Amperes per channel and Filter Caps?

[rockman85]

I've noticed this spec on a few other amplifiers, namely Adcom & Parasound. 

e.g. The A21+ has 60 amperes per channel and 108,000 uf filter cap.

What exactly do these specs mean, how much do they matter & why doesn't Emotiva list these specs?

[pedrocols]

Good question.

[bootman]

Caps are used as a filter and storage of power.

That is about as simple as I can state it without going in to the deep end of power amp design.

Keep in mind that in marketing, very large numbers are usually interpreted as "better" by the buying public.

[DYohn]

And a high current amp only means it can drive lower impedances.

[405x5]

Jul 18, 2019 14:27:27 GMT -5 rockman85 said:

I've noticed this spec on a few other amplifiers, namely Adcom & Parasound. 

e.g. The A21+ has 60 amperes per channel and 108,000 uf filter cap.

What exactly do these specs mean, how much do they matter & why doesn't Emotiva list these specs?

First part of your question is the amplifier’s spec. and all high end amps list their output.

Type and “value” of the caps is a feature of the design build and would be impractical to market comparison.

Bill

[Lonnie]

In a conventional transformer based power amplifier, secondary storage capacitors are needed to carry the load of the amp at frequencies above 60hz. The more the better. Now with a switching power supply, the secondary capacitance is simply there as a filter because the power supply itself is operating at a frequency well above any audio frequency. So this spec has no meaning to a switching power supply but does have value to an older transformer design.

The amperage output is meaningless and simply marketing hype because the output current of the audio stage is a factor of rail voltage to load impedance. For example, if the amplifier is producing 500 watts into a 4 ohm load the output current is roughly 11 amps. All quality audio stages are designed to produce well above whatever the rated power is as a safety margin to insure the longevity of the amp. So stating it can produce XX current is just marketing and meaningless in the real world. I designed an amp many years ago that had 36) 250 watt output devices. In theory it could produce several hundred amps of current, but in reality, the most it would produce into a 4 ohm load was only 47 amps. So don't let these kinds of things fool you.

Lonnie

[Lonnie]

While I'm on the subject of bogus or irrelevant figures lets talk about Signal to Noise. The average room noise floor is 50db. So anything better then that means its inaudible. To that end any decent piece of gear today will have a Signal to Noise ration well above the noise floor. So in my opinion its a spec that no longer has any meaning. Same goes for Cross Talk. Anything greater than 50db becomes inaudible.

OK, I'll get off my soapbox now.

Lonnie

[AudioHTIT]

Jul 19, 2019 8:27:28 GMT -5 Lonnie said:

While I'm on the subject of bogus or irrelevant figures lets talk about Signal to Noise. The average room noise floor is 50db. So anything better then that means its inaudible. To that end any decent piece of gear today will have a Signal to Noise ration well above the noise floor. So in my opinion its a spec that no longer has any meaning. Same goes for Cross Talk. Anything greater than 50db becomes inaudible.

OK, I'll get off my soapbox now.

Lonnie

You’re point is well taken that usually the noise floor of the room is above that of the gear, but I’ll push back a little, mostly because my rural location affords a very quiet environment often in the 20’s at night (when the crickets, frogs, and birds cooperate). I’m about to send an amp back who’s fan can be measured at 37 dBA SPL and is clearly audible between songs and behind quiet scenes (like water running in the next room). Mechanical noises like this need to somehow be incorporated into a dynamic range spec as they clearly infringe on ones enjoyment, and to your point make the Signal to Noise spec meaningless.

[geeqner]

Jul 19, 2019 7:55:55 GMT -5 Lonnie said:

In a conventional transformer based power amplifier, secondary storage capacitors are needed to carry the load of the amp at frequencies above 60hz. The more the better. Now with a switching power supply, the secondary capacitance is simply there as a filter because the power supply itself is operating at a frequency well above any audio frequency. So this spec has no meaning to a switching power supply but does have value to an older transformer design.

The amperage output is meaningless and simply marketing hype because the output current of the audio stage is a factor of rail voltage to load impedance. For example, if the amplifier is producing 500 watts into a 4 ohm load the output current is roughly 11 amps. All quality audio stages are designed to produce well above whatever the rated power is as a safety margin to insure the longevity of the amp. So stating it can produce XX current is just marketing and meaningless in the real world. I designed an amp many years ago that had 36) 250 watt output devices. In theory it could produce several hundred amps of current, but in reality, the most it would produce into a 4 ohm load was only 47 amps. So don't let these kinds of things fool you.

