TXB0102: Problem with level translation using TXB0102DCUR

If R7 and R30 are not there, then the output was completely lost., But I am going to try to use the TXU0102 device, since it has much lower output impedance.
The TXB0102 has a 4K output impedance, and the SN74LVC2G126 has a 4pF input capacitance. so at 42Mhz input, most of the voltage is lost.

The TXB has edge accelerators, so you cannot use the DC output impedance value of 4 kΩ to compute its switching characteristics. The datasheet guarantees 50 MHz with a 15 pF load, so your load is likely to be higher.

Well, I guess I am driving 2 inputs of the SN74LVC2G126, simultaneously so it's probably 4pF || 4pF || signal trace to gnd capacitance.
So I probably am above this 15pF limitation.  Do you think that replacing the TXB0102 with a TXU0102 will solve my problem?

The signal trace length of RF_OUT+ and RF_OUT- are approx 17.2mm

The TXU is much better suited for unidirectional signals.

Hey Craig,

I am also wondering what type of probe is being used to test the output- is it the same probe as the input? Can you reduce the ground loop of the probes if possible? 

Please also see the TXU-EVM for further evaluation as needed.

Regards,

Jack 

Thanks Jack, 
   I soldered a tiny little wire to the metal jacket around my oscope probe tip, to use as ground, so I have the minimum ground loop.
If I don't do this, then the measured signal is horrible since this is a 1500 watt RF amplifier.

Hey Craig,

I see. Please also make sure that the termination resistance is set similar to the oscope + the bandwidth is at least three times higher then the highest operating data rate (in this case, the 42MHz CLK signal). 

Feel free to drop in after evaluating with the TXU to see if issue mitigates as well. 

Regards,

Jack