Building a 3-way TL speaker

[Hari Iyer]

I could finally complete this project today and have installed the crossover components in a separate box with the Mundorf capacitors. I have just measured the impedance of the speakers to confirm if it agree with my simulated response and they do. Listened for an hour or so and they sound quite nice. Have invited some Mumbai based FMs for a listen over the weekend and some of them have tentatively confirmed visit. Will measure other parameters later on the weekend as i dont want my listening to get baised by what i measured.

 

[arunkvivek]

exciting weekend ahead.. the wood work is gorgeous.

 

[Hari Iyer]

I listened for a couple of hours and noticed that the sound stage was more warmer than my previous built with the same component values. The only change i made was with the speaker cables and adding Mundorf capacitors. I then measured to check why and found the FR showed a peak around 4KHz and after that a gradual dip in response till 20KHz. This was making the vocals too pronounced masking the lows a lot making them to loose detail. You can notice that in the minimum phase response too.



I then did a quick math and added a parallel notch on the tweeter to bring the level back and measured again to get the below response



The minimum phase is also around 0 deg till around 15KHz after which it goes towards 180 deg.
I would have loved to measure this speaker in an anechoic room but dont have that luxury.

 

[rikhav]

I am no expert but wouldn't it make sense to let the new capacitors play for around 20 to 30 hours and then make adjustments?

 

[Hari Iyer]

I am no expert but wouldn't it make sense to let the new capacitors play for around 20 to 30 hours and then make adjustments?

Me too no expert in this break-in areas for any components. I think after break-in for 20 to 30 hours the measured results still remain the same and you may notice only subjective improvements or your ears may get adjusted to the new sound presentation. I will leave them now as it is for the next few days / weeks and take inputs from FMs who are planning a listen of these speakers. Also my listening is not yet very well developed and am still learning subjective listening techniques. Hence i have to rely more on the objective measurements as i am quite ok in that area.

 

[mpw]

Best way to listen is to keep silent...

An old saying ...

 

[kapvin]

Best way to listen is to keep silent...

An old saying ...

trust @mpw to come up with an appropriate homily. :-D

 

[Hari Iyer]

Best way to listen is to keep silent...

An old saying ...

can you elaborate? I am no philosopher.

 

[ONEMANARMY]

Best way to listen is to keep silent...

Look who’s talking

 

[Hiten]

I would have loved to measure this speaker in an anechoic room but dont have that luxury.

I think it can be done in open ground when it is silent (Early morning).
regards.

 

[newlash09]

I think it can be done in open ground when it is silent (Early morning).
regards.

I guess it will need a lot of forum members to help here, as each speaker weighs more than 100 kgs. So that's atleast 200 kgs to lug to the morning park and back

 

[Hiten]

 

[Hari Iyer]

After somewhat settling on the freq response in post #87 i wanted to work on the time domain aspect of the speaker. The step response of this FR plot is as below. Those who have been following my post know that i am a measurement freak and i trust my mic more than my ears. Unless i can objectively co-relate my subjective findings, i wont know where to look for a tweak. It will be more like shooting arrows in the dark with trial and errors.



They look a bit disconnected at the midrange and the woofer with the midrange hovering a little longer and the woofer unable to keep pace with that. Though subjectively they sounded quite ok to me, it had a feeling like the singer and the musicians wanted to perform quickly and leave the scene. The bass punch was getting masked by this stronger midrange. I measured the delay in the impulse response and noted that it was around 18usec. I had a tough time figuring out which driver was faster ( or ahead) of the other. Physically the mids and the woofer were ahead of the tweeter and i had taken some effort to align them in time, but yet there was a 18usec (-6mm) delay between their individual arrivals.

This is the freq response with adding only the delay on the midrange driver -



I then figured out that it was the midrange that require to slow down and needed some kind of a delay or a speed breaker to be added. I added passive components (inductor and resistor) on the midrange and noted that the delay has gone from 18usec(-6mm) to 12usec(-4mm). Getting aggressive on the components was causing the FR to get affected and i had to stop there. But some delay was required to slow down the midrange.

During my conversation with @prem couple of months ago, i recalled him mentioning about silver wires not be used for speaker wires blindly as they are more used as a band-aid for response issues. By using these words of wisdom, i thought why not increase the speed of the tweeter by adding silver wire on them and retaining copper on the midrange and the woofer. I did and measured again but again found the same delay of 12usec with nothing happening to the impulse response.

I later researched on speed of electrons in various metals and discovered that the drift velocity of electrons varies in different metals and the cross-section area, density of the material and molecular weight has significant effect on its speed. I also discovered that drift velocity is inversely proportional to the cross-section area. The mateial which came to my mind was aluminium and i did have some 4 sqmm thick solid aluminium wires with me for experiment. I added initially one meter of wire to the midrange speaker in the pretext of adding delay and measured the impulse response and the delay moved from -4mm to +19mm. So the trick worked. I later reduced the wire length from 1 meter to around 1 feet and measured again to get a perfect delay of 0mm or 0sec.