Lonnie

Although I DO remember seeing an article in one of the trade-rags, back in the (1980's?) - when somebody did an interesting "experiment":

-took (2) High-End Mark Levinson Monoblocks  (One of their High-Current units with Passive Heatsinks / NO Noisy Fans)
-Ran Square Wave signals, 180-degrees out of phase into each Amp
-Parallel-Wired the Outputs together
-ARC WELDED with the Output

Relevant in the "Real World" = NO
Pretty darned COOL / IMPRESSIVE = YES

[DYohn]

S/N has a lot of meaning in headphone amps, but as my good friend Kevin Haskins said many years ago, "Inaudible is inaudible and you don't get extra brownie points for being twice as incredible as another solution."

[leonski]

No standards exist for this measurement. So, while it makes some coole numbers.....like 60 amps per channel, or whatever, it is also probably at a very low voltage and severaly time limited. Not much of musical value.

[vcautokid]

Be careful how you are rating room noise. As in what weighting are you using. A or C weighting. Also is your measuring device your unreliable smartphone are a real ANSI or OSHA approved or better Sound Level Meter. The specs are important to understand, and not fall into the Market-ese. Amazing Keith hasn't chimed in. Lonnie is spot on on this stuff. As time goes on, as nothing ever stands still, SMPS topologies become even better. There really isn't too much new under the sun on Linear Power Supplies though. I like them both when implemented well. As time goes on and my needs stay stringent and space and efficiency come into play, usually the SMPS topologies will be there. The PA-1 is such an example but not the only one. On the linear power front Emotiva's own A-300 and A-100 are my favorites. Reasonable without taking over the whole space, and reasonable on the money front too! That DOES matter.

So make sure you understand specs, and what matters, and what you will never miss.

[leonski]

Another part of the measurement nightmare.

I think if you just treat 'em as 'advisory, you'll be much better off!

[donh50]

Random thoughts...

Amperes per channel is a measure of the amplifier's output stage capability, mainly to drive low-impedance loads.

Power supply capacitance is a measure of how regulated the power supply is, e.g. how much the power supply voltage may sag under heavy loads.

Both are probably more marketing than engineering beyond a certain base level. Power output, steady-state (continuous) and dynamic (headroom, peak) will provide a pretty good indication as well. BTW, my old Gen1 amps listed power supply capacitance IIRC, not sure about maximum current output.

Room noise, meh... Mine is isolated and last time I looked was in the 20'ish dB SPL range if the miniSplit was off. Be careful with the noise floor; we can hear below the noise floor as our ear/brain system will average out the noise to recover signal from below the noise floor. Dither in your DAC does that, and the idea has been used for many decades in signal processing to recover the signal from the noise in systems like radios, radars, disc drive receivers, etc.

OSHA SPL guidelines are designed to be able to hear and understand conversation (converse) as you age. The guidelines are posted on the OSHA site and worth a read. If you want to retain "good" hearing, e.g. for music and such, you'd best be listening at far lower levels than OSHA decrees.

IME/IMO/FWIWFM/YMMV/my 0.000001 cent (microcent) etc. - Don

[leonski]

RE: Hearing and Hearing Loss

I was at an AirShow a couple years ago. Loud Jets common. People walking around with their KIDS on their shoulders.....neither with hearing protection.

I had my custom made earplugs and wife used some sound isolation 'cans'.

I was speaking with one of the vendors and mentioned the noise. His reply? 'A 55 year old Bushman in Africa has better hearing than an American HALF his age'.......

We are submerged in a constant blast of noise / sound which degrades our hearing starting when we are very young. When I was a 'yute', I could hear the ultrasonics

from store alarms. That went away by the time I was maybe 12 or so.


As for 'regulation'?   That is a separate function of a PS.   Caps don't really do anything for that, but DO provide ripple protection and smoothing of the waveform which would

ordinarily be 'choppy' after rectification.       Caps DO provide a reservoir in proportion to their Voltage and Capacitance ratings.   But are not 'regulation'.   The output of such a PS 

will look a little choppy on a scope.   

