why is nobody using capacitance multipliers in pedals?
why not use a capacitance multiplier in a Cry Baby 535q instead of a 6-way rotary with 6 different caps for example?
Probably because simple capacitance multipliers aren't good, and good capacitance multipliers aren't simple.
The name "capacitance multiplier"conjures up images of a neato chip or a transistor or two that magically make a capacitance bigger. Would it were so.
The simple capacitance multiplier is the circuit you get from placing a capacitor to ground on the base of a bipolar and taking the "capacitance" off the emitter. This does make things look like there is a bigger capacitor there, in terms of holding the voltage on the emitter steady, but it's not a capacitance exactly, and must emulate a capacitor to ground, not floating.
Most of the applications for switching several caps want floating caps, i.e. NOT with one lead tied to ground. So the simple circuit won't work.
There exist floating capacitor emulator circuits, but their complexity is generally higher than the pedal circuit you'd use them in.
Although I do wonder -- the feedback node of the crybaby circuit is seen as a variable capacitor by the inductor. What would happen if there were a parallel emitter follower with an adjustable pot coming off the collector of the first transistor? Seems you could use this to push the sweep range up and down.
Of course, as the second capacitor/pot begins to look large compared to the one connected to the treadle, then perhaps you lose overall range...
But maybe there is a sweetspot in there. This circuit is, in a sense, a capacitor multiplier -- so the the crybaby is a good example of something that already uses a capacitor multiplier even though it doesn't look like what I think the OP has in mind.
Just a thought -- now onto LTSpice to see what happens in my crybaby sim;)
What R.G. said.
In this case: this Wah circuit IS a "capacitance multiplier", or rather a cap-divider. When set to say 10nFd cap, and WAH pot is full up, it acts as 10nFd; but when WAH pot is set to 10%, it acts as 1nFd. It is an infinitely-variable capacitor action.
Well, not utterly infinite. Turn down a lot and the action is weak. Starting with a different cap gives you another range. The schematic I found looks like you get several octaves on the switch, plus whatever you get from ankle action.
Also the simple "capacitor multiplier" has one end of the "capacitor" grounded. The Wah circuit wants both ends of the cap off ground. This can be done; complications (and costs) happen.
Speaking of costs: a few nFd caps is no more expensive than the several op-amps needed to begin to make a good floating virtual capacitor.
Cleverer is not always better.
QuoteBut maybe there is a sweetspot in there...now onto LTSpice to see what happens in my crybaby sim
What happens is it works but it's not useful. The range of tweakability doesn't provide a sweet spot -- so you can either jack the range up and down and lose too much range with the treadle, or you can keep ok range with the treadle but the up/down tweak is more like how much you can tweak a 20% cap. The thought experiment was worth a try ;)
Quote from: PRR on July 20, 2015, 09:30:58 PM
When set to say 10nFd cap, and WAH pot is full up, it acts as 10nFd; but when WAH pot is set to 10%, it acts as 1nFd. It is an infinitely-variable capacitor action.
I hate to crawl out on a limb and disagree, but I have been analyzing the Crybaby in depth over the past few months trying to discover an inductorless strategy that is relatively faithful to the original.
When the pot is full forward it acts as 10nF as you say, but when it is 10%, it's more in the direction of 100nF. The simulated capacitance increases in value (resonant frequency goes down) as loop gain increases.
At full throttle gain (treadle all the way back) the simulated capacitance is near 300 nF.
> When the pot is full forward it acts
Matter of perspective. I now see that your perspective is more relevant here.
I was just looking at the pot and buffer. Unity-gain at most.
When you include the first transistor there IS significant voltage-gain, and thus capacitance Multiplication. As you say.
Sorry for the narrow view.
Quote from: PRR on July 20, 2015, 09:52:57 PM
Matter of perspective.
Yes, I think the main points you made still stand intact regardless. I didn't mean to detract from that. Some kind of variable gain feedback on the remote end of the capacitance multiplies its value whether up or down, and by definition, a capacitance multiplier.
