DS92LV0422: About TRI-STATE of TXCLKOUT_ver2

Hi Hide,

They are correct that when in tri state the pin should not be connected to a current source. I checked on one of our EVMs, and did not see this 1.5V when in tristate. Is TXCLKOUT connected to a receiver when this is tested? The pin might be biased this way.

Regards,

Ben

Hi Ben-san,

Thank you for kind support.

Our customer said that TXCLKOUT don’t connect to a receiver.

And, I have two questions.

[Q1]

Could you please share your measurement result(waveform) at EVM?

So, I would like to know behavior of TXCLKOUT.

[Q2]

Could you please let us know the measurement point?

I understand the following image.

Is my understanding correct?

<My image>

Regards,

Hide

Hi Hide,

1. I will have to set this up again in the lab. I will try to set it up tomorrow and show you what I am measuring.
2. I measured on the header on the EVM, although your diagram looks correct.

Regards,

Ben

Hi Ben-san,

Thank you for kind support.

I’m looking forward to hearing from you.

Regards,

Hide

Hi Hide,

Sorry, I was not able to redo this in the lab today. Please give me a couple more days to get this done. 

Regards,

Ben

Hi Ben-san,

I understand.

I kindly ask for your ongoing support.

And, the following are additional information.

Could you please let us know if you have any concern or advice?

-If they change " OSS_SEL pin=High" to "OSS_SEL pin=OPEN(So internal pull down)", " around 1.5V " will disappear(to GND level).

-There is no path connected to TXCLKOUT form other circuit because TXCLKOUT don’t connect to a receiver.

Regards,

Hide

Hi Hide,

Hide said:

-If they change " OSS_SEL pin=High" to "OSS_SEL pin=OPEN(So internal pull down)", " around 1.5V " will disappear(to GND level).

I believe this is expected, as shown in the datasheet. I will update you with my lab findings hopefully tomorrow.

Regards,

Ben

Hi Ben-san,

Thank you for your reply.

I apologize if I caused any confusion.

The intent of the report(additional information) is as follows.

[Their intention]

They guess that TXCLKOUT isn’t TRI-STATE even if the conditions are yellow line of Table7 and OSC_SEL is “000”.

Because, they guess that if TXCLK OUT is TRI-STATE, the voltage of TXCLK OUT does not change with the behavior of OSS_SEL pin.

> If they change " OSS_SEL pin=High" to "OSS_SEL pin=OPEN(So internal pull down)", " around 1.5V " will disappear(to GND level).

However, I believe that TXCLKOUT is TRI-STATE at yellow line of Table7 and OSC_SEL is “000”.

So, I’m waiting for your evaluation results and comments.

I’d greatly appreciate your verification.

Regards,

Hide

Hi Ben-san,

Thank you for always kind support.

I understand if OSC_SEL is “000”, TXCLKOUT is TRI-STATE even if OSS_SEL pin is High or Low, so TXCLKOUT is same condition.

However, if they change " OSS_SEL pin=High" to "OSS_SEL pin=OPEN(So internal pull down)", " around 1.5V " will disappear(to GND level).

Could you please let us know this cause if you have any information?

Regards,

Hide

Hello,

I don't believe TXCLKOUT is necessarily tristate if OSC_SEL is "000". In order to use the Oscillator output, OSC_SEL must not be "000".

OSC_SEL controls the OSC output highlighted above. So if OSC_SEL is "000", TXCLKOUT doesn't necessarily have to be tristate, it just won't be able to use the clock output generated by the deserializer. Does this answer your question?

Regards,

Ben

Hi Ben-san,

Thank you for your comment.

I understand your explanation.

We guess that if Case 1 and Case 2 are the same TRI-STATE, the output behavior of TXCLKOUT is the same.

Case1:If OSS_SEL is High, OSC_SEL is “000”

Case2:If OSS_SEL is Low, OSC_SEL is “000”

However, as a result of their evaluation, there is a difference in the output of TXCLKOUT.

Case1: TXCLKOUT is around 1.5V

Case2: TXCLKOUT is GND level

So, they guess that if OSS_SEL is High, OSC_SEL is “000”, TXCLKOUT isn’t TRI-STATE.

Are their guess correct?

My answer is “Their idea is not correct”.

I believe that if OSS_SEL is High, OSC_SEL is “000”, TXCLKOUT is TRI-STATE.

Because, I have your comment.

https://e2e.ti.com/support/interface-group/interface/f/interface-forum/1033721/ds92lv0422-about-tri-state-of-txclkout/3823521#3823521

Do Case1 and Case2 have any different states even if they are TRI-STATE?

And, I have one more question for [Your comment].

If OSC_SEL is "000", is TXCLKOUT unstable?

So, is it possible that TXCLKOUT is not TRI-STATE?

[Your comment]

So if OSC_SEL is "000", TXCLKOUT doesn't necessarily have to be tristate.

<Background>

Currently, they are evaluating DS92LV0422.

