LTSpice Automation
in [Design]
|
Prev: 240V AC power switch - based on current drawn from outlet
Next: Heatsinking surface mount high power MOSFETS
From: Joerg on 17 Feb 2010 20:37 Jim Thompson wrote: > On Wed, 17 Feb 2010 16:51:58 -0800, Joerg <invalid(a)invalid.invalid> > wrote: > >> Fred Bartoli wrote: >>> Joerg a �crit : >>>> Fred Bartoli wrote: >>>>> Joerg a �crit : >>>>>> Fred Bartoli wrote: >>>>>>> Fred Abse a �crit : >>>>>>>> On Wed, 17 Feb 2010 07:26:20 -0800, JosephKK wrote: >>>>>>>> >>>>>>>>> Geez, did everybody forget the three phase rectifier efficiency >>>>>>>>> that Fred Bartoli did for me? Note the stepped load on the right. >>>>>>>> T'warn't Fred Bartoli,t'was I :-) >>>>>>>> >>>>>>> Oh, I thought even I forgot about it :-) >>>>>>> >>>>>>> Well, I don't use much LTspice thanks to its poor GUI, so I >>>>>>> couldn't have done it... >>>>>>> >>>>>> You guys are spoiled, or too young. Back when I started the "GUI" of >>>>>> PSpice consisted of a rather small green CRT hanging off some CGA >>>>>> card. I had the deluxe edition, a CRT in nicotine-yellow :-) >>>>>> >>>>> Well, I did lots of *real design* before committing spice, which >>>>> happened on the late, maybe 10 years ago... >>>>> >>>>> I often do complicated projects where many design options have to be >>>>> studied, and when the GUI stands in the way, it's a no-no. >>>>> >>>>> The small sensor I'm finishing right now is a good example. >>>>> As was also the PSU that powers the DUTs on testers used to wafer >>>>> test the INTEL and AMD CPUs. (Yep, all your nice CPUs are tested by >>>>> my design :-) >>>> >>>> I always knew that the Romans never completely subdued you guys and >>>> you'd come back :-) >>>> >>>> >>>>> When I did that (circa 2000) I estimated it would have taken 3 to 4 >>>>> more time on the design pass with another spice, not even speaking of >>>>> LTspice. >>>>> >>>> Much of my stuff can't even possibly be simulated. Like a lot of >>>> transducers. >>>> >>> Sure it can. Well, most of it. That's just a matter of model development >>> and will it pay back to do so or not. And when it's done, it's done :-) >>> >> We are doing that as I write this, or rather, a client does. But it has >> its limits. If you want to perfect it you may slip the time-to-market >> schedule too much. In the old days it was even worse, once we blew up a >> Pentium-I during a sim (back when those were just out and really >> expensive). The client wasn't exactly happy when that happened. >> >> >>> I usually have big time developing adequate models (well, less and less >>> since I now have plenty). Then, with realistic models you gain a lot of >>> insight and have much better designs with less bench work. >>> For ex. it's easier to measure components at high temp, use spice and >>> have it right first time, than discovering pbs when baking the proto, >>> even without saying that investigating at high temp is slow and painful. >>> >> If you design super small stuff with lots of electronics, yes. Poring >> over one of those right now, custom chips with a few hundred channels >> that must fit into a little less than 2.5mm diameter tube. Flip-chip >> bonding these will be fun. > > I didn't do ANY simulation until around 1977-78. Prior to that was > all hand drawings and math solving; and bread-boarding... so that > covers all of my Motorola standard I/C products without a single > simulation. > > I still "think by hand", but I don't know how you can get by without a > simulator when you're talking literally thousands of transistors. > This one will have tens of thousands ... <bead of sweat dropping off forehead> -- Regards, Joerg http://www.analogconsultants.com/ "gmail" domain blocked because of excessive spam. Use another domain or send PM.
