I love the Aragon 8008 BB amp I've just acquired. To me it sounds like a true Class A amplifier, with layered depth, width, and detail that doesn't scream detail, it's just all there, exactly what I've noticed with real Class A amplifiers.
But the claim made by Mondial and/or Klipsch that this amp has 26 watts per channel of "Class A" power evades any kind of calculation I've been able to do.
I've measured the stable emitter resistor voltage in my early production unit and apparently unaltered and minty unit, and it's 26mV across each 0.33 ohm emitter resistor (way more than the Klipsch tech support memo says: 8mV for outside channel). That means 0.0788 amps per transistor pair, and there are 6 pairs of transistors, giving a total current of 0.473 amps. Staying within textbook linear Class A, twice that can be delivered at peak (one transistor goes to zero and the other doubles), or 0.946 amps. The square of that multiplied by 8 ohms is 7.16 watts peak power, which becomes 3.58 watts average (aka RMS) power per channel. I've read this kind of calculation now in many places and it makes perfect sense to me. The tech support memo recommendation would lead to 0.34 wpc average Class A power (somewhat more for the 12mV inside channel).
So is the factory claim of 26Wpc just a number chosen from thin air, or does it relate to something else? I figure they could do a measurement to cutoff, finding the maximum power so that neither transistor cuts off entirely to be around 26W. Of course that's no longer textbook definition quality, but it makes some sense, the number has some Class A-like meaning, but in a looser sense, because when you push a transistor to just before cutoff it's no longer linear. Therefore the 26Wpc claim relates to something you could call non-linear Class A, if not simply pure whatever.
Even fairly astute people I talk to can't believe that my amplifier which consumes 160W wall power at idle and gets quite warm only actually delivers 3.58 watts of Class A power. Somehow they think of Class A as being 20-35% efficient. But, as I see it, that kind of efficiency isn't possible with a Class AB amplifier which has much higher rail voltages than it would need for the indicated amount of Class A power, and all that additional rail voltage, which is nice to have for peak power delivery, chews up additional wall power at idle through the output transistors.
Am I missing something?
But the claim made by Mondial and/or Klipsch that this amp has 26 watts per channel of "Class A" power evades any kind of calculation I've been able to do.
I've measured the stable emitter resistor voltage in my early production unit and apparently unaltered and minty unit, and it's 26mV across each 0.33 ohm emitter resistor (way more than the Klipsch tech support memo says: 8mV for outside channel). That means 0.0788 amps per transistor pair, and there are 6 pairs of transistors, giving a total current of 0.473 amps. Staying within textbook linear Class A, twice that can be delivered at peak (one transistor goes to zero and the other doubles), or 0.946 amps. The square of that multiplied by 8 ohms is 7.16 watts peak power, which becomes 3.58 watts average (aka RMS) power per channel. I've read this kind of calculation now in many places and it makes perfect sense to me. The tech support memo recommendation would lead to 0.34 wpc average Class A power (somewhat more for the 12mV inside channel).
So is the factory claim of 26Wpc just a number chosen from thin air, or does it relate to something else? I figure they could do a measurement to cutoff, finding the maximum power so that neither transistor cuts off entirely to be around 26W. Of course that's no longer textbook definition quality, but it makes some sense, the number has some Class A-like meaning, but in a looser sense, because when you push a transistor to just before cutoff it's no longer linear. Therefore the 26Wpc claim relates to something you could call non-linear Class A, if not simply pure whatever.
Even fairly astute people I talk to can't believe that my amplifier which consumes 160W wall power at idle and gets quite warm only actually delivers 3.58 watts of Class A power. Somehow they think of Class A as being 20-35% efficient. But, as I see it, that kind of efficiency isn't possible with a Class AB amplifier which has much higher rail voltages than it would need for the indicated amount of Class A power, and all that additional rail voltage, which is nice to have for peak power delivery, chews up additional wall power at idle through the output transistors.
Am I missing something?
Last edited:
Are you sure that the amplifier is pure class A, not just a highly biased class AB, which sounds more logical. Measure the peak current rather an assuming it. Many amplifiers run class A for the first few watts.