[donh50]

Semantics. In The Olden Days it was common to spec power supply output voltage "regulation" with input voltage and load changes. The power supply was treated as a black box. Still done that way in a lot of specs, e.g. you can find myriads of power supplies that specify "load regulation" as a separate term from a regulated voltage design (which most are these days, but not in power amps, except for a class D and a few A/B, G, H...) Much different than the circuit topology of using an actual voltage regulator circuit (active with feedback and error amp or passive using e.g. zener diodes).

[leonski]

That;s what I'm saying.
Regulated is an actual feature. Sure, feedback and such but this is in addition to the other regular parts of a PS.

A basic Linear PS is just a transformer, bridge diode (4 diodes in a diamond) and filter caps. Other parts may be added, such as a resitor across the caps to discharge them when power is removed from the circuit.
Slow Start circuitry may be added to prevent surge and you may add overtemp or short protection in the PS. The BASIC LINEAR PS generally doesn't have Zener Diodes, either. Part of regulator circuitry.

My Basic LOW CURRENT schemaatic includes an LM78XX (XX is replaed by voltage). This type of circuit is good to maybe 1 amp or so. And needs an input voltage above the desired output voltage.
I think there are 2 types of these LM devices. the 7800 series are POSITIVE and the 7900 series are NEGATIVE. Both are 3 pole and may need a heatsink at higher powers.

I'd think that output voltage of a PS varies as a function of Input voltage. At least in an UnRegulated PS. but that shouldn't matter much. As long as the Ripple is low enough, that is. And that the PS caps ONLY
recharge when the voltage of the cap drops below PS voltage. A PS with No Load and at full charge will not be a big load, except for the bleed resistors....
Unless you are running a amp near its limits? That should matter, or matter much.

I will personally stick with the strict definition of 'regulated PS'.....

[KeithL]

If you really want to split semantic hairs.....

All power supplies, including simple ones with just rectifiers and filter capacitors, have a characteristic called "regulation".

More specifically separate characteristics for "supply regulation" or "source regulation" and "load regulation".

There is also a related noise spec....

As the filters charge to the peaks, the voltage goes up, then it sags between peaks, before rising again when the next peak arrives.

The resulting variation in the voltage is noise (as is any relatively quick variation in voltage).

In a standard linear power supply, consisting of a transformer, rectifiers, and filter capacitors....

- the output voltage drops if the input voltage drops

- the output voltage drops if the load current increases

- the noise on the output voltage increases as the load current increases

(Note that the type of circuitry used in most modern amplifiers is relatively insensitive to minor variations is supply voltage and small amounts of noise.)

All of the "regulation characteristics" of a simple filtered linear power supply are related to passive regulation.

What we're usually talking about when we use the term "regulated power supply" in modern circuitry is "active regulation".

(A zener diode is something called a "shunt regulator" which is technically active - but which falls between these two.)

However, both those little "tab regulators", and our big SMPS, are what we would call "fully active regulators".

Inside that 78xx or 79xx regulator is a precision voltage reference source... and an amplifier that amplifies that DC reference signal to produce the main output.

And our big SMPS has pretty much the same thing inside... a precision voltage reference and a BIG amplifier... which is optimized for DC output rather than audio frequencies.

(Strictly speaking our big SMPS board contains several of these... each configured to deliver a different output voltage.)

(Also, the amplifier component in the 78xx and 79xx tab regulators is a linear amplifier; in our SMPS, and many small modern regulators, it is a Class D amplifier instead.)

And, since the output voltages depend entirely on the reference voltages, they are immune to variations in both source voltage and load current.

(And, since the reference voltage is extremely clean, and the amplifier is designed to be immune to variations in its own supply voltage, the output is also very low in noise.)

May 10, 2020 14:45:00 GMT -5 leonski said:

That;s what I'm saying.
Regulated is an actual feature. Sure, feedback and such but this is in addition to the other regular parts of a PS.

A basic Linear PS is just a transformer, bridge diode (4 diodes in a diamond) and filter caps. Other parts may be added, such as a resitor across the caps to discharge them when power is removed from the circuit.
Slow Start circuitry may be added to prevent surge and you may add overtemp or short protection in the PS. The BASIC LINEAR PS generally doesn't have Zener Diodes, either. Part of regulator circuitry.

