On 05/09/2011 05:27 AM, Randy Yates wrote:> On 05/09/2011 12:56 AM, Rune Allnor wrote: >> On May 9, 2:34 am, Randy Yates<ya...@ieee.org> wrote: >>> Hello comp.dsp >>> >>> My current client has a requirement to obtain the the line-line >>> voltages of a three-phase power source inside a DSP given the >>> line-neutral inputs. He's thinking of doing it in the analog domain >>> using opamps to do the differencing, and feeding the resulting >>> line-line into the ADC (we're using the 2809, BTW). >>> >>> Further, there are at least two opamp architectures under consideration. >>> One is the standard difference amplifier between each of A-B, A-C, and B-C. >>> The other uses just three opamps to determine A-B and B-C, then relies on the >>> fact that A-C can be derived from those two. >> >> Is that a *fact*? >> >> Sure, the theory says so *provided* you know all there is >> to know about the system. >> >> In my experience, that assumption tends to break down when >> stuff gets interesting, like when there is a mechanical >> faliure somewhere. Like a short to ground, or some impedance >> straying off spec somewhere. > > Hi Rune, > > Thanks for your challenge! I can't think of a specific scenario in > which this assumption would break down, namely that, (A - B) - (B - C) > would not be equal to (A - C). Can you give me a concrete example?Correction: (A - C) - (A - B) = (B - C). -- Randy Yates % "So now it's getting late, Digital Signal Labs % and those who hesitate mailto://yates@ieee.org % got no one..." http://www.digitalsignallabs.com % 'Waterfall', *Face The Music*, ELO
Line-to-Line Sensing: Analog Vs Digital?
Started by ●May 8, 2011
Reply by ●May 9, 20112011-05-09
Reply by ●May 9, 20112011-05-09
On 05/08/2011 11:33 PM, Jerry Avins wrote:> On May 8, 8:34 pm, Randy Yates<ya...@ieee.org> wrote: >> Hello comp.dsp >> >> My current client has a requirement to obtain the the line-line >> voltages of a three-phase power source inside a DSP given the >> line-neutral inputs. He's thinking of doing it in the analog domain >> using opamps to do the differencing, and feeding the resulting >> line-line into the ADC (we're using the 2809, BTW). >> >> Further, there are at least two opamp architectures under consideration. >> One is the standard difference amplifier between each of A-B, A-C, and B-C. >> The other uses just three opamps to determine A-B and B-C, then relies on the >> fact that A-C can be derived from those two. >> >> The other alternative is to just input the line-neutral voltages directly >> (after appropriate voltage dividers) into the ADC. >> >> I'd be curious to learn what folks here think the "best" method would be >> given considerations of 1) noise, and 2) common-mode (the 3-phase center >> isn't necessarily the same as the ADC / opamp circuit ground). > > Draw the phasor diagrams and you will see that knowledge of phase is > critical. Phase is implicit with independent ADCs making simultaneous > measurements. Assunptions about phase symmetry are useless. If they > held, a single line-to-neutral RMS or peak amplitude would suffice.I agree. How is this relevent? -- Randy Yates % "So now it's getting late, Digital Signal Labs % and those who hesitate mailto://yates@ieee.org % got no one..." http://www.digitalsignallabs.com % 'Waterfall', *Face The Music*, ELO
Reply by ●May 9, 20112011-05-09
On 05/08/2011 10:19 PM, steveu wrote:>> Hello comp.dsp >> >> My current client has a requirement to obtain the the line-line >> voltages of a three-phase power source inside a DSP given the >> line-neutral inputs. He's thinking of doing it in the analog domain >> using opamps to do the differencing, and feeding the resulting >> line-line into the ADC (we're using the 2809, BTW). >> >> Further, there are at least two opamp architectures under consideration. >> One is the standard difference amplifier between each of A-B, A-C, and > B-C. >> The other uses just three opamps to determine A-B and B-C, then relies on > the >> fact that A-C can be derived from those two. >> >> The other alternative is to just input the line-neutral voltages directly >> (after appropriate voltage dividers) into the ADC. >> >> I'd be curious to learn what folks here think the "best" method would be >> given considerations of 1) noise, and 2) common-mode (the 3-phase center >> isn't necessarily the same as the ADC / opamp circuit ground). > > If you have a neutral connection, why not do things digitally?I'm confused on how that's going to help - see response to Vlad.> If you are using a single ADC there will usually be a phase > difference between samples. However, the C2000 ADC has two sample > and holds, so you can simultaneously sample two of the phases, and > simply take the difference of the results.Yes, that's what we're currently doing. The peripheral is nice in that respect.> As Vlad said, the ADC in the older C2000 parts isn't brilliant, so if you > need high precision, its probably not the best choice. You mentioned noise, > which suggests they may be looking for a fairly precise result.Yes, we are. Thanks for your input, Steve - I'll think about using an off-chip ADC. -- Randy Yates % "So now it's getting late, Digital Signal Labs % and those who hesitate mailto://yates@ieee.org % got no one..." http://www.digitalsignallabs.com % 'Waterfall', *Face The Music*, ELO
