Jump to content
IGNORED

Power Dynamics, Cables and Constant Current Sourced Amplifiers


jabbr
 Share

Recommended Posts

Something occurred to me with some of the recent talk about power cables, conditioners etc possibly limiting current inflow ... many outstanding Class A amplifiers are designed with cascodes and constant current sources. Similarly power supplies with input chokes all limit current inflow *by design*. I suppose if the current dynamics explanations for differences in AC cables, for example, are correct then this would only apply to specific amplifiers and other equipment which doesn't limit current inflow by design.

Custom room treatments for headphone users.

Link to comment
Share on other sites

Something occurred to me with some of the recent talk about power cables, conditioners etc possibly limiting current inflow ... many outstanding Class A amplifiers are designed with cascodes and constant current sources. Similarly power supplies with input chokes all limit current inflow *by design*. I suppose if the current dynamics explanations for differences in AC cables, for example, are correct then this would only apply to specific amplifiers and other equipment which doesn't limit current inflow by design.

 

I suppose it would depend on whether the piece was designed with the thought that external current flow and/or its rate of change would not be limited. (I.e., whether the designer rather than the external environment would control the relevant parameters, such as slew rate.)

One never knows, do one? - Fats Waller

The fairest thing we can experience is the mysterious. It is the fundamental emotion which stands at the cradle of true art and true science. - Einstein

Computer, Audirvana -> optical to EtherREGEN -> microRendu -> ISO Regen -> Pro-Ject Pre Box S2 DAC -> Spectral DMC-12 & DMA-150 -> Vandersteen 3A Signature.

Link to comment
Share on other sites

I suppose it would depend on whether the piece was designed with the thought that external current flow and/or its rate of change would not be limited. (I.e., whether the designer rather than the external environment would control the relevant parameters, such as slew rate.)

 

How about this type of design: https://passlabs.com/articles/cascode-amp-design

Custom room treatments for headphone users.

Link to comment
Share on other sites

How about this type of design: https://passlabs.com/articles/cascode-amp-design

 

From the tech article:

 

the common base device shields the gain transistor from voltage changes in the circuit

 

So what happens when the amp needs more current from the wall - the circuit in the amp reduces the resistance to hold the voltage constant? How would that be affected by an external current source that limited the rate of change in current?

 

I'd also be interested to see if Papa or Pass Labs have ever made/posted/included in manuals any recommendations for or against power conditioning.

One never knows, do one? - Fats Waller

The fairest thing we can experience is the mysterious. It is the fundamental emotion which stands at the cradle of true art and true science. - Einstein

Computer, Audirvana -> optical to EtherREGEN -> microRendu -> ISO Regen -> Pro-Ject Pre Box S2 DAC -> Spectral DMC-12 & DMA-150 -> Vandersteen 3A Signature.

Link to comment
Share on other sites

How about this type of design: https://passlabs.com/articles/cascode-amp-design

 

It is a very good question, but there are some serious misconceptions here as to what part of a circuit is "constant current" and what the time constants involved are.

 

In the listed article note that the circuit that is listed as constant current is the input buffer or gain stage, NOT the actual output circuit driving the speaker. The part using cascodes and constant current is a very good way to do things, I have done similar stuff in my designs for many years (I use cascodes all over the place). The output shown is NOT constant current, the cascode provides a constant VOLTAGE on the driven transistor, but the CURRENT drawn from the power supply to feed the speaker is still varying with the signal. This is still a good thing, it takes away one of the variables in the operating point, but it is not constant current.

 

Let's look at a class A amp, the simplest is the single ended amp with a transformer (such as the infamous SET), for ease of analysis lets do this with a plate choke and separate transformer (the infamous "parafeed" topology). The tube gets biased so that with no signal it is pulling half the maximum current. The voltage on the plate is essentially the supply voltage (minus the DCR drop of the choke). Now feed it a sine wave, the current through the tube goes from much less than half to somewhat lower than max current, the choke actually works such as to increase the voltage higher than PS voltage.

 

A neat part of this is that with a large enough choke the magnetic field in the choke actually supplies the difference in current between low and high. So this looks like it is a constant current amp.

 

But it isn't. The above only holds true if the signal is a constant sine wave. Real music goes up and down, this causes some PS current change when the signal is changing. So in reality for real music there ARE current changes that have to be dealt with by the power supply. Admittedly it is a much better situation than what you have for most solid state amps. (you can do the above topology with transistors, but that is REALLY rare) The amplitude of the current changes are proportional to how much the signal is changing. So for a lot of modern highly compressed recordings, it's not that big, but for highly dynamic recordings it can be pretty large.

