TXS0108E: looking for a TXS0108E IC capable of 12V

Thank you.  I did find this circuit while searching for a solution but I thought this only works with open drain and does not work with push/pull, do you know if that is correct?  I did order a handful of BS170's and put the circuit together but thus far haven't been able to make it work, on the RX side I see nothing but high.  I breadboarded 2 of these using the BS170s - on one I have the TX of my UART connect at LV1 (from your diagram), on the other I have the RX of my UART connected at LV1 and I have the two HV1s connected together.  Entirely possible I have something incorrect in my board but thus far it does not work, on my RX I only read high, never low.

For that reason I looked at the architecture of the TXS0108E to see if there were clues there as to how to do this in a way that would work for both.

The TXS has the same architecture, if you ignore the one-shot accelerators.

Such a MOSFET level shifter works well with open-drain signals, but it can also be used with push/pull signals. (The output of this shifter always is open drain.)

The BS170 wants to have a high gate voltage, but it should work, if the current is not too high.

LV and HV are the two power supplies, LV1 and HV1 are the signals. Please show a schematic.

This first schematic is the pair of BS170s with the high sides connected to each other, this is what I have on my bread board right now and was using to test if this MOSFET approach would actually work to shift between 12V & 3V3.  I do see the proper voltages on the high and low sides but thus far I have not been able to read (via RX) what I have sent (via TX) correctly.  I seem to be seeing all 1s/high.  I'm wondering if for this loopback testing I'm doing if I should just remove either R2 or R4?

This second schematic is how I would expect this to look "in production" with each BS170 serving a separate line.  Where I marked the lines as External Device TX and External Device RX are external connections between my device and the external device the circuit is intended to facilitate communications with.  I've included the TVS1400DRVR here because my device has 24 connections to external circuits that it uses for voltage monitoring and a TVS1400DRVR is used on each of those so I expect one will be used on these TX & RX communications lines as well.

At this point I am just trying to test that the MOSFET approach will work for the relatively low speed (4800 baud, perhaps up to 38400 baud) communications that I need to do.  Most of the external devices I'd be communicating with treat these 2 lines as TX & RX but some treat them as DATA & CLK meaning that at least one of the lines needs to capable of supporting bidirectional operations.  If it was always unidirectional I'd perhaps go with an optocoupler instead.

Let me know what you think, much appreciate the insight.

These look correct.

The sending device must be able to sink the current through all pull-up resistors; check the low-level voltage.

Can you look at the signal with an oscilloscope?

Hey Sean,

The LSF family of translators works exactly the same as a discrete MOSFET implementation, so you might find it helpful to review these videos (especially the up-translation and down-translation ones): 

Understanding the LSF family of bidirectional, multi-voltage level translators | TI.com Training Series