Curious, Mac is using transformer outputs in their Bi-polar amps, and still listing .003% distortion @ rated power and BW. Now they have a long history, maybe equaled by Luxman, for transformer quality, but I am used to graphs like posted here.Wow great to have Lars here!
@phase_accurate This paper seems to measure this effect and speculated about voice coil former and distortion effects in sector 11, by the auther of 《Current-Driving of Loudspeakers》:
https://acoustics.ippt.pan.pl/index.php/aa/article/view/1780/pdf_255
Personally I think hysteresis distortion in speakers seems pretty constant % to SPL at low level. However I'm not sure it's an effect rising % as level goes lower. Like transformer distortion. Hope Lars has an answer!
How Transformers Distort
https://www.soundonsound.com/techniques/analogue-warmth
I've seems transformer measurements shows lowering distortion as frequency goes higher. Maybe Mac uses 1khz or higher, but the graph I show uses 50hz.
Regards,
I've seen this on Voice coil magazine before, and was confused for quite a while, but now I think I understand.
A higher-end series of driver, whose 2nd order harmonics LF distortion is higher than its lower end drivers. Too much copper above the pole, but relatively not enough inside and below the gap.
Actually I own a pair of that driver myself, it's a good one with pretty good reputation. But purifi has noticeably better mid bass definition.
Thanks Lars for my sudden awareness!
Usually the 4 Ohm drivers have around half the inductance of 8 Ohm drivers. So nothing is gained as the pole frequency given by Re and Lvc is
f= Re / (2*Pi*Lvc)
And it is this pole frequency which is modulated via excursion and subsequent Lvc variation that leads to intermodulation distortion.
Regards
Charles
f= Re / (2*Pi*Lvc)
And it is this pole frequency which is modulated via excursion and subsequent Lvc variation that leads to intermodulation distortion.
Regards
Charles
yes, exactly. impedance transformation does not change anything but how much current and voltage is needed, whilst true power and distortion is the same
I don't know actually. 4R gives a higher voltage sensitivity than 8R so a speaker sounds louder at moderate levels. Any amp class can be designed for whatever impedance. It is true that because the way MOSFET tech scales with voltage then class D tends to be more efficient and makes less EMI using low impedance and higher current.Lars,
So, why the trend to 4 Ohm speakers. Very tough on amplifiers. Conventional A/B amps distortion can rise greatly. Most AVR's can't handle 4 Ohms at "enthusiastic" levels. Is this inherent in reducing and linearizing the inductance?
( A secret ploy to get us to buy class D?
Thanks. Slowly sinking into my old head.
So basically, all these 4 Ohm speakers are just plain not very smart considering most AVR's and mid-fi equipment can't safely drive them.
Of course, we need 4 Ohm drivers for MMT etc
What we really need is a killer 400 to 4K midrange with enough efficiency and smooth response.
"Sounds louder" Huh? 5W is 5W. If a 4 Ohm driver does 85 dB 1M 1W and and an 8 Ohm 85 1M 1W, no difference in output swapping current for voltage.
Yes, we can build low Z amps. My amp has sufficient power supply and heat sinking to do close to twice the current, but even with MOSFETs linearity is not as good. Transistors are better voltage amplifiers than current amplifiers. We can double up the outputs, but now our bias current doubles. You can also have stabilization issues unless careful trying to cover between 2 and maybe 12 Ohms. I don't know enough about class D to know how it deals with the load. I am guessing a little better. If I get brave, I may have MARCH send me an amp to find out if I think class D has arrived.
So basically, all these 4 Ohm speakers are just plain not very smart considering most AVR's and mid-fi equipment can't safely drive them.
Of course, we need 4 Ohm drivers for MMT etc
What we really need is a killer 400 to 4K midrange with enough efficiency and smooth response.
"Sounds louder" Huh? 5W is 5W. If a 4 Ohm driver does 85 dB 1M 1W and and an 8 Ohm 85 1M 1W, no difference in output swapping current for voltage.
Yes, we can build low Z amps. My amp has sufficient power supply and heat sinking to do close to twice the current, but even with MOSFETs linearity is not as good. Transistors are better voltage amplifiers than current amplifiers. We can double up the outputs, but now our bias current doubles. You can also have stabilization issues unless careful trying to cover between 2 and maybe 12 Ohms. I don't know enough about class D to know how it deals with the load. I am guessing a little better. If I get brave, I may have MARCH send me an amp to find out if I think class D has arrived.
I am a newbie here and looking to build the speakers with best drivers for my project, reading through this thread and analyzing content I have tried to search for a drivers (midwoofers) which correspond to the findings on the forum, criteria's were:
1) High Qms ( 4 or above)
2) Low Qes (range from 0.2 to 0.35)
3) Difference between Qes vs Qts around 0.02
4) Rms below or close to 1 Kg/s
and what I found was usually only well regarded, high price drivers:
1) Audio technology 18 H 52 06 13 SDKA-LR Midrange (Qms-4.92, Qes-0.27, Qts-0.25, Rms-0.54 Kg/s)
2) Seas E0074-08 W18EX003 (Qms-4.85, Qes-0.32, Qts-0.30, Rms-not listed )
3) Scan Speak 18WE/8542T00 (Qms-5.01, Qes-0.34, Qts-0.32, Rms-1 Kg/s)
All this are very expensive drivers, could anyone suggest drivers with listed specs but in cheaper price range?