There's many reasons to NOT use a capacitance multiplier, but there are some fun uses of them that may be worthwhile. For example strap a variable cap multiplier to a fuzz face in place of the emitter bypass cap. Or stick one on the NFB leg of an opamp stage. Kinda neat. Or make it sweepable via an LFO or envelope follower with a LDR, like the Spaceman WOW signal.
Also probably somewhat useful a in power supplies, as they may have better HF stability or transient response than a regulator or plain cap. Dunno, it's DIY, so why not try?
Imho cap. multipliers hold their place only where they really are useful. They aren't new circuits by any measure, so they tend to sick around only in those places where there are some actual reasons and benefits to use them.
It's rather simple when you think about it from this evolutionary perspective.
Quote from: R.G. on July 20, 2015, 07:46:02 PM
Probably because simple capacitance multipliers aren't good, and good capacitance multipliers aren't simple.
:icon_wink:
Same case with diode-cap voltage multipliers..
Quote from: Transmogrifox on July 20, 2015, 09:48:08 PM
I hate to crawl out on a limb and disagree, but I have been analyzing the Crybaby in depth over the past few months trying to discover an inductorless strategy that is relatively faithful to the original.
When the pot is full forward it acts as 10nF as you say, but when it is 10%, it's more in the direction of 100nF. The simulated capacitance increases in value (resonant frequency goes down) as loop gain increases.
At full throttle gain (treadle all the way back) the simulated capacitance is near 300 nF.
The Crybaby circuit uses an odd variant of the Miller capacitance effect for the variable capacitor. It makes the emitter follower cap look like different values to the inductor by changing the effective voltage gain of the stage it's hooked up around. I had been trying to reason out how the Crybaby worked for some years back in the 1980s and 1990s when that struck me. It was slow work back then, with the internet to work with. :(
The circuit varies the current available to charge the inductor by varying the AC voltage on the outside end of the resonance cap. The cap then seems to be different values to the inductor.
And inductorless version of the Crybaby is easy - but the performance is not good. There are simple "fake inductors" that can be made with gyrators - amplifiers that make a floating capacitor look like a grounded inductor. These are the basis of most graphic EQs. They work in the Crybaby circuit. The problem is that (1) they're noisy; hiss is a real problem with gyrators, and (2) they are completely free of the quirks of the real inductor, many of which are viewed as very valuable tonally by guitarists. The slightly sweetened sound of asymmetrical inductor saturation and the slight distortion when the inductor is pushed hard are completely missing. These things can, of course, be put into the inductor simulation, but the inductor simulation is then several times more parts and complication than the whole Crybaby circuit.
Other versions of an inductorless Crybaby are also easy - use an active filter. There are biquad, multiple feedback, and state variable circuits that do the same thing. But they also get complicated, and don't have the "valuable" side effects without a lot of tinkering.
I just used the 535q as an example...
could be a mid-frequency control on distortion pedal with a 3-band eq or really anything in a pedal...
Quote from: R.G. on July 20, 2015, 07:46:02 PM
Probably because simple capacitance multipliers aren't good, and good capacitance multipliers aren't simple.
The name "capacitance multiplier"conjures up images of a neato chip or a transistor or two that magically make a capacitance bigger. Would it were so.
The simple capacitance multiplier is the circuit you get from placing a capacitor to ground on the base of a bipolar and taking the "capacitance" off the emitter. This does make things look like there is a bigger capacitor there, in terms of holding the voltage on the emitter steady, but it's not a capacitance exactly, and must emulate a capacitor to ground, not floating.
Most of the applications for switching several caps want floating caps, i.e. NOT with one lead tied to ground. So the simple circuit won't work.
There exist floating capacitor emulator circuits, but their complexity is generally higher than the pedal circuit you'd use them in.
is this a floating multiplier?:
(http://www.next.gr/uploads/135-9317.png)
Quote from: nognow on July 21, 2015, 11:16:12 AM
I just used the 535q as an example...
could be a mid-frequency control on distortion pedal with a 3-band eq or really anything in a pedal...
