About DS10BR150

Greetings -

The DS10BR150 is a Buffer Repeater and supports DC to 1 Gbps (500MHz) operation.  100MHz Clock is fine.

I am checking on the failsafe feature and will advise.  If your source, is removed, powered-off, tri-state, and the DS10BR150 is on, failsafe might be desirable.  That can be done with an external pull up on the Plus, and a external pull down on the Minus, the 100 Ohm is internal.  Set for a nominal common-mode point of the driver you are using also and at +100mV bias.  This way the enabled 'BR150 device will sit at a stable known state, high in this case.

John Goldie

DPS APPS / SVA

Well, 

thanks for the answer, but I haven't understood well how I can develop the failsafe. Which are the values of pull down and pull up resistors?

Greetings -

Here is a link to a detailed AP NOTE on Failsafe (AN-1194) http://www.ti.com/lit/an/snla051b/snla051b.pdf

If you have a 3.3V rail, try 1.5k pull up (plus to VCC) and 1.5k pull down (minus to GND), and assuming you only have the 100 Ohm across the pair (no source termination). This should give you about 1.1mA bias, and that across the 100 Ohm is 110mV Bias for a known HIGH outputs state.

John

Yeah,

I've already read the Application note but it explains all the procedure starting from the knowledge of (look at Fig.1)

1) VTC of device;

2) value of Vfsb.

Now I want to know how I can find the Vfsb, what it is and what is its aim, assuming that VTC is the same also for DS10BR150, since it isn't shown on the datasheet.

Thank you 

Thank you very much, John! 

Do you think that a 50 mV bias Voltage could be enough? So, if the bias is >100 mV we are sure that output logic value is high but this causes more distortion on the signal. Since VTC is about 0V and noise is not so high, I think is better set a 50 mV bias. What's your opinion?

Thanks,

Hi -

Spec wise, biasing to 100mV is needed. (note 100mV comes from the LVDS standard not the part)

Since the thresholds are very tight, many do bias at >50mV without issue, this will distort less as you note, and is likely fine for the application.

John Goldie

DPS APPS / SVA

Does this part provide pre-emphasis? I am seeing significant overshoot with this part. I am driving

across 12" on a PCB and then a 3" LCD cable.

I have been using 1.5k failsafe resistors in my baseline design, but have also experimented

with 5k and 8k resistors per the AN-1994 app note.

It would also be nice if there was a version of this part with 100 Ohm resistor on the input but not

on the output. In my case the DS10BR150 is driving a LCD which has its own resistor.

Hi Brady,

The output resistor on this part helps to match the output impedance with the transmission line impedance.  This is necessary for good transmission characteristics.

The DS10BR150 does not include pre-emphasis.  Where are you measuring the waveform?  The channel characteristics between the BR150 output and the LCD could be causing the overshoot you are observing.

Regards,

Lee

I have probed the the output of the DS10BR150 buffer, and at both ends of the LCD cable. The signal looks good at the very end termination on the LCD but there's overshoot at the buffer output and at the start of the LCD cable. I am using a high BW active diff probe.

We are using .1uF and 4700pF decoupling caps for the buffers. The eval board for this part uses a .1uF and a 10uF.

If I disconnect the LCD cable which removes the end 100 Ohm termination, I still get a switching output since the buffer has an output resistor. Of course this signal is very distorted and has a larger amplitude.