"capacitance transformer", Is this (part of an attenuator) a ?

[dai h.]

(circuit claimed to be from a now defunct "transparent"(sounding) attenuator) :

http://i1237.photobucket.com/albums/...rRevCLight.jpg

What is it doing (the other parts--fixed R and 16ohm L-Pad--look pretty comprehensible)?

(originating thread) :

Rig-Talk • View topic - here are some marshall mod schematics!!

(background: Randall Aiken mentioned "capacitance transformer" in relation to attenuators some time ago) :

Speaker load

[R.G.]

It's most easily understood by replacing the transformer with its equivalent circuit, then referring all of the elements to the primary.

A transformer will convert a secondary load impedance to a primary load impedance, within the limits of its construction and equivalent circuit.

[g1]

From your link: Speaker load
"Note that the large value bipolar caps can be converted to smaller film
(or other type) caps if you use a capacitance transformer (if you don't
know what this is, you shouldn't be messing around with this circuit!).
In fact, if you do it right, you might be able to replace the parallel
inductor with the primary inductance of the capacitance transformer."

Apparently neither of us should be messing around with this circuit, I hope it's ok to look!
I get the impression the capacitance transformer is a way to use smaller type caps. rather than the larger value bipolars?
I am not familiar with the term and google doesn't seem to be either. Maybe there is a different technical name for it?

[Alan0354]

What capacitor? I don't even see a capacitor in the circuit!!! I see a voltage devider formed by the primary (30V) side of the transformer and the 33ohm 50W resistor. I take that the person use the transformer to emulate like a speaker that have inductance. The secondary is loaded with resistors to reflect back to the primary with the inductance. I don't know the L-pad and how it really hooked up. But I think it is a variable attenuator or something.

What is to debunk, it should work, but whether it sounds good or not is a totally different story.

I have the THD attenuator, it is the only one I'd seen that has more compensation and I took a chance on it. It work really good to me. It sounded very good until you try to put on highest attenuation PLUS dial down the volume pot, that's when it loss it.

[R.G.]

From your link:
"Note that the large value bipolar caps can be converted to smaller film
(or other type) caps if you use a capacitance transformer (if you don't
know what this is, you shouldn't be messing around with this circuit!).
In fact, if you do it right, you might be able to replace the parallel
inductor with the primary inductance of the capacitance transformer."
...
I get the impression the capacitance transformer is a way to use smaller type caps. rather than the larger value bipolars?
... Maybe there is a different technical name for it?

A transformer with primary turns Np and secondary turns Ns converts an impedance of X on its secondary to what appears to be an impedance of X * (Np/Ns)² in parallel with its primary, within the limits of its "imperfections".

The imperfections are
- a primary inductance
- leakage inductances Lp and Ls
- wiring resistances Rp and Rs
- an assortment of self capacitances

In theory, a 0.1uF capacitor on the secondary of a transformer which has a 10:1 turns ratio will appear to be 10uF at the primary terminals. If there is any power involved (and in a speaker load, there is) then both the transformer and the secondary capacitor must be able to handle the power the apparent cap at the primary would have to handle. There are other gotchas. One of them is that it's easier to buy/make capacitors than wideband power transformers.

[Enzo]

Alan, the mentioned capacitors are in the "speaker load" link schematics, not the photobucket schematic

[Alan0354]

Alan, the mentioned capacitors are in the "speaker load" link schematics, not the photobucket schematic

Now I see it, it makes a lot more sense.

[dai h.]

Thanks guys. So, since there is no capacitor, this is *not* a "capacitance transformer"?

