Yah that is one of the big confusions. There is an electrical resonant system and a mechanical resonant system. These may or may not be closely related and are not the same thing. The Qts as measured for the electrical system can be different than the Qts for the mechanical system by a factor of 2 or more in some drivers. Fs electrical and Fs mechanical can be much differnt also- as much as 30%. Only in a perfect driver are these the same. We do not have perfect drivers so these should be measured separately and determined separately. Part of the measure of a "good" driver is how close the electrical system is to the mechanical system. We listen to the mechanical system driven by the electrical system so the characteristics of the mechanical system are very important and should not be presumed to be near to the values of the electrical system. In my experience measuring drivers the typical errors run about 15% off on Qts mechanical and 10% off for Fs mechanical. This has resulted in using the mechanical values for system design which has proved to be far more accurate in the final result.
Fs and Qms are the electrical reflected values in the electrical system of what the mechanical system plus the electrical system are doing together at the electrical terminals. The electrical reflected values are not directly the mechanical system but rather the electrical systems output of what the electrical system "sees" of the mechanical system. It is a reflection and not the thing. The thing (mechanical system) is its own resonant system.
To improve the match between the voice coil assembly and the cone/diaphragm the first approximation suggest the mass of the two part be the same.
Then you agree with Thiele/Small's closed box paper on modeling drivers mechanical systems with a 2nd order electrical model. If there are known inaccuracies then the model can be corrected.
Yes but how does any of this relate to the differences YOU see concerning nominal 16 vs 8 electrical parameters? let's say measured accurately in your set-up. You seem to be dodging the question at hand see post 18 & 20?
Yes but how does any of this relate to the differences YOU see concerning nominal 16 vs 8 electrical parameters? let's say measured accurately in your set-up. You seem to be dodging the question at hand see post 18 & 20?
broken models
The voice coil diaphragm system ends up being lighter in the 16 ohm version than the 8 ohm with less moving mass resulting in lower Q and slightly higher resonant frequency. The BL product of the 16 ohm is usually higher than the 8 ohm. This means the 16 ohm is a little more efficient than the 8 ohm version due to the lower mass and higher BL. As was written manufactures sometimes add mass to the 16 ohm to make it behave the same as the 8 ohm.
Modeling the electrical system plus modeling the mechanical system has been written about by many. Whether it is written out as a spring mass damper or an inductor capacitor resistor has no effect on the final result as both lead to the same model for the simple system. My point was in the physical beast of a driver there is a inductor with capacitance and resistance plus mechanical spring of the air and suspension, mass of the driver, and damping of several factors with both these systems in operation at all times in the transducer. These are two systems, one electrical and one mechanical and both need to be included in any useful model. AES journal published one paper "How to give your woofer a lie detector test." I co-authored a paper on this very subject which was rejected by the AES with the review board saying "Everyone has known about this for many years so the paper has no use to anyone." Allow me to translate- nah better skip that- The paper really blew up the simple popular model used by so many that is often so wrong. Assuming the electrical system analysis tells us clearly about the mechanical system is simply not so- except in a perfect transducer. Do you recall the "fudge factor" and box loading error adjustments of later Small papers? Well that is because Small never got around to discovery of why there was errors in the designs based on the electrical impedance measurements. There was that fudge factor and box correction errors and several others I forget now. All these were feeble attempts to correct the basic error of the fact the electrical system does not necessarily represent the mechanical system.
I measure BL directly, moving mass very carefully making certain there is no change in amplitude using the delta mass method, Re directly, and resonant frequency of the mechanical system and resonant frequency of the electrical impedance curve. Using this data to calculate Q and box tuning works fine with results within 5% of theoretical with no fudge factor or need for iffy box loss equations.
If you use an oscillator that has a harmonic series like an 8 step synthesizer the electrical resonance is when the tiny peak on the 8 step curve is right at the top of the wave. The mechanical resonance is when the impedance reaches its maximum value. These two are always different. Sometimes as much as 20%. Just a note though. It is the mechanical resonance that is of interest for frequency response and box design and the electrical resonance is used for crossover design and compensations in the electrical curve like an RLC to equalize the electrical resonant peak for use with a high pass passive filter.
I hope this does not dodge the question you have posed. Was not trying to. In the end the 16 ohm driver has always worked better within my experience.
The voice coil diaphragm system ends up being lighter in the 16 ohm version than the 8 ohm with less moving mass resulting in lower Q and slightly higher resonant frequency. The BL product of the 16 ohm is usually higher than the 8 ohm. This means the 16 ohm is a little more efficient than the 8 ohm version due to the lower mass and higher BL. As was written manufactures sometimes add mass to the 16 ohm to make it behave the same as the 8 ohm.
