Jump to content

How a casual comment has confused me.....


maxg

Recommended Posts

In a recent thread Mark mentioned, laughingly, how amp manufacturers boast that their amp doubles in power going from 8 ohms to 4 ohms but never mention what happens when the impedance climbs to 16 or 32 ohms.

This got me thinking and then confused the hell out of me.

As follows:

Lets take a Khorn with a type A crossover. I have seen that the impedance of this speaker varies quite dramatically:

5.5 KHz - 3.5 ohms

35 Hz - 15 ohms

2 KHz - 40 ohms

All approximate values.

Now - let us imagine we are playing a frequency sweep at 104 dB (nominally stated to be 1 watt of power using 104 dB/w/m)

I had thought that amps hold constant voltage and it is the current that goes up (and the power). Now this doesnt make sense according to my calculations:

If we suppose the 104 dB figure is into an 8 ohm load when needing 1 watt:

Power = Amps squared * Impedance

If the impedance is 8 then amperes = root (1/8) = 0.35 amps (approx)

Voltage, therefore (Power = volts * amps) = 2.85 Volts

So far so good: That figure looks quite familiar.

Now - if we take the impedance at 3.5 ohms and the voltage is constant:

Power = V squared / Resistance

= 8.12 / 3.5

= 2.3 Watts (amperes have climbed to 0.8 A)

On the other hand, at 40 ohms power is a measely 8.12/40 = 0.203 watts (amperes have fallen to 0.07 A)

So I am holding a constant 104 dB and my power requirements are varying by 10 fold?

Huh?

Where did I go wrong?

Link to comment
Share on other sites

maxg 

Some random thoughts.

There's no mention of sensitivity of each of the drivers in your analisys. When evaluating SLP from xxhz to xxKhz, one important factor is the sensitivity of each of the drivers used.

While amp spec's usally report power from xxHz to xxKhz, speaker systems are certianly not rated for  fixed max power handling from xxhz to xxkhz.  No manufacture of speakers states their speakers can handle any amount of watts from xxHz to xxKhz.  Imagine if they had to....they would have to use the max successful number which in the case of a barebone 3 way (gentle slope with not diodes, polyswitch, bulbs, etc) using a K-77 tweeter, k-55, k-33, would sound like 5 watts contiunous from 20Hz to 20Khz.

The tweeter diaprhagm in the k-77 according to EV spec's can only handle 5 watts continuous and 25 - 50 watts peak.  It is true that the K-77 can produce a high SPL with 5 watts.  It is true that polyswitches, diodes, bulbs, and other devices will help absorb excess power.  It is true that higher order xovers consume more power than lower order ones.

If you want to evaluate how much power a speaker system can handle form xxHz to xxKhz your really evaluating the power absorbtion of all the componets in the system collectively.

Music power, MAX power, peak power, all are obscure terms.






Link to comment
Share on other sites

Thanks for that - you just made things even more complicated - just when I thought I was getting somewhere.

However - doesn't the cross-over serve to equalize the loads the amp sees between the different drivers enabling us to treat the speaker as a single unit when assessing senstiviity?

If my original question was not clear (taking the KHorn Type A X-over as my example):

We see a figure quoted for sensitivity as 104 dB/w/m with no mention of impedance. I am assuming that this measure is at impedance of 8 ohms - as most measures are. What confused me is that to hold this constant level (104 dB) - taking into account impedance variations across the frequency spectrum appears to vary the power need by a factor of 10 (assuming voltage to be constant).

Is this correct? Does voltage actually vary? Are my calculations wrong somewhere?

Link to comment
Share on other sites


Max, 

q- However - doesn't the cross-over serve to equalize the loads the amp sees between the different drivers enabling us to treat the speaker as a single unit when assessing senstiviity?

a- not really.  basic crossover is a voltage based architecture (there are current based versions of xovers which behave differently)  which means it prone to variaitons based on reactance and capacitance.  the load the amp sees has a direct bearing on the load of the drivers, plus the DCR of the inductors in the case of low pass and band pass sections and ESR in the case of band pass and hi pass sections, minus the impedeance drop at the crossover points due to overlaps in drivers and minus the drop due to shunt devices that are placed between crossover componets and the drivers.  What this means is that in a 6db per octave xover which uses a crossover frequency of 500hz and a 4 ohm woofer and a 16 ohm mid driver, at the overlap point, the impedance is lower.  There are crossovers that attempt to provide constent impedance but they do this by introducing componets to help average out the load such as in the case of the swamping resistor. Current based crossovers are even more complicated since they introduce variations as a result of the reactance of all the drivers due to the serial path the music takes vs the parallel path used in a voltage derived xover.

Q- We see a figure quoted for sensitivity as 104 dB/w/m with no mention of impedance. I am assuming that this measure is at impedance of 8 ohms - as most measures are. What confused me is that to hold this constant level (104 dB) - taking into account impedance variations across the frequency spectrum appears to vary the power need by a factor of 10 (assuming voltage to be constant).