Reply by ●May 9, 20112011-05-09
On 05/09/2011 01:27 AM, Mitja Nemec wrote:> On May 9, 12:56 pm, Rune Allnor<all...@tele.ntnu.no> wrote: >> On May 9, 2:34 am, Randy Yates<ya...@ieee.org> wrote: >> >>> Hello comp.dsp >> >>> My current client has a requirement to obtain the the line-line >>> voltages of a three-phase power source inside a DSP given the >>> line-neutral inputs. He's thinking of doing it in the analog domain >>> using opamps to do the differencing, and feeding the resulting >>> line-line into the ADC (we're using the 2809, BTW). >> >>> Further, there are at least two opamp architectures under consideration. >>> One is the standard difference amplifier between each of A-B, A-C, and B-C. >>> The other uses just three opamps to determine A-B and B-C, then relies on the >>> fact that A-C can be derived from those two. >> >> Is that a *fact*? >> >> Sure, the theory says so *provided* you know all there is >> to know about the system. >> >> In my experience, that assumption tends to break down when >> stuff gets interesting, like when there is a mechanical >> faliure somewhere. Like a short to ground, or some impedance >> straying off spec somewhere. >> >> Nah, measure all three phases relative to some earth locally >> at the DSP board. That way you cover all the bases - you might >> even be able to detect problems with the earth cable that you >> would miss if you only measured differences between phases. >> >> Rune > > I'd also go this way, unless you have high harmonics present, and > sequential sampling would influence the results. In that case > measuring A-B and B-C, sampling them simultaneously and deriving A-C > is probably the only way with internal ADC.I don't understand what high harmonics have to do with anything. If there are high harmonics in the line-neutrals, they'll probably be in the line-lines too. Depends on their amplitudes and phases, of course. And how high is "high?" We're sampling at 100 kHz. -- Randy Yates % "So now it's getting late, Digital Signal Labs % and those who hesitate mailto://yates@ieee.org % got no one..." http://www.digitalsignallabs.com % 'Waterfall', *Face The Music*, ELO
Reply by ●May 9, 20112011-05-09
On 05/08/2011 11:29 PM, Richard Owlett wrote:> Randy Yates wrote: >> Hello comp.dsp >> >> My current client has a requirement to obtain the the line-line >> voltages of a three-phase power source inside a DSP given the >> line-neutral inputs. He's thinking of doing it in the analog domain >> using opamps to do the differencing, and feeding the resulting >> line-line into the ADC (we're using the 2809, BTW). >> >> Further, there are at least two opamp architectures under consideration. >> One is the standard difference amplifier between each of A-B, A-C, and B-C. >> The other uses just three opamps to determine A-B and B-C, then relies >> on the >> fact that A-C can be derived from those two. >> >> The other alternative is to just input the line-neutral voltages directly >> (after appropriate voltage dividers) into the ADC. >> >> I'd be curious to learn what folks here think the "best" method would be >> given considerations of 1) noise, and 2) common-mode (the 3-phase center >> isn't necessarily the same as the ADC / opamp circuit ground). > > > *MISSING* info > > "What is client's goal?"I gave the requirement: My current client has a requirement to obtain the the line-line voltages of a three-phase power source inside a DSP given the line-neutral inputs. Goal? His goal is to make money. :) --Randy -- Randy Yates % "So now it's getting late, Digital Signal Labs % and those who hesitate mailto://yates@ieee.org % got no one..." http://www.digitalsignallabs.com % 'Waterfall', *Face The Music*, ELO
Reply by ●May 9, 20112011-05-09