From: Phil Hobbs on 17 Feb 2010 20:43 On 2/17/2010 7:35 PM, Joerg wrote: > Jim Thompson wrote: >> On Wed, 17 Feb 2010 16:10:13 -0800, Joerg <invalid(a)invalid.invalid> >> wrote: >> >>> Jim Thompson wrote: >>>> On Wed, 17 Feb 2010 14:25:26 -0800, Joerg <invalid(a)invalid.invalid> >>>> wrote: >>>> >>>>> Fred Bartoli wrote: >>>>>> Fred Abse a �crit : >>>>>>> On Wed, 17 Feb 2010 07:26:20 -0800, JosephKK wrote: >>>>>>> >>>>>>>> Geez, did everybody forget the three phase rectifier efficiency >>>>>>>> that Fred Bartoli did for me? Note the stepped load on the right. >>>>>>> T'warn't Fred Bartoli,t'was I :-) >>>>>>> >>>>>> Oh, I thought even I forgot about it :-) >>>>>> >>>>>> Well, I don't use much LTspice thanks to its poor GUI, so I >>>>>> couldn't have done it... >>>>>> >>>>> You guys are spoiled, or too young. Back when I started the "GUI" >>>>> of PSpice consisted of a rather small green CRT hanging off some >>>>> CGA card. I had the deluxe edition, a CRT in nicotine-yellow :-) >>>> Sheeeesh! When I started using Spice I drew schematics on paper pads, >>>> numbered the nodes, typed in the netlist and ran it under DOS. >>>> >>>> Aaron eased my pain by writing a pre/post version controller which >>>> numbered all the .CIR and .DAT files so I could keep track of all the >>>> changes. >>>> >>>> Data spewed forth from a tractor feed printer: >>>> >>>> .001 * >>>> .002 * >>>> .003 * >>>> .004 * >>>> >>>> etc. Anyone else remember those days? >>>> >>> A friend of mine tried this with a Commodore daisy wheel printer, >>> same that I used to have. It could do microstepping and he just used >>> the dot. Which consequently wore out real fast ... >>> >>> My first forays in to computing was writing Fortran. Using a Juki >>> punch card machine. >> >> Summer of 1961 I took a sequence of courses on the Institute's first >> transistorized mainframe (IIRC IBM 709 ?), FAP (machine language) and >> FORTRAN... lots of punched cards... don't drop your stack ;-) >> > > Old trick: _Immediately_ number the cards after taking off the machine. > Else on turfing during the bicycle ride and it's all over. > You guys obviously went to the wrong school. We just shoved the cards through the verifier, and it not only numbered but decoded and printed them. Of course that was 1977, and I was the last one using cards at the entire school--I thought they were fun and didn't want to miss the party. Cheers Phil Hobbs -- Dr Philip C D Hobbs Principal ElectroOptical Innovations 55 Orchard Rd Briarcliff Manor NY 10510 845-480-2058 email: hobbs at electrooptical dot net http://electrooptical.net
From: Jim Thompson on 17 Feb 2010 21:02 On Wed, 17 Feb 2010 17:37:57 -0800, Joerg <invalid(a)invalid.invalid> wrote: >Jim Thompson wrote: >> On Wed, 17 Feb 2010 16:51:58 -0800, Joerg <invalid(a)invalid.invalid> >> wrote: >> >>> Fred Bartoli wrote: >>>> Joerg a �crit : >>>>> Fred Bartoli wrote: >>>>>> Joerg a �crit : >>>>>>> Fred Bartoli wrote: >>>>>>>> Fred Abse a �crit : >>>>>>>>> On Wed, 17 Feb 2010 07:26:20 -0800, JosephKK wrote: >>>>>>>>> >>>>>>>>>> Geez, did everybody forget the three phase rectifier efficiency >>>>>>>>>> that Fred Bartoli did for me? Note the stepped load on the right. >>>>>>>>> T'warn't Fred Bartoli,t'was I :-) >>>>>>>>> >>>>>>>> Oh, I thought even I forgot about it :-) >>>>>>>> >>>>>>>> Well, I don't use much LTspice thanks to its poor GUI, so I >>>>>>>> couldn't have done it... >>>>>>>> >>>>>>> You guys are spoiled, or too young. Back when I started the "GUI" of >>>>>>> PSpice consisted of a rather small green CRT hanging off some CGA >>>>>>> card. I had the deluxe edition, a CRT in nicotine-yellow :-) >>>>>>> >>>>>> Well, I did lots of *real design* before committing spice, which >>>>>> happened on the late, maybe 10 years ago... >>>>>> >>>>>> I often do complicated projects where many design options have to be >>>>>> studied, and when the GUI stands in the way, it's a no-no. >>>>>> >>>>>> The small sensor I'm finishing right now is a good example. >>>>>> As was also the PSU that powers the DUTs on testers used to wafer >>>>>> test the INTEL and AMD CPUs. (Yep, all your nice CPUs are tested by >>>>>> my design :-) >>>>> >>>>> I always knew that the Romans never completely subdued you guys and >>>>> you'd come back :-) >>>>> >>>>> >>>>>> When I did