Modeling the electrical system plus modeling the mechanical system has been written about by many. Whether it is written out as a spring mass damper or an inductor capacitor resistor has no effect on the final result as both lead to the same model for the simple system. My point was in the physical beast of a driver there is a inductor with capacitance and resistance plus mechanical spring of the air and suspension, mass of the driver, and damping of several factors with both these systems in operation at all times in the transducer. These are two systems, one electrical and one mechanical and both need to be included in any useful model. AES journal published one paper "How to give your woofer a lie detector test." I co-authored a paper on this very subject which was rejected by the AES with the review board saying "Everyone has known about this for many years so the paper has no use to anyone." Allow me to translate- nah better skip that- The paper really blew up the simple popular model used by so many that is often so wrong. Assuming the electrical system analysis tells us clearly about the mechanical system is simply not so- except in a perfect transducer. Do you recall the "fudge factor" and box loading error adjustments of later Small papers? Well that is because Small never got around to discovery of why there was errors in the designs based on the electrical impedance measurements. There was that fudge factor and box correction errors and several others I forget now. All these were feeble attempts to correct the basic error of the fact the electrical system does not necessarily represent the mechanical system.
I measure BL directly, moving mass very carefully making certain there is no change in amplitude using the delta mass method, Re directly, and resonant frequency of the mechanical system and resonant frequency of the electrical impedance curve. Using this data to calculate Q and box tuning works fine with results within 5% of theoretical with no fudge factor or need for iffy box loss equations.
If you use an oscillator that has a harmonic series like an 8 step synthesizer the electrical resonance is when the tiny peak on the 8 step curve is right at the top of the wave. The mechanical resonance is when the impedance reaches its maximum value. These two are always different. Sometimes as much as 20%. Just a note though. It is the mechanical resonance that is of interest for frequency response and box design and the electrical resonance is used for crossover design and compensations in the electrical curve like an RLC to equalize the electrical resonant peak for use with a high pass passive filter.
I hope this does not dodge the question you have posed. Was not trying to. In the end the 16 ohm driver has always worked better within my experience.
Reality check:
This thread was originally about compression drivers.
With decent horn loading, the acoustical load on a compression driver will be predominantly resistive over much of it's range.
For optimum efficiency, this does have to be matched to the driver's mechanical impedance. However this has nothing to do with matching the mass of the coil to the mass of the diaphragm. It is a function of BL, Rvc, diaphragm area and throat area.
The total moving mass serves only to set the high-frequency breakpoint. In this respect lighter is better. Extra moving mass in the coil, it's former or the diaphragm is undesirable.
With respect to the question of 8ohm vs 16ohm coils:
Any competent manufacturer will be able to manufacture drivers with different electrical impedances having otherwise identical parameters e.g. Fs, Qes, Qms, Qts, Vas.
Witness, for example, car subwoofers with 2 separate 2ohm voicecoils. These may be connected in series (for four ohms nominal impedance) or in parallel (for one ohm nominal impedance). Either configuration will give identical results in the same enclosure (aside from electrical impedance and sensitivity).
To get the same performance with coils of different impedances in any kind of driver, the total volume of wire needs to be the same e.g. for 16 ohm vs 8ohm coils: the 16ohm coil will need to be about 1.4 times the length, and about 0.7 times the cross sectional area of the 8ohm coil.
The idea that the electrical resonance and mechanical resonance are two separate things is complete nonsense, as is the idea that mechanical Qts is different to electrical Qts.
Regards - Godfrey
This thread was originally about compression drivers.
With decent horn loading, the acoustical load on a compression driver will be predominantly resistive over much of it's range.
For optimum efficiency, this does have to be matched to the driver's mechanical impedance. However this has nothing to do with matching the mass of the coil to the mass of the diaphragm. It is a function of BL, Rvc, diaphragm area and throat area.
The total moving mass serves only to set the high-frequency breakpoint. In this respect lighter is better. Extra moving mass in the coil, it's former or the diaphragm is undesirable.
With respect to the question of 8ohm vs 16ohm coils:
Any competent manufacturer will be able to manufacture drivers with different electrical impedances having otherwise identical parameters e.g. Fs, Qes, Qms, Qts, Vas.
