• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Stax Tube Amplifier Distortion vs Solid State

JustAnandaDourEyedDude

Addicted to Fun and Learning
Joined
Apr 29, 2020
Messages
518
Likes
819
Location
USA
My chart was 4 ohms. the (I think) "standard" 2.83V signal yields 2 watts at the speaker via increased current.

If my chart is 8 ohms, it shifts a row, with 1 watt as the power at the 2.83V sensitivity..

View attachment 102004

It's not meant to be anything more than a general example of "louder requires more power and much louder requires much more power".
@Degru (Edit: @trl) is right. The volts and ohms are immaterial if the power efficiency of each of the two speakers is quoted, rather than the voltage efficiency.
The following lines in Post #27 are inconsistent with one another:
Sensitivity 85
85 2.83 0.707 2

Edit: the power sensitivity of 85 dBSPL / 1 W / m is equivalent to a voltage sensitivity of 85 dBSPL / 2 V / m with a 4 ohm speaker.
 
Last edited:
  • Like
Reactions: trl

Degru

Active Member
Joined
Feb 19, 2019
Messages
230
Likes
254
Location
Beaverton, OR
@Degru is right. The volts and ohms are immaterial if the power efficiency of each of the two speakers is quoted, rather than the voltage efficiency.
The following lines in Post #27 are inconsistent with one another:
Sensitivity 85
85 2.83 0.707 2

Edit: the power sensitivity of 85 dBSPL / 1 W / m is equivalent to a voltage sensitivity of 85 dBSPL / 2 V / m with a 4 ohm speaker.
Right about what? I don't recall partaking in that discussion specifically :p
 

JustAnandaDourEyedDude

Addicted to Fun and Learning
Joined
Apr 29, 2020
Messages
518
Likes
819
Location
USA

Rottmannash

Major Contributor
Forum Donor
Joined
Nov 11, 2020
Messages
2,968
Likes
2,604
Location
Nashville
Your meter is not fast acting enough to give you proper peak rates. And it is not calibrated to boot. Really, there is no correlation between what you think the SPLs are and what I am reporting with calibrated instruments.

I remember being at an audio show where Andrew Jones was demoing his speakers. Someone asked him what level he was playing at. He asked the audience to provide guesses. People were throwing numbers like 85 or 90 dB. At the end he said that the peaks are likely over 105 dB or something like that.

As I said, there seems to be massive confusion regarding what levels you all listen to. Either that you are not representing the rest of us who using the power and dynamics we have to get realistic sound levels.
So you're saying those free dB meter phone apps aren't state of the art, calibrated instruments? And the tooth fairy isn't real either? The horrors never cease...
 
OP
amirm

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
44,368
Likes
234,381
Location
Seattle Area
Here is another measurement this time at 20 Hz:

Stax SRM-007t tube amplifier versus SRM-313 Solid State 20 Hz Distortion Measurements.png


We are in much worse shape now. The tube amp can't get past 94.7 dBSPL. The solid state ekes out a few more to the tune of 97 dB -- far cry from the 112 dB we were showing for 1 kHz tone. Since highest amplitude in our music is going to be at these low frequencies, you will absolutely be limited by these amps when it comes to distortion. Given the threshold of hearing of 73 dBSPL, we need every dB we can get which the solid state produces.

As I said and pleaded, please, please don't throw out single number SPLs. They have no value when it comes to determining loudness.
 

pma

Major Contributor
Joined
Feb 23, 2019
Messages
4,591
Likes
10,727
Location
Prague
Is there a difference in clipping voltage at electrical side between 20Hz and 1kHz, or is it just because of FR decay at acoustic side? Curious, I will try Senn HD 598.
----------------

P.S.: I made the measurement with HD598, however it was pointless at 20Hz, for the reason of extreme distortion. The measurement on acoustical side has made some sense at 100Hz. Electrical side was measured at 50Hz and 100Hz, acoustical side only at 100Hz. There is no clipping or dramatic rise of distortion on electrical side. 1Vrms according to company specs means 112dB SPL/1V/1kHz. I am ending about 2V not to damage the transducers.

It seems that the dynamic headphone + amp have more predictable behaviour than the electrostats.

HD598_el_ac_100Hz_distLs.png
 
Last edited:

Degru

Active Member
Joined
Feb 19, 2019
Messages
230
Likes
254
Location
Beaverton, OR
Is there a difference in clipping voltage at electrical side between 20Hz and 1kHz, or is it just because of FR decay at acoustic side? Curious, I will try Senn HD 598.
----------------

P.S.: I made the measurement with HD598, however it was pointless at 20Hz, for the reason of extreme distortion. The measurement on acoustical side has made some sense at 100Hz. Electrical side was measured at 50Hz and 100Hz, acoustical side only at 100Hz. There is no clipping or dramatic rise of distortion on electrical side. 1Vrms according to company specs means 112dB SPL/1V/1kHz. I am ending about 2V not to damage the transducers.

It seems that the dynamic headphone + amp have more predictable behaviour than the electrostats.

View attachment 102182
303 should have flat bass extension if sealed properly, though there will be severe rolloff if not sealed. However, poor seal would also mean horrendously bad distortion (much higher than seen here) so I don't think that is the case and we are just seeing the distortion of the amp. For reference the driver should be capable of below 0.5% at 95db at 20hz. If the measurement was set up correctly then this is one dirty amp lol.
 
  • Like
Reactions: pma

Degru

Active Member
Joined
Feb 19, 2019
Messages
230
Likes
254
Location
Beaverton, OR
Here is another measurement this time at 20 Hz:

View attachment 102173

We are in much worse shape now. The tube amp can't get past 94.7 dBSPL. The solid state ekes out a few more to the tune of 97 dB -- far cry from the 112 dB we were showing for 1 kHz tone. Since highest amplitude in our music is going to be at these low frequencies, you will absolutely be limited by these amps when it comes to distortion. Given the threshold of hearing of 73 dBSPL, we need every dB we can get which the solid state produces.

As I said and pleaded, please, please don't throw out single number SPLs. They have no value when it comes to determining loudness.
Thanks for measuring that. I found it pretty easy to make my SRM-3 hard clip with bass transients, and looks like the 313 and 007t are no exception.
 

PeteL

Major Contributor
Joined
Jun 1, 2020
Messages
3,303
Likes
3,838
Here is another measurement this time at 20 Hz:

View attachment 102173

We are in much worse shape now. The tube amp can't get past 94.7 dBSPL. The solid state ekes out a few more to the tune of 97 dB -- far cry from the 112 dB we were showing for 1 kHz tone. Since highest amplitude in our music is going to be at these low frequencies, you will absolutely be limited by these amps when it comes to distortion. Given the threshold of hearing of 73 dBSPL, we need every dB we can get which the solid state produces.

As I said and pleaded, please, please don't throw out single number SPLs. They have no value when it comes to determining loudness.
May I ask where how you assert this treshold of earing at 20 Hz? I assumed it's based on Fletcher Munson Curve curve, but I wonder how we come up with such a precise (73 dB SPL) number, Is it the treshold of earing for the best human ears out there, or an average human ear? Is it based on experimental studies, data from audiologists, or something arithmetically determined. I'm genuinely curious on this subject, if you could give us some background, reading material. To be clear I am certainly not disputing frequency weighting, but always tought "treshold of hearing" and "treshold of pain" slightly arbitrary. Wondering if anyone, and in what conditions someone can actually hear 1 dB SPL at 1k (or 73 dBs at 20). I notice that it's well documented that we lose some of the highest frequencies with age or other factors, but what about levels, does this 0 dB SPL A weighted have an actual real meaning. I'm quite certain my own treshold is higher than this, and frequency wise I've proven it to be quite good.
 
Last edited:

Francis Vaughan

Addicted to Fun and Learning
Forum Donor
Joined
Dec 6, 2018
Messages
933
Likes
4,697
Location
Adelaide Australia
The tube amp can't get past 94.7 dBSPL. The solid state ekes out a few more to the tune of 97 dB
Hmm, that is most interesting. The gap in clipping levels is pretty consistent with what one would expect due to the tube's inherent voltage drop and the circuit's inability to swing as far as the SS design can. Not that either are stellar. Looks like the headphone's FR has dropped off a cliff. Looking at one test result, the SR-303 is 30dB down at 20Hz. So, if anything, these results are a bit better. (Which is faint praise to say the least.) 20db down. EQ'ing done at these frequencies isn't going to be a starter for these headphones.