On May 9, 5:47�pm, Randy Yates <ya...@ieee.org> wrote:> On 05/09/2011 01:27 AM, Mitja Nemec wrote: > > > > > > > > > > > On May 9, 12:56 pm, Rune Allnor<all...@tele.ntnu.no> �wrote: > >> On May 9, 2:34 am, Randy Yates<ya...@ieee.org> �wrote: > > >>> Hello comp.dsp > > >>> My current client has a requirement to obtain the the line-line > >>> voltages of a three-phase power source inside a DSP given the > >>> line-neutral inputs. He's thinking of doing it in the analog domain > >>> using opamps to do the differencing, and feeding the resulting > >>> line-line into the ADC (we're using the 2809, BTW). > > >>> Further, there are at least two opamp architectures under consideration. > >>> One is the standard difference amplifier between each of A-B, A-C, and B-C. > >>> The other uses just three opamps to determine A-B and B-C, then relies on the > >>> fact that A-C can be derived from those two. > > >> Is that a *fact*? > > >> Sure, the theory says so *provided* you know all there is > >> to know about the system. > > >> In my experience, that assumption tends to break down when > >> stuff gets interesting, like when there is a mechanical > >> faliure somewhere. Like a short to ground, or some impedance > >> straying off spec somewhere. > > >> Nah, measure all three phases relative to some earth locally > >> at the DSP board. That way you cover all the bases - you might > >> even be able to detect problems with the earth cable that you > >> would miss if you only measured differences between phases. > > >> Rune > > > I'd also go this way, unless you have high harmonics present, and > > sequential sampling would influence the results. In that case > > measuring A-B and B-C, sampling them simultaneously and deriving A-C > > is probably the only way with internal ADC. > > I don't understand what high harmonics have to do with anything. If > there are high harmonics in the line-neutrals, they'll probably be in the > line-lines too. Depends on their amplitudes and phases, of course. > > And how high is "high?" We're sampling at 100 kHz. > -- > Randy Yates � � � � � � � � � � �% "So now it's getting late, > Digital Signal Labs � � � � � � �% � �and those who hesitate > mailto://ya...@ieee.org � � � � �% � �got no one..."http://www.digitalsignallabs.com% 'Waterfall', *Face The Music*, ELOYou will see harmonics anyhow, it is the sequential sampling that can add some error into your results, as you wiill/would be subtracting a value measured at one point in time by the value measured in another point in time. I don't know the speed of your ADC, but if it is configure to run at top speed, sequential samples should be around 500ns apart. Now if have harmonics up to 10kHz (note that this is quite high - 200 harmonic) it basically mean 1.8 degree of error if you don't sample simultaneously. If this is a problem or not you are the best judge. For me (and I usually have to deal with harmonics just up to 50, and they are usually below 5% of first) this is not an issue. By my experience 28xx ADC is more 12 going on 10, though it really helps if you use external voltage reference. Especially if the device will operate at variable temperatures. Mitja
Reply by ●May 9, 20112011-05-09
On May 9, 5:34�pm, Randy Yates <ya...@ieee.org> wrote:> On 05/08/2011 09:21 PM, Vladimir Vassilevsky wrote: > > > > > Randy Yates wrote: > >> Hello comp.dsp > > >> My current client has a requirement to obtain the the line-line > >> voltages of a three-phase power source inside a DSP given the > >> line-neutral inputs. He's thinking of doing it in the analog domain > >> using opamps to do the differencing, and feeding the resulting > >> line-line into the ADC (we're using the 2809, BTW). > > > Ouch. The built-in ADC of 28xx series is *horrible*. I wouldn't rely on it for any accuracy. > > Oh really? I wasn't aware of that. (12 bits going on 8 - LOL!) > > >> Further, there are at least two opamp architectures under consideration. > >> One is the standard difference amplifier between each of A-B, A-C, and B-C. > >> The other uses just three opamps to determine A-B and B-C, then relies on the fact that A-C can be derived from those two. > > >> The other alternative is to just input the line-neutral voltages directly > >> (after appropriate voltage dividers) into the ADC. > > >> I'd be curious to learn what folks here think the "best" method would be > >> given considerations of 1) noise, and 2) common-mode (the 3-phase center > >> isn't necessarily the same as the ADC / opamp circuit ground). > > > I would use "star" connection and do all differences in the DSP. Since you are interested in the differences, it doesn't matter > > where the center point is. > > What do you mean by "star" connection? Do you mean using a Y of three > caps, one to each phase, and taking the center point? We've thought > about that too but one thing I don't like about that is there will be > some ripple. Or is that what you mean? That any ripple is irrelevent > if you're taking differences? > > And where would you use this star connection? You can't just connect > it to circuit ground - thinking (possible) ground currents. > > Just worried about common mode. Sticking 10 volts into a 3V referenced > ADC isn't going to go over very well. > -- > Randy Yates � � � � � � � � � � �% "So now it's getting late, > Digital Signal Labs � � � � � � �% � �and those who hesitate > mailto://ya...