What happens when the Aleph-X "crosses over"?

nania

Member
2002-06-26 11:51 pm
NY Area
I am presuming that the Aleph-X is a Class AB amplifier in that it operates electrically Class A by the amount of bias current and then "crosses over" to Class B operation thereafter. For a set of speakers with a modest 86db per watt sensitivity, an 8 watt amp will drive them to 95dB! That seems to be about the limit for any sensible listening so why do we need the extra watts of an Aleph?

Mr. Pass has stated that the amp sounds significantly better while operating within the bias current so why wouldn't a single stage amp (such as an X-version of a penultimate Zen) sound better than an Aleph-X? The Zen will be less powerful but it will be operating pure Class A and with less complexity and no need to switch out of the bias current. Does the dynamic nature of music require a multiple of the available power to keep the soundstage from collapsing at the 95 dB level? If I am wrong in my initial premise and presumption please explain what actually happens when we run out of current bias and why the problem of dynamic power requirements couldn't be solved by a larger, faster capacitance?

Thanks all,
nania
audio newbie
 
I'm not sure about the pwoer thing, but when your amp pushes more current than the bias, it will switch to class AB, I don't think right over to class B though. You see, when the amp starts going into the class AB region of operation, it means that the ouput voltage swing is large enough to cause the output transistors to turn off during portions of the audio wave. This doesn't mean both will turn off at the same time, but essentiall only one at a time. In other words, in class AB, one device takes over the current load before the other turns of, but in class B, one turns on and starts working immediately after the other turns off. But in class A, both devices carry current always and never turn off. So in your Aleph, the bias is set so it operates class A until the output voltage swing is great enough to push into class AB but the effect of it switching isn't done by any special circuit and has no latency. It is also a dynamic switch, in that it can vary from class A to AB and not just "switch over" I hope this answers half of your question.
 

nania

Member
2002-06-26 11:51 pm
NY Area
My understanding of Class AB operation may be flawed but if the output devices aren't fed by a constant current (current biased) they are operating Class B. When the output devices are activated by the audio signal they are operating Class B. When the output devices are activated by a current bias (or full open and ready current) the ampliier is operating Class A. So Class AB is defined by this hybrid activation of the output devices. Is this incorrect?
 
I always thought the sole advantage of Aleph is that it has a variable bias enabling class A operation over a wider range without the inefficiency of simply upping the standing bias, rather than simply crossing over to A/B.

Only the three-stage Alephs did that (crossing to A/B), or am I completely wrong?

Jake
 
Alephs do not cross to class AB. Because the variable current source is able to supply twice the peak current relative to the standing bias current after which it will clip hence no class AB operation. Clipping is also asymetric as the positive side will clip while the -ve side continues to operate into class B until either the protection circuit engages or the rail (less losses) is reached.
 
nania, you need to study classes of operation a bit more.

When an amplifier operates class AB, the current is set so that each device in the output conducts for more than 50% of the signal and less than 100% in other words, neither device switches off until the signal fed into them is greater than the bias current. In class B, the devices aren't biased to conduct and only conduct if the audio signal feeds more than about .7 volts into them. In class A, both devices are running for the whole waveform and never switch off. Essentially, class AB is the same as class B, but both transistors are kept on so that there is no point of crossover between the transistors, thus eliminating crossover distortion if done properly. Also, the Aleph is biased full class A as far as I know and it doesn't switch to class AB anyway.
 

Nelson Pass

The one and only
Paid Member
2001-03-29 12:38 am
If the gain of an Aleph current source is set to values
greater than 1/2 the output current, it is in fact possible
to operate beyond class A with one half of the output
stage in cutoff.

This is not a very desirable mode, as it is pretty nonlinear
compared to normal operation, but I suppose it is better
than clipping.

Since we set the current source gain at approximately 1/2
in production Alephs, we do not attempt to take advantage
of this effect.
 

nania

Member
2002-06-26 11:51 pm
NY Area
Thank you Mr. Pass for the explanation of the Aleph operating above its bias current. Would the same limitation occur with the Aleph-X as designed by Mr. Grey Rollins?

Thank you Duo, AudioFreak and Jackeh but I should like to add that my intial questions were answered in a way that leaves me unreconciled. Is Duo correct when he states that Class AB is an entirely different mode that is not the hybrid as I understand it? I would like to know if my understanding of Class A, Class AB and Class B operation are correct and if not, where I am wrong. I would also like to know if a penultimate Zen operating at 8 watts will drive a low impedendence speaker with the same authority as an X-250 that is only pushed to 8 watts of power and why wouldn't the Zen sound better if it had the same capacitance as the X-250 for "dynamic reserve"? Does an X-Zen described above even need "dynamic reserve" if it is operating with a wide open (full current) output stage?
 
