Why Stax headphone being so small can go down to 6Hz?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
An esl of any type will go as low as you want it too, as with any speaker.
It all depends on what type of method is used to drive them.
If a if a transformer is used to drive them it would take a rather large core and high primary inductance to achieve a good response below 10 or 20 hz but it can be done as I have done this ,but at a very low level.
The stax system uses a direct drive system ,and, that low of a response is easily achievable because there is no transformer to limit the low frequency response of the amplifier.
I hope that helps you to understand. jer
 
Dear Geraldryjr,

Your explanation definitely enlightened me, thank you!
I know I missed something.
But I do still have some questions.

Tube amps use output transformers too, although in step down rather than step up as used in ESL amp.
And some ESL amp are using old tube amp's output transformer in the opposite direction.
So why did output transformer in ESL amp attenuate the bass, but not in tube amp?

I am also under the impression that ESL speaker with larger membrane do better into the low frequency.
If transformer is the limit, how does larger membrane help?
It is that, because larger membrane need smaller movement to create the same SPL, it needs lesser voltage and higher current to drive to the same SPL, thus allow for lower transformer ratio, which cause less attenuation in the bass?
If the above guess is right (lower transformer ratio means less attenuation in the bass), then every ESL builder should be looking for high native voltage amps (which need less step up ratio), I'm not sure if this is the case.

A lot of questions. Would appreciate any answer.
 
"Tube amps use output transformers too, although in step down rather than step up as used in ESL amp.
And some ESL amp are using old tube amp's output transformer in the opposite direction.
So why did output transformer in ESL amp attenuate the bass, but not in tube amp?"

A little more info as to which amps and transformers are being compared.
Most tube amp OPT's are only designed for a maximum lowest of 40hz to 80hz or so at -3db so the flat resposne is higher than that by at least an octave for an average to mediocre hi-fi tube amp and is the norm for guitar amps.
For for higher quality OPT that goes down to at least 20hz flat requires at 4 times or more iron than a power transformer with the same power rating at 60hz.
This is the reason that a that a 50 watt high quality transformer that goes to 20hz or better are so expensive and large.
For instance if you take a look at some of the plitron transformer spec's such as a 70 watt opt that goes to 20hz you would see that by the size of it, one would think it should handle at least 300 watts of power or more.
In which it could but not at 20hz.

"I am also under the impression that ESL speaker with larger membrane do better into the low frequency.
If transformer is the limit, how does larger membrane help?
It is that, because larger membrane need smaller movement to create the same SPL."

Yes this true,but ther are two factors to take in consideration here.
First it takes (x) amount of displacment to produce (y) amount of spl at (z) frequency.
So in order to produce alot of bass (or more spl at any frequency for that matter) one must be able to displace a large amount of volume of air and is linear as per octave.
Something like 2*(x) for (z)/2=(y)spl
(some one can jump in and correct me if I am wrong but I am pretty sure it is right)
So increasing the surface area is the best way to combat the problem of a limited excursion of an esl.
I have found that a suface area 4 to 6 times that of a single dynamic driver, is equivalent to the single driver(due to the limited excursion)as a good giudeline.

Second, due to the the nature of a dipole radiator the lower frequency's get cancled by the backwave and there are several ways to combat this.
Increasing the panel width and/or adding wings on the sides to effectively increase the width and by bass boost filtering or both.
If you were to mount the panel in a wall (with the back side open of course) eliminating the back wave cancellation you would find plenty of bass and spl's only limited by the surface area and maximum excursion just like any other speaker.
This was meant for as an example,and there have been many discussions about this type of mounting and I don't wish to open that can of worms at the moment.

"It is that, because larger membrane need smaller movement to create the same SPL, it needs lesser voltage and higher current to drive to the same SPL, thus allow for lower transformer ratio, which cause less attenuation in the bass?
If the above guess is right (lower transformer ratio means less attenuation in the bass), then every ESL builder should be looking for high native voltage amps (which need less step up ratio), I'm not sure if this is the case."