I see a voltage devider formed by the primary (30V) side of the transformer and the 33ohm 50W resistor. I take that the person use the transformer to emulate like a speaker that have inductance. The secondary is loaded with resistors to reflect back to the primary with the inductance. I don't know the L-pad and how it really hooked up. But I think it is a variable attenuator or something.

yes I think this is supposed to be an attenuator. (Did a bit more googling, and) I suspect (not absolutely sure) it's a copy of this one:

Motherload - speaker simulator, loadbox, amp attenuator, motherload, sequis, palmer, thd, hotplate, weber mass, recording guitar, direct box, guitar di

(see post around middle by same poster:

Rig-Talk • View topic - What's the best attenuator out there?? )

So basically a voltage divider overall (attenuator = taking a portion of the whole and passing it on to the speaker while the amp output sees the "whole" (is loaded)), but what is the advantage of the transformer? (I guess) it looks to the amp output like the reflected load from the side loaded w/the higher value power Rs, has a series inductance (since it's a winding around a core), (and low DCR?), handles some of the power dissipated (through the "670R" 50W x2 resistors). Just a cheaper cleverer way than using a application specific "transformer attenuator"? (As an aside, the other parts (Rs, L-Pad--probably doesn't actually handle 100W) just look to enable continuous attenuation, and it makes sense for those to be located there in the circuit since the L-Pad being an electro-mechanical part that could fail, there should not be catastrophic failure since the load the amp sees should just go up a bit it it does or something happens and connections from speaker not being plugged in, failing, etc. go open on that side.)

[R.G.]

Thanks guys. So, since there is no capacitor, this is *not* a "capacitance transformer"?

I'd say it differently. It is an impedance transformer, like all transformers. It may notionally be called a capacitance transformer if what it transforms is a capacitance. It's a way to get a high capacitance (that is, a low capacitive impedance) from a capacitor which is smaller and probably tighter tolerance.

The reason this matters is that electrolytics have poor tolerance and their capacitance drifts, while non-electros have tighter tolerance, but are large and expensive to make in high capacitances. In power attenuators and crossovers, you need high capacitance at a lowish voltage, but you don't want the value drifting around. Some companies make film capacitors which have acceptable tolerances and power handling capability, but generally at a breathtaking price.

The price you pay for the "capacitor transformer" is that you have to both buy the transformer, and the quality/parasitics of the transformer dirty up the capacitance you're trying to get. That may be OK. Details matter. Cheap transformer => lots of parasitics; low parasitics => expensive transformer, unless you get lucky and **want** the parasitics you get.

what is the advantage of the transformer? (I guess) it looks to the amp output like the reflected load from the side loaded w/the higher value power Rs, has a series inductance (since it's a winding around a core), (and low DCR?), handles some of the power dissipated (through the "670R" 50W x2 resistors).

As I said, replace the transformer at the primary leads with a network which contains:
- a series inductor; this is the sum of the primary and secondary leakage inductances referred to the primary by the impedance ratio
- a series resistor; this is the sum of the primary and secondary wiring resistances referred to the primary by the impedance ratio
- an impedance which is the transform of any secondary load through the square of the turns ratio like any transformer does; think of an output transformer here
- a big inductance in parallel with the transformed secondary impedance; this is the primary inductance of the transformer
- a large and nonlinear resistance in parallel with the transformed secondary impedance, representing the core losses
- parasitic capacitors of various values everywhere, representing the fact that everything has a capacitance to everything else.

In the picture you show, they appear to have used a standard power-line transformer, judging from the voltages;this would be done to get it cheaply. They specify it for 160VA, and rate the secondary load at 50W, so it is likely that the transformer is being used well below its nominal voltages, which keeps it more linear than you'd normally expect for a power line transformer. The ratio of transformation from secondary to primary is the square of the voltage ratio; this is 30/240, or 1/8 squared. Ignoring parasitics on the transformation, the 670R resistance in the secondary appears like a 670/(8*8) = 10.46 ohm resistor in the primary, which is then series'd and paralleled by the transformer parasitics as noted above.

Just a cheaper cleverer way than using a application specific "transformer attenuator"?

Cheaper anyway.

[dai h.]

I'd say it differently. It is an impedance transformer, like all transformers. It may notionally be called a capacitance transformer if what it transforms is a capacitance. It's a way to get a high capacitance (that is, a low capacitive impedance) from a capacitor which is smaller and probably tighter tolerance.

