Detecting the peak value
in [Design]
|
Prev: Changing the color of objects/primitives only ? (flatshading...) (Re-using verteces, is user supplied data considered "state" aswell ?)
Next: Ronald Reagan socialism commercial
From: hal on 1 Aug 2010 13:07 On 1-Aug-2010, Jim Thompson <To-Email-Use-The-Envelope-Icon(a)On-My-Web-Site.com> wrote: > > > >On 1-Aug-2010, Jim Thompson > ><To-Email-Use-The-Envelope-Icon(a)On-My-Web-Site.com> wrote: > > > >> On Sun, 1 Aug 2010 15:29:19 GMT, hal(a)nospam.com wrote: > >> > >> >Hello all, > >> > > >> >I've been following this thread and it has brought up some issues I'm > >> >currently working on. First of all, I'm a hobbyist that likes to > >> >learn > >> >by > >> >doing; as such, I'm interested in making a circuit do what I want > >> >without > >> >regards to a BOM - especially in regards to performance. In other > >> >words, > >> >I > >> >like "best" rather than "cheapest" or "easiest." > >> > > >> >Currently I'm working on a LCR meter for personal use with better than > >> >.1% > >> >accuracy (closer to .02% if possible) for all functions. Also, all > >> >calibration will be internal and automatic; I have quite a few of > >> >Vishay's > >> >S102C series resistors (.01%, 2ppm) so I have a ready internal > >> >reference > >> >to > >> >use, keeping all measurements ratiometric if possible, and use > >> >resistance > >> >as > >> >the "known" in all equations. > >> > > >> >I'm measuring capacitance using constant-current charge time, so I > >> >already > >> >will know the capacitance > >> > >> Are you not deluding yourself that you know "constant-current" to 0.1% > >> accuracy? > > > >Probably so - hence my request for suggestions. I suppose that I should > >have emphasized that this is not only something that I do for enjoyment, > >but > >to _learn_. As such, I would welcome feedback on not only that I *am* > >going > >wrong, but *how* as well. > > Yep. A welcome change in SED direction. > > Pondering that, it's going to be tricky to make a loop-stabilized > current source that tracks a slewing capacitor's change in voltage > (Early-effect or channel-length modulation). > > Perhaps _not_ use a classic current source, instead sense capacitor > current (into a virtual ground) and adjust a voltage drive? > > I presume you're measuring time-to-voltage to get capacitance? Yes, it seemed the best way to use the high precision components that I have. I was going to charge through one of the precision Rs and measure to get the charging current. As long as the current source is linear over a small range (a couple volts starting at .5 or 1V - NOT in-circuit, of course), with an accurate clock, and using Kelvin sensing for both the reference and DUT, I felt that I should be able to achieve very good accuracy. I realize that this is a one-dimensional measurement of the capacitance, but it suits for what I am trying to accomplish. The fun part, for me, is trying to see how accurate I can get that one measurement - ditto for ESR. One issue I'm still working on is a current source with the required linearity. There are quite a few designs that are accurate at a set current, but I don't have the expertise or equipment to determine if they are accurate over a sweeping voltage range. Suggestions, anyone? > > > > > > >> > >> >before I measure the ESR and can account for Z. I > >> >originally planned to use a peak-detection (measurement) scheme with a > >> >100khz low distortion sine wave, however 1-2% accuracy was the best it > >> >seemed that I could do (Yes, I realize that for ESR, that is good > >> >enough > >> >- > >> >but remember, I'm doing this to learn and just to see if I can for the > >> >fun > >> >of it). So, as it stands, I am planning on using a S/H or tracking > >> >ADC > >> >with > >> >a high enough sample rate to detect both high and low peaks of both > >> >the > >> >DUT > >> >and a reference resistance and calculate the ESR from there. > >> > > >> >As the required additional circuitry to measure the ESR is almost as > >> >much > >> >as > >> >all of the rest put together, I can't help but wonder if there's an > >> >easier > >> >way that I've missed and I'm using a sledge hammer to drive a nail. > >> >So, > >> >I > >> >would welcome any suggestions, tips or hints. > >> > > >> >Thanks, > >> > > >> >Hal > >> > >> ...Jim Thompson > >> -- > > ...Jim Thompson