Witness, for example, car subwoofers with 2 separate 2ohm voicecoils. These may be connected in series (for four ohms nominal impedance) or in parallel (for one ohm nominal impedance). Either configuration will give identical results in the same enclosure (aside from electrical impedance and sensitivity).
To get the same performance with coils of different impedances in any kind of driver, the total volume of wire needs to be the same e.g. for 16 ohm vs 8ohm coils: the 16ohm coil will need to be about 1.4 times the length, and about 0.7 times the cross sectional area of the 8ohm coil.
The idea that the electrical resonance and mechanical resonance are two separate things is complete nonsense, as is the idea that mechanical Qts is different to electrical Qts.
I'm not surprised
They were being polite.
Regards - Godfrey
You seem to believe you have this all figured out and I do not care to argue. As a loudspeaker manufacture and physicist I suppose all the information I have presented is incorrect within the limited knowledge available to you in the land of theory.
My key point- have you tried anything I suggested in the lab? Obviously not. To say the electrical system is the same as the mechanical system... is like saying a transistor is the same as a lever. Neither are true.
I was simply trying to bring to light the real effects which you have rejected without trial or empirical evidence. I will defer to your expertise in decisions without trial. There are many experts of this variety. I have presented fairly clear descriptions and named a paper which addresses the subject and mentioned several other works that address the error in the concept you are holding as a singular truth.
Yes I agree other manufacturers may adjust parameters so 8 and 16 ohm drivers have similar performance. That was never even a topic. With all things being otherwise equal the 16 ohm will be better.
My key point- have you tried anything I suggested in the lab? Obviously not. To say the electrical system is the same as the mechanical system... is like saying a transistor is the same as a lever. Neither are true.
I was simply trying to bring to light the real effects which you have rejected without trial or empirical evidence. I will defer to your expertise in decisions without trial. There are many experts of this variety. I have presented fairly clear descriptions and named a paper which addresses the subject and mentioned several other works that address the error in the concept you are holding as a singular truth.
Yes I agree other manufacturers may adjust parameters so 8 and 16 ohm drivers have similar performance. That was never even a topic. With all things being otherwise equal the 16 ohm will be better.
If you actually are a manufacturer or work for one, then you should be explicitly stating that bit of information.
Clarification
Did not see that in the terms of service for DIY audio- did I violate something? Please let me know.
The electrical resonance is when the current and voltage waveforms measured at the voice coil are exactly in phase with one another.
The mechanical resonance occurs at the very peak in the reflected impedance curve as measured at the voice coil.
I have never seen these two measurements be the same frequency in testing over 1000 drivers of all shapes and variety. Some have been very near within 3% and others different by as much as 100%.
Go to the lab and try it. You will find this statement is correct.😀
Did not see that in the terms of service for DIY audio- did I violate something? Please let me know.
The electrical resonance is when the current and voltage waveforms measured at the voice coil are exactly in phase with one another.
The mechanical resonance occurs at the very peak in the reflected impedance curve as measured at the voice coil.
I have never seen these two measurements be the same frequency in testing over 1000 drivers of all shapes and variety. Some have been very near within 3% and others different by as much as 100%.
Go to the lab and try it. You will find this statement is correct.😀
Interesting
Thanks for your interesting information. This thread has actually morphed into some high level engineering discussions and your experience is appreciated. This is a good example of how a simple question has sparked a higher level discussion and how DIYaudio has grown to be the cutting edge place to learn of all the internet sites. I don't understand why this comment had to be thrown in from left field.
.
"If you actually are a manufacturer or work for one, then you should be explicitly stating that bit of information. "
.
You are making nice, civil discussion of high level design parameters that I wouldn't have considered. Are we all supposed to where a sign around our neck?
Back to my original question it appears to be true as I had hoped, that for home audio where max spl and voltage efficiency isn't required, the 16 ohm driver has no draw backs.
Thanks for your interesting information. This thread has actually morphed into some high level engineering discussions and your experience is appreciated. This is a good example of how a simple question has sparked a higher level discussion and how DIYaudio has grown to be the cutting edge place to learn of all the internet sites. I don't understand why this comment had to be thrown in from left field.
.
"If you actually are a manufacturer or work for one, then you should be explicitly stating that bit of information. "
.
You are making nice, civil discussion of high level design parameters that I wouldn't have considered. Are we all supposed to where a sign around our neck?
Back to my original question it appears to be true as I had hoped, that for home audio where max spl and voltage efficiency isn't required, the 16 ohm driver has no draw backs.