@ieee.org � � � � �% � �got no one..."http://www.digitalsignallabs.com% 'Waterfall', *Face The Music*, ELOHe might have meant star connection of voltage dividers and the middle point goes to the ground of your circuit. Assuming your circuit is otherwise isolated, the ground of the circuit becomes "virtual" neutral point (I really don't know how to translate the term we use back here). You might want to measure also the voltage between "virtual" neutral point and neutral point of the grid in order to determine if there is any unbalance. You can avoid this additional measurement if the ground of your circuit (middle point of star connection of voltage dividers) is connected to the grid neutral point. But then you loose the advantage of having the circuit isolated. This is probably redundant, but never assume grid neutral point to be close or equal to grounding point. Mitja
Reply by ●May 9, 20112011-05-09
On May 9, 11:27=A0am, Randy Yates <ya...@ieee.org> wrote:> On 05/09/2011 12:56 AM, Rune Allnor wrote: > > > > > > > On May 9, 2:34 am, Randy Yates<ya...@ieee.org> =A0wrote: > >> Hello comp.dsp > > >> My current client has a requirement to obtain the the line-line > >> voltages of a three-phase power source inside a DSP given the > >> line-neutral inputs. He's thinking of doing it in the analog domain > >> using opamps to do the differencing, and feeding the resulting > >> line-line into the ADC (we're using the 2809, BTW). > > >> Further, there are at least two opamp architectures under consideratio=n.> >> One is the standard difference amplifier between each of A-B, A-C, and=B-C.> >> The other uses just three opamps to determine A-B and B-C, then relies=on the> >> fact that A-C can be derived from those two. > > > Is that a *fact*? > > > Sure, the theory says so *provided* you know all there is > > to know about the system. > > > In my experience, that assumption tends to break down when > > stuff gets interesting, like when there is a mechanical > > faliure somewhere. Like a short to ground, or some impedance > > straying off spec somewhere. > > Hi Rune, > > Thanks for your challenge! I can't think of a specific scenario in > which this assumption would break down, namely that, (A - B) - (B - C) > would not be equal to (A - C). Can you give me a concrete example?Assume a 4-wire system, A,B,C,N. If insulation shatters and one of the phases arcs to earth (not N) and there is an impedance between N and earth (might happen; that's wht the N-wire is there), the difference sensor is in trouble. Rune
Reply by ●May 9, 20112011-05-09
Randy Yates <yates@ieee.org> wrote: (snip)>> Thanks for your challenge! I can't think of a specific scenario in >> which this assumption would break down, namely that, (A - B) - (B - C) >> would not be equal to (A - C). Can you give me a concrete example?> Correction: (A - C) - (A - B) = (B - C).That sounds right, though the error will be slightly higher. If there is an uncertainty (amplifier noise, non-linearity, quantization noise, etc.) in (A-C) and (A-B) then it will be slightly higher in (B-C) obtained from the subtraction. (Maybe sqrt(2) times higher.) -- glen
Reply by ●May 9, 20112011-05-09
On May 8, 11:33=A0pm, Jerry Avins <j...@ieee.org> wrote:> On May 8, 8:34=A0pm, Randy Yates <ya...@ieee.org> wrote: > > > > > > > > > > > Hello comp.dsp > > > My current client has a requirement to obtain the the line-line > > voltages of a three-phase power source inside a DSP given the > > line-neutral inputs. He's thinking of doing it in the analog domain > > using opamps to do the differencing, and feeding the resulting > > line-line into the ADC (we're using the 2809, BTW). > > > Further, there are at least two opamp architectures under consideration=.> > One is the standard difference amplifier between each of A-B, A-C, and =B-C.> > The other uses just three opamps to determine A-B and B-C, then relies =on the> > fact that A-C can be derived from those two. > > > The other alternative is to just input the line-neutral voltages direct=ly> > (after appropriate voltage dividers) into the ADC. > > > I'd be curious to learn what folks here think the "best" method would b=e> > given considerations of 1) noise, and 2) common-mode (the 3-phase cente=r> > isn't necessarily the same as the ADC / opamp circuit ground). > > Draw the phasor diagrams and you will see that knowledge of phase is > critical. Phase is implicit with independent ADCs making simultaneous > measurements. Assunptions about phase symmetry are useless. If they > held, a single line-to-neutral RMS or peak amplitude would suffice.Here's an example: A-to-neutral =3D 120V B-to-neutral =3D 120V C-to-neutral =3D 120V Among the many solutions for line-to-line are A-to-B =3D 208V B-to-C =3D 208V C-to-A =3D 208V and A-to-B =3D 0V B-to-C =3D 0V C-to-A =3D 0V Phase matters! Jerry -- Engineering is the art of making what you want from things you can get.
