A high current Hoist Controller
2 January 2007: First publication.
This project is in the BETA testing phase, and so far, is published as an idea only!
This circuit was designed to replace the forward/reverse control switch used to control a hoist that lifts a elderly lady's wheelchair into and out of her motorcar. The bought system used a changeover switch which consistently failed after a couple of uses. A electronic switch was to be designed, one that can switch 100A under peak load.
MOSFET power transistors.
N channel MOSFETs for the high and low side switches.
Protection against conduction of both top and bottom FETS of the same half of the circuit.
A high quality image of the circuit can be found here ( 13KB ).
The circuit consists of a H bridge drive, three switches and a voltage doubler.
A 100KHz oscillator (U1A + U1B) feeds a set of current amplifying transistors ( Q1 and Q4 ). These in turn feed C1,D1,D2 and C2 to produce a 24V supply which is used to switch the high side FETs on.
H bridge driveN channel MOSFETS are used throughout because of their lower on resistance, cost and availability. The low side MOSFETS are directly controlled by the input signals via a 10K resistor. This resistor, combined with the large gate capacitance of the MOSFETS ensures that the voltage rises gradually. That makes this circuit perform badly under PWM conditions, but has other benefits. As the voltage passes 0.7V, long before the FET switches on, a 100K resistor enables a BJT to shut down the upper MOSFET transistor. That is how the H bridge circuit is protected against accidental current shunting.
Another BJT also shuts down the upper MOSFET unless the opposite side low end is driven on. This enables us to switch all the MOSFETS off, allowing for a free running motor.
If both channels are turned on, the low side FETS conduct, shorting the load and acting as an electronic break to dc motors.
MOSFETS have a inherent diode across the drain-source pins. Those diodes ensures that any spikes are fed back into the supply. As these diodes may not be fast enough to absorb the initial spike of a inductive load being switched off, a 1 Ohm and 100nF surge suppressing circuit is added to each load connection. In addition a 18V tranzorb is placed across the load. A couple of 1000uF capacitors shunts the next fastest spike and the supply is expected to shunt the rest.
First. heres how NOT to build the circuit.
The biggest problem facing the construction of this circuit is the connection of large current carrying conductors. This circuit requires the soldering of those conductors which is just asking for problems.
Allowing the power connectors to be bolted to the unit simplifies overall construction of the unit and simplifies maintenance significantly.
This circuit also separates the control and power circuits again simplifying construction.
The power circuit is produced by etching small traces of copper away from a off cut double sided board. Simply use tape to mask off the copper you need to keep.
Next etch the copper away using your favorite etchant.
Remember to tint the board when you're done as this will add valuable thickness to the conductors.
Large diameter copper wire soldered onto both the top and bottom of the PCB helps to carry the currents.
Because the FETs are not switched rapidly, this circuit is NOT suited for PWM or any other fast repetitive switching operation. Slow switching means lower noise generation and a huge up scaling potential by simply paralleling MOSFETS. The IRFZ48 would be the first choice but we had IRFZ44s in stock. :-)