how "robust" a drive does a BJT require

[Simon]

Say I'm using a BJT to do PWM drive of a fan motor. Does a BJT have the same fast rise/fall requirements of a MOSFET ? The driver will be in a control box and the fan in a box some distance away. So I'd put the dive BJT in the fan case and leave the drive electronics in the control box. I could use an NPN drive BJT with a "pullup" PNP to drive it controlled from an NPN in the control box.

Mosfets are attractive but this needs to stay simple at the motor end with no fancy drive electronics and I'm concerned that mosfets may not be as robust in a 24V automotive environment.

[SeanB]

Not much required. You need a base emitter resistor to keep it off at high temperature, around 470R 0.5W will do. You will probably be using a darlington transistor as well, to have a 20A or so output, with a sub 50mA drive. You will probably use the same value as base resistor, it will allow 50mA to flow, which should saturate almost any darlington with a gain of over 1000. If the transistor you use has a lower gain drop the resistor value, but watch power dissipation in the resistor.

Add a thermal cutout to the heatsink, as it will get hot dissipating a few watts of heat.

All in all a mosfet looks good, just use a bypass diode for the body diode, and use a 1k resistor to tie the gate down, and a 1k resistor to limit input voltage to the gate. Add a 15V 0.4W zener across the gate as protection, and a 100V transorb across the drain source as well. Use a 100V mosfet as well with low on resistance and you need minimal heatsink, and the input will survive quite a lot of abuse. A one time 125C thermal cutout on the heatsink will help with faults, failing safe. Drive it with 24V from a totem pole output stage, and it will work. Choode a mosfet that will give desired current saturation though at 10V Vgs.

[Simon]

Yes a PNP darlington would be a good idea, I was hoping to just have the one driving transistor near the fan saving someone from the excuse to dream up another expensive electronics box (I won't be designing the electronics - they won't let me but I can lay the grounds for favorable solutions).

[Codemonkey]

How far away from the BJT/MOSFET are you planning on locating its driver circuit ? I'd be more converned about ensuring the drive signals (I'm assuming PWM) don't get distorted due to the cable lengths and any other interference in an automotive environment. Also, think about the EMC issues of having a long bit of wire with a fast rising/falling edge signals on it, unless sheilded its likely to be a good interference generator!

[Simon]

Yes that had been my worry. Suddenly however things have changed. I'm being asked to design the electronics and I now know that the electronics sits in the box next to the motor and there is a pot control signal so a constant voltage and no issue.

[SeanB]

So you can do low side switching now, which will make life a little easier. Route power through the box to put the fuse and thermal cutout in the positive line for added protection.

[Simon]

now that I have been shown what we have that is how the original unit works so I have to replicate that and it is of course the best option