The transformation ratio does not determine the low frequency response of a transformer.
The primary inductance does.
From this (primary turns) the transformation ratio determines how much of a load the panel presents to the amplifier.
Yes it is logical to say that high voltage swing from the amp and lesser turns ratio is the way to go,but in fact the opposit is true.
As stated in another thread and in my transformer testing has shown that in order to keep the core form saturating at a higher voltage at (x)frequency then one must use more primary turns and this is directly refelcted to the amount of secondary turns needed to keep the same ratio which greatly increases the secondary winding capacitance,not good.
But, yes, the trade off is the higher the voltage swing the lesser transformation ratio can be used.
The other trade off is raising the lowest usable frequency,but then we wouldn't have a fullrange system anymore,would we?
This is a subject where compromise plays a big part and I have found that in order to get around these issues is to use a very large core and I have recently found a vendor where I can get such larger cores rather cheaply.
starting with a 1800va torroidal at about $20 to $30.
Using a large core will allow me a nice low frequency and still be able to use a high transformation ratio as it will require less turns on the primary by as much as a factor of 4 times or more compared to the 200va core I am using at the moment greatly reducing any added secondary capacitance.
I think I have covered your most of your questions.
If not,keep asking . jer
 
This is a subject where compromise plays a big part and I have found that in order to get around these issues is to use a very large core and I have recently found a vendor where I can get such larger cores rather cheaply.
starting with a 1800va torroidal at about $20 to $30.

I think I have covered your most of your questions.
If not,keep asking . jer

Hello

Can you tell me the vendor name ?

Thank you

Bye

Gaetan
 
Hi,

If I did not get it wrong most Stax headphones are of ESL nature.

All that I know.

From what I read ESL cannot go low enough for full range unless it is a few meter in diameter?

That is not strictly correct. ESL's can produce very low frequencies without problem. However as they are dipoles these low frequencyies will be partially cancelled by the dipole action. This of course can be equalised.

However as ESL Diaphragms cannot move much it is not really possible to apply much equalisation, unless the diphragm is made very large.

So how does Stax, being 10cm or less in size, can go 6 to 41,000 Hz?

First, headphones differ greatly from any speakers. If you have open backed headphones (such as the Stax), you have a situation where the headphones cushion closes the system against your ear, so in effect you are sitting inside the Box with the other side of the headphone venting in the room! Under those conditions the LF response can be quite extended on the "inside".

If you listen to the sound emanating from headphones sitting on spomeone elses head the sound is typically tinny with no bass. This changes dramatically when the headphones are on your ears.

Secondly, based on measured performance (Boston Audio Society - November 2002 Meeting Headphone Test Clinic) the Stax ESL headphones are at least 10dB down at 20Hz, one may speculate that at 6 Hz they would be at least 20dB down.


Stax specify neither the actual measurement methode nor the limits for this frequency response, so the number is meaningless.

I have a pair of small active PC speakers that rolls off below 100Hz at 12dB/Octave and barely reaches 10KHz with any flatness.

If I wanted to though, I could claim it performs 25Hz - 40KHz.

I would only need to perform my measurements with a -24dB flatness and omit stating this anywhere (or at best state it in the most well hidden small print).

I did some search but did not find any result.
There must be something obvious that I missed out.

Yes. You first missed that the numbers are pure marketing specmanship (see the measurement in above link) and secondly you are missing the fundamental differences between how headphones and speakers operate.

As for all that talk about transformers, pulllleeeessse!

Gerald, you are so far off, it is severely unfunny. The transformers in ESL's are not the limiting factor. At any extent, they can be removed (a friend is running a 3-Way system that uses "DIRECT DRIVE" ESL Panels for high frequencies and upper bass/midrange. The LF limitations of the Midrange Panel are not down to the transformer, it ain't using one.

But because of the limited size it rolls off at around 150Hz and is equalised only down to around 60Hz to avoid overdriving the panel (which produces arcing and burns holes into the film).

Ciao T
 
"Gerald, you are so far off, it is severely unfunny. The transformers in ESL's are not the limiting factor. At any extent, they can be removed (a friend is running a 3-Way system that uses "DIRECT DRIVE" ESL Panels for high frequencies and upper bass/midrange. The LF limitations of the Midrange Panel are not down to the transformer, it ain't using one. "

If a transformer isn't a limiting factor then why is your friend not using them instead of a direct drive system?
I personaly would rather use a direct drive system and will have working in the near future.
I simply stated as to why an average opt does not have a very good low frequency response.
I did not state that the transformer is the only limiting factor

"Second, due to the the nature of a dipole radiator the lower frequency's get cancled by the backwave and there are several ways to combat this.
Increasing the panel width and/or adding wings on the sides to effectively increase the width and by bass boost filtering or both."

"So in order to produce alot of bass (or more spl at any frequency for that matter) one must be able to displace a large amount of volume of air and is linear as per octave"


"A little more info as to which amps and transformers are being compared."

Obviously you missed the quotes and too what was quoted and what I stated to what was quoted. jer
 
Hi,

If a transformer isn't a limiting factor then why is your friend not using them instead of a direct drive system?

Because we are already using tube Amp's that swing enough voltage and direct coupling saves the cost, weight, size of two transformers?

However, my point stands, the Transformers are not the reason LF output from ESL's is limited and direct-drive cannot help to extend LF response, as the point where the limitations of a competently designed transformer are happens far below the cutoff enforced by dipole action and limited excursion of ESL panels.

I personaly would rather use a direct drive system

Sure, but it will not get you an iota of LF extension, unless your transformers where very badly designed. I'm very much in favour of direct drive to ESL's, as mentioned, I worked with my Friend on his system and designed the Amplifiers.

I simply stated as to why an average opt does not have a very good low frequency response.

Funny. I routinely work with tube amplifiers where the frequency response is essentially flat to 20Hz and with -3dB points in the 10Hz region or so (mind you, not at full power, but few amplifiers are running near full power with music).

So, again, in practice the transformers do not HAVE TO BE a limiting factor.

Ciao T
 
Those numbers are all but meaningless but a well driven Stax SR-007mk1 (not the Mk2 since the baffle isn't sealed) will go very low. Stax aren't alone in giving out such meaningless measurements as when Sennheiser manufactured electrostatics (HE60 and HE90 aka The Orpheus) they claimed for the 12Hz-65kHz former and 7Hz-100kHz for the latter but at -10dB. Beyer did it too with the only electrostatic they made, the ET1000 (10-25000hZ).

Secondly, based on measured performance (Boston Audio Society - November 2002 Meeting Headphone Test Clinic) the Stax ESL headphones are at least 10dB down at 20Hz, one may speculate that at 6 Hz they would be at least 20dB down.

Thanks for that, something regarding Stax headphones that I hadn't seen before. :cool: Not sure what driver they used for the Stax headphones but if it was the SRM-T1 they keep talking about then no wonder they ran into issues. That amp uses 6CG7 tubes which are pushed way past their limits and the simple CRC PSU is far from ideal. Still not nearly as bad as the Koss E/90 amp, the ESP950 would have fared a lot better off the SRM-T1. I doubt they could have found a Stax SRM-T2 for the comparison but that would have made it truly interesting. :)
 
Hi,

Those numbers are all but meaningless but a well driven Stax SR-007mk1 (not the Mk2 since the baffle isn't sealed) will go very low. Stax aren't alone in giving out such meaningless measurements

Yes, I agree, Stax can go quite low, compared to many other headphones.

Shame all the numbers have to be made up and ridiculous, just because "everyone does it".

And I'd add that Stax Headphones with modified tube driver (for more SPL and for crossfeed to get the musicians out of the head) are the only 'can's I tolerate for long periods.

I generally have Speakers, the only 'can's I own right now are some Audio Technica noise-cancelling ones for my frequent airtravel... I might get another pair of Stax now, given that we have a new Baby...

Though as long as the Neighbour plays movies very loud and very late I doubt I'll be turning down my system any time soon... :p

Ciao T
 
Even my smallest panel produces low frequency's nearly flat to 40hz.
Although fuzzy and hard to read this a pink noise test at 1 meter at 6 watts input.
The first set of bold numbers on the bottom of the chart from the left marks 10hz.
The test was done with no eq and the dips in the graph are room reflections.
I am running 6kv of bais on this panel and have measured the diagphram clipping (slaping against the stators) and have not experienced any arc or burned holes in the diagphram with this panel ,with a d/s spacing of .065".
the second set of pics was done the same way except that it was a nearfield measurement.
And as you can see I'm not using some cheezy $2 condenser microphone from radio shack. jer
 

