> > Over the past couple of weeks, I've randomly noticed on several > > occasions that some of my various ICs (MOSFET gate drivers, 74HC595, > > opamps) are getting very hot to the touch. > Occam's Razor says that the preferred explanation is that the voltage > regulator is oscillating. (1) An aside: Looking at the discussion since I wrote that I see that I misread the original statement. I read " some of my various ICs ... are getting very hot to the touch" as including the gate drivers but not only the gate drivers. If the heating is in fact specific to one type of IC (as appears to be the case) Occam may change his opinion :-). (2) MAIN POINT POWER SUPPLIES DO NOT SUCK (usually) A number of people have subsequently noted that "cheap bench supplies" are able to source but not sink current at their set voltage". This is usually true and is an easy trap to fall into BUT This is not necessarily limited to cheap supplies. You have been sold a "power supply". The sign of the power supplied is usually considered to be positive unless explicitly stated in the specification. On a little consideration this failure to absorb energy AND maintain set point voltage AND not emit smoke should come as no surprise - but often does. I have been caught out by this in the past. On occasion the supply may have 'guts enough' to emit smoke rather than allow the voltage to rise above some point. I've seen that happen :-). Sometimes a circuit which is not designed to return energy to the supply and which is not obviously going to do so, will do so out of spite. This can happen when eg load current surges exceed immediate capability to supply them (whether from the supply or an external filter capacitor) and the supplied voltage droops briefly and then restores when the load peak passes. An inductive element handling significant energy may be made to "ring" under these conditions and may return energy to its local power rail and thence to the psu. I have had circuits that will cause the power supply to go temporarily insane with the inbuilt voltmeter reading horrendous and hopefully incorrect values when too small a filter capacitor is used external to the power supply. This reaction may introduce the need for a delicate balance to avoid the psu and circuit dancing together. eg if the psu is set to limit current to protect a prototype circuit BUT the circuit interacts with the power supply if the current limit is exceeded SO that an external filter cap must be used to handle current peaks THEN the capacitor may allow the "protected" circuit to die from over current peaks :-). Solutions must be on a case by case basis. _____________ If you want a +/- power supply you need to buy one or make one. If you don't want your + power supply to fail at being a - power supply you need to check that it can handle excess energy gracefully or, if it can't, take steps to prevent it being exposed to the need to act as a load. If the supply handles offers of power by ignoring them and you need to deal with regenerated or other "negative" sourced power then you need to make explicit arrangements to deal with the energy. Failure to do so can lead to death of equipment, power supply or (just possibly) you. As VG (I think it was) said, you can add a series diode to the PSU output. This of course adds a voltage drop and somewhat wrecks the regulation. How much this matters is application dependant. If you do this you must then deal with the excess energy. Diode voltage droop can be handled by extending the Vout sense point to outside the diode - best done by mounting the diode inside the supply. This may be as simple as adding the diode in place of a hard circuit connection internally but this also may mess up the supply's sensing circuitry so must be on a supply by supply bases. The behaviour of the supply when the now undriven but externally sensed voltage rises above set point also needs checking. An easy way to handle regenerative return up to a known power level is to add external load resistors on the bus. These are usually supplied by the psu but when it shuts down under small over-voltage the resistors dissipate all the regenerative energy. Doing this wrecks the internal psu current meter reading due to resistor current and increases the power level the supply is required to supply. It is however often quite an effective solution and the negatives may be acceptable. If mild rise is acceptable then a zener diode or electronic load bank may be acceptable. An electronic load which draws current only when Vout suddenly exceeds Vrecent (eg sense line via an RC filter) may suffice and is easily implemented. If Vpsu is usually fixed or seldom varied this may be an excellent solution. If Vsupply is always fixed during any one operation an electronic load can be adjusted to operate "just above" Vsupply. --=20 http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive View/change your membership options at http://mailman.mit.edu/mailman/listinfo/piclist .