One notices that the newer Stax amplifiers claim to swing 450 volts. The first measurements above suggest the 313 seems to swing about 320, the 007t maybe 250 - at best. No doubt one is going to need big swings to get the best out of these headphones. Even 450 may be a bit anaemic. Claws back a bit of headroom. But getting the inherent bass response of the headphones up is really the only cure.

The difference in clip points is 2.3dB = a voltage difference of 1.3. So, SRM-313 at 320volts :: 320 / 1.3 = 245 volts for the 007t. Again, consistent answer with above. However the specs say the 007t can swing 340 volts. I really doubt that. The SRM-313 spec says it gets 350 volts. It seems to get closer, which I would expect. In both cases the quoted spec seems an unattainable best case.

At 20Hz the headphones will present as near an open circuit as one could measure. So this should be a best case in terms of voltage swing. But worst case for frequency response. Testing at say 10kHz should be a very different question. There the load is becoming significant, and I would expect the 007t to fare quite poorly compared to the 313. Indeed I would expect it to deliver awful THD, saved only by the fact that the harmonics will be inaudible.

I assume the SR-303 is plugged into a "pro" high bias voltage socket (580v) , and not the low (230v) socket.
 

the_brunx

Senior Member
Joined
Apr 7, 2020
Messages
341
Likes
859
This is why people have made tube amps with much more voltage swing and greater linearity like the Blue Hawaii and the D.I.Y. T2. But the Stax SR-T1 and SR-707 can be improved substantially - adding constant current loads instead of plate resistors helps quite a lot, and some have gone to the trouble of replacing the power transformer for one with greater capacity and changing to a different tube type (6S4A whose plates can dissipate more than the stock 6FQ7.)

Listening at levels above about 85~90 dB will destroy your hearing, Amirm. The Stax amps under test seem to be able to reach these levels, or close to them (albeit with higher levels of distortion than is deemed "HiFi.") Why would anyone want an amplifier that is capable of ruining one's hearing?

i think you need to read previous posts.
 

pma

Major Contributor
Joined
Feb 23, 2019
Messages
4,591
Likes
10,727
Location
Prague
Hmm, that is most interesting. The gap in clipping levels is pretty consistent with what one would expect due to the tube's inherent voltage drop and the circuit's inability to swing as far as the SS design can. Not that either are stellar. Looks like the headphone's FR has dropped off a cliff.

IMO the only way to discover what has happened is to measure distortion at amplifier voltage output and to find the point of clipping. I think both electrical and acoustical measurements should have been done simultaneously always. It indicates to possible method issues.
 

solderdude

Grand Contributor
Joined
Jul 21, 2018
Messages
15,891
Likes
35,912
Location
The Neitherlands
P.S.: I made the measurement with HD598, however it was pointless at 20Hz, for the reason of extreme distortion. The measurement on acoustical side has made some sense at 100Hz.

Is this done with a current drive amp or voltage drive amp ?
 

Francis Vaughan

Addicted to Fun and Learning
Forum Donor
Joined
Dec 6, 2018
Messages
933
Likes
4,697
Location
Adelaide Australia
IMO the only way to discover what has happened is to measure distortion at amplifier voltage output and to find the point of clipping. I think both electrical and acoustical measurements should have been done simultaneously always. It indicates to possible method issues.
To some extent that is where we came in. An electrostat presents pretty much a pure capacitive load. At lower frequencies it looks more like an open circuit than much else. Measuring performance, and measuring with accuracy whilst not disturbing the operation is a tall ask. The tube based amplifiers are really not robust enough to survive even a moderate load. The SS amps are easier. Much lower output impedance and robust output stage. Measuring the SRM-313 driving the SR-303 would probably produce useful results. The SRM-007t would, I suspect, just collapse unless heroic efforts are made. The designers have taken significant advantage of the nature of the load presented by an electrostatic headphone to allow the design to work at all.