For simplicity let's assume that we're talking about an ordinary push-pull output stage. Now, zero in on one output device and watch the signal as it goes through.
If the signal is present at all times during the cycle (think of a sine wave), then that's class A.
If the signal is there for 50% of the time (i.e. the sine wave is split in half at the X axis, so that you have two sets of disconnected humps; one set swinging positive, the other swinging negative), that's class B.
If the signal is present for more than 50%, but less than 100% of the cycle, that's class AB.
Class A is easy to envision, with both halves of the circuit working all the time. Class B switches the negative half of the amp off when the signal swings positive, then switches it back on when the signal goes negative (turning off the positive half as it drops below 0V). This is efficient, in the sense that the half of the amp that's not doing any work isn't drawing current. That means that a class B amp runs cooler and uses less electricity. So far, so good, but there's a price to pay: the devices don't switch on and off perfectly. As one side is turning on and the other is turning off, a glitch appears as the signal crosses the X axis. To an extent, this can be reduced with negative feedback, but that just opens another can of worms. (Not being a bird, I don't enjoy the thought of worms...others seem think it sounds tasty.) In an amp running class AB, there's a bit of overlap. The positive side of the amp carries the signal a short distance into the negative realm to give the negative side more time to get going. As an admittedly imperfect analogy, think of relay runners--as the first runner approaches, the second begins running so that the baton is passed while both are in motion. This is a lot smoother than if the first runner were to run up to the second, stop, hand over the baton, and the second runner begin running--clearly a clumsy way to get the job done. Some feel that an AB amp also has trouble handing the signal from one half to the other, but that's another matter.
Class isn't assessed on the amount of bias current present per se, it's how much of the signal is carried by the output devices during a duty cycle. Bias current is just how the job gets done.
The Aleph output stage is class A by virtue of the fact that (under normal circumstances) both halves of the circuit carry 100% of the signal 100% of the time. Once a circuit is in class A, more bias current doesn't make it any more "A" than it was. I run my Aleph 2s about 10% harder than the stock version in terms of bias, but that doesn't make them more class A. They're still just class A. (Okay, you could tempt me to say they rate an A+...but there isn't any such operating class.)
Operating class is a sliding scale with class A on one end, class B on the other, and everything between being class AB.
As to the sound of different circuits all delivering the same amount of power...they will sound different. Some of the factors are known and understood. Others are purely a matter of conjecture. Some people like to think that distortion numbers can describe the performance of a circuit. Their viewpoint becomes difficult to maintain when faced with two amps with the same measured distortion that sound different. You can get off on various tangents by measuring the sundry distortion components and comparing them, but even so, things still don't add up. It's not likely that any of the Zens will sound like an X. In turn, the X will not sound like an Aleph.
The Aleph-X is, for all intents and purposes, still an Aleph. Your best bet is to set the output (the bottom MOSFETs) and the current sources (the top MOSFETs) to each handle about 50% of the load.

Grey
 

nania

Member
2002-06-26 11:51 pm
NY Area
Mr. Grey Rollins

Thank you for the "relay race" analogy. It was quite eloquent and very much in line with how I envisioned the idea of bias current. I have been attributing the electrical Class designation based on how and when the output stage(s) reacted to (amplified) the audio input signal. I pictured Class A operation as the full current flow available immediately for any audio input signal whether the signal be null or a symbol crash. Hence by my definition, if a greater percentage of the full output current was available immediately to amplify the audio signal the more Class A with the pure Class A designation reserved for operation with 100% of the full output current avaiable at all times. If the signal engaged the output device to generate current then it would be Class B. What I apparently missed was the idea of there being a positive and negative signal. I was assuming the amplification as a gain from zero (null to full signal) instead of positive to negative wave amplitude signals. Your explanation describes a push pull scenario. Do any of the Aleph or Zen amps operate push pull or are they all "single ended" and is my description above of any use in describing single ended Class AB operation?

Now, as far Aleph-X sounding different then Zen-X than current commercial X products I am not concerned with nuance and sublety. I want to know if a Zen-X will have the same output current to my speakers at 1 watt +9dB gain as an Aleph-X at the same gain level. Will it grip my bass the same way? Will it create the same soundstage? Will it resolve the image as well or better? If it can't, I want to identify the why it can't? My goal in all this is to decide if an X-Zen will be as good as an Aleph-X at the maximum 95dB operating level. I know the questions that I will ask in this pool of talent will come to know the answer if they don't know the answer already.
 

Nelson Pass

The one and only
Paid Member
2001-03-29 12:38 am
If you are not concerned with nuance and subtlety, then
they will be the same, otherwise I can assure you that
they are different. Partly if really depends on your taste.

I have not finished working out the tweaks associated
with the XZen, and so it is a bit early to get into the
comparisons, but in general it seems to be the case that
if you X any simple class A circuit, it gets a lot better.
 