Yes, that is how I understood your question, and that was how I answered.
"Floating" means that one of the two terminals is not connected to ground, or "AC ground" at a fixed voltage or power supply. This is harder to do because instead of having to emulate only the voltage/current at one of the device terminals, the emulator circuit has to emulate the voltage/current at both ends of the cap.
Quoteis this a floating multiplier?:
http://www.next.gr/uploads/135-9317.png
Don't know. I get "not found" when I try that link.
There is one pedal people use that could be better changed to a capacitance multiplier implementation and that is the noise gate. The noise gate cuts off a signal below a selected threshold to avoid having noise swamp the signal at low levels. But back in 1948, RCA was faced with the same thing with phonographs and developed a more interesting circuit to do the same thing without the sudden cutoff that is really jarring to an audience. They called the circuit "Magic Monitor" and it used a 6AV6 and a 6BA6 as the two tubes in the circuit. The 6AV6 amplified the incoming signal and used its diode section to rectify it with the increasing negative value used to bias the 6BA6 which was connected as a reactance tube. At high signal levels, the 6BA6 gain was reduced and it had no effect on the signal arriving at its plate. As the signal level dropped, the rectified bias dropped and the 6BA6 was connected as a variable capacitance across the audio. The audio was fed into the circuit via a parallel RC circuit so if the reactance of the tube was low, it would be two parallel RC circuits in series with little change in bandpass, like an oscilloscope input with parallel RC circuits in the input attenuator. If the signal level dropped, the reactance would increase, shunting the high frequencies and acting like a Dolby circuit to reduce the high frequency component.
I will see if I can find a way to scan the circuit.
Quote from: amptramp on July 21, 2015, 12:50:50 PM
There is one pedal people use that could be better changed to a capacitance multiplier implementation and that is the noise gate. The noise gate cuts off a signal below a selected threshold to avoid having noise swamp the signal at low levels. But back in 1948, RCA was faced with the same thing with phonographs and developed a more interesting circuit to do the same thing without the sudden cutoff that is really jarring to an audience. They called the circuit "Magic Monitor" and it used a 6AV6 and a 6BA6 as the two tubes in the circuit. The 6AV6 amplified the incoming signal and used its diode section to rectify it with the increasing negative value used to bias the 6BA6 which was connected as a reactance tube. At high signal levels, the 6BA6 gain was reduced and it had no effect on the signal arriving at its plate. As the signal level dropped, the rectified bias dropped and the 6BA6 was connected as a variable capacitance across the audio. The audio was fed into the circuit via a parallel RC circuit so if the reactance of the tube was low, it would be two parallel RC circuits in series with little change in bandpass, like an oscilloscope input with parallel RC circuits in the input attenuator. If the signal level dropped, the reactance would increase, shunting the high frequencies and acting like a Dolby circuit to reduce the high frequency component.
I will see if I can find a way to scan the circuit.
Quote from: R.G. on July 21, 2015, 11:33:17 AM
Quote from: nognow on July 21, 2015, 11:16:12 AM
I just used the 535q as an example...
could be a mid-frequency control on distortion pedal with a 3-band eq or really anything in a pedal...
Yes, that is how I understood your question, and that was how I answered.
"Floating" means that one of the two terminals is not connected to ground, or "AC ground" at a fixed voltage or power supply. This is harder to do because instead of having to emulate only the voltage/current at one of the device terminals, the emulator circuit has to emulate the voltage/current at both ends of the cap.
Quoteis this a floating multiplier?:
http://www.next.gr/uploads/135-9317.png
Don't know. I get "not found" when I try that link.
here is the source:
http://www.next.gr/inside-circuits/capacitance-multiplier-l9317.html
Quote from: R.G. on July 21, 2015, 10:05:51 AM
I had been trying to reason out how the Crybaby worked for some years back in the 1980s and 1990s when that struck me. It was slow work back then, with the internet to work with. :(
I have been in a sense duplicating your efforts, but with the advantage of hindsight and your "technology of the ..." page to give me the answer up front.
