half bridge, noise supression and proper ettiquette

I need to control two 12v motors that draw ~8 amps each. They are going to be throttled way down using PWM and PID control, I only really need 20% or so of their output, but I can not change them for better motors. Now I already have some nice half-bridge chips in my parts drawer, BTN7970P. Since I don't need to reverse the motors I will use one chip per motor. My question is is it better to do the PWM and such on the high side or low side? I assume it makes a magical difference I am unaware of. The whole motor assembly will likely be a few feet from my control electronics, so I am thinking I may put the chips in a seperate enclousure mounted directly to the motors w/ ultrfast recovery diodes on nice short leads and run the PWM control logic over the long distance, rather than power, to keep noise in check.

thanks folks, the real question is, high side or low side?

the real question is, high side or low side

I would say it doesn’t matter.

For that device, the high-side on-resistance is slightly lower than the low-side on-resistance, according to the data sheet. So I would use high-side switching. But it makes very little difference.

permnoob: so I am thinking I may put the chips in a seperate enclousure mounted directly to the motors w/ ultrfast recovery diodes on nice short leads and run the PWM control logic over the long distance, rather than power, to keep noise in check.

Mounting near motor sounds good - but be careful with ground noise issues.

The devices already have fast recovery diodes built in as its a MOSFET half bridge surely? You might want to augment with power shottky rectifiers to reduce freewheel losses though. Transient suppression diode(s) and decoupling can all be in the enclosure too?

MarkT: The devices already have fast recovery diodes built in as its a MOSFET half bridge surely? You might want to augment with power shottky rectifiers to reduce freewheel losses though. Transient suppression diode(s) and decoupling can all be in the enclosure too?

The device is designed for active freewheeling so that the mosfet body diodes conduct for a very short time, keeping freewheel losses to a minimum. And I agree that decoupling should certainly be included in the enclosure.

Fantastic, I was leaning towards high-side due to the on resistance, good to hear there are no hidden “gotchas” in selecting hi/lo. gonna attempt to once again understand the losses due to freq of switching (I understand the theory, its the digging the values out of the data sheet that still evade me) and see if I should just go to the minimum (8hz iirc) or if the losses are low enough I can just keep it in whatever freq range keeps it quietest, mechanically and electrically.