From: John Larkin on 1 Aug 2010 13:06 On Sun, 1 Aug 2010 16:00:54 GMT, hal(a)nospam.com wrote: > >On 1-Aug-2010, Jim Thompson ><To-Email-Use-The-Envelope-Icon(a)On-My-Web-Site.com> wrote: > >> On Sun, 1 Aug 2010 15:29:19 GMT, hal(a)nospam.com wrote: >> >> >Hello all, >> > >> >I've been following this thread and it has brought up some issues I'm >> >currently working on. First of all, I'm a hobbyist that likes to learn >> >by >> >doing; as such, I'm interested in making a circuit do what I want without >> >regards to a BOM - especially in regards to performance. In other words, >> >I >> >like "best" rather than "cheapest" or "easiest." >> > >> >Currently I'm working on a LCR meter for personal use with better than >> >.1% >> >accuracy (closer to .02% if possible) for all functions. Also, all >> >calibration will be internal and automatic; I have quite a few of >> >Vishay's >> >S102C series resistors (.01%, 2ppm) so I have a ready internal reference >> >to >> >use, keeping all measurements ratiometric if possible, and use resistance >> >as >> >the "known" in all equations. >> > >> >I'm measuring capacitance using constant-current charge time, so I >> >already >> >will know the capacitance >> >> Are you not deluding yourself that you know "constant-current" to 0.1% >> accuracy? > >Probably so - hence my request for suggestions. I suppose that I should >have emphasized that this is not only something that I do for enjoyment, but >to _learn_. As such, I would welcome feedback on not only that I *am* going >wrong, but *how* as well. > > You don't really need an accurate current source, just a fairly stable one. Dump the current into a reference resistor (you need one of those anyhow) and measure it, then use the same current on your cap. Actually, you don't even need a current source. Just a voltage and a reference resistor... A B V+---------Rref---------Rx---------gnd or Cx Just measure VA and VB relative to ground and do a little math. It would get tricky for small values of Cx. This works great for resistors; I use it for RTDs all the time. Modern ADC are so linear that it pays to do ratiometric stuff whenever you can. Here's another idea: Sine wave goes through Cref or Rref into the summing point of a transimpedance opamp. Digitize the opamp output. Substitute Rx or Cx or Lx for Rref, digitize again, do some math. That can handle very small Cx values and avoids a lot of strays problems. John
From: krw on 1 Aug 2010 13:14 On Sun, 1 Aug 2010 17:07:46 GMT, hal(a)nospam.com wrote: > >On 1-Aug-2010, Jim Thompson ><To-Email-Use-The-Envelope-Icon(a)On-My-Web-Site.com> wrote: > >> > >> >On 1-Aug-2010, Jim Thompson >> ><To-Email-Use-The-Envelope-Icon(a)On-My-Web-Site.com> wrote: >> > >> >> On Sun, 1 Aug 2010 15:29:19 GMT, hal(a)nospam.com wrote: >> >> >> >> >Hello all, >> >> > >> >> >I've been following this thread and it has brought up some issues I'm >> >> >currently working on. First of all, I'm a hobbyist that likes to >> >> >learn >> >> >by >> >> >doing; as such, I'm interested in making a circuit do what I want >> >> >without >> >> >regards to a BOM - especially in regards to performance. In other >> >> >words, >> >> >I >> >> >like "best" rather than "cheapest" or "easiest." >> >> > >> >> >Currently I'm working on a LCR meter for personal use with better than >> >> >.1% >> >> >accuracy (closer to .02% if possible) for all functions. Also, all >> >> >calibration will be internal and automatic; I have quite a few of >> >> >Vishay's >> >> >S102C series resistors (.01%, 2ppm) so I have a ready internal >> >> >reference >> >> >to >> >> >use, keeping all measurements ratiometric if possible, and use >> >> >resistance >> >> >as >> >> >the "known" in all equations. >> >> > >> >> >I'm measuring capacitance using constant-current