disturbance of BIAS current becomes the disturbance of APC.
However, the voltage between BIAS and GND is not disturbed.

I am very troubled with this issue.

Nasser san

thank you for your reply.

Below figure ① After removing 220nF
Previously, when '1' was continuous, the waveform was broken, but it has been improved.
This 220nF needs to be connected between the BIAS pin (secondary inductor side) and the anode of the LD, while I was connecting it to the primary side of the inductor.

Comparing the input signal to DIN + and DIN- with the floodlight waveform,
'1' such as 010010 could not be observed.
It is estimated that this is because the set modulation current is large.
The resistance constant of ② was increased and the modulation current was reduced.
However, '1' such as 010010 could not be observed.

By removing 1nF of ③, '1' such as 010010 could be observed.

④ has been changed to 200nF recommended by the data sheet.

However, the waveform is still disturbed.

To improve,

(1) Adjustment of RD
(To make the impedance of MOD- and MOD + the same

(2) Inductor adjustment
(To separate MOD + BIAS

Please tell me if there are any other effective adjustment points for waveform shaping.

Nasser san

thank you for your reply.

Thank you for your advice and references on ringing.

There is another problem with ringing.

The data input to DIN + and DIN- of ONET4211LD is K28.5.
10-Bit Code RD– [0011111010]
10-Bit Code RD + [1100000101]

However, the waveform to be observed is lacking in '1',
The part that should be 0 is "1",
It has become different data.

(1) When R MODSET is relatively small and the modulation current is large
⇒ '1' is not enough

(2) When R MODSET is made several times as large as (1) and the modulation current is relatively small.
⇒The part that is originally '1' or the part that is '0' does not correspond

There is no problem with the differential current waveform between MOD + and MOD-.
Regarding LD specifications, it supports frequencies up to 1.5GHz.
(The case 625MHz

I don't think these are due to ringing due to impedance mismatch.

Before the quality of the waveform, I'm wondering why the output doesn't correspond to the input.

Impact of ringing?
(Implementation of Snubber circuit

Impact of MOD and BIAS separation?
(Inductor change

Is the APC or MOD setting wrong?
(Too much modulation current is flowing for the output limit of APC, etc.

Impedance shift between MOD + and MOD-?
(Fine adjustment of RD resistance

Regards,,Takahashi

Hi Takahashi-San,

1). For impedance shift, perhaps you can fine tune Rd or 20-ohm Resistor and bring it closer to 15-ohm to see if you see any improvement.

2). Is it possible to operate the device - as an experiment - in open-loop mode to make sure we are getting a stable operation? Then afterward we can enable APC using values we used for open-loop?

3). Would it be possible to operate the device in mid range or MOD and BIAS? Do we see any issue under this condition?

Regards,,Nasser

Nasser san

thank you for your reply.

> 1). For impedance shift, perhaps you can fine tune Rd or 20-ohm Resistor and bring it closer to 15-ohm to see if you see any improvement.

⇒I will try it.

> 2). Is it possible to operate the device --as an experiment --in open-loop mode to make sure we are getting a stable operation Then afterward we can enable APC using values ​​we used for open-loop?

⇒ Non-APC mode is Fix the APCSET pin to VCC with a 100kΩ resistor and I think it can be obtained by opening the PD pin. In non-APC mode, I would use RMODSET and RBIASMAX to determine the modulation and bias currents. At this time, IBIASMAX = ITHMAX, so I think RBIASMAX will be a fixed value. In other words, is it okay to recognize that the RMODSET is adjusted for the optimum output waveform?

> 3). Would it be possible to operate the device in mid range or MOD and BIAS? Do we see any issue under this condition?

⇒ Specifications of ONET4211LD.
Try to check the operation in the middle of the usable range of the modulation current and the bias current.
Does that mean that?

Regards, Takahashi

Nasser san

I'm sorry often.
Please give me the following two points of advice.

(1)
I disabled APC and set the bias current and modulation current to obtain the optimum projection current, but the optical output drops and the waveform changes as the operating time elapses.
I think this is due to the temperature characteristics.
To avoid this, I used the measured light output value and enabled APC, but the light output drops.
What is the cause of this?

(2)
I want to minimize the ringing of the floodlight waveform.
I implemented a Snubber circuit (R, C) between the cathode of the LD and GND, but it doesn't work.
As a method of setting constants,
The ringing frequency is measured, and the time constants of C and R are determined with reference to it.
Please tell me the best method.

Regards, Takahashi

Hi Takahashi-San,

I thought you had this working in APC mode and you were mainly working on optimizing the snubber circuit.

1). I agree perhaps either because of aging or temperature it cannot maintain either extinction ratio or the required bias current to stay within the laser threshold current. I think perhaps it is mostly because of the extinction ratio. Perhaps you can increase extinction ratio so you would have some margin over time - if possible.  

Given when you enabled APC the light output dropped, perhaps PD pin may not be connected. You had APC working before and now the overall setup perhaps has changed. If possible it would be good to compare.

2). I understand your reasoning for selecting R&C values but from what i understand these values are determined experimentally. These could highly depend on the Ferrite Bead between the Cathode and bias pin.

Regards,,Nasser

Nasser san

thank you for your reply.

> I thought you had this working in APC mode and you were mainly working on optimizing the snubber circuit.

⇒Yes. That's right.

> 1). I agree perhaps either because of aging or temperature it cannot maintain either extinction ratio or the required bias current to stay within the laser threshold current. I think perhaps it is mostly because of the extinction ratio. Perhaps you can increase extinction ratio so you would have some margin over time --if possible.

⇒ Does this indicate the possibility of improvement by reducing the modulation current?

> Given when you enabled APC the light output dropped, perhaps PD pin may not be connected. You had APC working before and now the overall setup perhaps has changed. If possible it would be good to compare.

⇒ PDpin is connected.
However, in order to proceed with various evaluations efficiently, the wiring between LD and PDpin is made a little longer than the original mounting.

> 2). I understand your reasoning for selecting R & C values ​​but from what i understand these values ​​are determined experimentally. These could highly depend on the Ferrite Bead between the Cathode and bias pin.

⇒ Is it okay to understand that the values ​​of C and R can only be tested in kind from the influence of the parasitic capacitance of the board and the optimum value is searched for?

Is it okay to recognize that the inductor between the BIAS and LD cathodes also affects ringing, and changing the inductor can improve ringing?

Regards, Takahashi

Hi Yoshihito,

Nasser is out until Tuesday.  I see this is marked as resolved.  Do you still require support?

Thanks,

Nicholaus

Nicholaus san

Thank you for contacting me.

It's not completely resolved,
I will proceed.

Thank you very much.

takahashi