When you neglect input coupling capacitance and DC blocking caps around the pot it's pretty quick work to get a typical second order low-pass transfer function with a gain term acting on w0, and the effect of the pot on the center frequency is readily apparent. That much of it doesn't take much work, and is a satisfactory explanation for why it works. However, going back to the original patent, the designer saw this as a variable impedance, so I have set myself to manipulate the transfer function until I can see a [gain X capacitor] term jump out somewhere. The synthesis of a variable capacitance apparent when you measure Q in simulation and watch what happens to Q vs w0 in a textbook RLC circuit when you vary only the capacitor.
This might also reveal a way to implement the variable cap in a useful way in other circuits -- especially if a second inductor and capacitor tacked on could be woven into a loop to get something similar to a 4-stage phaser. I have already played with the Crybaby in a feedback loop to get a single notch. A second LC tank would give the second notch and if it could be kept stable it may be a platform for a simple 3-transistor inductor-based phaser effect.
Quote from: Transmogrifox on July 21, 2015, 05:06:07 PM
... However, going back to the original patent, the designer saw this as a variable impedance...
Yeah - I'd love to have had that. When I was messing with this, I didn't have patent searches available. :icon_cry: I did have access to the corporate patent departments who would do patent searches for you based on written descriptions; that was less than no help. I got a pile of paper on things not nearly related to the topic at hand. :icon_lol:
Quote from: nognow on July 21, 2015, 12:56:51 PM
is this a floating multiplier?:
here is the source:
http://www.next.gr/inside-circuits/capacitance-multiplier-l9317.html
No. That causes the output terminal to look like a capacitor to ground.
Worse yet, the capacitor has a resistor in series with it equal to R3, as an inherent part of the circuit operation. You'd really like for a "capacitor" to look as though it had zero Equivalent Series Resistance (ESR), although in many circuits you can stand some resistance especially if you replace part of the resistances that are already there with an ESR. It just makes the design tricky (-ier!).
But it's a neat circuit, similar to the gyrator making the output look like an inductor.
> is this a floating multiplier?:
I believe it is a grounded inductor (disagreeing with R.G., at my risk; but agree about the likely faults).
It also has one more resistor than it needs for most audio work; and higher hiss therefrom.
Some people do use something close (quite an English version :) ):
http://www.sugardas.lt/~igoramps/article57.htm
Brief illustration:
(http://www.sugardas.lt/~igoramps/article57/image003.gif)
vaugely on topic - what effect does a cap mult have on the cap's ESR? do the benefits of the low ESR cap also multiply?
Quote from: PRR on July 22, 2015, 12:01:26 AM
I believe it is a grounded inductor (disagreeing with R.G., at my risk; but agree about the likely faults).
It also has one more resistor than it needs for most audio work; and higher hiss therefrom.
I didn't go compare it part for part against the standard gyrator circuit. I (at my own risk! :icon_biggrin: ) just took their word that it's a capacitance multiplier. I've seen similar so I let it go with what it said it was.
Quote from: duck_arse on July 22, 2015, 09:23:31 AM
vaugely on topic - what effect does a cap mult have on the cap's ESR? do the benefits of the low ESR cap also multiply?
I think it varies tremendously with the cap multiplier circuit. The one circuit referred to in the link (if it is a cap multiplier, not an inductor sim; I guess I gotta to look that up) makes no bones about it - it inserts R3 in series with the cap, whatever the capacitance's ESR is. I think that will be the case for all cap multipliers, and probably gyrators.
If you think about it, there has to be a current-sensing part in the multiplier circuit. A capacitance differs from a resistor in that the voltage and current are not directly related, they're time-related, as in I = C* dv/dt. So to fake both the voltage and current, you have to sense the current through some element in the circuit, and a resistor is the right way to do that. That sensing resistor is likely to appear in series with the "multiplied" capacitance no matter what you do.