A- in your example, hidden factors is the SPL rating of the drivers and what was done to lower the SPL of mid drivers and tweeters so that they output SPL approximate to the woofer.  In the case of the mid driver, in order to level match the woofer, an autoformer is used which introduces impedance in 2X, 4X, 8X and 16X steps.  This equates to -3db, -6db, -9db, and -16db.  Most folks are using the 2x/-3db or 4x/-6db tap.  At 2X the load is 28ohms from 500hz - 5000hz assuming no swamping resistor is used.

Q - Is this correct? Does voltage actually vary? Are my calculations wrong somewhere?

A - Your basic question can best be answered by evaluating the performance of a full range driver and what changes occurs to input voltage as the input frequency changes.  Assuming a driver can output 70hz - 12khz +- 3db, you can expect the calculations to be consistent within +-3db from 70hz - 12khz.  The impedance will typically increase as you scale up the frequecny ladder.  However, the impedance increase is off set by the greater efficency drivers have at higher frequency than lower frequencies, resulting in SPL levels that belnd together withing the +- rating of the driver.  Now if you introduce 3 drivers, consideration has to be to how each driver will behave base don the manufacures specs within the rated range of the drivers.  Now introduce crossovers, and more factors are introduced.  Add polyswitches, diodes, lamps, and other driver protection devices, and even more factors are introduced.





Link to comment
Share on other sites


Max, 

q- However - doesn't the cross-over serve to equalize the loads the amp sees between the different drivers enabling us to treat the speaker as a single unit when assessing senstiviity?

a- not really.  basic crossover is a voltage based architecture (there are current based versions of xovers which behave differently)  which means it prone to variaitons based on reactance and capacitance.  the load the amp sees has a direct bearing on the load of the drivers, plus the DCR of the inductors in the case of low pass and band pass sections and ESR in the case of band pass and hi pass sections, minus the impedeance drop at the crossover points due to overlaps in drivers and minus the drop due to shunt devices that are placed between crossover componets and the drivers.  What this means is that in a 6db per octave xover which uses a crossover frequency of 500hz and a 4 ohm woofer and a 16 ohm mid driver, at the overlap point, the impedance is lower.  There are crossovers that attempt to provide constent impedance but they do this by introducing componets to help average out the load such as in the case of the swamping resistor. Current based crossovers are even more complicated since they introduce variations as a result of the reactance of all the drivers due to the serial path the music takes vs the parallel path used in a voltage derived xover.

Q- We see a figure quoted for sensitivity as 104 dB/w/m with no mention of impedance. I am assuming that this measure is at impedance of 8 ohms - as most measures are. What confused me is that to hold this constant level (104 dB) - taking into account impedance variations across the frequency spectrum appears to vary the power need by a factor of 10 (assuming voltage to be constant).

A- in your example, hidden factors is the SPL rating of the drivers and what was done to lower the SPL of mid drivers and tweeters so that they output SPL approximate to the woofer.  In the case of the mid driver, in order to level match the woofer, an autoformer is used which introduces impedance in 2X, 4X, 8X and 16X steps.  This equates to -3db, -6db, -9db, and -16db.  Most folks are using the 2x/-3db or 4x/-6db tap.  At 2X the load is 28ohms from 500hz - 5000hz assuming no swamping resistor is used.

Q - Is this correct? Does voltage actually vary? Are my calculations wrong somewhere?

A - Your basic question can best be answered by evaluating the performance of a full range driver and what changes occurs to input voltage as the input frequency changes.  Assuming a driver can output 70hz - 12khz +- 3db, you can expect the calculations to be consistent within +-3db from 70hz - 12khz.  The impedance will typically increase as you scale up the frequecny ladder.  However, the impedance increase is off set by the greater efficency drivers have at higher frequency than lower frequencies, resulting in SPL levels that belnd together withing the +- rating of the driver.  Now if you introduce 3 drivers, consideration has to be to how each driver will behave base don the manufacures specs within the rated range of the drivers.  Now introduce crossovers, and more factors are introduced.  Add polyswitches, diodes, lamps, and other driver protection devices, and even more factors are introduced.





Link to comment
Share on other sites

According to my pea brain, and what I remember when my betters were talking about this, the thing that prevents these variations from making a speaker sound absolutely horrible is that as frequency changes (and the impedance), so does the ratio of power to output, so that when it is 'hard' for the amp to drive the speaker, simultaneously it becomes easier for the speaker to create output, and when it becomes 'easy' to drive the speaker, the speaker does not create as much output. So it all evens out in the end.

Anybody looking at impedance graphs will quickly surmise that a) the impedance of any speaker is all over the map, so B) something else must be at work to prevent wild changes in output.

Link to comment
Share on other sites

LousyTourist

"as frequency changes (and the impedance), so does the ratio of power to output, so that when it is 'hard' for the amp to drive the speaker, simultaneously it becomes easier for the speaker to create output, and when it becomes 'easy' to drive the speaker, the speaker does not create as much output. So it all evens out in the end."


