phase-shift-follower idea, NOW: `BarberPole-Rectifier`

[puretube]

free information block:

[bioroids]

I think I understand the concept, but how can this be implemented?
Could this be used for ultra clean power supply ac/dc?
Am I missing the point enterely? it's friday...

Luck

Miguel

[puretube]

it`s intended for signals, controlling filters, compressors, etc., only;
in power supplies you won`t mind a long/large time constant...

(it`s late - yer rite...

[bioroids]

Oh I see, the idea is to get clean envelopes with fast attack/decay times?  that sounds more like it

May be better to save the idea for next week ha ha!

Regards

Miguel

[Chico]

Puretube:

Glad to hear that Gerti's party went well!  

As for your idea, I also still need to read through all that material, and more particularly, ponder this idea over the weekend.

That said, I am trying to envision this circuit, and quite frankly,  am struggling a little bit.  

Are you thinking of sweeping/modulating an all pass filter network against a reference signal to get variable phase delay, then using parallel peak hold detectors to select the largest peak of the two?  Would this require carefully controlling the RC values of each all pass filter stage in the network that is being swept/modulated?

Now that would be something.  You could effectively account smooth envelope detection, even for low frequencies.

I look forward to hear what you have cooking up here.

Best regards

Tom

[StephenGiles]

Perzacerly!
Stephen

[DDD]

puretube,
you are going to:
- split the signal into two identical signals
- shift one signal by 90 degrees
- separately rectify the signals
- sum rectified signals
- filter the ripple that is some times lower than usual.
Did I understand you properly?

[StephenGiles]

Ton, would you be using 2 peak hold or 1 peak hold and 1 average hold, then selecting through a nor gate or an or gate or even a whore gate! Then perhaps into a window detector to provide adaptive high or low control voltages to drive 4016 gates as required. Carajo Ché, I theeeeenk
I like theeeeeeeees!!
Stephen

[puretube]

free information block:

[puretube]

Stephen: I intend to keep it all analogue... - no hoes -  :lol:

picturesque imagination: a (body-) surfer right on top of the wave...

[Vsat]

I remember seeing an Orban schematic where a Dome filter was being used (IIRC) as part of an (again IIRC....) envelope follower for a compressor.

He has some very unusual patents... using a freq shifter to compress audio bandwidth into 500 Hz to 1 KHz range for feeding into a spring reverb... then a 2nd complementary freq shift to restore the signal from the spring back to full 20 Hz to 20 KHz bandwidth. Neat idea, a bit heavy on the components...
Regards, Mike

[puretube]

I love that last one: nice case for LM1496...

[puretube]

free information block:

[Vsat]

Made even simpler by sharing the same quadrature oscillator between the two shifters. And the quadrature osc can be fixed in frequency. Still, you can buy a lot of BBDs or memory chips for the cost involved in building/aligning a pair of freq shifters... nevertheless, frequency compression/bandwidth compression  is an interesting concept.
Regards, Mike

[Vsat]

Say you wanted a BBD delay (with 10 KHz cutoff freq) to pass a zero to 20 KHz signal bandwidth... first downshift the audio by 10 KHz... this compresses bandwidth to 10 KHz (to 1/2 original bandwidth in Hertz, 20 KHz becomes 10 KHz, 10 KHz becomes zero Hz, zero Hz becomes minus 10 KHz)... feed this into BBD... then shift up delayed signal (do a reverse shift) of 10 KHz... this restores the original zero to 20 KHz bandwidth.

So this downshift/upshift technique would give either twice the original bandwidth for a given delay time, or twice the delay time for a given bandwidth.... lot of extra parts and possible signal corruption when  it would be simpler to just install a 2nd BBD or memory chip. OTOH with the Orban patent the idea is to provide extended bandwidth beyond what a normal spring can provide... adding a second spring will not extend the bandwidth!
Regards, Mike

[Chico]

Puretube:

Can you flush out this idea of a barberpole rectifier a little more?

My understanding of the dome filter is that a sort of phase difference is being performed.  That is, two "channels" are set up where each channel is a cascaded chain of allpass filters.  By considered selection of the RC components, the phase effect of each stage of each chain can be controlled such that a predetermined broadband phase difference is realized between the outputs of the respective channels.

