Is my understanding correct that output impedence (of a source)can be modelled by fir/iir filters. The application is none other than voip slic who tries to adjust its output impedence so that it becomes the conjugate of telephone impedence , thereby maximising power transfer to the telephone.I was struggling to get the idea..the pressing questions are does a slic behave like a constant current/voltage source?Is the impedence is in parallel or series? how a resistor and inductor in series can be modelled by a filter? Any insight is greatly appreciated.
impedence matching by fir / iir filter
Started by ●March 6, 2008
Reply by ●March 6, 20082008-03-06
On 6 Mar, 07:35, "pal.debabrata123" <pal.debabrata...@gmail.com> wrote:> Is my understanding correct that output impedence (of a source)can be > modelled by fir/iir filters.The frequency responses of analog systems can be modeled by discrete-time filters. 'Impedance' is a concept which does not exist in the discrete-time world.> The application is none other than voip slic > who tries to adjust its output impedence so that it becomes the conjugate > of telephone impedence , thereby maximising power transfer to the > telephone.I was struggling to get the idea..You mix concepts from the analog and digital worlds. Do you try to match impedances on a physical line which is connected to the input of a digital system? If so, it is the input stages of the ADC (the anti-alias filter) which should be matched to the line impedace. Again, there is no such thing as impedance in digital systems, only numbers.> how a resistor and inductor in series can be > modelled by a filter?Use the Laplace transform. Then use BLT to find a discrete-time filter which has a similar frequency response as the RLC network. Rune
Reply by ●March 6, 20082008-03-06
Rune Allnor wrote:> On 6 Mar, 07:35, "pal.debabrata123" <pal.debabrata...@gmail.com> > wrote: > >>Is my understanding correct that output impedence (of a source)can be >>modelled by fir/iir filters. > > > The frequency responses of analog systems can be modeled by > discrete-time filters. 'Impedance' is a concept which does not > exist in the discrete-time world. > > >>The application is none other than voip slic >>who tries to adjust its output impedence so that it becomes the conjugate >>of telephone impedence , thereby maximising power transfer to the >>telephone.I was struggling to get the idea.. > > > You mix concepts from the analog and digital worlds. > > Do you try to match impedances on a physical line which is > connected to the input of a digital system? If so, it is the > input stages of the ADC (the anti-alias filter) which should > be matched to the line impedace. > > Again, there is no such thing as impedance in digital systems, > only numbers. > > >>how a resistor and inductor in series can be >>modelled by a filter? > > > Use the Laplace transform. Then use BLT to find a discrete-time > filter which has a similar frequency response as the RLC network. > > RuneI wonder if what the OP is actually interested in would be "telephone line equalization networks"?
Reply by ●March 6, 20082008-03-06
Richard Owlett wrote: ...> I wonder if what the OP is actually interested in would be "telephone > line equalization networks"?Maybe he'd like to replace the transformer on his utility pole with a small computer that calculates the voltage to be delivered to his house. Jerry -- Engineering is the art of making what you want from things you can get. �����������������������������������������������������������������������
Reply by ●March 6, 20082008-03-06
On Thu, 06 Mar 2008 06:03:17 -0600, Richard Owlett <rowlett@atlascomm.net> wrote:>Rune Allnor wrote: >> On 6 Mar, 07:35, "pal.debabrata123" <pal.debabrata...@gmail.com> >> wrote: >> >>>Is my understanding correct that output impedence (of a source)can be >>>modelled by fir/iir filters. >> >> >> The frequency responses of analog systems can be modeled by >> discrete-time filters. 'Impedance' is a concept which does not >> exist in the discrete-time world. >> >> >>>The application is none other than voip slic >>>who tries to adjust its output impedence so that it becomes the conjugate >>>of telephone impedence , thereby maximising power transfer to the >>>telephone.I was struggling to get the idea.. >> >> >> You mix concepts from the analog and digital worlds. >> >> Do you try to match impedances on a physical line which is >> connected to the input of a digital system? If so, it is the >> input stages of the ADC (the anti-alias filter) which should >> be matched to the line impedace. >> >> Again, there is no such thing as impedance in digital systems, >> only numbers. >> >> >>>how a resistor and inductor in series can be >>>modelled by a filter? >> >> >> Use the Laplace transform. Then use BLT to find a discrete-time >> filter which has a similar frequency response as the RLC network. >> >> Rune > >I wonder if what the OP is actually interested in would be "telephone >line equalization networks"?I recall seeing these when I used to work on phone exchange equipment design for Fujitsu in the '90s. The POTS lines are required to be terminated in a particular complex impedance at the exchange. However, the value of the impedance varied from country to country. It is a PITA to have to change a bunch of passive analog components to suit each customer. It is possible to synthesise a complex impedance by filtering the input signal and feeding it back to the line via a reference impedance (which could be real, i.e. a cheap resistor). By changing the filter taps, you can change the effective impedance. IIRC, this was implemented in codecs from Analog Devices. The one part could control: - termination impedance - input equalisation - output equalisation - hybrid balance as well as implementing the codec (i.e. A/D and D/A conversion, antialiasing and reconstruction filtering) function. Regards, Allan
Reply by ●March 6, 20082008-03-06
On Thu, 06 Mar 2008 00:35:25 -0600, "pal.debabrata123" <pal.debabrata123@gmail.com> wrote:>Is my understanding correct that output impedence (of a source)can be >modelled by fir/iir filters. The application is none other than voip slic >who tries to adjust its output impedence so that it becomes the conjugate >of telephone impedence , thereby maximising power transfer to the >telephone.I was struggling to get the idea..the pressing questions are >does a slic behave like a constant current/voltage source?Is the impedence >is in parallel or series? how a resistor and inductor in series can be >modelled by a filter? > >Any insight is greatly appreciated.BTW, you probably don't want maximum power transfer. You do want a particular frequency response and hybrid balance though. Regards, Allan
Reply by ●March 6, 20082008-03-06
"pal.debabrata123" <pal.debabrata123@gmail.com> wrote in message news:POydnWsNAPMwDVLanZ2dnUVZ_qSonZ2d@giganews.com...> Is my understanding correct that output impedence (of a source)can be > modelled by fir/iir filters. The application is none other than voip slic > who tries to adjust its output impedence so that it becomes the conjugate > of telephone impedence , thereby maximising power transfer to the > telephone.I was struggling to get the idea..the pressing questions are > does a slic behave like a constant current/voltage source?Is the impedence > is in parallel or series? how a resistor and inductor in series can be > modelled by a filter? > > Any insight is greatly appreciated. >The standard source (and hence termination) impedance for a 2 wire loop is 600ohms in series with a 2.15uF capacitor (don't ask me where the precision cap value came from). There is not need to provide "maximum power transfer" as you describe it. The need to match source, loop, and termination impedance is about reducing reflections and NOT about maximizing power transfer. Unless you are doing something complex like creating conferencing features, there is little need to know or to be able to model the impedance of a given telephone loop. PBXs with conference features need to maximize the 4 wire loss so as to avoid oscillation when many conference 'legs' are added to a particular conference, so they will have loop modeling circuitry in their 2-to-4-wire circuits (aka hybrids or transhybrids). However, in these cases, the PBX SLICs still provide a simple source/termination impedance to the 2 wire loops as described above. Getting back to your question: There is a company called Tollgrade Communications that created a system that would *present* the impedance seen into a telephone loop back to a remote location, and this location was connected through a digital link. However, as far as I know, there was no signal processing (e.g. IIR or FIR structures) involved in the digital part of this system. If, however, you do ever find a way to model impedances using FIR/IIR techniques, we would certainly like to hear about it. Bob
Reply by ●March 6, 20082008-03-06