4 & 8 ohm version woofers show exactly as claimed
Here are a 4 and 8 ohm woofer that are the same save for the voice coil resistance and winding. Please note the Q is shown as the same even though by the number there for Qms the 4 ohm has a higher Q and more mass. 8 ohm has higher BL and all other things claimed.
http://www.parts-express.com/pdf/290-520s.pdf
http://www.parts-express.com/pdf/290-521s.pdf
Here are a 4 and 8 ohm woofer that are the same save for the voice coil resistance and winding. Please note the Q is shown as the same even though by the number there for Qms the 4 ohm has a higher Q and more mass. 8 ohm has higher BL and all other things claimed.
http://www.parts-express.com/pdf/290-520s.pdf
http://www.parts-express.com/pdf/290-521s.pdf
Good example.
"8ohm" driver: BL = 18.9 and Re = 5.3
"4ohm" driver: BL = 16.5 and Re = 3.8
Ratio of BLs = 18.9 / 16.5 = 1.145
Ratio of Res = 5.3 / 3.8 = 1.395
Assuming the same field strength for both, the higher BL of the "8ohm" unit suggests a longer voice coil wire.
However that's not enough to account for the resistance ratio on it's own, so the voice coil wire must be smaller diameter as well.
Assuming the same poleplate and Xmax, ofcourse thats how it is
Would be nice if they offered options fore different Xmax instead, but that would be more of a woofer thing 🙂
Knowing from the other thread that you use Tripath amps, the answer is not as clear cut.
I know that the Tripath amps like smaller impedances, so if I was you I would go with
the 8 ohm version, especıally if you use digital crossovers.
Greets,
Klaus
I know that the Tripath amps like smaller impedances, so if I was you I would go with
the 8 ohm version, especıally if you use digital crossovers.
Greets,
Klaus
I used to use 16 ohm, then I switched to 8 ohms. Can't say that there was any notable differences in performance of the two. The 16 ohm allows for smaller caps and inductors for a little lower cost on the passive comps. The 16 ohm can be a lot harder to get and to get replacement parts for since they are not as common.
Bottom line to me is that there isn't much difference - both have pluses and minuses.
Bottom line to me is that there isn't much difference - both have pluses and minuses.
It is very easy to throw a resistor in parallel with the driver to adjust the impedance down. My vote is still the 16 ohm... with a resistor if needed for the amp.
I do not get it. This gentleman is using active crossover so..."what does this mean?????" Seems like non-sense. Sorry but do not mean to be rude. Adding a resistor in parallel with the driver is an excellent way to lower Z of the driver and make the dominant character resistive. All that is good for passive or active system.
Could you please explain what part is not so simple? -Thanks
Member
Joined 2003
Speakers are not resistors. They have an impedance that changes with frequency. At some frequency they will be 12 ohms, at some other frequency maybe only 6 ohms.
Example:
So, lets say you put a 5 ohm resistor in parallel with a speaker:
-At the point where it is 12 ohms, you will now have ~3.5 ohms, or a 8.5 ohm reduction
-At the point where it is 6 ohms, you will now have ~2.7 ohms, or a 3.3 ohm reduction.
-Reduction of impedance over frequency is not uniform.
Since that image posted above is not a flat horizontal line, attenuation is not applied in a uniform manner over frequency, and both impedance and frequency response of the driver will change in a non-uniform manner over frequency.
There's a reason why attenuators are ususally L-pads, and not just a single resistor in parallel.
Example:
So, lets say you put a 5 ohm resistor in parallel with a speaker:
-At the point where it is 12 ohms, you will now have ~3.5 ohms, or a 8.5 ohm reduction
-At the point where it is 6 ohms, you will now have ~2.7 ohms, or a 3.3 ohm reduction.
-Reduction of impedance over frequency is not uniform.
Since that image posted above is not a flat horizontal line, attenuation is not applied in a uniform manner over frequency, and both impedance and frequency response of the driver will change in a non-uniform manner over frequency.
There's a reason why attenuators are ususally L-pads, and not just a single resistor in parallel.
Last edited:
Speakers are not resistors. They have an impedance that changes with frequency. At some frequency they will be 12 ohms, at some other frequency maybe only 6 ohms.
Example:
So, lets say you put a 5 ohm resistor in parallel with a speaker:
-At the point where it is 12 ohms, you will now have ~3.5 ohms, or a 8.5 ohm reduction
-At the point where it is 6 ohms, you will now have ~2.7 ohms, or a 3.3 ohm reduction.
-Reduction of impedance over frequency is not uniform.
Since that image posted above is not a flat horizontal line, attenuation is not applied in a uniform manner over frequency, and both impedance and frequency response of the driver will change in a non-uniform manner over frequency.