Nico nailed it. Although the Aragon 8008 is biased more heavily toward Class A than most, it's a Class AB amplifier, the same as the majority of modern hifi amps (pure Class B amps are unsuitable for hifi use; in essence they don't exist in the market).
Last edited:
Charles,
It's not even a highly biased Class AB. This category belongs to Class AB with generally more than 125mA per output pair.
Classic case of a feisty marketing dept and a submissive engineering team.....
Ain't nuttin' new.
Hugh
It's not even a highly biased Class AB. This category belongs to Class AB with generally more than 125mA per output pair.
Classic case of a feisty marketing dept and a submissive engineering team.....
Ain't nuttin' new.
Hugh
Of course it is a Class AB amplifier, not Class A, I wasn't claiming otherwise. The total rail to rail current per channel is 0.473 amps, which exceeds 125ma per output pair if there were only one pair (which is quite common, even recommended by some). I'm determining that current from a measurement of the emitter resistors in my unit, which is as good as any other kind of measurement I think.
I've seen the "watts of Class A power" description for Class AB amplifiers used by many people, not just out and out hypesters. People such as Nelson Pass and John Atkinson seem to calculate it just as I do here, for Class AB amps such as the Pass X250 and Krell KSA 250. And I wanted to make sure that I am doing it as correctly as they do, when I calculated it as 3.58 watts per channel. That seems to me the correct calculation, not the factory claim of 26 watts. It would also be interesting to know if there were any kind of meaning to that 26 watt claim, but that wasn't actually my main question.
I just want to be sure I didn't err in my 3.58 watts calculation, or if there isn't some other way of doing this. My friends are not believing me, they think my amp has to have more Class A power than that. I actually got the clearest explanation of how to make that calculation from a 2005 post by Nelson Pass in DIYAudio. I believe his current amps are all Class AB also, though the XA series has substantial Class A power, 30 watts and up, but with even more power in Class AB. And I believe his "watts of Class A power" are calculated correctly, and conservatively for the XA series.
I've seen the "watts of Class A power" description for Class AB amplifiers used by many people, not just out and out hypesters. People such as Nelson Pass and John Atkinson seem to calculate it just as I do here, for Class AB amps such as the Pass X250 and Krell KSA 250. And I wanted to make sure that I am doing it as correctly as they do, when I calculated it as 3.58 watts per channel. That seems to me the correct calculation, not the factory claim of 26 watts. It would also be interesting to know if there were any kind of meaning to that 26 watt claim, but that wasn't actually my main question.
I just want to be sure I didn't err in my 3.58 watts calculation, or if there isn't some other way of doing this. My friends are not believing me, they think my amp has to have more Class A power than that. I actually got the clearest explanation of how to make that calculation from a 2005 post by Nelson Pass in DIYAudio. I believe his current amps are all Class AB also, though the XA series has substantial Class A power, 30 watts and up, but with even more power in Class AB. And I believe his "watts of Class A power" are calculated correctly, and conservatively for the XA series.
Last edited:
Charles, if I understand the first part of your statement, then you calculation for class A power delivery into the load is correct. And I take it you measured the current (drop across Re) under DC (static) or no signal conditions.
However, I do not understand why you do not accept the factory claim of 26 watt, it may also be quite accurate, because the statement refer to dynamic conditions, i.e. there is a signal.
To do this would be just as simple.
Firstly select AC Volt on the meter and place it across the loudspeaker terminals (while a loudspeaker or suitable load is connected). Apply a 100 Hz sine wave signal to the input of the amp (use your PC sound card) and increase the volume to the point where you can just hear that the signal distorts (this is about 1% distortion but it will be very loud) and take the meter reading in AC volts.
Most meters approximates an RMS reading in AC setting (some are true rms measuring). Now apply the formula P=V^2/R and you will find that it most probably is 26 watt RMS.
However, I do not understand why you do not accept the factory claim of 26 watt, it may also be quite accurate, because the statement refer to dynamic conditions, i.e. there is a signal.
To do this would be just as simple.
Firstly select AC Volt on the meter and place it across the loudspeaker terminals (while a loudspeaker or suitable load is connected). Apply a 100 Hz sine wave signal to the input of the amp (use your PC sound card) and increase the volume to the point where you can just hear that the signal distorts (this is about 1% distortion but it will be very loud) and take the meter reading in AC volts.