The original idea was to reduce ripple (presumptively by a factor of 4) because there would be 4 detectors per period compared to one (half wave rectifier).  Thus a smaller time constant smoothing filter and more accurate envelope.

The idea that we are discussing here, is essentially an infinite number of detectors per period using a "barberpole phase shifter"

This brings me to a few thoughts/questions.

The way I see your idea is that the outputs from each channel of a dome filter start out in phase, then drift up to +/-45 degrees out of phase.  The outputs of each dome filter are rectified, and diodes select the output having the highest amplitude.

So is the barber pole really moving in two directions?  Are there two barberpoles?  A first one drifts from 0 to 45 degrees and the second one drifts from 0 to -45 degrees?

Also, wouldn't there be an awful lot of components to adjust to smoothly vary the phase over the broad band?  Would you get "lumps" or "clusters" in the phase response?

Finally (for now) because the dome is a differential phase circuit comprised of capacitors (and thus causal), isn't the the output of each dome channel inherently delayed with respect to the original signal?  Would you use a third all pass network to create a delay of the original signal so that your envelope is actually "lookahead".

If not, is the cost (in delay time, not component count) of a smooth envelope detection short enough to yield a worthwile circuit (that goes for the fixed 4 detector version as well)

Well, I need to head to a party.  

Best regards

Tom



S

[Chico]

Puretube:

Can you flush out this idea of a barberpole rectifier a little more?

My understanding of the dome filter is that a sort of phase difference is being performed.  That is, two "channels" are set up where each channel is a cascaded chain of allpass filters.  By considered selection of the RC components, the phase effect of each stage of each chain can be controlled such that a predetermined broadband phase difference is realized between the outputs of the respective channels.

The original idea was to reduce ripple (presumptively by a factor of 4) because there would be 4 detectors per period compared to one (half wave rectifier).  Thus a smaller time constant smoothing filter and more accurate envelope.

The idea that we are discussing here, is essentially an infinite number of detectors per period using a "barberpole phase shifter"

This brings me to a few thoughts/questions.

The way I see your idea is that the outputs from each channel of a dome filter start out in phase, then drift up to +/-45 degrees out of phase.  The outputs of each dome filter are rectified, and diodes select the output having the highest amplitude.

So is the barber pole really moving in two directions?  Are there two barberpoles?  A first one drifts from 0 to 45 degrees and the second one drifts from 0 to -45 degrees?

Also, wouldn't there be an awful lot of components to adjust to smoothly vary the phase over the broad band?  Would you get "lumps" or "clusters" in the phase response?

Finally (for now) because the dome is a differential phase circuit comprised of capacitors (and thus causal), isn't the the output of each dome channel inherently delayed with respect to the original signal?  Would you use a third all pass network to create a delay of the original signal so that your envelope is actually "lookahead".

If not, is the cost (in delay time, not component count) of a smooth envelope detection short enough to yield a worthwile circuit (that goes for the fixed 4 detector version as well)

Well, I need to head to a party.  

Best regards

Tom



S

[StephenGiles]

Tom - did you ever see my Royal Filter circuit?   Here:
http://www.diystompboxes.com/sboxforum/viewtopic.php?t=23101&highlight=royal+filter

This is basically the front end of an EH Crashpad driving a dual bandpass filter based on Craig Anderton's Bi Filter Follower. Depending on the setting of the start and stop frequency pots, both rising and falling sweeps can be obtained. These sweeps are ripple free - at least they sounded ripple free when I last breadboarded it a couple of years ago. Now it seemed to me that if the voltage at which the start frequency is set could be made as a measure of the strength of the string plucked, and thereafter an independent rate of decay  to a preset voltage controlling the stop frequency of the filters.

Puretube suggested a sample and hold function to pick off the initial peak voltage and it may be that this fairly simple approach could achieve a ripple free auto wah - not wishing, of course to detract from the magnificent discussions already developing on this thread.
Stephen

[puretube]

free information block:

[puretube]

free information block: