On Thu, 20 Nov 1997 16:50:24 -0300 Walter Markiw writes: > I need to build an H bridge to drive a 1/8 HP DC 220 volt motor. >I started trying with >mosfets from International Rectifier (IR350).But I'm now facing >another >problem,probably because of some mistake I'm making in the circuit.I'm >driving the transistor with 15 volt in the gate,but somehow I can't >get the >voltage drop between collector an emiter to go under 12 volts The drive voltage is measured between the gate and source (there's no such thing as "emitter" and "collector" on a FET). It should be positive 12V or so to turn the device on, and zero or negative to turn it off. However, if the gate-source voltage is allowed to exceed +- 20V, the gate oxide will break down and the part will be ruined. I don't know what FETs damaged by such abuse act like, I suppose the primary symptom would be DC leakage through the gate. Test the resistance from gate to source with an ohmmeter, and if not "infinite", the FET is bad. Also the drain to source should measure practically "open" with the gate shorted to source. ,even >when it's >only 35 volts I am using to power a small 12V permanent magnet motor >with no >load,and I keep having this drop even if I'm turning a led on through >a 2,2 >kohm resistor. > You can imagine how hot the mosfet gets with that voltage drop and >the >inductive load of the motor. > Can you show me how the circuit should be like? I intend to build an >H >bridge for driving 220 volts of DC current.The only components I have >are some >opto-photo-darlingtons (the optocouplers work fine) which opto isolate >the >microprocessor signal from the mosfets Optodarlingtons are very slow. They will not pass a PWM signal well. In order to switch the FETs efficienty, they need a gate drive circuit capable of heavy currents. Though there are various chips built specifically for this function, the output stage of a 555 timer is a good readily-avaialable cheap substitute. Be sure to use a small (47 ohm) resistor between the 555 and the FET to prevent oscillation. It isn't very practical to drive the FETs directly through the optocouplers, unless very slow switching speed is acceptable. The drain voltage can't be used directly, as it is too high (and goes away when the FET comes on). Isolated power supplies for the drive will be needed. Fortunately the average current demand from them is small. ,a 16c74,and the mosfets of >course. > How should I build that bridge? > There is a fifth mosfet,between the bridge and ground,which is the >only one >working at 10 kherz (doing the Pulse width modulation).In that way,the >bridge >conducts pulsing current either on one direction or the other,to make >the motor run clock or counterclockwise at the desired speed. Usually the bridge itself is PWMed to control the speed. If you don't need to reverse the motor more often than a couple of times per minute, seriously consider using a mechanical relay to replace the H-bridge, and control the speed with a single-ended FET. Burning of the relay contacts can be eliminated by making sure the drive FET is off while changing the relay setting. In either case, diodes will be required to catch the voltage pulses resulting from the motor's inductance. With your proposed setup, 3 sources of FET drive will be required. The FETs at the top of the bridge (drains to 220 V rail) act as voltage followers, the drive voltage needs to vary from 0 to 232 (220+12V) volts. Note that the gate voltage needs to be raised above the drain voltage to get full conduction, and fall along with the source voltage all the way to 0 to stay in full cutoff. All the while, being sure not to blow out the gate with more than 20V positive or negative. This is a tough problem, considerably simplified by using P-channel FETs. With the sources now connected to the "constant" 220V, a drive of 220 to 208 V is used (0 to -12V relative to the 220V line). A single -12V supply relative to the 220V line can serve both. In very high current designs, since P-channel FETs have higher on resistance than N-channel ones, it is worth the extra complexity to have two "floating" supplies, one for each FET. Not so for the 1/8 hp motor. The FETs in the bottom of the bridge have the complication that the source voltage is controlld by the fifth FET (which is a bad idea). If you go with a full-bridge rather than relay reversing, get rid of the fifth FET and just PWM the two bottom ones. Since their sources are now grounded, a simple 0-12V driver will work them. In most cases, the drivers to the top FETs can still be controlled through slow optocouplers. They can stay saturated until the motor needs to reverse. Be sure not to drive the bottom FETs for a couple hundred us after changing the top ones. (If you're going for precise low speed control, where reverse drive needs to be available instantly on demand, then fast drivers on the top FETs would be needed as well). > This configuration worked fine for a 12 volt supply,but it somehow >didn't >for a 35 volt supply. The FET gates and optocouplers will withstand 12V, maybe 35V, but definitely not 220V. So different (more complicated) drive circuits are needed.