Most meters approximates an RMS reading in AC setting (some are true rms measuring). Now apply the formula P=V^2/R and you will find that it most probably is 26 watt RMS.
Last edited:
Make no mistake 26 watt is very loud and your family members and neighbors will curse you while fooling around measuring the amp output.
26W corresponds to ~2.5A peak Class A output current.
At 70Vdc rails in the 8008, that makes >175W idle heat per channel, more than 350W stereo, which would require stupendous size heatsinks.
Build a few Class A power amps, and just a glance at an amplifier at a show plus the power specs are enough to spot BS.
It gets even worse : a lot of the real Class A amps were not real Class A, by the classic definition.
More than a few Class A amplifier models are biased to the continuous power level, not peak, means ~30pc less heat.
(and the brochures sure do not include idle dissipation numbers
)
Nico, my Aragon 8008 BB puts out about 250W into 8 ohms with minimal distortion (just below the clipping point). That's the Class AB power. I measured that on delivery, to be sure I wasn't getting a dud. The question is how much "class A power" it puts out. My calculation based on bias current measurements is 3.58W, the Mondial factory claimed 26W, which appears to be an inflated figure...I'm not sure how you would measure the dynamic and linear "Class A Power" in a Class AB amplifier, my first post suggested a means to measure the oxymoronic "non-linear Class A Power" which is how I suspect the factory spec was derived.
But actually, 3.58W true class A power is far more than most audio amplifiers have. A fairly small proportion of audio amplifiers dissipate more than 160W at idle. It may well be "just enough" or even more than enough to get a nice Class-A-like sound. In fact, that is my listening observation, it is the only amp I have with Class-A-like sound other than my Krell FPB 300. And the Aragon is smaller, lighter, far less complex, consumes on far less power during use, and is currently working. 160W idle dissipation is more than most amps, and even with the deceptively large heat sinks (mostly two inches deep across the depth of the amp) they are pretty much taxed as much as I would want...or more actually...with idle temperatures reaching 132 degrees F after 10 hours of idling. So clearly it becomes increasingly less cost effective to push for more Class A power than this. Fortunately the bias apparently cuts back as the heat sinks get hotter, and the sinks do have a lot of thermal mass also, so in moderate use I haven't seen heat sink temperatures above 136 degrees F (but I'm going to keep measuring, see my avatar).
And with my extremely inefficient Acoustat 1+1 speakers, I'm measuring power consumption during moderately high playback levels of 300W, almost double the idle power. I could be reaching average speaker power levels of 30W, and peaks above 250W (into 4 and lower ohms). Ordinary Class A power meeting these requirements would be outlandish. The Krell FPB had previously been my workaround, and even it is pretty outlandish.
But actually, 3.58W true class A power is far more than most audio amplifiers have. A fairly small proportion of audio amplifiers dissipate more than 160W at idle. It may well be "just enough" or even more than enough to get a nice Class-A-like sound. In fact, that is my listening observation, it is the only amp I have with Class-A-like sound other than my Krell FPB 300. And the Aragon is smaller, lighter, far less complex, consumes on far less power during use, and is currently working. 160W idle dissipation is more than most amps, and even with the deceptively large heat sinks (mostly two inches deep across the depth of the amp) they are pretty much taxed as much as I would want...or more actually...with idle temperatures reaching 132 degrees F after 10 hours of idling. So clearly it becomes increasingly less cost effective to push for more Class A power than this. Fortunately the bias apparently cuts back as the heat sinks get hotter, and the sinks do have a lot of thermal mass also, so in moderate use I haven't seen heat sink temperatures above 136 degrees F (but I'm going to keep measuring, see my avatar).
And with my extremely inefficient Acoustat 1+1 speakers, I'm measuring power consumption during moderately high playback levels of 300W, almost double the idle power. I could be reaching average speaker power levels of 30W, and peaks above 250W (into 4 and lower ohms). Ordinary Class A power meeting these requirements would be outlandish. The Krell FPB had previously been my workaround, and even it is pretty outlandish.
If you want a further opinion.