Since a current sense resistor is almost certain to be bigger than even a normal cap's ESR, the effect of the multiplier on the cap's own ESR is immaterial - the inserted resistance is many, many times bigger and unrelated to the actual cap's ESR. So no, the benefits of a low-ESR cap are not multiplied; in fact, they're completely lost.
This schematic of the RCA VRA 141 amplifier and Magic Monitor shows the use of the Magic Monitor with its reactance tube capacitance multiplier to reduce treble frequencies for low level signals. This is a preferred alternative to a noise gate:
(http://i224.photobucket.com/albums/dd222/amptramp1/VRA141_zpsfwwfvdsu.jpg)
There is no amplification of the signal from the phono input to the Magic Monitor output - the circuit acts as an attenuator with a capacitance that increases as signal level decreases to keep noise and hiss out of the output. The input is amplified and rectified and drives a reactance tube to increase the shunt capacitance to ground as the signal gets quieter.
I found this:
(http://www.tradeofic.com/uploadfile/ic-circuit/200972322381853.gif)
if it is not ok, could you please help me find a schematic that fits?...
The National Semiconductor app note capacitance multiplier is good, but it shows obsolete devices - the LM102 is not an op amp, it is a unity-gain buffer and the LM101A requires 150 pF for C2 whereas more recent amplifiers would not require C2 at all. The thing you have to note about this design is that (1+(Rb/Ra)) input signal has to remain within the rails so for this circuit operating at ±15 volt rails, the input has to remain within 1/11 of this or 1.36 volts and this does not take into account the fact that the op amp input range is not rail-to-rail, so this is reduced even further.
Let's look at how this works: the signal is applied at C-> and with R2 adjusted to the right, the output of the LM101A is effectively grounded and the value of C looks like 0.1 µF, as you would expect. As you move the slider of the pot to the left, you get inverting gain out of the LM101A so the capacitor has the input on the left and an inverted proportional signal on the right, causing more current to flow. But more current through a capacitor at a given frequency looks like an increased value of capacitor. That is the capacitance multiplier effect.
Quote from: nognow on July 22, 2015, 02:57:20 PM
I found this:
...
if it is not ok, could you please help me find a schematic that fits?...
(1) The value of "OK" has to be defined. A horse is OK for a single person transportation device if and only if the person has the skills to ride it and means to keep it fed and watered. A horse is not OK as a means for flying. Well, at least not for flying more than a few meters. :icon_lol:
If your value of "OK" includes a capacitance which has one lead grounded, that circuit is OK. If your definition of "OK" has to have the multiplied capacitance with both leads not connected to ground, that circuit is "not OK".
(2) A good way to figure out whether a capacitance multiplier is grounded or floating is to count the output leads. If there is only one output lead (as in arrow pointing to "C" in this one), then a grounded second lead is implied, and it's not a floating capacitance multiplier.
So, let's write down the objectives: do you or do you not need a floating capacitance for your multiplied capacitance? It's really a big deal in figuring out what will be OK.
has any of you guys found a good floating multiplier?
Quote from: nognow on July 23, 2015, 09:47:29 AM
has any of you guys found a good floating multiplier?
So, let's write down the objectives: do you or do you not need a floating capacitance for your multiplied capacitance? It's really a big deal in figuring out what will be OK.
Quote from: R.G. on July 23, 2015, 09:53:59 AM
Quote from: nognow on July 23, 2015, 09:47:29 AM
has any of you guys found a good floating multiplier?
So, let's write down the objectives: do you or do you not need a floating capacitance for your multiplied capacitance? It's really a big deal in figuring out what will be OK.
I'm not sure...the stock caps that I want to replace are obviously floating...
how can I decide?
The decision should be simple: If it isn't grounded, you can't use it.
Quote from: bool on July 23, 2015, 11:13:31 AM
The decision should be simple: If it isn't grounded, you can't use it.