Knowledge of the nature of the load and the internal circuitry of the amplifiers does allow us to gain a reasonably clear idea of what is going on. It is pretty clear that the amplifiers are being driven into clipping, and the numbers are consistent with this, and even the differences in numbers are consistent. Had there been some disparity, there would be places to question, but overall I think we can say that these measurements are exactly what one would expect. To a large extent they show up deficiencies in the headphone. There is no reason to imagine that the amplifiers themselves are in the least bit worried by being asked to reproduce 20Hz. But asking them to deliver the massive swings needed to get the headphone to reproduce 20Hz at any useful volume is a just too much. Indeed I doubt there is an amplifier made that could. And even if one could, there is little chance it could do so without the headphones madly arcing and destroying themselves.
 

pma

Major Contributor
Joined
Feb 23, 2019
Messages
4,591
Likes
10,727
Location
Prague
@Francis Vaughan , we are able to measure very accurately with charge-like meauring amplifiers and probes with both very high input resistance (100 Meg) and low capacitance (10 pF). I am absolutely sure that the task has technical solution.
 

John_M

Member
Forum Donor
Joined
Jun 17, 2020
Messages
62
Likes
51
Thanks for doing these reviews - hope to see more electrostatic amps/ headphones reviewed. I would send mine but I live in the UK!
 

Ron Texas

Master Contributor
Joined
Jun 10, 2018
Messages
6,074
Likes
8,908
Proof positive that tubes really sound different. Some out there must like the sound of distortion.
 

Degru

Active Member
Joined
Feb 19, 2019
Messages
230
Likes
254
Location
Beaverton, OR
Looks like the headphone's FR has dropped off a cliff. Looking at one test result, the SR-303 is 30dB down at 20Hz. So, if anything, these results are a bit better. (Which is faint praise to say the least.) 20db down.
That measurement has bad seal; the bass should extend flat on these headphones. The pads are shaped to fit a human head without having to clamp much, which makes sealing on an artificial head problematic. Plus, with bad seal and rolloff like that the harmonic distortion goes absolutely through the roof (20-50% or higher at much lower volumes) so this is not the case for Amir's measurements.

> I assume the SR-303 is plugged into a "pro" high bias voltage socket (580v) , and not the low (230v) socket.

If it was plugged into normal bias, Amir would have a dead lambda on his hands right now :p You can really only run it at extreme low (barely audible) volumes on normal bias without too much risk of permanent damage. I know someone who got impatient waiting for a piece of gear to arrive and plugged a Lambda Signature into normal bias, and ended up damaging it after a few minutes of low volume listening. Squealing, imbalance, bass completely messed up sounding, basically totally broken.
 
Last edited:

Degru

Active Member
Joined
Feb 19, 2019
Messages
230
Likes
254
Location
Beaverton, OR
To some extent that is where we came in. An electrostat presents pretty much a pure capacitive load. At lower frequencies it looks more like an open circuit than much else. Measuring performance, and measuring with accuracy whilst not disturbing the operation is a tall ask. The tube based amplifiers are really not robust enough to survive even a moderate load. The SS amps are easier. Much lower output impedance and robust output stage. Measuring the SRM-313 driving the SR-303 would probably produce useful results. The SRM-007t would, I suspect, just collapse unless heroic efforts are made. The designers have taken significant advantage of the nature of the load presented by an electrostatic headphone to allow the design to work at all.

Knowledge of the nature of the load and the internal circuitry of the amplifiers does allow us to gain a reasonably clear idea of what is going on. It is pretty clear that the amplifiers are being driven into clipping, and the numbers are consistent with this, and even the differences in numbers are consistent. Had there been some disparity, there would be places to question, but overall I think we can say that these measurements are exactly what one would expect. To a large extent they show up deficiencies in the headphone. There is no reason to imagine that the amplifiers themselves are in the least bit worried by being asked to reproduce 20Hz. But asking them to deliver the massive swings needed to get the headphone to reproduce 20Hz at any useful volume is a just too much. Indeed I doubt there is an amplifier made that could. And even if one could, there is little chance it could do so without the headphones madly arcing and destroying themselves.
And here is where transformer setups come in to save the day :p

Stats should be able to handle levels up to twice their bias voltage without being destroyed right away, though of course wear and tear on the driver is not out of the question at those levels, and that's assuming the headphone is on your head and not unsealed. It's safe to say any non-suicidal listening volumes (even Amir's) shouldn't damage the headphone itself, though may stretch the diaphragm over time a bit.
 
Top Bottom