My Basic LOW CURRENT schemaatic includes an LM78XX (XX is replaed by voltage). This type of circuit is good to maybe 1 amp or so. And needs an input voltage above the desired output voltage.
I think there are 2 types of these LM devices. the 7800 series are POSITIVE and the 7900 series are NEGATIVE. Both are 3 pole and may need a heatsink at higher powers.

I'd think that output voltage of a PS varies as a function of Input voltage. At least in an UnRegulated PS. but that shouldn't matter much. As long as the Ripple is low enough, that is. And that the PS caps ONLY
recharge when the voltage of the cap drops below PS voltage. A PS with No Load and at full charge will not be a big load, except for the bleed resistors....
Unless you are running a amp near its limits? That should matter, or matter much.

I will personally stick with the strict definition of 'regulated PS'.....

[leonski]

Wow, Keith, you have split the hair into invisible segments. A passive regulated supply is UNRegulated. It floats. How rubustly it is designed helps what you see at the
output, but in no case could it be considered regulated except by an advertising copywriter. Otherwise known as a simple Linear PS.

If by INSENSITIVE you mean that as long as the demand voltage is lower than the cap voltage, you should be OK. A conventional Linear Recharges as a function of power line frequency
and if the output of the bridge is above the current voltage level of the caps. If the caps are at bridge output voltage, no charging will occur and the amp is probably 'idle'.....

Keith also refers to the LM78XX or LM79XX series of regulators. NOTE that they will not 'regulate' a 5 volt input to a 9 volt regulated output.
The INPUT must ebe above the desired output voltage. So, if you want 24 volts OUT, you should feed it somewhat aove that. Maybe as much as 30 volts.

[KeithL]

Even though a passively filtered power supply is not regulated... the proper term for how much it sags under load is still "supply regulation".

How quickly the capacitors drain will depend on their capacitance and the load; how rapidly they refill will depend on the source impedance of the transformer, rectifiers, and line itself.

With this type of supply, for a given load, and other components, larger capacitors offer "better supply regulation" (meaning that the voltage changes less under changing loads).

If you read carefully you'll see that what I said was that:

  "the type of circuitry used in most modern amplifiers is relatively insensitive to minor variations is supply voltage and small amounts of noise."

What I was referring to was the amplifier circuitry.

The gain of most modern amplifiers is set by the ratio of two resistors.... (I'm referring to the majority of both modern solid state Class A/B and Class D amplifiers.)

Although the maximum output is limited by the supply rails.... as long as you remain below that limit.... and above certain minimums.... the gain is INDEPENDENT of the rail voltage.

Likewise, modern amplifier circuitry is also generally designed to be largely immune to noise on the supply rails.

(The relevant specification is called "supply rejection"... and is quite high for most modern designs.)

Both changes in rail voltage, and noise on the voltage rails, will still have some small effect on performance, which is why a regulated supply is considered to be superior.

(And, in a poor design, they may have a significant effect.)

You are also quite correct.....

Both 78xx and 79xx "tab" regulators are "linear series pass regulators"....

They means that they are designed to produce a regulated output voltage that is lower than their input voltage - and always impose a minimum loss between the two.

However, when it comes to modern switching regulators, some output a voltage lower than their input voltage, some a higher voltage, and some can do either if configured to do so.

You can read about a bunch of them here if you like (these are mostly little ones):  

www.maximintegrated.com/en/products/power/switching-regulators/step-down-up-buck-boost.html

The SMPS we use in our XPA amps actually incorporates multiple output regulators - each delivering a different regulated rail voltage.

May 11, 2020 22:54:50 GMT -5 leonski said:

Wow, Keith, you have split the hair into invisible segments. A passive regulated supply is UNRegulated. It floats. How rubustly it is designed helps what you see at the
output, but in no case could it be considered regulated except by an advertising copywriter. Otherwise known as a simple Linear PS.

If by INSENSITIVE you mean that as long as the demand voltage is lower than the cap voltage, you should be OK. A conventional Linear Recharges as a function of power line frequency
and if the output of the bridge is above the current voltage level of the caps. If the caps are at bridge output voltage, no charging will occur and the amp is probably 'idle'.....

Keith also refers to the LM78XX or LM79XX series of regulators. NOTE that they will not 'regulate' a 5 volt input to a 9 volt regulated output.
The INPUT must ebe above the desired output voltage. So, if you want 24 volts OUT, you should feed it somewhat aove that. Maybe as much as 30 volts.