 

Class A just means that the output transistors never turn off, not that they are constant current. With a common solid state class A design, the current from the PS is still following the signal waveform, going up and down during the waveform. Much worse than the above which only has current change when the signal gets louder of softer. Now you CAN put a large choke in a solid state amp to get the same behavior as above, but it will be a BIG HEAVY EXPENSIVE choke, I don't know of anybody that does this commercially (I have built an amp using this myself).

 

There IS a way to get truly constant current from a solid state design, but I don't know of anybody has ever sold such a thing. You put a solid state CCS in each supply rail, it works, but is insanely inefficient. I actually built such a thing years ago, it sounds amazing, but the efficiency was around 5%, yep that means that for a max 10 watt output you are putting in 200 watts, all the time, since it is "constant", 200 watts if the output is 10 watts, 200 watts if the output is half a watt.

 

It IS certainly doable to have a constant current PREAMP, but not a power amp.

 

So even in the best existing amps you have varying current loads on the PS. It still has to be able to supply those changes and deal with things like repetitive transients.

 

So does the FEED to the PS make any difference to this? Yes it does. The AC feeding the PS has an impedance VS frequency just like any other system does, you want that to be low over the whole frequency range of interest. A normal household AC feed is going to be pretty low. Everything in-between the feed and the amp increases the impedance somewhat, meters, panels, breakers, branch circuits, outlets, plugs, power cords, all can play a part in this. If all this is done well, it doesn't cause too much of a change, but it CAN if care is not taken in the implementation.

 

An interesting tidbit is that many systems designed to "cleanup the line" actually make this much worse than it should be by adding inductance to the line, this can significantly decrease the PSes capability to recover from transient events.

 

So the upshot is that well done tube amps probably have less of an impact from AC delivery than solid state amps, but are still not immune to it. Pretty much any solid state amp can be adversely affected by increases in the impedance of the AC delivery system.

 

John S.

Link to comment
Share on other sites

Hi Jabbr

Does something like the attached schematic of my highly modified Silicon Chip magazine design come close enough?

The typical current source does not fully take into account the loading effects of the Bias Current drawn by the next stage, which is why I included a series diode and trimpot in one side of the differential pair.

An E.E. like John Swenson could achieve similar by adding an additional transistor to take these differences into account though.

Note also the JLH PSU add-on for the front end of the amplifier, and also that the power for both channels is obtained via individual voltage regulators powered by separate transformers, or separate secondary windings

A friend nicknamed this amplifier "The Holo Amp", as in Holographic.

The DIY Audio Current Mirror thread refers to the original version before adding an offset corrector to make it fully DC coupled.

 

Regards

Alex

 

 

 

 

http://imageshack.com/a/img538/1862/HVvdeQ.jpg

 

Current Mirror Discussion - Page 15 - diyAudio

 

How a Digital Audio file sounds, or a Digital Video file looks, is governed to a large extent by the Power Supply area. All that Identical Checksums gives is the possibility of REGENERATING the file to close to that of the original file.

PROFILE UPDATED 13-11-2020

Link to comment
Share on other sites

Jabbr

Re above.

 

I meant Current Mirror, not Current Source.

 

How a Digital Audio file sounds, or a Digital Video file looks, is governed to a large extent by the Power Supply area. All that Identical Checksums gives is the possibility of REGENERATING the file to close to that of the original file.

PROFILE UPDATED 13-11-2020

Link to comment
Share on other sites

It is a very good question, but there are some serious misconceptions here as to what part of a circuit is "constant current" and what the time constants involved are.

 

In the listed article note that the circuit that is listed as constant current is the input buffer or gain stage, NOT the actual output circuit driving the speaker.

 

The design is fully cascoded. From the Nelson Pass 1978 patent: https://www.google.com/patents/US4107619 in which:

The approach of the present invention has been to devise a transistor power amplifier with nearly constant voltage, constant current across and through the device, respectively. Voltage is controlled by using a common base transistor having a constant voltage source connected between its base and a reference line, with the common base transistor having its emitter connected to the collector of a common emitter or common collector transistor, with the emitter of the latter connected to the same reference line. This circuit configuration is a cascode connection which is used in gain stages and output stages of the present transistor amplifier for maintaining a constant voltage.

A second aspect of the transistor amplifier of the present invention involves pulling a constant amount of current through the amplifier output. This is achieved by employing the current output of a first low power amplifier having a relatively distortion free, voltage stabilized signal, as described above as the signal bearing portion of a larger constant current amplifier. Current from the low power amplifier is fed through a load to a second, high power amplifier of constant current output and gain less than unity. The second amplifier is maintained outside of the feedback loop of the first amplifier but the currents of the two are summed, so that the second amplifier does the greater amount of work through the load, but the sonic quality from the load is derived from the first amplifier. By maintaining nearly constant current through the load and constant voltage across the amplifier, distortion may be reduced by 10-30 times.

