Let's Take A Vote...
in [Design]
|
From: John Devereux on 29 Jul 2010 10:10 John Larkin <jjlarkin(a)highNOTlandTHIStechnologyPART.com> writes: > On Thu, 29 Jul 2010 13:38:51 +0300 (EEST), Okkim Atnarivik > <Okkim.Atnarivik(a)twentyfout.fi.invalid> wrote: > >>John Larkin <jjlarkin(a)highnotlandthistechnologypart.com> wrote: >>: Do thermals matter to you? Latching relays are fabulous. As analog >>: switches, no semiconductor comes close. We recently measured the wiper >> >> Interestingly, in the LHe temperature OptoMOS switches can be closed >>but not opened. Switch-off relies on the charge leaking away from the >>MOSFET gate, and this leak obviously freezes. > > Maybe you're just not waiting long enough. A 2N7002 will keep itself > on or off, gate floating, for days. A cryo temps, that might extend to > a few million years. I'm impressed that they work at all. > > Possibly they use a silicon resisor for the pulldown, and the > resistance goes way, way up when it's cold. So it might turn off in a > few weeks. > > Optomos SSRs are great signal switches too. I recently blew up a bunch > of Clare parts, to find their voltage:current destruct limits. The > datasheets are horrible about that. Yes, pathetic for something clearly intended as an I/O component. I got no answer from them either about it. [...] -- John Devereux
From: John Larkin on 29 Jul 2010 10:14 On Thu, 29 Jul 2010 01:53:29 -0400, Phil Hobbs <pcdhSpamMeSenseless(a)electrooptical.net> wrote: >Robert Baer wrote: >> John Larkin wrote: >>> On Wed, 28 Jul 2010 00:52:53 -0700, Robert Baer >>> <robertbaer(a)localnet.com> wrote: >>> >>>> John Fields wrote: >>>>> On Tue, 27 Jul 2010 07:01:58 -0700, John Larkin >>>>> <jjlarkin(a)highNOTlandTHIStechnologyPART.com> wrote: >>>>> >>>>>> On Tue, 27 Jul 2010 04:00:22 -0500, John Fields >>>>>> <jfields(a)austininstruments.com> wrote: >>>>>> >>>>>> >>>>>>>> It has to do with >>>>>>>> getting SI units right. Did you ever read the wiki piece on >>>>>>>> dimensional analysis? Do you think it is smoke and mirrors? >>>>>>>> >>>>>>>> So, where did I say that charges can't generate forces? If you can't >>>>>>>> find such a statement, YOU are the one with emotions clouding your >>>>>>>> reason. >>>>>>> --- >>>>>>> Nonsense. >>>>>>> >>>>>>> All it means is that its location has slipped my mind, that the >>>>>>> message has been deleted or, who knows??? >>>>>> Who knows??? I know. You are deluded or just a liar. I would never say >>>>>> anything so silly. >>>>> --- >>>>> You would, you have, and you will again, so you're the liar. >>>>> >>>>> "Latching relays have infinite gain." is a pretty silly thing to say, >>>>> yes? >>>>> >>>>> >>>>> JF >>>> I think i "made a case" that the "gain" was not too hot, using >>>> rough numbers for input power to switch states, and power handling >>>> capability. >>>> For an infinite "gain", either the power to switch states must be >>>> zero, and/or the power handling capability must be infinite. >>>> Clearly, NEITHER exists. >>> >>> Power gain is Pload/(Pcoil*DutyCycle), where Dutycycle is the fraction >>> of time that the coil is energized. In plain English, power gain is >>> averaged load power divided by averaged coil power. That has no upper >>> bound as duty cycle approaches zero. In, say, a home thermostat that >>> uses one AA battery, Dutycycle might be a few tens of PPM, which is >>> why the battery will last a year or two. Probably the clock/LCD run >>> the battery down more than the relay does. >>> >>> So the argument devolves to whether a number that is unboundedly large >>> can be referred to as "infinite." Go for it. >>> >>> John >>> >>> >> By your own statement, you admit that the duty cycle IS NOT ZERO, and >> therefore there IS a bound. >> And "duty cycle" does not cut it; if so, one could take a very large >> (latching, if that "helps") relay and operate it *once* using its >> required 200KW of power, to control one microwatt of load - and >> "therefore" have an absurdly large "gain" based on the "duty cycle" of >> almost zero. >> > >The amount of bandwidth we've been wasting recently on exactifussitudes >like this makes arguing about angels dancing on the head of a pin seem >positively practical. > >Personally, I make an average of about six stupid mistakes before >breakfast, so I'm used to it by now. Fields has a private meaning for >the word 'force', and Larkin is using 'infinite' in a loose sense. As working engineers, we use a lot of terms in a loose sense. Like charge, average, infinite, heat, "Gaussian", power factor, Q, impedance, noise, exponential, "final", linear, all sorts of stuff that's mathematically imprecise. Because it's good enough to make things work. Somebody accused me here of not being a good scientist: guilty! John
From: John Larkin on 29 Jul 2010 11:39 On Thu, 29 Jul 2010 15:10:41 +0100, John Devereux <john(a)devereux.me.uk> wrote: >John Larkin <jjlarkin(a)highNOTlandTHIStechnologyPART.com> writes: > >> On Thu, 29 Jul 2010 13:38:51 +0300 (EEST), Okkim Atnarivik >> <Okkim.Atnarivik(a)twentyfout.fi.invalid> wrote: >> >>>John Larkin <jjlarkin(a)highnotlandthistechnologypart.com> wrote: >>>: Do thermals matter to you? Latching relays are fabulous. As analog >>>: switches, no semiconductor comes close. We recently measured the wiper >>> >>> Interestingly, in the LHe temperature OptoMOS switches can be closed >>>but not opened. Switch-off relies on the charge leaking away from the >>>MOSFET gate, and this leak obviously freezes. >> >> Maybe you're just not waiting long enough. A 2N7002 will keep itself >> on or off, gate floating, for days. A cryo temps, that might extend to >> a few million years. I'm impressed that they work at all. >> >> Possibly they use a silicon resisor for the pulldown, and the >> resistance goes way, way up when it's cold. So it might turn off in a >> few weeks. >> >> Optomos SSRs are great signal switches too. I recently blew up a bunch >> of Clare parts, to find their voltage:current destruct limits. The >> datasheets are horrible about that. > >Yes, pathetic for something clearly intended as an I/O component. I got >no answer from them either about it. > >[...] I have some crude point-of-destruction SOAR graphs if you're interested, on their CPC1008N part. After blowing a bunch up, I learned that you can look at waveforms and pretty accurately anticipate second breakdown (or whatever makes them fail) just before it happens. Plotting graphs is much faster when you don't have to replace the part every data point, like the civil engineering students breaking concrete beams. John
From: Phil Hobbs on 29 Jul 2010 11:46 John Larkin wrote: > On Thu, 29 Jul 2010 15:10:41 +0100, John Devereux > <john(a)devereux.me.uk> wrote: > >> John Larkin <jjlarkin(a)highNOTlandTHIStechnologyPART.com> writes: >> >>> On Thu, 29 Jul 2010 13:38:51 +0300 (EEST), Okkim Atnarivik >>> <Okkim.Atnarivik(a)twentyfout.fi.invalid> wrote: >>> >>>> John Larkin <jjlarkin(a)highnotlandthistechnologypart.com> wrote: >>>> : Do thermals matter to you? Latching relays are fabulous. As analog >>>> : switches, no semiconductor comes close. We recently measured the wiper >>>> >>>> Interestingly, in the LHe temperature OptoMOS switches can be closed >>>> but not opened. Switch-off relies on the charge leaking away from the >>>> MOSFET gate, and this leak obviously freezes. >>> Maybe you're just not waiting long enough. A 2N7002 will keep itself >>> on or off, gate floating, for days. A cryo temps, that might extend to >>> a few million years. I'm impressed that they work at all. >>> >>> Possibly they use a silicon resisor for the pulldown, and