This isn't the same thing, but many years ago I did a H-bridge at=20 28Vdc , 15A using 2- T90 style relays for the positive DC rail and 2-=20 N-channel current-sense MOSFETs for the negative DC rail. The=20 application was a large DC linear actuator acting as a servo that=20 needed to operate anywhere from several times per minute to a few=20 times per hour. Timing was such that the relay closed first, then PWM the MOSFET to=20 full-on. Turn off was similar: turn the MOSFET off, wait, turn the=20 relay OFF, wait, turn both MOSFETs on for braking. We built many=20 hundreds of these units and all worked reliably. The major failure=20 point was the feedback position sensor, which we had no control over. dwayne At 05:35 AM 1/2/2015, RussellMc wrote: >Summary: What are the issues when using a semiconductor switch across an >electromechanical contact to (seek to) obtain higher voltage operation tha= n >the mechanical contact would otherwise allow and substantially lower power >dissipation than a a semiconductor switch alone will allow for a given all >up cost? DC operation is of primary interest, but AC gains should also be >possible. > >_____________________________ > >Electromechanical power relays are low cost for typical "domestic" AC >switching situations, have "long enough" lifetimes when used within spec >(100,000 + operations) and have an Rdson [tm] contact resistance) that mos= t >semiconductors struggle to meet - or do so only at very substantially >greater cost. . > >At 230 VAC (or 110) switching a 15A load a 10 milliOhm "contact resistance= " >dissipates 2.25 Watts. >And around 12 Watts at 50 milliOhm. >Semiconductor device suitable for these VI combinations are liable to need >substantial heatsinking and the all up cost starts to climb (device, >heatsink, mounting, housing, ...), but an eg 15A 230 VAC relay from a >reputable manufacturer is well under $10 in volume and needs no heatsinkin= g >or special feeding - they can be eg DIN rail mounted with push in >connections or eg lug mounted (2 bolts) with eg push on tags - quick, easy= , >low cost - choose any 3. PCB mounting is also common but slightly less >maintenance friendly if things do go wrong. > >The DC situation is not so good. >A 230 VAC 30A rated relay may be specified at 20A up to 28 VDC and then >derated very substantially as voltage increases - and is good for perhaps = a >few amps at 250 VDC or not rated at all for 250VDC operation. >A 230 VDC 20A relay is both rare and expensive. > >Once operated and the contact solidly closed a relay good for 20A at 28 VD= C >should (it seems) be good for 20A at 250 VDC if insulation issues are not = a >factor. >The problem is arc formation on break and, to a lesser but still >significant extent on make. Even with pure resistive loads this is a major >issue and any sort of inductive load component adds to the problem. >[[I can "arc weld" with feed from a single 30Vmp 250 Wmp PV panel - only >about 8A !!!. At eg 400 VDC (or higher) high Wattage DC is a somewhat >fearsome beast]]. > >An "obvious" solution (and many decades old) is to place a semiconductor >switch, say a MOSFET (or probably an IGBT at higher voltages), across the >contact to handle make/break (with high Rdson compared to the relay contac= t >resistance and short term large dissipation) with the mechanical contact >operating just after the FET and the FET being deactivated just after the >contact opens. The FET needs to be able to handle any inrush current, shor= t >term dissipation and and turn off transients. Even allowing for this, the >device 'size', cost and heatsinking should be modest compared to using a >semiconductor only solution. >Relay voltage ratings should be much enhanced. Even if the FET was left >activated throughout its dissipation would be minimal due to shunting by >the relay contact - allowing an extremely simple switching sequence. > >A web search for examples using various terms turned up very little. >Here is an example of AC switching with a TRIAC bypassing a relay contact > > http://www.echola.com/power/Hybrid-Relay.html > >Questions: > >How common is this? > >Any examples? > >What sort of performance gains are realised in practice for DC and AC >switching? > >Why is it a bad idea compared to alternatives? / What have I missed? > > > > Russell >-- >http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive >View/change your membership options at >http://mailman.mit.edu/mailman/listinfo/piclist --=20 Dwayne Reid Trinity Electronics Systems Ltd Edmonton, AB, CANADA (780) 489-3199 voice (780) 487-6397 fax www.trinity-electronics.com Custom Electronics Design and Manufacturing --=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 .