Open drain inout to FPGA voltage translator

Here is a schematic for TXB0106, and it is exactly the same for TXS0108. As a said this works only if the direction of the line remains always the same B->A. The 1k resistor is only there to avoid oscillations in the line. The picture I attached on the first post is for TXS0108 with 1k resistor between both sent_in, sent_out and chip.

I am starting to have some progress with the above configuration but the problem is that the line doesnt change direction from B->A so the 3V3 line (blue) remains constant as you can see in the picture.

Actually the minimum pulse width is defined by SENT Protocol and it is 8.55 us. Also in the oscilloscope you can see this. The rise time is some how distorted cause I said in the first post the pull-up resistor for 5 V should be 1.6 Kohm, but with 1.6 Kohm there are oscillations on the line. Another problem is that the chip doesn't bring the line to 0 V for the 5V region. By the way I tested the system using a FET as voltage translator and works fine.

Hi,

In order to diagnose the issue we should try to go through a few debug steps. If possible, could you try replacing the 3.3k pullup and 2.2k resistors with 0ohm resistors. Then, try testing the circuit using the 3.3k pullup and a 0ohm resistor in place of the 2.2k. Also, if you could provide scope shots of these two tests it would help us to resolve the issue. Thanks. 

Hi aozer,

I performed the described experiments but there was no success. Below the results (The first picture is with both resistors 0ohm and the second only with 2k2 replaced by 0ohm).

Hello,

Is it possible to force a DC high from the sensor side and a DC low from the sensor side and see how the line responds on the A side? It seems like with both resistors in place the translation from A to B functions but the translation from B to A does not. This will allow us to see if the B to A translation can respond with simple DC signals. 

Also, try leaving the 3.3k open to disconnect the pullup and see how the line responds. 

Actually I forced the sensor line to low and high by removing the sensor. And connect the line to power supply. You were right the B to A translation doesnt respont at all its always constant to 3.3V.

Without a pull-up there is a response but as you can see on the picture the signal on the sensor side looks terrible and is not what we are expecting. As I mentioned in mz first post the pull-up resistor on Sent line should be 1.6K and is defined by the sensor chip supplier. TXS0108 has two different values for the pull up resistor depending if the line is low or high and this is in my opinion what causes the problems. By putting a 3.3K pull-up in parallel with the internal resistor I thought it could solve the problem but it seems we have problems with the diriction regognition.

I am using an Engineering sample the chip is not in production yet. As I wrote before the system works fine when using FET as voltage translator even with 10k pull-up resistors. I'll give it a try with LSF010X, otherwise unfortunately I'll go with the FET translators.

Hi Christos ,

Attached app note for LSF explaining the theory of operation : www.ti.com/.../slva675b.pdf
it is a simple FET which is controlled using a master FET inside biased at the gate to VrefA+Vt. When input is high(VrefA) , the FET is off and pullup at the output pulls to any voltage between VrefA and VrefB.
Here is EVM user guide in case you want to purchase :www.ti.com/.../sdlu003a.pdf

I would be curious to know how it goes with the LSF family. LSF0101, LSF0102, LSF0108 can be considered. LSF0204 integrates 200kohm into the device. 

I have performed measurements with LSF10 and there is now a changing in direction of the signal, and I am also able to recognize the signals as high and low on the FPGA. The only problem as you can see from the graphs is that the line doesn't go to 0V but the lowest voltage is about 0.5 V.  I used LSF0101 from the evaluation board LSF010XEVM-001. This is the picture of the signal on the FPGA side.

That means that I have to increase the pull-up resistor. At the moment  am using the pull-up resistor provided by the board which is 300ohm. I'll give it a try on Monday. Thanks.

I have tested the system with 1.5k pull-up and works fine. Thank you very much for the help.