The reason this matters is that electrolytics have poor tolerance and their capacitance drifts, while non-electros have tighter tolerance, but are large and expensive to make in high capacitances. In power attenuators and crossovers, you need high capacitance at a lowish voltage, but you don't want the value drifting around. Some companies make film capacitors which have acceptable tolerances and power handling capability, but generally at a breathtaking price.

The price you pay for the "capacitor transformer" is that you have to both buy the transformer, and the quality/parasitics of the transformer dirty up the capacitance you're trying to get. That may be OK. Details matter. Cheap transformer => lots of parasitics; low parasitics => expensive transformer, unless you get lucky and **want** the parasitics you get.

(Sorry if I'm being annoying with my thick-headedness, but just to be clear) I think I understand that a "capacitor transformer" is a way to achieve the large capacitor (capacitance) with a smaller capacitor for the LC tank circuit portion of a reactive spk. emulating load/network (for the reasons outlined above), but in the schematic shown, does it exist in any form?

As I said, replace the transformer at the primary leads with a network which contains:
- a series inductor; this is the sum of the primary and secondary leakage inductances referred to the primary by the impedance ratio
- a series resistor; this is the sum of the primary and secondary wiring resistances referred to the primary by the impedance ratio
- an impedance which is the transform of any secondary load through the square of the turns ratio like any transformer does; think of an output transformer here
- a big inductance in parallel with the transformed secondary impedance; this is the primary inductance of the transformer
- a large and nonlinear resistance in parallel with the transformed secondary impedance, representing the core losses
- parasitic capacitors of various values everywhere, representing the fact that everything has a capacitance to everything else.

In the picture you show, they appear to have used a standard power-line transformer, judging from the voltages;this would be done to get it cheaply. They specify it for 160VA, and rate the secondary load at 50W, so it is likely that the transformer is being used well below its nominal voltages, which keeps it more linear than you'd normally expect for a power line transformer. The ratio of transformation from secondary to primary is the square of the voltage ratio; this is 30/240, or 1/8 squared. Ignoring parasitics on the transformation, the 670R resistance in the secondary appears like a 670/(8*8) = 10.46 ohm resistor in the primary, which is then series'd and paralleled by the transformer parasitics as noted above.


Cheaper anyway.

(okay, sadly I'm not getting most of this , only so far as what seems to be the internal impedance.)
The spec. sheet gives 10 ohms DCR (5ohms*2 windings) for the PRI, 0.2754 ohms DCR for SEC (0.1377ohms *2 windings). The impedance ratio is 670/10.46, so 64:1. The winding ratio is sqrt (670/10.46) = 8, or 8:1. Internal impedance of the entire power transformer (looking from the perspective of the secondary is 0.2754 + ((10 ohms + fixed R (330 ohms x2?))/64) = 0.2754 + (670/64) = 10.46875 (The series L can be calculated similarly? As in L value of winding + L value of other winding according to the impedance ratio?)

note: a 670 ohm resistor isn't a standard value and looks a bit strange, so I guessed it was two series 330 ohms (standard value) + the DCR of the winding (10 ohms) = the 670 ohms figure.

If the total internal impedance is 10.47 ohms, and this is in series with 33 ohms in parallel with 16 ohms ( = 10.776 ohms--giving the assumed to be 16 ohms spk. a 16 ohm DCR value for simplicity), the total the 16 ohm amp output sees would be about 21.25 ohms.

Also, the PT seems to be something with a standard specification, since I found a couple that appeared extremely similar (I guess it's good if you can just use an "off the shelf" standard item). One example in the RS Components catalog:

http://docs-asia.electrocomponents.c...6b80da0de3.pdf

[R.G.]

I think I understand that a "capacitor transformer" is a way to achieve the large capacitor (capacitance) with a smaller capacitor

Yes; I believe the term refers to the idea of making a high voltage, low-value capacitor appear as a high capacitance/low voltage capacitance on the other side of the transformer.

for the LC tank circuit portion of a reactive spk. emulating load/network (for the reasons outlined above), but in the schematic shown, does it exist in any form?

