Forums

polymer aluminum cap

Started by John Larkin September 17, 2012
There are nonlinearities when you really whack a cap.

The goal is to avoid the op amp with the 50 ohm series resistor. That is 
not remotely how the cap will be used in the application.

Your eval circuit doesn't reflect the application, thus testing the part 
in the actual application is the way to go if you don't want to be 
bothered building a pulser.


On Wed, 19 Sep 2012 12:44:09 -0700, miso <miso@sushi.com> wrote:

>There are nonlinearities when you really whack a cap.
Is ESR nonlinear? Is ESL nonlinear? How?
> >The goal is to avoid the op amp with the 50 ohm series resistor.
Your goal maybe. Certainly not mine. My goal was to put a current step into the cap and scope the resulting voltage step. That is
>not remotely how the cap will be used in the application.
Current is current. The cap could be used in all sorts of applications. Will ESR change as a function of nearby components?
> >Your eval circuit doesn't reflect the application, thus testing the part >in the actual application is the way to go if you don't want to be >bothered building a pulser. >
Testing the part in the application gives me no baseline for design and simulation, tells me nothing about margins, makes it hard to optimize after the parts are selected and soldered to the board. The tests that I did gave me a pretty good idea about the ESL and ESR of these caps. Those are not things I'd prefer to be ignorant of. -- John Larkin Highland Technology, Inc jlarkin at highlandtechnology dot com http://www.highlandtechnology.com Precision electronic instrumentation Picosecond-resolution Digital Delay and Pulse generators Custom laser drivers and controllers Photonics and fiberoptic TTL data links VME thermocouple, LVDT, synchro acquisition and simulation
On 9/19/2012 1:41 PM, John Larkin wrote:
> On Wed, 19 Sep 2012 12:44:09 -0700, miso <miso@sushi.com> wrote: > >> There are nonlinearities when you really whack a cap. > > Is ESR nonlinear? Is ESL nonlinear? How? > >> >> The goal is to avoid the op amp with the 50 ohm series resistor. > > Your goal maybe. Certainly not mine. My goal was to put a current step > into the cap and scope the resulting voltage step. > > That is >> not remotely how the cap will be used in the application. > > Current is current. The cap could be used in all sorts of > applications. Will ESR change as a function of nearby components? > >> >> Your eval circuit doesn't reflect the application, thus testing the part >> in the actual application is the way to go if you don't want to be >> bothered building a pulser. >> > > Testing the part in the application gives me no baseline for design > and simulation, tells me nothing about margins, makes it hard to > optimize after the parts are selected and soldered to the board. > > The tests that I did gave me a pretty good idea about the ESL and ESR > of these caps. Those are not things I'd prefer to be ignorant of. > >
I simply can't get through to you. Let me try this one more time. You DO NO, I repeat DO NOT, want a circuit with feedback driving the cap. Feedback circuits have rather complicated settling characteristics. You want to whack the cap with a very simple source follower using a low inductance resistor. I = c * dv/dt. You need a small dt to get a significant I. You seem to think I'm pulling this out of my ass, but I gave you links to National's app note. It even mentions building the resistor by paralleling resistor rather than using the traditional wirewound power resistor. Even ADI references the National article. This is how the gurus do it. You can chose your own route. I say to try the application circuit because, let me repeat this again, your test circuit does not replicate the application. Do you think I made up putting devices in series to get more measurable data? That is how it is done. I think I'm beginning to see why Thompson has issues with you.
On 09/19/2012 05:02 PM, miso wrote:
> On 9/19/2012 1:41 PM, John Larkin wrote: >> On Wed, 19 Sep 2012 12:44:09 -0700, miso <miso@sushi.com> wrote: >> >>> There are nonlinearities when you really whack a cap. >> >> Is ESR nonlinear? Is ESL nonlinear? How? >> >>> >>> The goal is to avoid the op amp with the 50 ohm series resistor. >> >> Your goal maybe. Certainly not mine. My goal was to put a current step >> into the cap and scope the resulting voltage step. >> >> That is >>> not remotely how the cap will be used in the application. >> >> Current is current. The cap could be used in all sorts of >> applications. Will ESR change as a function of nearby components? >> >>> >>> Your eval circuit doesn't reflect the application, thus testing the part >>> in the actual application is the way to go if you don't want to be >>> bothered building a pulser. >>> >> >> Testing the part in the application gives me no baseline for design >> and simulation, tells me nothing about margins, makes it hard to >> optimize after the parts are selected and soldered to the board. >> >> The tests that I did gave me a pretty good idea about the ESL and ESR >> of these caps. Those are not things I'd prefer to be ignorant of. >> >> > > I simply can't get through to you. Let me try this one more time. You DO > NO, I repeat DO NOT, want a circuit with feedback driving the cap. > Feedback circuits have rather complicated settling characteristics. You > want to whack the cap with a very simple source follower using a low > inductance resistor. > > I = c * dv/dt. > > You need a small dt to get a significant I. > > You seem to think I'm pulling this out of my ass, but I gave you links > to National's app note. It even mentions building the resistor by > paralleling resistor rather than using the traditional wirewound power > resistor. Even ADI references the National article. This is how the > gurus do it. You can chose your own route. > > I say to try the application circuit because, let me repeat this again, > your test circuit does not replicate the application. > > Do you think I made up putting devices in series to get more measurable > data? That is how it is done. > > I think I'm beginning to see why Thompson has issues with you. > >
Here we go again. Can't you folks understand the difference between an argument and a quarrel? You guys are talking past each other. John is using a pulse generator, not a function generator. Pulse generators IME almost all have a very low impedance output amplifier driving a real genuine physical 50-ohm resistor. They work fine open or shorted, or with any arbitrary cable reflection you like. They really really look like a 50-ohm source. Do you have a pulse generator that isn't like that? Which one? Cheers Phil Hobbs -- Dr Philip C D Hobbs Principal Consultant ElectroOptical Innovations LLC Optics, Electro-optics, Photonics, Analog Electronics 160 North State Road #203 Briarcliff Manor NY 10510 hobbs at electrooptical dot net http://electrooptical.net
On Wed, 19 Sep 2012 14:02:07 -0700, miso <miso@sushi.com> wrote:

>On 9/19/2012 1:41 PM, John Larkin wrote: >> On Wed, 19 Sep 2012 12:44:09 -0700, miso <miso@sushi.com> wrote: >> >>> There are nonlinearities when you really whack a cap. >> >> Is ESR nonlinear? Is ESL nonlinear? How? >> >>> >>> The goal is to avoid the op amp with the 50 ohm series resistor. >> >> Your goal maybe. Certainly not mine. My goal was to put a current step >> into the cap and scope the resulting voltage step. >> >> That is >>> not remotely how the cap will be used in the application. >> >> Current is current. The cap could be used in all sorts of >> applications. Will ESR change as a function of nearby components? >> >>> >>> Your eval circuit doesn't reflect the application, thus testing the part >>> in the actual application is the way to go if you don't want to be >>> bothered building a pulser. >>> >> >> Testing the part in the application gives me no baseline for design >> and simulation, tells me nothing about margins, makes it hard to >> optimize after the parts are selected and soldered to the board. >> >> The tests that I did gave me a pretty good idea about the ESL and ESR >> of these caps. Those are not things I'd prefer to be ignorant of. >> >> > >I simply can't get through to you. Let me try this one more time. You DO >NO, I repeat DO NOT, want a circuit with feedback driving the cap.
That's what *you* don't want. Beats me why.
>Feedback circuits have rather complicated settling characteristics. You >want to whack the cap with a very simple source follower using a low >inductance resistor.
I can't get through to you. The B&K makes a nice clean, fast pulse with a 50 ohm source impedance, no complicated settling that I've ever seen. I don't know or much care what's inside the box. I wanted a 100 mA current step, and I got it.
> >I = c * dv/dt. > >You need a small dt to get a significant I.
Look at my waveform. It has an obvious fast step (Ipulse * ESR) and an obvious ramp (Ipulse integrated into 180 uF.) The information to calculate ESR is in plain sight. Note that the generator pulse is 5 volts open-circuit, and the scope is 2 mV/cm.
> >You seem to think I'm pulling this out of my ass, but I gave you links >to National's app note. It even mentions building the resistor by >paralleling resistor rather than using the traditional wirewound power >resistor. Even ADI references the National article. This is how the >gurus do it. You can chose your own route. > >I say to try the application circuit because, let me repeat this again, >your test circuit does not replicate the application.
Of course not. My test circuit measures ESR.
> >Do you think I made up putting devices in series to get more measurable >data? That is how it is done. > >I think I'm beginning to see why Thompson has issues with you. >
Well, he is a redneck asshole. I bet he has trouble with lots of people. At least you haven't insulted my wife, so far. I measured a part. I posted the results for public use. The numbers are good. Polymer aluminum caps are great. Hey, it's still set up over on my bench. I blasted the cap with freeze spray and it looks like ESR went *down* a bit. Try that with a regular 'lytic cap! -- John Larkin Highland Technology, Inc jlarkin at highlandtechnology dot com http://www.highlandtechnology.com Precision electronic instrumentation Picosecond-resolution Digital Delay and Pulse generators Custom laser drivers and controllers Photonics and fiberoptic TTL data links VME thermocouple, LVDT, synchro acquisition and simulation
On Wed, 19 Sep 2012 17:19:51 -0400, Phil Hobbs
<pcdhSpamMeSenseless@electrooptical.net> wrote:

>On 09/19/2012 05:02 PM, miso wrote: >> On 9/19/2012 1:41 PM, John Larkin wrote: >>> On Wed, 19 Sep 2012 12:44:09 -0700, miso <miso@sushi.com> wrote: >>> >>>> There are nonlinearities when you really whack a cap. >>> >>> Is ESR nonlinear? Is ESL nonlinear? How? >>> >>>> >>>> The goal is to avoid the op amp with the 50 ohm series resistor. >>> >>> Your goal maybe. Certainly not mine. My goal was to put a current step >>> into the cap and scope the resulting voltage step. >>> >>> That is >>>> not remotely how the cap will be used in the application. >>> >>> Current is current. The cap could be used in all sorts of >>> applications. Will ESR change as a function of nearby components? >>> >>>> >>>> Your eval circuit doesn't reflect the application, thus testing the part >>>> in the actual application is the way to go if you don't want to be >>>> bothered building a pulser. >>>> >>> >>> Testing the part in the application gives me no baseline for design >>> and simulation, tells me nothing about margins, makes it hard to >>> optimize after the parts are selected and soldered to the board. >>> >>> The tests that I did gave me a pretty good idea about the ESL and ESR >>> of these caps. Those are not things I'd prefer to be ignorant of. >>> >>> >> >> I simply can't get through to you. Let me try this one more time. You DO >> NO, I repeat DO NOT, want a circuit with feedback driving the cap. >> Feedback circuits have rather complicated settling characteristics. You >> want to whack the cap with a very simple source follower using a low >> inductance resistor. >> >> I = c * dv/dt. >> >> You need a small dt to get a significant I. >> >> You seem to think I'm pulling this out of my ass, but I gave you links >> to National's app note. It even mentions building the resistor by >> paralleling resistor rather than using the traditional wirewound power >> resistor. Even ADI references the National article. This is how the >> gurus do it. You can chose your own route. >> >> I say to try the application circuit because, let me repeat this again, >> your test circuit does not replicate the application. >> >> Do you think I made up putting devices in series to get more measurable >> data? That is how it is done. >> >> I think I'm beginning to see why Thompson has issues with you. >> >> > >Here we go again. Can't you folks understand the difference between an >argument and a quarrel? > >You guys are talking past each other. John is using a pulse generator, >not a function generator. Pulse generators IME almost all have a very >low impedance output amplifier driving a real genuine physical 50-ohm >resistor. They work fine open or shorted, or with any arbitrary cable >reflection you like. They really really look like a 50-ohm source. > >Do you have a pulse generator that isn't like that? Which one? > >Cheers > >Phil Hobbs
The B&K 4003A is a classic analog function generator, but it does have a "pulse" setting. It sure behaves like a voltage source behind a 50 ohm resistor. Geez, all I wanted was a 100 mA current step. The B&K can sure manage that. The waveform that I posted looks right. -- John Larkin Highland Technology, Inc jlarkin at highlandtechnology dot com http://www.highlandtechnology.com Precision electronic instrumentation Picosecond-resolution Digital Delay and Pulse generators Custom laser drivers and controllers Photonics and fiberoptic TTL data links VME thermocouple, LVDT, synchro acquisition and simulation
A 50 ohm source is not low enough to jam current into a cap. Consider 
the impedance associated with a switch mode power supply fet and you 
will come to the same conclusion.

Hence those in the biz build these drivers.



On Wed, 19 Sep 2012 16:58:24 -0700, miso <miso@sushi.com> wrote:

> >A 50 ohm source is not low enough to jam current into a cap.
Did somebody repeal Ohm's Law? Did I fake this waveform? https://dl.dropbox.com/u/53724080/Parts/Polymer_ESR.JPG -- John Larkin Highland Technology, Inc jlarkin at highlandtechnology dot com http://www.highlandtechnology.com Precision electronic instrumentation Picosecond-resolution Digital Delay and Pulse generators Custom laser drivers and controllers Photonics and fiberoptic TTL data links VME thermocouple, LVDT, synchro acquisition and simulation
John Larkin wrote:

> On Wed, 19 Sep 2012 16:58:24 -0700, miso <miso@sushi.com> wrote: > > >>A 50 ohm source is not low enough to jam current into a cap. > > > Did somebody repeal Ohm's Law? Did I fake this waveform? > > https://dl.dropbox.com/u/53724080/Parts/Polymer_ESR.JPG >
Of course you faked it. Paint shop goes a long ways. :) Jamie
On Wed, 19 Sep 2012 22:30:31 -0400, Jamie
<jamie_ka1lpa_not_valid_after_ka1lpa_@charter.net> wrote:

>John Larkin wrote: > >> On Wed, 19 Sep 2012 16:58:24 -0700, miso <miso@sushi.com> wrote: >> >> >>>A 50 ohm source is not low enough to jam current into a cap. >> >> >> Did somebody repeal Ohm's Law? Did I fake this waveform? >> >> https://dl.dropbox.com/u/53724080/Parts/Polymer_ESR.JPG >> > >Of course you faked it. Paint shop goes a long ways. :) > >Jamie
On reconsideration, it looks more like 12 milliohms. -- John Larkin Highland Technology Inc www.highlandtechnology.com jlarkin at highlandtechnology dot com Precision electronic instrumentation Picosecond-resolution Digital Delay and Pulse generators Custom timing and laser controllers Photonics and fiberoptic TTL data links VME analog, thermocouple, LVDT, synchro, tachometer Multichannel arbitrary waveform generators