charge time, so I >> >> >already >> >> >will know the capacitance >> >> >> >> Are you not deluding yourself that you know "constant-current" to 0.1% >> >> accuracy? >> > >> >Probably so - hence my request for suggestions. I suppose that I should >> >have emphasized that this is not only something that I do for enjoyment, >> >but >> >to _learn_. As such, I would welcome feedback on not only that I *am* >> >going >> >wrong, but *how* as well. >> >> Yep. A welcome change in SED direction. >> >> Pondering that, it's going to be tricky to make a loop-stabilized >> current source that tracks a slewing capacitor's change in voltage >> (Early-effect or channel-length modulation). >> >> Perhaps _not_ use a classic current source, instead sense capacitor >> current (into a virtual ground) and adjust a voltage drive? >> >> I presume you're measuring time-to-voltage to get capacitance? > >Yes, it seemed the best way to use the high precision components that I >have. I was going to charge through one of the precision Rs and measure to >get the charging current. As long as the current source is linear over a >small range (a couple volts starting at .5 or 1V - NOT in-circuit, of >course), with an accurate clock, and using Kelvin sensing for both the >reference and DUT, I felt that I should be able to achieve very good >accuracy. I realize that this is a one-dimensional measurement of the >capacitance, but it suits for what I am trying to accomplish. The fun part, >for me, is trying to see how accurate I can get that one measurement - ditto >for ESR. One issue I'm still working on is a current source with the >required linearity. There are quite a few designs that are accurate at a >set current, but I don't have the expertise or equipment to determine if >they are accurate over a sweeping voltage range. Suggestions, anyone? How are you going to calibrate your equipment (test your accuracy)? <...>
From: Jim Thompson on 1 Aug 2010 13:19 On Sun, 01 Aug 2010 10:06:49 -0700, John Larkin <jjlarkin(a)highNOTlandTHIStechnologyPART.com> wrote: >On Sun, 1 Aug 2010 16:00:54 GMT, hal(a)nospam.com wrote: > >> >>On 1-Aug-2010, Jim Thompson >><To-Email-Use-The-Envelope-Icon(a)On-My-Web-Site.com> wrote: >> >>> On Sun, 1 Aug 2010 15:29:19 GMT, hal(a)nospam.com wrote: >>> >>> >Hello all, >>> > >>> >I've been following this thread and it has brought up some issues I'm >>> >currently working on. First of all, I'm a hobbyist that likes to learn >>> >by >>> >doing; as such, I'm interested in making a circuit do what I want without >>> >regards to a BOM - especially in regards to performance. In other words, >>> >I >>> >like "best" rather than "cheapest" or "easiest." >>> > >>> >Currently I'm working on a LCR meter for personal use with better than >>> >.1% >>> >accuracy (closer to .02% if possible) for all functions. Also, all >>> >calibration will be internal and automatic; I have quite a few of >>> >Vishay's >>> >S102C series resistors (.01%, 2ppm) so I have a ready internal reference >>> >to >>> >use, keeping all measurements ratiometric if possible, and use resistance >>> >as >>> >the "known" in all equations. >>> > >>> >I'm measuring capacitance using constant-current charge time, so I >>> >already >>> >will know the capacitance >>> >>> Are you not deluding yourself that you know "constant-current" to 0.1% >>> accuracy? >> >>Probably so - hence my request for suggestions. I suppose that I should >>have emphasized that this is not only something that I do for enjoyment, but >>to _learn_. As such, I would welcome feedback on not only that I *am* going >>wrong, but *how* as well. >> >> > >You don't really need an accurate current source, just a fairly stable >one. Dump the current into a reference resistor (you need one of those >anyhow) and measure it, then use the same current on your cap. > [snip] Larkin misses the Early-effect/channel-length-modulation errors as the voltage changes. ...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 | Spice is like a sports car... Performance only as good as the person behind the wheel.