On Thu, 6 Mar 2008 07:56:59 -0800, "BobW" <nimby_NEEDSPAM@roadrunner.com> wrote:> >"pal.debabrata123" <pal.debabrata123@gmail.com> wrote in message >news:POydnWsNAPMwDVLanZ2dnUVZ_qSonZ2d@giganews.com... >> Is my understanding correct that output impedence (of a source)can be >> modelled by fir/iir filters. The application is none other than voip slic >> who tries to adjust its output impedence so that it becomes the conjugate >> of telephone impedence , thereby maximising power transfer to the >> telephone.I was struggling to get the idea..the pressing questions are >> does a slic behave like a constant current/voltage source?Is the impedence >> is in parallel or series? how a resistor and inductor in series can be >> modelled by a filter? >> >> Any insight is greatly appreciated. >> > >The standard source (and hence termination) impedance for a 2 wire loop is >600ohms in series with a 2.15uF capacitor (don't ask me where the precision >cap value came from).The good thing about standards is that there are so many of them. Each country has a different standard POTS termination impedance. It's been a while since I've looked at the exact values, but I do recall that most of them were about 600ohm in the voice frequency range, with a capacitive component causing the impedance to fall at higher frequencies. A quick google came up with "In the United States, the loop is nominally 600R, whereas in Britain it is a 300R resistance in series with a parallel combination of 1 kR and 220 nF." Which country does the 600R + 2.15uF come from?>If, however, you do ever find a way to model impedances using FIR/IIR >techniques, we would certainly like to hear about it.This has been done already. Have you searched for "POTS termination impedance" in patents? I would post a link to a datasheet, however all the parts I know that used to do this don't seem to exist any more :( I did find (in an old workbook from the '90s) pages of equations regarding the design of impedance synthesisers. The only really tricky bit is realising that you have to do some of it in continuous time circuits to get accuracy at higher frequencies. The latency of the digital filter will be too high for that. It is all too easy to make an oscillator if presented with an unusual line. Regards, Allan
Reply by ●March 6, 20082008-03-06
"Allan Herriman" <allanherriman@hotmail.com> wrote in message news:6p90t352ok1ngoc6vbr3e0tgkpcase06ms@4ax.com...> On Thu, 6 Mar 2008 07:56:59 -0800, "BobW" > <nimby_NEEDSPAM@roadrunner.com> wrote: > >> >>"pal.debabrata123" <pal.debabrata123@gmail.com> wrote in message >>news:POydnWsNAPMwDVLanZ2dnUVZ_qSonZ2d@giganews.com... >>> Is my understanding correct that output impedence (of a source)can be >>> modelled by fir/iir filters. The application is none other than voip >>> slic >>> who tries to adjust its output impedence so that it becomes the >>> conjugate >>> of telephone impedence , thereby maximising power transfer to the >>> telephone.I was struggling to get the idea..the pressing questions are >>> does a slic behave like a constant current/voltage source?Is the >>> impedence >>> is in parallel or series? how a resistor and inductor in series can be >>> modelled by a filter? >>> >>> Any insight is greatly appreciated. >>> >> >>The standard source (and hence termination) impedance for a 2 wire loop is >>600ohms in series with a 2.15uF capacitor (don't ask me where the >>precision >>cap value came from). > > The good thing about standards is that there are so many of them. Each > country has a different standard POTS termination impedance. It's been > a while since I've looked at the exact values, but I do recall that > most of them were about 600ohm in the voice frequency range, with a > capacitive component causing the impedance to fall at higher > frequencies. > A quick google came up with "In the United States, the loop is > nominally 600R, whereas in Britain it is a 300R resistance in series > with a parallel combination of 1 kR and 220 nF." > Which country does the 600R + 2.15uF come from? >Allan -- I think that you're confusing the line-modeling network's impedance (used in the transhybrid circuit) with the line circuit's source/termination impedance. The 600ohm+2.15uF is from Bell Pub 48002. I have never seen an R||C used for a source/termination in a line circuit. [snip]> Regards, > AllanBob
Reply by ●March 6, 20082008-03-06
On 6 Mar, 06:35, "pal.debabrata123" <pal.debabrata...@gmail.com> wrote:> Is my understanding correct that output impedence (of a source)can be > modelled by fir/iir filters. The application is none other than voip slic > who tries to adjust its output impedence so that it becomes the conjugate > of telephone impedence , thereby maximising power transfer to the > telephone.I was struggling to get the idea..the pressing questions are > does a slic behave like a constant current/voltage source?Is the impedence > is in parallel or series? how a resistor and inductor in series can be > modelled by a filter? > > Any insight is greatly appreciated. �Patent No US 6990191 by Legerity includes the fundamentals of what you want. Personaly, I used a voltage source via resistive feed to realise a constant current dc path plus a programmable ac telco impedance done by FIR filters to get the international matching networks although my app was not a consumer chip but a precision lab measurement system. HTH Steve
