TUSB544: I2C Configuration for 4 Channel Custom Alt Mode, Source side

Gary

Does the lane ordering matter in your application? Is LN0/LN1 one pair and LN2/LN3 the second pair? If LN0/LN1 is independent from LN2/LN3, can you please set to two lane customer alt mode with flip (HLHHH)?

Thanks
David

David,

The lanes are independant between LN0/LN1 and LN2/LN3. In fact, for our testing, we connect DTX2 to DRX1 and DTX1 to DRX2 to wrap back the signals.

As stated I have tried HLHHL and this works, by setting register 0x0A to 0x13, I cannot see why changing FLIP would be appropriate as our source is connected to the upstream pins, and connector to downstream.

My question is why does HLHLL not work at all for our configuration?

Regards
Gary

Gary

For HLHHL, do you have DTX1 wrap back to DRX1? For HLHLL, can you repeat the same thing, DTX1 wrap back to DRX1, and DTX2 wrap back to DRX2?

Thanks
David

David,

As stated above: for our testing, we connect DTX2 to DRX1 and DTX1 to DRX2 to wrap back the signals. There is no easy method for us to change this setup.

Is this going to be a problem?

Regards

Gary

Gary

By proposing wrap DTX1 back to DRX1, I was trying to separating the issue between USB and Custom mode.

When configuring HLHHL, and with DTX2 wrap back to DRX1, and DTX1 wrap back to DRX2, can you please probe the output of URX2 and URX1 and see if there are data on them? I want to make sure the direction is correct in the configuration and whether the issue is no data or data corruption.

Thanks
David

Hi David,

When setup in HLHHL I can see data on URX2 and URX1, the eye looks good and IBERT testing reports there are no errors.

When configured as HLHLL, and in wrap back,  I have probed URX1 and URX2 and there is nothing on those pin: stuck low

Also, when configured as HLHLL. we drove Gbs into DRX2, UTX2, UTX1, DRX1 and saw zero output on URX2, DTX2, DTX1 and URX1

Regards

Gary

Gary

Are you AC or DC coupled the input and output of TUSB544?

'Also, when configured as HLHLL. we drove Gbs into DRX2, UTX2, UTX1, DRX1 and saw zero output on URX2, DTX2, DTX1 and URX1', can you please repeat the same experiment for configuration of HLHHL?

Thanks
David

David,

All inputs and outputs are AC coupled, 100nF.

The same experiment was repeated for configuration HLHHL and worked as expected

Regards

Gary

Gary

Can you please change URX1 to be driving and UTX1 to be receiving?

Thanks
David

David,

That is not possible for our design, we have URX1 directly connected to receiver only pins of a device, and same fixed tx function for the UTX connection

Regards

Gary

David,

With the register settings: 0x0A = 0x12, 0x0C = 0x52, 0x10 = 0x80, 0x11 = 0x80, 0x13 = 0x00, 0x20 = 0x0D, 0x21 = 0x0D

Then changing 0x0B = 0x00 ->0x08, to change RX1 direction, does not make any difference when we measure the signal on our expected output side of the redriver: no signal is coming out.

We did also try swapping rest of the channels changing, 0x0B -> 0x0F, and this also made no difference to the other channels. No signals are getting through the device.

Also note that in the working setup of 0x0A = 0x12, 0x0B = 0x00, 0x0C = 0x52, 0x10 = 0x80, 0x11 = 0x80, 0x13 = 0x00, 0x20 = 0x0D, 0x21 = 0x0D;

I changed 0x0B to 0x0F, which stopped our link from working, as expected, but when we put the value of 0x0B back to 0x00 the link did not recover, data was being sent to the redriver but there was no longer any signal being driven out?? Not sure why this is the case and the only way to recover is to power cycle the device. We did try reseting all registers by setting the I2C_RST bit in 0x1B but this did not help to recover the device.

Regards

Gary

Gary

I check this in the lab and verified all four lane paths work when 0x0A = 0x12 and 0x13, and 0x0C = 0x02.

Would you please make sure

HPDIN = H
SLP_SO# = H, you may want to toggle L as well
I2C_EN = H

Thanks
David