Mr. Pass

judging by the amount of reads on the Aleph-X thread as compared to the Zen-X thread it appears that I am in the minority in my opinion that a single stage Zen-X will be the ultimate last word in Audio amplification. For all the many landmarks you have achieved I believe that the Zen-X will be your greatest legacy and the thing you will be most remebered for in what you have given the world. Forgive me for not having the delicacy of hearing that many "audiophiles" claim to have but I very often do reliably catch the obvious problems that many so called great ears miss. I have quite excellent measured hearing but I often don't get to hear the magical changes attributed to some gear, hence my apparent discount of the terms "nuance" and "subtley". I hope I haven't offended anyone in this forum. I was only trying to express what was important to me in my quest for audio enlightment. What I have found is that the ability to deliver high sustained and peak current to the speaker is the best predictor of the meat and potatos of what stereo audio is all about, imaging. Does my observation conflict with your experiences? If you expect the current delivery profile of the Zen-X working speakers at 95dB to be identical to an Aleph-X doing the same work, where does the difference in sound come from?
 
nania: It doesn't matter what kind of amp you have, they will all have to ouput the same amount of current to make 1 watt of power in the same speaker. For example, say you are using an 8ohm speaker, if you power it to 1 watt, you need 2.89 volts, thus having a current of 0.35 amps no matter what. This is however only peak power, if you have rms power of 1 watt, then you need 4.08 volts and a current of 0.51 amps all by ohm's law calculations.RMS power of 1 watt is essentially a peak power of 2.0808 watts... I hope this helps to explain how an amp delivers current to a speaker to make power.;)


EDIT:BTW, if one amp sounds better than another at the same power, it usually means the better amp has superior control over the current it feeds to the speaker and probably less distortion as well. High current capability is only a third of the game to get good punchy bass, the other third is power supply current headroom, and the last third is the ability of the amp to use that current effectively and quickly without distorting the original waveform...:cool:
 

nania

Member
2002-06-26 11:51 pm
NY Area
Duo,

I know that the popular knowledge states the amount of current is determined by the speaker impedence and not by the amplifier but I don't think that this simplistic idea tells the right story because a speakers impedence (resistance) is only fixed when the drivers are not receiving a signal. When the drivers and crossovers are interacting with the signal, Ohms law has serious limitations in predicting what is really going on. So I respectfully disagree with your statement that it doesn't matter which amp delivers the signal and that the current will be constant and dependent only on the speaker impedence (resistance). I assert that the audio signal is delivered as power and that power has a profile of current x amplitude. In other words, a given time portion of an identical audio signal can be delivered with a profile of 2 amps and 8 mV or 8 amps and 2mV. I believe that the amplifier (and the interconnects to a lesser extent) are the determining factor of how much current controls the speaker drivers. A high current sounds like a tighter grip on the speaker drivers which results in less driver transient motion and a tighter image. Volts will slap the drivers into motion but lose control relatively quickly and the residual motion (transient) from the inertia results in loose control and a poorer image. To make an analogy, voltage is a punch and current is a grab and pull. This is the only way that I can reconcile what I hear to what I understand about electronics. I would propose that if my esteemed colleagues in this forum find no objection that we refer to this unorthodox view of amp and speaker interaction as the Nania audio power theorum. Even though I haven't been able to prove it mathematically...yet, I think it will be easier to refer to it that way than explaining what is meant by it. One last point, I believe Duo meant 2.83V and not 2.89 but poor interconnects could very well make that the right number if we use the Nania audio theorum 8^p
 
Nania,

The amount of power (Current x Voltage) present at the driver terminals IS arbitrarily determined by the impedance of the driver. Ohms Law is completely capable of determining this so long as you consider the reactive parts of the drivers impedance and not just the DC resistance. So for a given driver that receives 1W @ 1kHz AT THE TERMINALS OF THE DRIVER the combination of current and voltage IS determined by ohms law. For Example a driver that measures 8ohms at 1kHz receives 2.83 Volts RMS and 0.3534 Amps RMS. However due to the impedance of the speaker cable and the impedance of any passive crossover used, the voltage / current combination output from the amp will be higher than the above figures again determined by ohms law. The output impedance of the amplifier serves to determine damping factor, max power into low impedance loads and to some extent the distortion signature of the amp.
 
