rusty@cadnetix.COM (Rusty Carruth) (06/20/89)
In article <4100@tekig4.LEN.TEK.COM> jime@tekig3.LEN.TEK.COM (James W. Edwards) writes: >I posted a description of a switching control problem a few days ago. >I have gotten quite a few responses. Several people have asked for >additional info. Also, I have been informed that without this additional >detail that I would probably get overdesigned solutions. To refresh >your memory I have included the original request along with the additional >detail. > > >>I have a problem for you electronics wizards. I would like to control >>the DC power being supplied to two loads as shown in the following diagram: >> >> +--[Load 1]-------------[CP1]-------------+-------+ >> | | | >> | | DC >> +--[Load 2]-------------[CP2]-------------+ POWER >> | | >> | | >> | | >> +-------------------------------------------------+ >> | ...etc..., you will need this info to make sense out of my reply Quick question: Why not just use a choke and a relay? Like so: cp<x> == ____()()()()()()()____ | \~ \ \o----------- ______ ) ) ) ______) where ()()... is the choke, and the other mess is the relay, use NO contacts if you want relay fail to leave loads on. I await all the reasons I know will be forthcoming to explain why this won't work :-). Oh, I suppose we WILL need to know if the load is purely resistive or not, won't we? (and finding the proper value of the choke is "left as an exercise for the reader" :-), translated, I'm too lazy right now to find the info I'd need to figure it out for you, and besides, I don't have all the info I'd need in order to figure it out anyway.... :-). Good luck. ---------- Rusty Carruth UUCP:{uunet,boulder}!cadnetix!rusty DOMAIN: rusty@cadnetix.com Daisy/Cadnetix Corp. (303) 444-8075\ 5775 Flatiron Pkwy. \ Boulder, Co 80301 Radio: N7IKQ 'home': P.O.B. 461 \ Lafayette, CO 80026
tomb@hplsla.HP.COM (Tom Bruhns) (06/30/89)
rusty@cadnetix.COM (Rusty Carruth) writes: >In article <4100@tekig4.LEN.TEK.COM> jime@tekig3.LEN.TEK.COM (James W. Edwards) writes: >>I have a problem for you electronics wizards. I would like to control >>the DC power being supplied to two loads as shown in the following diagram: >> >> +--[Load 1]-------------[CP1]-------------+-------+ >> | | | >> | | DC >> +--[Load 2]-------------[CP2]-------------+ POWER >> | | >> | | >> | | >> +-------------------------------------------------+ >> | >...etc..., you will need this info to make sense out of my reply > >Quick question: Why not just use a choke and a relay? Like so: > >cp<x> == ____()()()()()()()____ > | > \~ > \ > \o----------- > ______ > ) > ) > ) > ______) > >where ()()... is the choke, and the other mess is the relay, use >NO contacts if you want relay fail to leave loads on. > >I await all the reasons I know will be forthcoming to explain why this >won't work :-). Happy to oblige! (:-) :-) :-)) Actually, I have nothing against relays, but wonder why you recommend a _choke_ in series with the contacts! After all, an inductance tries very hard to maintain a constant current, and will accomodate large voltage changes if you try to change the current through it quickly. So what happens when you open the relay contacts? You get an arc. I learned this lesson the hard way with relays, many years ago. The relays were sealed surplus jobs, and I couldn't see what was going on. But I took one apart and found vaporized contact re-plated all over the inside! Admittedly, I was controlling current through a rather large inductance, with 120V open-circuit, which is far worse than what you suggest, but still, you would do well to put some arc-suppression across the contacts. One way to do this is to put a diode across the inductor, in this case (assuming the load itself is resistive). > >Oh, I suppose we WILL need to know if the load is purely resistive or not, >won't we? > >(and finding the proper value of the choke is "left as an exercise for >the reader" :-), translated, I'm too lazy right now to find the info >I'd need to figure it out for you, and besides, I don't have all >the info I'd need in order to figure it out anyway.... :-). > >Good luck. >---------- >Rusty Carruth UUCP:{uunet,boulder}!cadnetix!rusty DOMAIN: rusty@cadnetix.com >Daisy/Cadnetix Corp. (303) 444-8075\ 5775 Flatiron Pkwy. \ Boulder, Co 80301 >Radio: N7IKQ 'home': P.O.B. 461 \ Lafayette, CO 80026 >----------
jime@tekig3.LEN.TEK.COM (James W. Edwards) (07/23/89)
I posted a description of a switching control problem a few days ago. I have gotten quite a few responses. Several people have asked for additional info. Also, I have been informed that without this additional detail that I would probably get overdesigned solutions. To refresh your memory I have included the original request along with the additional detail. >I have a problem for you electronics wizards. I would like to control >the DC power being supplied to two loads as shown in the following diagram: > > +--[Load 1]-------------[CP1]-------------+-------+ > | | | > | | DC > +--[Load 2]-------------[CP2]-------------+ POWER > | | > | | > | | > +-------------------------------------------------+ > | > GND > >CP1/2 = Control Points 1/2 > >I would like to be able to turn on/off the power to either of >the loads through CP1/CP2. CP1/CP2 are simply whatever electrical >component(s) that you feel would be necessary to achieve this objective. >The switching action of CP1/2 must be controllable through a digital >signal. CP1/2 must each be able to switch 15A @ 12/24 V to their loads. >The switching action is actually more complicated than just switching >on/off: In the off position, the associated load must still >recieve some power (say 10%), and in the on position it will >recieve full power. Furthermore, when switching from off to on or >from on to off, the rise or fall in current/voltage to a load must >be controlled (for example, it could rise/fall exponentially to the >final voltage/current). The intended circuit must be very >reliable and operate in a harsh environment (-70F to 150F and >high, non-condensing humidity). If it matters, the power source >is regulated, but probably fairly noisy due to the surrounding >environment. Finally, it is very desirable that the most >likely failure mode will result in both loads recieving full >power (i.e. CP1/2 switched on). Thanks in advance! > Additional Detail: ================= 1). CP1 and CP2 will both use the same circuit design. I will be sequencing them on/off. I only showed both CP1 and CP2 for completeness. 2). It makes no difference if CP1/2 operate between the DC power source and the load, or the load and ground. Apparently, due to the popularity of N-type mosfets, the load and ground configuration is more desirable. 3). The load will be switched on/off at a regular rate (~1Hz) for extended periods of time. 4). A fairly large voltage drop across CPx is tolerable - for the intended purpose of this circuit I don't see any problem with up to a 1V drop. 5). Even though CPx is switching a regulated DC power source to the load, the load is not particularly sensitive to slow variations in the voltage. I'm just trying to avoid transients which could affect the load life. Thus, the ramp up/down spec in the voltage during switch on/off is not particularly tight - a resistor/cap network to slow down the edges is sufficient. 6). Reliability, simplicity, and low cost are all very desirable (aren't they always). Highest priority is for reliability. Think of this circuit in terms of controlling the headlights on your car (this is not the actual application) - if the circuit were to fail in the wrong way while you were driving at night, you might be in potentially dangerous situation. Thanks! Jim Edwards