From: hal on 1 Aug 2010 13:21
On 1-Aug-2010, John Larkin <jjlarkin(a)highNOTlandTHIStechnologyPART.com> wrote: > >Hello all, > > > >I've been following this thread and it has brought up some issues I'm > >currently working on. First of all, I'm a hobbyist that likes to learn > >by > >doing; as such, I'm interested in making a circuit do what I want without > >regards to a BOM - especially in regards to performance. In other words, > >I > >like "best" rather than "cheapest" or "easiest." > > > >Currently I'm working on a LCR meter for personal use with better than > >.1% > >accuracy (closer to .02% if possible) for all functions. Also, all > >calibration will be internal and automatic; I have quite a few of > >Vishay's > >S102C series resistors (.01%, 2ppm) so I have a ready internal reference > >to > >use, keeping all measurements ratiometric if possible, and use resistance > >as > >the "known" in all equations. > > > >I'm measuring capacitance using constant-current charge time, so I > >already > >will know the capacitance before I measure the ESR and can account for Z. > > I > >originally planned to use a peak-detection (measurement) scheme with a > >100khz low distortion sine wave, however 1-2% accuracy was the best it > >seemed that I could do (Yes, I realize that for ESR, that is good enough > >- > >but remember, I'm doing this to learn and just to see if I can for the > >fun > >of it). So, as it stands, I am planning on using a S/H or tracking ADC > >with > >a high enough sample rate to detect both high and low peaks of both the > >DUT > >and a reference resistance and calculate the ESR from there. > > > >As the required additional circuitry to measure the ESR is almost as much > >as > >all of the rest put together, I can't help but wonder if there's an > >easier > >way that I've missed and I'm using a sledge hammer to drive a nail. So, > >I > >would welcome any suggestions, tips or hints. > > > >Thanks, > > > >Hal > > Really precise capacitance measurement (and "really precise" starts at > 1% maybe) generally uses a bridge driven by a sine wave. One problem > with caps is that C varies with frequency because of dielectric > absorption and other effects, so the frequency is usually specified > for accurate measurement. DA can make a ramp-type C measurement > nonlinear. > > If you have a fast ADC, I'd suggest just taking a lot of samples and > absolute value+averaging. That would be better, lower noise data than > looking for/at the peak, and you could include a lot more samples. > > If you can use a capacitor for the reference, you can make a voltage > divider with Cref and Cx, digitize the overall voltage and the > midpoint voltage, and do the math. That's ratiometric on the > excitation source, and works for resistors, too. I guess you could > make a divider from a precision resistor and Cx, given a good sine > reference. > > A decent 16-bit SAR ADC is cheap nowadays and they tend to be > amazingly linear. Sampling all up and down the waveform makes it even > better, and averaging lots of samples can extend the resolution beyond > 16 bits. You can do synchronous detection tricks to reduce noise. The > nice thing about an ADC approach is that the tweaks are software. > > To do fancy stuff, like 3/4 terminal measurements, you might buy a > couple of old HP or GR RLC bridge manuals and study their topologies. > Those old manuals generally had schematics and good theory sections. > > > John A bridge was my first inclination; however, I'm wanting an all-in-one LCR meter with one (Kelvin) set of test leads. To do so will involve fairly complex switching circuitry and trying to make a bridge topology work with that arrangement was a little too far beyond my comfort level. For the sine source, I was going to use one of the ADI DDS chips. Do you think I would achieve better results using a sine-measurement technique for the capacitance as well as the ESR? Most of what I do is low frequency or DC so I don't really have a need to measure the cap at 50 or 100khz. I don't want to use a reference cap as the best I have is 1% and I was hoping to better that. Hal |