Inherited legacy L99H01 / L99H02 design TS/ACT_OFF pin is N/C

[sdtbb]

I have inherited a legacy H bridge motor control hardware and firmware based on L99H01 (I think L99H02 might be drop in replacement). I’m a firmware developer - not a hardware designer, and I’m new to motor control, so I’m familiarizing myself with the hardware and firmware, and I’m having a “how can this have ever worked” moment.

The TS/ACT_OFF pin is not connected (floating). 

The value 0xC4 is written to REG3 by the firmware.

EXT_TS-bit = high (thermal sensor interface), and EXTTH[5:0] is set to 000100b (n=7, m=3) , giving

Vth = 7* (0.31 + 3*0.03) = 2.8V.

Vcc is 3.3V.

Apparently this design has functioned correctly for over 10 years, but since we will shortly be spinning a new board and porting the firmware to a new (ST) MCU I’d like to understand if there is a fundamental problem with this design.  All of the original hardware and firmware developers are long departed (probably retired 😑).

Thanks in advance if for any feedback.

 

 

 

[Peter BENSCH]

Well, a floating TS/ACT_OFF pin is risky because its function depends on EXT_TS. With EXT_TS = 1, the pin is in thermal sensor mode; without a real diode chain, that is not the intended use. It may have “worked” thanks to leakage, noise margins, or accidental biasing of the floating node.

Recommendation: do not leave TS/ACT_OFF floating on the new board, but define it explicitly.

The value 0xC4 in REG3 is important because it sets RWD = 1 and EXT_TS = 1 and programmes EXTTH[5:0] = 000100b, which explicitly puts the device into thermosensor mode. However, without a proper diode chain, this is not usually the correct application.

If you are already doing a redesign, you should consider replacing this ancient L99H0x with something more modern, for example one with integrated MOSFETs?
What do you think about taking a look at the DC motor driver website or asking your local distributor?

Regards
/Peter

[sdtbb]

Peter, thanks for the quick reply. I agree that the next hardware spin should either include diodes, or tie TS/ACT_OFF to Vcc if it continues to be unused.

For the existing hardware would it be worth changing the REG3 configuration from 0xC4 to 0xFF, so that EXT_TS=1, but Vth evaluates to 0V ? Perhaps this would give the greatest margin against noise based triggering of the thermal sensor protection ?

I’m pretty certain that setting EXT_TS=0 is a non-starter, because there is no mention of an internal pullup on the TS/ACT_OFF pin, so I expect the motor driver would randomly cut out.