the >>> resistance goes way, way up when it's cold. So it might turn off in a >>> few weeks. >>> >>> Optomos SSRs are great signal switches too. I recently blew up a bunch >>> of Clare parts, to find their voltage:current destruct limits. The >>> datasheets are horrible about that. >> Yes, pathetic for something clearly intended as an I/O component. I got >> no answer from them either about it. >> >> [...] > > I have some crude point-of-destruction SOAR graphs if you're > interested, on their CPC1008N part. After blowing a bunch up, I > learned that you can look at waveforms and pretty accurately > anticipate second breakdown (or whatever makes them fail) just before > it happens. Plotting graphs is much faster when you don't have to > replace the part every data point, like the civil engineering students > breaking concrete beams. > > John > It's amazing that they remain civil--I mean, War between the States, sorry. Cheers Phil Hobbs (Recently saw the Confederate Correct-orrr episode of Rocky & Bullwinkle again. Brilliant.) -- Dr Philip C D Hobbs Principal ElectroOptical Innovations 55 Orchard Rd Briarcliff Manor NY 10510 845-480-2058 hobbs at electrooptical dot net http://electrooptical.net
From: John Larkin on 29 Jul 2010 12:00
On Thu, 29 Jul 2010 11:46:08 -0400, Phil Hobbs <pcdhSpamMeSenseless(a)electrooptical.net> wrote: >John Larkin wrote: >> On Thu, 29 Jul 2010 15:10:41 +0100, John Devereux >> <john(a)devereux.me.uk> wrote: >> >>> John Larkin <jjlarkin(a)highNOTlandTHIStechnologyPART.com> writes: >>> >>>> On Thu, 29 Jul 2010 13:38:51 +0300 (EEST), Okkim Atnarivik >>>> <Okkim.Atnarivik(a)twentyfout.fi.invalid> wrote: >>>> >>>>> John Larkin <jjlarkin(a)highnotlandthistechnologypart.com> wrote: >>>>> : Do thermals matter to you? Latching relays are fabulous. As analog >>>>> : switches, no semiconductor comes close. We recently measured the wiper >>>>> >>>>> Interestingly, in the LHe temperature OptoMOS switches can be closed >>>>> but not opened. Switch-off relies on the charge leaking away from the >>>>> MOSFET gate, and this leak obviously freezes. >>>> Maybe you're just not waiting long enough. A 2N7002 will keep itself >>>> on or off, gate floating, for days. A cryo temps, that might extend to >>>> a few million years. I'm impressed that they work at all. >>>> >>>> Possibly they use a silicon resisor for the pulldown, and the >>>> resistance goes way, way up when it's cold. So it might turn off in a >>>> few weeks. >>>> >>>> Optomos SSRs are great signal switches too. I recently blew up a bunch >>>> of Clare parts, to find their voltage:current destruct limits. The >>>> datasheets are horrible about that. >>> Yes, pathetic for something clearly intended as an I/O component. I got >>> no answer from them either about it. >>> >>> [...] >> >> I have some crude point-of-destruction SOAR graphs if you're >> interested, on their CPC1008N part. After blowing a bunch up, I >> learned that you can look at waveforms and pretty accurately >> anticipate second breakdown (or whatever makes them fail) just before >> it happens. Plotting graphs is much faster when you don't have to >> replace the part every data point, like the civil engineering students >> breaking concrete beams. >> >> John >> > >It's amazing that they remain civil--I mean, War between the States, sorry. > >Cheers > >Phil Hobbs > >(Recently saw the Confederate Correct-orrr episode of Rocky & Bullwinkle >again. Brilliant.) I have the complete R&B set on DVDs, but I haven't had time to see them all yet. I just got the full Monte Python set as a gift, further compounding my problems. I use to rush home from school to catch "Rocky the Flying Squirrel" before they changed the name to R&B. "How long did you scrooch them for, Gidney?" "Oh, just 30 seconds. Or was it 30 years...." John |