I agree with your 3.58W of ClassA into 8r0.
That is 7.16W of ClassA into 16r and 14.32W of ClassA into 32r and 28.64W of ClassA into 64r
These higher resistances require much higher drive voltages.
58r1 with 0.946Apk passing is equivalent to 26W.
that requires a drive voltage of 54.96Vpk.
250W into 8r0 requires 63.25Vpk.
It seems entirely possible that 26W of ClassA is around the maximum ClassA this amplifier can deliver.
22.6mV of device bias would give 26W of ClassA @ ~63.3Vpk
If this is the case then this is a most unusual way of specifying maximum ClassA power. But equally it is telling the truth, by hiding the actual specification. Naughty !
I agree with your 3.58W of ClassA into 8r0.
That is 7.16W of ClassA into 16r and 14.32W of ClassA into 32r and 28.64W of ClassA into 64r
These higher resistances require much higher drive voltages.
58r1 with 0.946Apk passing is equivalent to 26W.
that requires a drive voltage of 54.96Vpk.
250W into 8r0 requires 63.25Vpk.
It seems entirely possible that 26W of ClassA is around the maximum ClassA this amplifier can deliver.
22.6mV of device bias would give 26W of ClassA @ ~63.3Vpk
If this is the case then this is a most unusual way of specifying maximum ClassA power. But equally it is telling the truth, by hiding the actual specification. Naughty !
Last edited:
Now that I understand your what you are trying to say, your comment is very valid and what Andrew says makes sense, it seems a very misrepresented claim by the manufacturer.
I don't want to try and cover for the manufacturer, but is it possible that they employ some sliding bias arrangement?
I don't want to try and cover for the manufacturer, but is it possible that they employ some sliding bias arrangement?
Brochure states 120W idle power consumption, irrc.
(though, same one that says 10mV bias setting across 0.33 Ohm)
(though, same one that says 10mV bias setting across 0.33 Ohm)
If what you are saying is accurate it may make it just fall into the category having 26 watt class A, but what is questionable is bias current of 10 mV. Could he have fooled around with the bias setting - or maybe someone before him and turned it down because of the somewhat high heat being suspicious?
8008 MKII, 8008BB, and the Indy 8008 MKIII are likely next to equivalent.
=> www.aragonav.com/aragon product page.html
I've yet to see an Aragon/Klipsch/Indy audio power amp schematic with sustained bias.
=> www.aragonav.com/aragon product page.html
I've yet to see an Aragon/Klipsch/Indy audio power amp schematic with sustained bias.
The early 8008's where designed by Dan Agostino, i dont recall them running hot at all , when at idle.
This may sound silly but maybe the idle power consumption has nothing to do with the amp bias, it may be what all the other paraphernalia needed to operate.
Who knows, it may use a quad core microprocessor for it protection circuitry and communications. Or a shunt regulator, there could be many explanations other than the amplifier.
By the sound of it our mate has measured the amp bias current and calculated the idle power of the amp, maybe measuring the AC input current from the wall socket may shed some light.
Who knows, it may use a quad core microprocessor for it protection circuitry and communications. Or a shunt regulator, there could be many explanations other than the amplifier.
By the sound of it our mate has measured the amp bias current and calculated the idle power of the amp, maybe measuring the AC input current from the wall socket may shed some light.
Seriously doubt it.
With drivers (about 7W total dissipation) and front end bias added, power relay consumption, bridge rectifier + transformer losses, the 120W spec number doesn't sound that farfetched. (not the smartest heatsink arrangement, imo, but smart looking sells)
8008 Mark-2 channel =>
With drivers (about 7W total dissipation) and front end bias added, power relay consumption, bridge rectifier + transformer losses, the 120W spec number doesn't sound that farfetched. (not the smartest heatsink arrangement, imo, but smart looking sells)
8008 Mark-2 channel =>
Attachments
Last edited:
I'd say (only) 2.53 watts RMS.
The square of 0.946 A multiplied by 8 ohms is 7.16 watts peak-peak power, which is 2.53 watts RMS (divided by 2.83), but correct me if I am wrong 😱
Are the emitter resistors 0.33 ohms (did you check)?
- Status
- Not open for further replies.