Then, when OSS_SEL is High, OSC_SEL is “000” (This state is required for their application), excessive noise is observed on TXCLKOUT.

Therefore, they are investigating the cause of noise.

Regards,

Hide

Hi Ben-san,

Thank you for your prompt reply and detail information.

Could you reproduce Case2 at EVM?

Case1:If OSS_SEL is High, OSC_SEL is “000” ->TXCLKOUT is around 1.5V (You have already reproduced)

Case2:If OSS_SEL is Low, OSC_SEL is “000” -> TXCLKOUT is GND level

I guess that if can’t reproduce, this phenomenon is caused by their application.

If can reproduce, this phenomenon is caused by the internal circuit of device.

And, I understand that the meaning “X” of Table7 is “don’t care” so "H H L L" is yellow line.

Is my understanding correct?

If my understanding is wrong, is “H H L L” a deprecated setting?

Regards,

Hide

Hi Ben-san,

Thank you for detail information.

I understood as follows.

Is my understanding correct?

<My understanding>

If OSC_SEL is "000", TXCLKOUT is TRI-STATE(same condition) but output voltage level depends on OSS_SEL.

This is the correct behavior.

Case1:If OSS_SEL is High, OSC_SEL is “000” ->TXCLKOUT = TRI-STATE =around 1.5V(High)

Case2:If OSS_SEL is Low, OSC_SEL is “000” -> TXCLKOUT= TRI-STATE=GND level(Low)

[Datasheet comment]

the PDB pin is tied high and the deserializer enters power-down when the serial stream stops.

When the serial stream starts up again, the deserializer locks to the input stream, asserts the LOCK pin, and outputs valid data.

In power-down mode, the LVDS data and clock output states are determined by the OSS_SEL status.

Regards,

Hide

Hi Hide-san,

Yes your understanding looks mostly correct. "If OSC_SEL is "000", TXCLKOUT is TRI-STATE" is not fully correct, as there are other parameters for a TRISTATE output (as we discussed before with the state table). But generally I think you have the correct understanding of it. Do you have any further questions?

Regards,

Ben

Hi Ben-san,

I have one more question for your comment; “If OSC_SEL is "000", TXCLKOUT is TRI-STATE" is not fully correct”.

(Sorry for the repetition. I don't understand the nuances.)

I believe that TXCLKOUT is TRI-STATE at “OSC_SEL is 000”.

Is this thread comment correct?

https://e2e.ti.com/support/interface-group/interface/f/interface-forum/1033721/ds92lv0422-about-tri-state-of-txclkout/3823521#3823521

And, I understand that it is not desirable to check the status of output voltage of TXCLKOUT in the status of OSS_SEL(H or L).

The reason is <your comment>, so “If OSC_SEL is "000", TXCLKOUT is TRI-STATE" is not fully correct”.

Is my understanding correct?

<your comment>

In these cases, I don't think that simply changing OSS_SEL from H to L is necessarily a good way to determine a difference between states.

See the combinations in the below table which you previously referenced:

The inputs "H H H L" are listed, however "H H L L" is not a listed configuration, and I don't think it should be used to test the state of TXCLKOUT.

Regards,

Hide

Hello Hide-san,

No problem, I will try to explain more fully.

I think what you are referring to is this image:

The purpose of the above chart is actually to show the ability of TXCLKOUT to use the Oscillator as an output, rather than the clock received by the serial stream. If OSC_SEL was '000' the sections of "OSC Output" would not have any valid clock. So, if OSC_SEL is '000', TXCLKOUT would output the received input clock if there is one available, as you can see in this table (TXCLKOUT 'Active'):

This paragraph also describes what I am talking about:

Hide said:

And, I understand that it is not desirable to check the status of output voltage of TXCLKOUT in the status of OSS_SEL(H or L).

The reason is <your comment>, so “If OSC_SEL is "000", TXCLKOUT is TRI-STATE" is not fully correct”.

Is my understanding correct?

What I mean when I say not to use OSS_SEL  or OSC_SEL to determine whether TXCLKOUT is TRISTATE, I mean you should take into account the other parameters that impact the state of TXCLKOUT. These are shown in the table I posted earlier in this reply.

So, in summary:

• OSC_SEL controls the oscillator option, and not TXCLKOUT as a whole
• OSS_SEL controls the state of data and clock (TXCLKOUT) output when in power-down mode

I hope this helps!

Regards,

Ben

Hi Ben-san,

Thank you for always kind support.

My understanding is as follows.

-If PDB=H, OEN=H, OSS_SEL=H, LOCK=L, “OSC_SEL is 000”, TXCLKOUT is TRI-STATE.

-If TXCLKOUT is TRI-STATE at the above conditions, output voltage is changed by OSS_SEL.

 This is the correct behavior.

 Case1:If OSS_SEL is High, around 1.5V(High)

Case2:If OSS_SEL is Low, GND level(Low)

Regards,

Hide

Hi Ben-san,

Thank you for always kind support.

I greatly appreciate your cooperation.

Regards,

Hide