Not that I can see. What appears there seems to be a resistive transformer - making a low-ohms resistor out of a high-ohms one. I realize I wasn't very clear about that.

The spec. sheet gives 10 ohms DCR (5ohms*2 windings) for the PRI, 0.2754 ohms DCR for SEC (0.1377ohms *2 windings). The impedance ratio is 670/10.46, so 64:1. The winding ratio is sqrt (670/10.46) = 8, or 8:1. Internal impedance of the entire power transformer (looking from the perspective of the secondary is 0.2754 + ((10 ohms + fixed R (330 ohms x2?))/64) = 0.2754 + (670/64) = 10.46875 (The series L can be calculated similarly? As in L value of winding + L value of other winding according to the impedance ratio?)

Look at this article on transformers and look down at "equivalent circuit". Any secondary loading on the secondary of Zsec will be transformed to the primary terminals of the ideal transformer as Zsec* (Np²/Ns²).

So in the schematic, Zsec = 670R. This transforms to the primary as 670/64 = 10.47R. For the equivalent impedance at the actual physical terminals, you would add to this the real resistance of the primary windings (5+5 = 10R) and the transformed resistance of the secondary wire (0.2754 /64 = 0.0043R) for a total resistance seen at the primary of 10R + 0.0043R + 10.47R = 20.47R.

The primary inductance is what it is - which is unfortunate, as it's not easy to control in manufacture. It's big. In general, power transformers are designed so the no-load current in the primary is 1% to 5% of the full load current. Since the primary is (probably) the 240V side, and the little note says "160VA", we can estimate the full load primary current at 160VA/240Vac = 0.667A, and a guess at the no-load current might be 0.007A to 0.035A. We can very rawly estimate the inductive portion of that as maybe half (there's some core loss), and so the inductance may be Xl = 240V/[0.0035A to 0.0175A] = 68.57K to 13.7K; the frequency is (probably) 50 or 60Hz, so the inductance might be 68571/(2*pi*F) to 13700/(2*pi*F) or something like 36H to 182H. This is the crudest sort of estimation, hardly better than a WAG, and may be way off, but it illustrates the point that the primary inductance is BIG. Core loss resistance is both big and nonlinear with voltage excitation.

That primary inductance would appear on the 30V side as 1/64 as big, or about 1/2H to 3H to get even rawer numbers.

The leakage inductances are purely determined by the physical locations of the wires in the coils, and may vary hugely depending on whether and how the primary and secondary are interleaved. There are formulas to estimate them, but only guys who are trying to design output transformers bother. And there is no way to tell what you get in a commercial power transformer but to measure them with a short-circuit test.

note: a 670 ohm resistor isn't a standard value and looks a bit strange, so I guessed it was two series 330 ohms (standard value) + the DCR of the winding (10 ohms) = the 670 ohms figure.

If the total internal impedance is 10.47 ohms, and this is in series with 33 ohms in parallel with 16 ohms ( = 10.776 ohms--giving the assumed to be 16 ohms spk. a 16 ohm DCR value for simplicity), the total the 16 ohm amp output sees would be about 21.25 ohms.

Yep.

Also, the PT seems to be something with a standard specification, since I found a couple that appeared extremely similar (I guess it's good if you can just use an "off the shelf" standard item).

As I mentioned, it's certainly cheaper, and good if you're the builder and seller. As to what the unknowns do to sound quality - well, if you get lucky, it may be OK. If you're designing to use the quirks and nonlinearities of a commercial power transformer for audio, and the accounting and purchasing department buys something with the same (but non-audio) specs, you might get anything in the next batch.
Designing to non-specified quirks of commercial gear is a time-honored way to lose an engineering job.

[dai h.]

R.G., thanks. Sorry I haven't replied! I'm having another tendonitis flareup (seems to be a delayed reaction to forcing open a stuck rice cooker lid--dried rice makes for very effective glue apparently), taking lots of ibuprofen and trying to let the hands rest as much as I can. Been trying to read a bit more hopefully to make a bit more sense out of the info.