There is a lot of truth to this.  I

At the low band, watts used are high, ohms are low, SPL level requires movement of a lot of mass

At the higher band, watts used is low, ohms are high, SPL levels require the  movement of less mass



Link to comment
Share on other sites

First off, amps are not voltage constant with current changing. The voltage changes and that is your volume change. As the voltage changes so does the current. So, the calcualtions are easier than you think and the current required for Klipsch speakers is not all that high. Highly efficient speakers do not require a lot of current or voltage. Use your plain jane Power formula of voltage times amps, derived from the voltage and impedance, thereby giving the amps. The only time you need a lot of current and a lot of voltage is if you playing the sound at very high levels. At the 1 watt point, it is a miniscule load on your amp.

If you have inefficient speakers of 88db/watt, then you need a monster amp that has high voltage and high current. Luckily most high voltage amps have high current capability such as Sunfires etc.

Also, unless you playing with a tone generator, the low ohm dip in most speakers is not relavent. Music and movies are playing many tones all at once so your amp sees a more normalized average load. It will supply all the current you need for the low ohm dip. Remember if it didn't, you would know because it would burn out your output transistors. So, all in all, you get what you pay for![:)]

Link to comment
Share on other sites

Max, you are mostly correct.

Most modern amplifiers act as voltage sources until you put too much of a load on them. "Voltage source" is a theoretical concept but means that as load conditions change, their output voltage is constant (assuming a constant input of course, you can always turn down the volume).

Therefore, in testing a speaker, a swept frequency tone of constant rms voltage, say 2.82 volt is used. Then the spl in front of the speaker box is measured with the microphone often at 1 meter distance. Some adjustments can be made mathematically if you use more or less voltage or distance.

= = = = =

The question may be, what do we do when various bass, mid, and tweeter systems are more or less efficient. E.g. say a sealed box puts out 97 dB at that 2.82 volts , but a mid puts out 107 dB, and a tweeter puts out 104 dB. Obviously we have to reduce the drive to the tweeter and reduce the drive to the mid even more.

One way of doing this is with an L-pad made up of adjustable resistors. The L-Pad can present a fairly constant resistance to the amp when the L-pad is adjusted. Depending on the twist of an L-Pad knob, more and more power is diverted to the resistors within the L-pad and less and less to the driver hooked up to it. The overall load of the L-pad input often doesn't vary very much (it is not suppose to, anyway) and does stay around 8 ohms.

One problem is the the driver "looks" backwards into the L-pad too. It sees a higher and higher impedance and this leads to less of a damping factor. I had posted PWK's "The Trouble with Attenuators" some time ago where this is explained. You can probably find it in a search of Technical Questions.

PWK, used transformers, or actually auto-transformers, instead of L-pads. Essentially, if you want to reduce the amount of power going to a driver, you use the transformer as a step-down transformer to reduce the voltage applied to the driver. One thing about a step-down transformer is that the input impedance is raised relative to the load (driver) connected to it. Essentially, less power is absorbed because the input impedance is raised. Remember that Ohms = Volts / current. If ohms go up, current goes down, if V is unchanged. More ohms, less current, less power absorbed.

Just as an aside, house hold lightbulbs are pretty much ordinary wire wound resistors. You may have a 10 watt lightbulb, and a 20 watt and a 40 watt. You plug them into "the mains" which has a constant rms voltage of 120 or 240. How do the lightbulb and mains know to take off 10 watts, or 20, or 40 watts? The answer is that the resistance of the various bulbs determine the power consumption. Therefore, if you can change a lightbulb, you have adjusted power by varying resistance!

- - - - -

Attached is the schematic and input impedance of the Heresy as shown in two different Dope from Hope. (I had posted my collection in "Free The Dope From Hope". As you can see from the schematic, various taps on the autotransformer are used to reduce the voltage going to the mid and tweeter to reduce their ultimate output down to that of the bass sealled box.

But, this causes big changes in the input impedance. Here it is shown on a log graph. You could turn the graph up side down and it would equate to attenuation and the frequency response of the crossover network.

You can also roughly assign areas of the frequency axis to the various drivers. You'll see the mid is where the impedance is much greater than the bass though then it drops down a bit in the tweeter. PWK pointed out that the impedance actually does vary by almost an order of magnitude (power of 10).

Best,

Gil

post-2552-13819329852148_thumb.gif

Link to comment
Share on other sites

In a recent thread Mark mentioned, laughingly, how amp manufacturers boast that their amp doubles in power going from 8 ohms to 4 ohms but never mention what happens when the impedance climbs to 16 or 32 ohms.

Because if an amp doubles from 8 into 4 ohms, then it's always going to half from 8 into 16 and from 16 into 32 and so forth...It's harder to double down into lower impedances because you start nearing the output impedance of the amplifier (which will then break down the low impedance driving the high impedance "rule of thumb" required for an ideal voltage source to work properly).

In other words, lower output impedance is better and the way you see that is by the lowest impedance that it can drive properly. Also, the higher power handlings at lower impedances indicate that the amp is very good at dissipating heat and has very strong power supplies. Current is what generates heat, not power...

Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

×
×
  • Create New...