 

But of course the main point here is that a proper power supply design will insulate the AC mains from the transistors. As you note:

A neat part of this is that with a large enough choke the magnetic field in the choke actually supplies the difference in current between low and high. So this looks like it is a constant current amp.

An choke input filtered PS *does* in fact limit the inrush current, and *does* increase impedance but of course this is followed by capacitance which decreases output impedance. There need not be strong correlation between the input impedance of a specific power supply and the output impedance, e.g. capacitance multipliers and other active means of reducing output impedance.

 

In the setting of a choke input filtered cap multiplier PS how much does the AC mains impedance matter?

 

 

There IS a way to get truly constant current from a solid state design, but I don't know of anybody has ever sold such a thing. You put a solid state CCS in each supply rail, it works, but is insanely inefficient.

 

Like: (1978) http://www.firstwatt.com/pdf/art_a40.pdf,

(1994) http://www.firstwatt.com/pdf/art_zen_amp.pdf

http://www.firstwatt.com/pdf/art_zv2.pdf ?

 

Do you mean "exactly" constant current as opposed to "reasonably" constant current?

 

Aren't robust pi filters using chokes and caps a sign of a well made amp? part of the correlation between amp cost and weight?

Custom room treatments for headphone users.

Link to comment
Share on other sites

So does the FEED to the PS make any difference to this? Yes it does. The AC feeding the PS has an impedance VS frequency just like any other system does, you want that to be low over the whole frequency range of interest. A normal household AC feed is going to be pretty low. Everything in-between the feed and the amp increases the impedance somewhat, meters, panels, breakers, branch circuits, outlets, plugs, power cords, all can play a part in this. If all this is done well, it doesn't cause too much of a change, but it CAN if care is not taken in the implementation.

 

An interesting tidbit is that many systems designed to "cleanup the line" actually make this much worse than it should be by adding inductance to the line, this can significantly decrease the PSes capability to recover from transient events.

 

So the upshot is that well done tube amps probably have less of an impact from AC delivery than solid state amps, but are still not immune to it. Pretty much any solid state amp can be adversely affected by increases in the impedance of the AC delivery system.

 

Ok, this is the gist of the discussion. Rather than discuss perfectly constant power draw amplifiers, instead lets look at amplifier designs that have differing degrees of current draw depending on music. Class A, for example, will have a certain minimum current draw at rest and so the ratio of current with a given dynamic to silence is clearly lower than Class B. Similarly different power supply designs will have differing input and output impedances. A PS design with a relatively high input impedance and a low output impedance will be less sensitive to variations in AC power cable impedance.

 

The amp design with a *relatively* large static current draw (or less change in current draw as a function of signal), as well as a *relatively* high input and low output impedance, should have less SQ effect with different AC cables. If that *were actually* the case, then it would lend support to the idea that the SQ effects of different AC power cables are due to different impedances.

Custom room treatments for headphone users.

Link to comment
Share on other sites

Hi Jabbr

Does something like the attached schematic of my highly modified Silicon Chip magazine design come close enough?

 

The current mirror is another design which has a similar effect. Used in shunt regulators e.g. Salas Reflektor-D, the concept of the shunt regulator is that a relatively constant amount of power is used regardless of draw, the rest being shunted away.

 

I don't know enough about your design to say, but it looks like the current mirror is on the input and isn't there a cascoded output stage?

Custom room treatments for headphone users.

Link to comment
Share on other sites

Jabbr

In the supplied schematic, the Current Mirror is used to supply roughly equal current to both halves of the Differential Pair, however many typical Current Mirror designs aren't perfect, and there is often >500mV difference between the collector voltages of the Differential Pair.

The VAS (Voltage Amplifier Stage) of the amplifier uses a Cascode for wide bandwidth and linearity.

Personally, I think we have come a long way since those original Nelson Pass designs.

Many of the problem areas in an amplifier have been addressed in books by Douglas Self.

Have you read any of his Power Amplifier design books ?

 

Regards

Alex

 

How a Digital Audio file sounds, or a Digital Video file looks, is governed to a large extent by the Power Supply area. All that Identical Checksums gives is the possibility of REGENERATING the file to close to that of the original file.

PROFILE UPDATED 13-11-2020

Link to comment
Share on other sites

.

Personally, I think we have come a long way since those original Nelson Pass designs.

Many of the problem areas in an amplifier have been addressed in books by Douglas Self.

Have you read any of his Power Amplifier design books ?

 

Yes, the point I am trying to make is that at the very least a 1978 patent describes a constant voltage-constant current amp design and aside from efficiency as well as sound quality issues, as well as whether the practical implementation is entirely constant -- that such amps should be less or not at all subject to differences in AC power cable impedance.

Custom room treatments for headphone users.