that (circa 2000) I estimated it would have taken 3 to 4 >>>>>> more time on the design pass with another spice, not even speaking of >>>>>> LTspice. >>>>>> >>>>> Much of my stuff can't even possibly be simulated. Like a lot of >>>>> transducers. >>>>> >>>> Sure it can. Well, most of it. That's just a matter of model development >>>> and will it pay back to do so or not. And when it's done, it's done :-) >>>> >>> We are doing that as I write this, or rather, a client does. But it has >>> its limits. If you want to perfect it you may slip the time-to-market >>> schedule too much. In the old days it was even worse, once we blew up a >>> Pentium-I during a sim (back when those were just out and really >>> expensive). The client wasn't exactly happy when that happened. >>> >>> >>>> I usually have big time developing adequate models (well, less and less >>>> since I now have plenty). Then, with realistic models you gain a lot of >>>> insight and have much better designs with less bench work. >>>> For ex. it's easier to measure components at high temp, use spice and >>>> have it right first time, than discovering pbs when baking the proto, >>>> even without saying that investigating at high temp is slow and painful. >>>> >>> If you design super small stuff with lots of electronics, yes. Poring >>> over one of those right now, custom chips with a few hundred channels >>> that must fit into a little less than 2.5mm diameter tube. Flip-chip >>> bonding these will be fun. >> >> I didn't do ANY simulation until around 1977-78. Prior to that was >> all hand drawings and math solving; and bread-boarding... so that >> covers all of my Motorola standard I/C products without a single >> simulation. >> >> I still "think by hand", but I don't know how you can get by without a >> simulator when you're talking literally thousands of transistors. >> > >This one will have tens of thousands ... > ><bead of sweat dropping off forehead> Characterize, then convert the digital to behavioral models, then it'll be easy ;-) ...Jim Thompson -- | James E.Thompson, CTO | mens | | Analog Innovations, Inc. | et | | Analog/Mixed-Signal ASIC's and Discrete Systems | manus | | Phoenix, Arizona 85048 Skype: Contacts Only | | | Voice:(480)460-2350 Fax: Available upon request | Brass Rat | | E-mail Icon at http://www.analog-innovations.com | 1962 | I love to cook with wine. Sometimes I even put it in the food.
From: JosephKK on 17 Feb 2010 22:46 On Wed, 17 Feb 2010 12:14:27 -0800, Fred Abse <excretatauris(a)invalid.invalid> wrote: >On Wed, 17 Feb 2010 07:26:20 -0800, JosephKK wrote: > >> Geez, did everybody forget the three phase rectifier efficiency that >> Fred Bartoli did for me? Note the stepped load on the right. > >T'warn't Fred Bartoli,t'was I :-) OOOPS! And thanks again, i am getting my Fourier analyses and THD numbers versus load now.
From: JosephKK on 17 Feb 2010 22:50
On Wed, 17 Feb 2010 15:33:11 -0700, Jim Thompson <To-Email-Use-The-Envelope-Icon(a)My-Web-Site.com> wrote: >On Wed, 17 Feb 2010 14:25:26 -0800, Joerg <invalid(a)invalid.invalid> >wrote: > >>Fred Bartoli wrote: >>> Fred Abse a écrit : >>>> On Wed, 17 Feb 2010 07:26:20 -0800, JosephKK wrote: >>>> >>>>> Geez, did everybody forget the three phase rectifier efficiency that >>>>> Fred Bartoli did for me? Note the stepped load on the right. >>>> >>>> T'warn't Fred Bartoli,t'was I :-) >>>> >>> >>> Oh, I thought even I forgot about it :-) >>> >>> Well, I don't use much LTspice thanks to its poor GUI, so I couldn't >>> have done it... >>> >> >>You guys are spoiled, or too young. Back when I started the "GUI" of >>PSpice consisted of a rather small green CRT hanging off some CGA card. >>I had the deluxe edition, a CRT in nicotine-yellow :-) > >Sheeeesh! When I started using Spice I drew schematics on paper pads, >numbered the nodes, typed in the netlist and ran it under DOS. > >Aaron eased my pain by writing a pre/post version controller which >numbered all the .CIR and .DAT files so I could keep track of all the >changes. > >Data spewed forth from a tractor feed printer: > >.001 * >.002 * >.003 * >.004 * > >etc. Anyone else remember those days? > > ...Jim Thompson Yep, but my spice versions were ISspice 1.94 and ISSpice 386/4 I also had tabular numeric data. Core engine Spice 2f4 and 2g6 IIRC. |