1) High Qms ( 4 or above)
2) Low Qes (range from 0.2 to 0.35)
3) Difference between Qes vs Qts around 0.02
4) Rms below or close to 1 Kg/s
and what I found was usually only well regarded, high price drivers:
1) Audio technology 18 H 52 06 13 SDKA-LR Midrange (Qms-4.92, Qes-0.27, Qts-0.25, Rms-0.54 Kg/s)
2) Seas E0074-08 W18EX003 (Qms-4.85, Qes-0.32, Qts-0.30, Rms-not listed )
3) Scan Speak 18WE/8542T00 (Qms-5.01, Qes-0.34, Qts-0.32, Rms-1 Kg/s)
All this are very expensive drivers, could anyone suggest drivers with listed specs but in cheaper price range?
Some are slightly out of your criteria but maybe this: http://www.dibirama.altervista.org/home-page/mid-woofer/721-audax-hm170c0-mid-woofer-6-5-8-ohm-120-wmax-2.html
Or this: http://www.dibirama.altervista.org/...x-hm170mn0-mid-woofer-6-5-8-ohm-120-wmax.html
Another: http://www.dibirama.altervista.org/...udax-hm170z18-mid-woofer-6-8-ohm-120wmax.html
This: http://www.dibirama.altervista.org/...b17nrx2c35-4-mid-woofer-6-4-ohm-120-wmax.html
Or that: http://www.dibirama.altervista.org/...are-hwb160-mid-woofer-6-5-8-ohm-160-wmax.html
That: http://www.dibirama.altervista.org/...are-hwg160-mid-woofer-6-5-8-ohm-260-wmax.html
Yet another: http://www.dibirama.altervista.org/home-page/mid-woofer/70-monacor-sph165kep.html
You might over constrain the problem: Specifying Qms and Qes and Qts is given as Qts=(Qms•Qes)/(Qms+Qes). Similarly, Qms and the lumped parameter Rms are directly linked (like Mms and Cms are given from the T/S parameters fs Qms Qes and Qts and vice versa).
What is the background for these requirements if I may ask? The specs are quite typical for midwoofers suitable for vented boxes. The most important parameter might be fs and that is a don’t care?
People who select drivers only on basic motor-suspension-cone mass qualities (and not exactly the right ones too) could maybe do some homework first.
Define your demands. You want a midwoofer. For starters: what frequency range and what level? Maybe: what directivity (polar radiation pattern)? Which distortion limits at which levels have to be met? What crossover (technique)? What amp power is available? These are only a few criteria to start with, not 'what Qms or Rms'. Have you even considered the relation between those two (Qms = (1/Rms) *SQR (Mms/Cms))?
Define your demands. You want a midwoofer. For starters: what frequency range and what level? Maybe: what directivity (polar radiation pattern)? Which distortion limits at which levels have to be met? What crossover (technique)? What amp power is available? These are only a few criteria to start with, not 'what Qms or Rms'. Have you even considered the relation between those two (Qms = (1/Rms) *SQR (Mms/Cms))?
Yes, you are right it is given in formula, but was is lacking is the ratio/values of mentioned variables... that's why I have specified the ranges. As mentioned earlier I am searching midwoofer for 2 way bookshelf speaker, budget for the project is low, I do not have chance to listen various drivers in different combination, so this way will not work for me. I found an article: https://www.speakerplans.com/index.php?id=faq1
which inspired me, maybe there is some "magic" formula in ratio of T/S parameters which will enable me to find the driver I am looking for, so I dig further in this forum as well and what I found was more or less in correlation with parameters above, like:
Quote:
"Drivers with a very high mechanical Q can sound more open, cleaner and have a better dynamic range. This is because they have less loss."
"A driver with a low Qts of around 0.20 would have a large magnet and be able to move the cone with a lot of force." and so far for other parameters as well.
I understand that Fs, Frequency response and radiation pattern are very important but this are more easy to find: radiation pattern is defined by size of the driver, low Fs and good frequency response nowadays you can find in very affordable drivers as well ...
So back to my quest what I have found:
- More or less all top manufacturers high end drivers like (Audio Technology's most praised mid range drivers, SEAS Graphene new Excel line midwoofers, Scan Speak Elipticor and Illuminator range of midwoofers and SB Acoustics Satori line midwoofers) fit in the same parameter's range as mentioned above and I have found very few in lower pries category drivers that have the same T/S parameters.
The Cheapest one Identified was SB Acoustics SB17NRX2C35 (Thanks to forum member YSDR). So what I would like to do next is to search past "great" and "famous" midrange drivers / speakers and compare their T/S and if there will be correlation, for me it will be the indication to what parameters look for in affordable drivers, when baying them future hobby DIY projects.
If you use flawed arguments to choose your drivers you are limiting yourself. Myself, the drivers i have liked best tend to have had low Qm.
The whole premise of this thread is flawed IMO.
dave
In the end, to get the desired result, the sum of the damping provided by the amplifier and that provided by the speaker are the same. In the case of a high Qms driver one needs a amplfier with high damping (ie low Rour), meaning you limit yourself from trying some of the bes amplifiers out there.
And as such cannot go that low except maybe in a big horn. The force has to be considered in conjunction with the mass to be mived and how hard it is to get moving. The magnet, cone, surround, and many, many other factors play in.
Low Qts means that either Qm or Qe pr both are also small.
… and the shape of the cone, the material, the compliance, the dustcap...
dave
What would qualify as a high Qms in my experience would be something above 10 or so. Not above 4.
As the Qms alone does not say much, more interesting is actually the Rms factor as it puts in relation several other interesting parameters (lower moving mass, lower resonance frequency AND a high Qms) to achieve a low Rms. That is: a detailed and articulate punchy bass also at lower listening levels, not just with several watts input or loudness activated.