Good job, wich brand are the drives and did you use them already?
But I don't understand how that idle current reduction can work.
It's based on C2 wich is charged instantly to 5V by Q1 during the step pulse and then discharged by the 10K resistor.
The TQ input will be driven high by the 2 inverters when the voltage of C2 reaches +-1.6V, the chip is now in 100% current mode.
But the timing constant for a 100n and 10K resistor is approx. 1msec. This means that the drive is only in 100% mode for a period of 1msec after reception of a step pulse. I think this is way to short for slow stepping, TQ will only remain high at step rates above +-1Khz. At 1/8 microstepping this would correspond to 0.6 RPS or 36 RPM before full torque is available.
I guestimate that it takes +- 100msec for the motor to complete and stabilise the actual move of a single step so the timing constant for the RC network should be at least 100 times higher.
All the above values are not exactly calculated but should be close, I ordered a drive to check this a few days ago.
I can't wait to see the waveform on the current resistors when stepping at something like 50-100 Hz and compare the low speed torque with the THB6064 drive.