Excuse me, I did intend to write 2.83 volts, but my point is, if either amplifier is driving the same speaker, either amplifier WILL push the same amount of current for a certain amount of power at a certain frequency. A speaker will not work without voltage and current, in simplicity, voltage * current is power and that is how it always works. As the frequency changes, the impedance of the speaker will change, and quite dramatically, however, if either amplifier is outputting the same frequency, it will see the same impedance on that same speaker. Thus, either amplifier will simply have to achieve the same amount of current output for a desired voltage to reach a specific level of power in the speaker at the same frequency. Certain limiting factors will cause either amplifier to sound totally different from one another, for example, if one amplifier is trying to feed the speaker 1 one watt of power, it must always put a certain amount of current out with whatever voltage is happens to feed it. Things like output devices, etc will have limiting effects on how much current it outputs and how well it's controlled. You are wrong in stating that current simply means tighter bass, as current will only work with voltage through a resistance to produce power. The determining factor is that you need current to achieve tighter better bass, but, an amp may not be able to control that current as well as another as it's demanded by the speaker. Also, if your amplifer feeds 8mv into an 8 ohm impedance, it will have to exert 1ma and will produce .000008 watts peak, however, it the same amp tries to deliver only 1mv, it will only ever have to exert 8ma IF the load impedance was .125ohms, and I surely doubt a speaker would do that!! So, your statement about changing current and voltage around is utterly wrong with a load like a speaker! Essentially, if your amp fed 1mv into the speaker, it would generate a current of
.000125 amps. Also, your stating that a higher current makes a tighter grip is wrong. If you want a speaker to load to a higher current, you must feed it more voltage or reduce it's impedance, thus, a higher current means more power and a higher amplitude in decibels out of the speaker itself. Essentially, you need more control over the current delivered and you need to BE ABLE to supply that current when it's needed, that is what makes bass tight and punchy!
 

nania

Member
2002-06-26 11:51 pm
NY Area
Duo

<b>...either amplifier WILL push the same amount of current for a certain amount of power at a certain frequency</b>

There is the underlying problem! The audio signal is a composite of many different simulataneous frequencies so the power is not a factor of a nominal resistive load of the speaker. Something much more complex is happening and I say that it can be better explained if we look at the situation from the source (signal) rather than the terminals (speaker). When I proposed the Nania audio power theory, it was to invite this forum to look at a better way to predict what is actually making the stereo image and what is f**king it up. Ohm's law may be adequate to describe a fixed frequency electronic tone but is inadequate to describe the stereo image of music.

<b>You are wrong in stating that current simply means tighter bass, as current will only work with voltage through a resistance to produce power.</b>

You seem to have missed my point. The resistance is in flux as the speaker drivers are in motion and the motion appears to be determined by the power profile as described by the Nania audio power theory.

<b>The determining factor is that you need current to achieve tighter better bass, but, an amp may not be able to control that current as well as another as it's demanded by the speaker.</b>

Here you appear to contradict your earlier statement and seem to understand why we need to explain what is happening in another way.

AudioFreak

we have all been taught Ohms Law as gospel to explain voltage, resistance and current but just as Newtonian Physics was inadequate to explain lightspeed energy we must attempt to find another way to explain what is happening to our music. I will be starting another thread to introduce the Nania audio power theory to the forum because it clearly is off the original topic and I think it deserves the full scrutiny of this Forum. If it can hold up against the scrutiny and critical eye of the peers in this forum, it may prove to have value. Even if it gets shot down in flames I believe it may become an important step to better understanding and that is what I am here for.
 
You simply do not understand that I mean both amplifiers will exert the same power for a certain voltage into the same speaker at the SAME frequency! I know that the power level varies as the impedace varies, but, if both amplifiers play the same music, then the frequencies of all the passages are the same essentially. You must also test your "nania theorum"
and further explain it.

Your question was about which amplifier delivers more current for a given amount of power, are you trying to say that you meant one amplifier was to exert a completely different frequency than the other in this test? That could bring about very in-accurate answers!

Are you also saying that your using different speaker cables and different speakers for this theory while testing either amp(in math)

As speaker motion is simply a derivation of output power, then I must exemplify that the amplifier must have control of some sort over this power. This is where the tightness of bass can come in, one amplifier may have too low a damping factor or a weak power supply or slow transistors, but more current does not ever simply mean more bass! I have not contradicted myself, I'm simply stating that punchy well placed bass is best had when the amplifier has 1: A good availability of current supply and 2: The ability to easily keep that current under control!

I suggest you get some education in this field before you testify things you simply don't understand. You've also stated earlier that you are a beginner in the audio field, did you mean this or was it a joke? You mentioned that you couldn't yet prove your theorum mathematically, thus, may get inaccurate results by using it...

Either way you go about it, if either amplifer tries to push a certain frequency or mix of them into a speaker at a certain voltage, current will flow. Any amplifier trying to drive the speaker will still see the same loading conditions caused by the speaker and the speaker will draw the same amount of current from either amp for same passages of music or frequencies. If an amplifer hasn't got the current head room it needs to drive the load, then it will start to flake out losing control of the bass since the voltage drops as a result of current loading, rather like a linear PSU with too small a transformer.