Attachments

  • single panel setup.jpg
    single panel setup.jpg
    65.3 KB · Views: 244
  • single panel pink noise response.jpg
    single panel pink noise response.jpg
    36.3 KB · Views: 239
  • open setup(new).jpg
    open setup(new).jpg
    443.3 KB · Views: 233
  • open no eq.jpg
    open no eq.jpg
    866.5 KB · Views: 233
  • new eq'ed ,+6db  200hz,+9db 12khz.jpg
    new eq'ed ,+6db 200hz,+9db 12khz.jpg
    942.6 KB · Views: 217
I agree with you on your points as I am trying to figure out where the confusion has started.
The test that I have just posted were done with a transformer.
I'm sure we need not have to get into transformer theory as to how to get to 20 hz.
Sure esl's can get down to 20hz but it is not going to be a chest pounding 20hz by any means. jer
 
Thank you ThorstenL and geraldfryjr for such long and detailed answer!
The summary I get is that:
- Stax did not mension the FR is at what dB crititier, although it can be sure it is full-range class.
- Headphone, because they are near the ear, has the effect that help to enhance the bass.
- Headphone does not suffer from dipole effect, which helps the bass.
- bad transformer might cause some LF roll-off

ThorstenL:
you mensioned about direct drive have sufficient voltage to go load enough.
May I know what is the set-up, including RMS voltage, biasing voltage, diaphragm size/thickness and insulation thickness?
I'm certainly interested in direct drive, and avoide the X-former if possible.
I'm have some Semisouth 1.7kV silicon carbide power JFET on hand, it might be able to output 1kV RMS (1.4kV pp) if driven in balanced mode, I hope it will be sufficient.
 
I have stated before that with the advent of todays HV solid state devices that one could easily build a pair of direct drive amplifiers for much less than what one transformer would cost.
I have a design that I created in circuitmaker that simulates with out any errors using a basic common amplifier topology and cascading devices to handle the higher voltages using npn horz. output transistors for the output.
It does 800vp-p and can easily be scaled up for more voltage swing,with one on each side of the panel would of course double the voltage to 1600vp-p.
with a +- 575v supply.
I am planning on revamping the circuit to use HV fet's on a much higher supply voltage.
Even though I have not actualy built it yet, the fact that it simulates without errors is a good sign.
I have built a low voltage version of Neil S. Mckean's design which is based on david hermeyer's tube design to learn how to stack fet's.
It worked.
I used four irf510's on a 180v supply with a single TL082 driving the stack with no issues.
Only I haven't applied it to an esl yet because I didn't have any working panels or bias supply at the time,this was two years ago.
But it did drive a little piezo drive quite nicley.
I came with this concept four years ahead of this thread,http://www.diyaudio.com/forums/plan...age-type-transistors-directly-drive-esls.html
The circuit I designed used 4 IRFbg30's stacked using a MOT for 3kv supply with an 8k load resistor.
Trying to find and/or the cost of building 4 900watt 8Kohm resistors is what has kept me from building it.
And trying to discuss this subject on DIYaudio has been proven to be a bit touchy for some. jer
 
Hi,

you mensioned about direct drive have sufficient voltage to go load enough. May I know what is the set-up, including RMS voltage, biasing voltage, diaphragm size/thickness and insulation thickness?

I do not have data on the Panels, as I did not design them. My friend derived most of the designs form information in Roger Sanders Book.

The Amplifiers where 845 Push-Pull (1200V HT, around 1200V RMS output) for the midrange panel and 300B Push-Pull (400V HT, around 400V RMS output) for the high frequency panel. No idea on Bias, but likely a lot more than the peak signal value.

The HF panel was intended for use above 4 or 5KHz IIRC and had a very narrow state/film spacing. The Midrange panel had a wider spacing, broadly in line with many commercial ESL's.

The system was completed with a pair of dipole dynamic woofer systems using a pair of 18" Pro Audio woofers per side and 500W RMS Amplifiers.

All equalisation and crossover was line level passive.

Quite an impressive system and capable of quite high SPL's.

Ciao T
 
I had electrostatic headphones about 20 years ago. Best sound I ever heard! When I listen to AM radio stations over the Internet now I can hear the (sub-sonic) 20Hz signal between the program and comercials with cheap full size headphones because they are headphones! Speakers drive air, headphones drive eardrums. No can splain it simpler.

Jim
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.