Quote from: R.G. on July 23, 2015, 09:53:59 AM
Quote from: nognow on July 23, 2015, 09:47:29 AM
has any of you guys found a good floating multiplier?
So, let's write down the objectives: do you or do you not need a floating capacitance for your multiplied capacitance? It's really a big deal in figuring out what will be OK.
I'll go with floating I guess...
OK then, floating it is.
Google gives over 150,000 results for "floating capacitance multiplier". A quick look at the top results and the images show that really floating capacitance multipliers are possible, but require lots of active elements and suffer from various ills. Many of the results recount the use of differential voltage current conveyors (DVCCs) and current controlled current conveyors (CCCs) as well as multiple OTAs, and list this as problems in making floating capacitance multipliers. Here's an excerpt from one of the papers:
QuoteUnfortunately, the circuits in [19] [21] consist of many different active elements while the reported circuit in [20] comprises 4 OTAs and the reported circuit in [22] comprises 2 CCCCTAs. There can be found that the recently proposed topologies need to employ a lot of elements, which is not convenient to either further fabricate in IC or practically implement and providing higher power consumption.
From this I make:
(1) the current state of the art for floating capacitance multipliers is that they're complex (4 x OTA!)
(2) they're not easily amenable to reduction to an IC
There is also a fair amount of discussion on temperature sensitivity and the need for matching components and so on - the bane of precision analog work.
I have no delusion that my knowledge is comprehensive, but as far as I can tell - as I said earlier - the good ones aren't simple, and the simple ones aren't good. And in circuit design for pedals, that's often enough to rule them out. Pedal circuit design is inherently limited in how much circuitry can be used, just the nature of the beast.
If the real question is can we provide you with a simple, easy capacitance multiplier for you to use in some unstated circuit you're working on, I can't. Maybe someone else has run into a simple, easy, economic circuit that does it.
Quote from: R.G. on July 23, 2015, 12:16:00 PM
OK then, floating it is.
Google gives over 150,000 results for "floating capacitance multiplier". A quick look at the top results and the images show that really floating capacitance multipliers are possible, but require lots of active elements and suffer from various ills. Many of the results recount the use of differential voltage current conveyors (DVCCs) and current controlled current conveyors (CCCs) as well as multiple OTAs, and list this as problems in making floating capacitance multipliers. Here's an excerpt from one of the papers:
QuoteUnfortunately, the circuits in [19] [21] consist of many different active elements while the reported circuit in [20] comprises 4 OTAs and the reported circuit in [22] comprises 2 CCCCTAs. There can be found that the recently proposed topologies need to employ a lot of elements, which is not convenient to either further fabricate in IC or practically implement and providing higher power consumption.
From this I make:
(1) the current state of the art for floating capacitance multipliers is that they're complex (4 x OTA!)
(2) they're not easily amenable to reduction to an IC
There is also a fair amount of discussion on temperature sensitivity and the need for matching components and so on - the bane of precision analog work.
I have no delusion that my knowledge is comprehensive, but as far as I can tell - as I said earlier - the good ones aren't simple, and the simple ones aren't good. And in circuit design for pedals, that's often enough to rule them out. Pedal circuit design is inherently limited in how much circuitry can be used, just the nature of the beast.
If the real question is can we provide you with a simple, easy capacitance multiplier for you to use in some unstated circuit you're working on, I can't. Maybe someone else has run into a simple, easy, economic circuit that does it.
any other options to adjust capacitance perhaps?...
Quote from: nognow on July 23, 2015, 03:00:16 PM
any other options to adjust capacitance perhaps?...
Switches, including high frequency switching where the duty cycle makes a capacitor appear duty-cycle-times smaller than it really is.
There are adjustable capacitors, where a set of capacitor plates is inserted or removed from between another set of plates. Big, mechanically clumsy, fragile, etc.
Here's an idea - tell us what you're trying to do, and we can hypothesize some alternate ways to do it that may be simpler and easier.
Unless the idea is just to use adjustable capacitors and/or capacitor multipliers. We're close to exhausting that one, I think.