There's a reason why attenuators are ususally L-pads, and not just a single resistor in parallel.
I think you've indavertantly shot yourself in the foot 😀 as you have illustrated quite clearly why you might want to parallel a resistor across a driver VC in some instances.
The impedance over frequency was not uniform before the addition of the resistor, but it is much more so after.
Using your specific example notice something about the % of reactance related impedance deviation of one versus the other. You failed to notice the ratio (the important thing here) in one case is 2:1 and in the other case 1.22:1 which in the case of either a passive XO or an imperfect and not quite voltage-source-like amp driving it happens to be a big improvement.
Paralleling the VC with a resistor does not change the voltage sensitivity of the driver and there is no attenuation if the device driving it is a perfect voltage source. In the case of a driving device that looks more like a current source it should reduce the amplitude at the impedance maxima considerably and hence hopefully the acoustic output. (resonance, etc) In the case of an X-O that wants to be terminated by a predictable impedance it should reduce unpredicted interactions within the X-O pass band.
Note that driving a driver with an Lpad accomplishes much the same thing, but in some cases may not be necessary if level adjustment is not needed..
I think this technique helps a lot with amplifiers that do not behave like ideal voltage sources, and I confess I do use it from time to time.
Hopefully in my haste to post something before I have to head upstairs to dinner I haven't further confused the issue. 😀
Last edited:
In parallel not in series + active crossover
The resistor is in parallel and not in series. With active crossover used (as specified in an earlier post) there will be no attenuation and the load will never ever be greater than the resistor value. Indeed it will flatten that curve you show a lot anyway with any pad or passive crossover. The key point was to load the amplifier as the chosen amp likes loads less than 16 ohms.
I have used an 8 ohm resistor in parallel with an 8 ohm driver making it nominal 4 ohms and boy does that tame the Z curve as far as a passive crossover is concerned. With the parallel resistor that curve varies from 4 to 8 ohms and not 8-80 ohms. It is a great trick as many amps drive 4 ohms just fine. Further the 8 ohm parallel resistor damps the 8 ohm driver throughout the DC to light entire range. Damping otherwise not available with a passive crossover. More damping of driver is almost always better, yes? Of course this does not work with bad amps that cannot drive 4 ohm load well.😀
Yah, what Kevinkr said...kudos.
The resistor is in parallel and not in series. With active crossover used (as specified in an earlier post) there will be no attenuation and the load will never ever be greater than the resistor value. Indeed it will flatten that curve you show a lot anyway with any pad or passive crossover. The key point was to load the amplifier as the chosen amp likes loads less than 16 ohms.
I have used an 8 ohm resistor in parallel with an 8 ohm driver making it nominal 4 ohms and boy does that tame the Z curve as far as a passive crossover is concerned. With the parallel resistor that curve varies from 4 to 8 ohms and not 8-80 ohms. It is a great trick as many amps drive 4 ohms just fine. Further the 8 ohm parallel resistor damps the 8 ohm driver throughout the DC to light entire range. Damping otherwise not available with a passive crossover. More damping of driver is almost always better, yes? Of course this does not work with bad amps that cannot drive 4 ohm load well.😀
Yah, what Kevinkr said...kudos.
And you are quite certain that a paralel component doesnt affect phase bahaviour in respect to other drivers, delay etc
Anyway, only few 8ohm driver are only just 8ohm, but often has actual 6ohm impedance, or even 5ohm
Which means that making 8 ohm driver into 4ohm with 8ohm resistor, might actually be closer to 3ohm
But doesnt matter much if 16ohm
Well, I have always believed strongly in true Lpads fore attenuation, with passive filters
But I have lately found that it does seem to sound better with a series resistor
I have also believed in zobels on tweeters, but have become less convinced
Its clear to me that all paralel components are to be handeled with exstreme care
Maybe its exclusively with passive filters, cant say
Anyway, only few 8ohm driver are only just 8ohm, but often has actual 6ohm impedance, or even 5ohm
Which means that making 8 ohm driver into 4ohm with 8ohm resistor, might actually be closer to 3ohm
But doesnt matter much if 16ohm
Well, I have always believed strongly in true Lpads fore attenuation, with passive filters
But I have lately found that it does seem to sound better with a series resistor
I have also believed in zobels on tweeters, but have become less convinced
Its clear to me that all paralel components are to be handeled with exstreme care
Maybe its exclusively with passive filters, cant say
- Status
- Not open for further replies.