Link to comment
Share on other sites

From the tech article:

 

 

 

So what happens when the amp needs more current from the wall - the circuit in the amp reduces the resistance to hold the voltage constant? How would that be affected by an external current source that limited the rate of change in current?

 

I'd also be interested to see if Papa or Pass Labs have ever made/posted/included in manuals any recommendations for or against power conditioning.

 

He does talk about it in the manual. He recommends not using any power products with his amps. Given the design, you shouldn't hear a difference. That said, I've always been able to hear differences after market PC's make on the Pass amps I had.

Link to comment
Share on other sites

Something occurred to me with some of the recent talk about power cables, conditioners etc possibly limiting current inflow ... many outstanding Class A amplifiers are designed with cascodes and constant current sources. Similarly power supplies with input chokes all limit current inflow *by design*. I suppose if the current dynamics explanations for differences in AC cables, for example, are correct then this would only apply to specific amplifiers and other equipment which doesn't limit current inflow by design.

 

Have a look at ESP's website. Once there, you can try some listening tests they make available. You listen through headphones while they cycle back and forth between stock, and the PC's they make. I wouldn't make a decision based just on those tests, of course, but it is interesting, and may find it useful.

Link to comment
Share on other sites

He does talk about it in the manual. He recommends not using any power products with his amps. Given the design, you shouldn't hear a difference. That said, I've always been able to hear differences after market PC's make on the Pass amps I had.

 

 

I own two Pass Labs products the XA30.5 and the INT-60. Neither manual address power conditioners.

The Truth Is Out There

Link to comment
Share on other sites

I own two Pass Labs products the XA30.5 and the INT-60. Neither manual address power conditioners.

 

I'd ask how they sound, but it would just make me jealous. ;)

One never knows, do one? - Fats Waller

The fairest thing we can experience is the mysterious. It is the fundamental emotion which stands at the cradle of true art and true science. - Einstein

Computer, Audirvana -> optical to EtherREGEN -> microRendu -> ISO Regen -> Pro-Ject Pre Box S2 DAC -> Spectral DMC-12 & DMA-150 -> Vandersteen 3A Signature.

Link to comment
Share on other sites

I own two Pass Labs products the XA30.5 and the INT-60. Neither manual address power conditioners.

 

I haven't read every manual put out by Pass Labs. But the ones I have read say something like this.

 

"The amplifier is powered by a toroidal transformer which charges .16 Farad capacitance. Thisunregulated supply feeds the output transistors only with a full power ripple of about .2 volt.The power draw of this system is constant regardless of the music playing through theamplifier. As such, it does not depend on a high quality AC outlet or special power cords,since the dynamic performance does not create a variation in AC line draw. If the AC line isrunning low, the output stage will bias to a higher current level by way of compensation.The amplifier is stable into any load impedance or reactance including a direct short, and willdeliver clean audio signal into as low as 2 ohms at 120 watt peaks."

Link to comment
Share on other sites

I haven't read every manual put out by Pass Labs. But the ones I have read say something like this.

 

"The amplifier is powered by a toroidal transformer which charges .16 Farad capacitance. Thisunregulated supply feeds the output transistors only with a full power ripple of about .2 volt.The power draw of this system is constant regardless of the music playing through theamplifier. As such, it does not depend on a high quality AC outlet or special power cords,since the dynamic performance does not create a variation in AC line draw. If the AC line isrunning low, the output stage will bias to a higher current level by way of compensation.The amplifier is stable into any load impedance or reactance including a direct short, and willdeliver clean audio signal into as low as 2 ohms at 120 watt peaks."

 

Interesting and thanks. I personally don't use a power conditioner. Since the personnel at Reno HiFi noted its not necessary due to the design and build of the amp.

The Truth Is Out There

Link to comment
Share on other sites

I'd ask how they sound, but it would just make me jealous. ;)

 

LOL, They sound pretty darn good. The only issue with the little xa30 is its in a small room and the room temps raises after a long listening sessions. What comes in handy is the AC outflow is 8 feet above it. Ps comes in great in the winter months.

The Truth Is Out There

Link to comment
Share on other sites

Personally, I think we have come a long way since those original Nelson Pass designs.

Many of the problem areas in an amplifier have been addressed in books by Douglas Self.

Have you read any of his Power Amplifier design books ?

 

Nelson has also come a long way since 1978. Self, Cordell and others. I'm not trying to start a debate on the "best" amplifier design, though I absolutely love Pass, not just because of the sound, but even more so his passion for teaching and continuous publication of multiple circuits. I pointed out a patent from 1978 which claims to be constant current-constant voltage. John Swenson had said that such a beast had not been commercialized.

Custom room treatments for headphone users.

Link to comment
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
 Share



×
×
  • Create New...