I was amazed with my measurement of how much a simple material type can add or subtract velocity of different frequencies. So blindly using any cables or wires for your amplifier power supply / speaker cables wont make sense and it has to be for the specific speed that you require. Sometime even a simple cheap aluminium wire could be good enough in some case.

This is the FR and the step response after both the delay ckt and the aluminium wire on the midrange, silver wire on the tweeter and copper wire on the woofer,




In this step response each driver picks-up from where the other leaves. The woofer pick-up is where the midrange leaves.

So how do they sound subjectively. The singers and musicians sounded much relaxed and not in a hurry and wanted to stay back for more.
Well its for others to listen and decide if they agree with the objective measurements.

 

[arunkvivek]

That's an interesting way to achieve time delay! I would love to see some listening experiences of the new setup.

 

[Spinekar]

After somewhat settling on the freq response in post #87 i wanted to work on the time domain aspect of the speaker. The step response of this FR plot is as below. Those who have been following my post know that i am a measurement freak and i trust my mic more than my ears. Unless i can objectively co-relate my subjective findings, i wont know where to look for a tweak. It will be more like shooting arrows in the dark with trial and errors.

View attachment 44191

They look a bit disconnected at the midrange and the woofer with the midrange hovering a little longer and the woofer unable to keep pace with that. Though subjectively they sounded quite ok to me, it had a feeling like the singer and the musicians wanted to perform quickly and leave the scene. The bass punch was getting masked by this stronger midrange. I measured the delay in the impulse response and noted that it was around 18usec. I had a tough time figuring out which driver was faster ( or ahead) of the other. Physically the mids and the woofer were ahead of the tweeter and i had taken some effort to align them in time, but yet there was a 18usec (-6mm) delay between their individual arrivals.

This is the freq response with adding only the delay on the midrange driver -
View attachment 44192


I then figured out that it was the midrange that require to slow down and needed some kind of a delay or a speed breaker to be added. I added passive components (inductor and resistor) on the midrange and noted that the delay has gone from 18usec(-6mm) to 12usec(-4mm). Getting aggressive on the components was causing the FR to get affected and i had to stop there. But some delay was required to slow down the midrange.

During my conversation with @prem couple of months ago, i recalled him mentioning about silver wires not be used for speaker wires blindly as they are more used as a band-aid for response issues. By using these words of wisdom, i thought why not increase the speed of the tweeter by adding silver wire on them and retaining copper on the midrange and the woofer. I did and measured again but again found the same delay of 12usec with nothing happening to the impulse response.

I later researched on speed of electrons in various metals and discovered that the drift velocity of electrons varies in different metals and the cross-section area, density of the material and molecular weight has significant effect on its speed. I also discovered that drift velocity is inversely proportional to the cross-section area. The mateial which came to my mind was aluminium and i did have some 4 sqmm thick solid aluminium wires with me for experiment. I added initially one meter of wire to the midrange speaker in the pretext of adding delay and measured the impulse response and the delay moved from -4mm to +19mm. So the trick worked. I later reduced the wire length from 1 meter to around 1 feet and measured again to get a perfect delay of 0mm or 0sec.

I was amazed with my measurement of how much a simple material type can add or subtract velocity of different frequencies. So blindly using any cables or wires for your amplifier power supply / speaker cables wont make sense and it has to be for the specific speed that you require. Sometime even a simple cheap aluminium wire could be good enough in some case.

This is the FR and the step response after both the delay ckt and the aluminium wire on the midrange, silver wire on the tweeter and copper wire on the woofer,
View attachment 44193

View attachment 44194

In this step response each driver picks-up from where the other leaves. The woofer pick-up is where the midrange leaves.

So how do they sound subjectively. The singers and musicians sounded much relaxed and not in a hurry and wanted to stay back for more.
Well its for others to listen and decide if they agree with the objective measurements.

Just reading this is fun
keep it up

 

[kapvin]

Hi Hari,

the maths of this escapes me. (and the physics, which, of course is derived maths )

you reduced the length of the wire by 2.28ft (1m to 1 ft) and your delay went for +19mm to 0mm. At 346m/s (speed of sound at 25c) - thats 0.00005sec.

1st question - how can you be sure that this is not a measurement artifact?
2nd question- given that temperature will impact conductance, what are you doing to ensure that this repeatable at different atmospheric conditions?
3rd question - i am assuming that this measurement is on design axis. as you go off axis,(either horizontally or vertically) time of flight from each driver to listener changes, and therefore so how do you plan to compensate for that? so even if point 1 and 2 are demonstrably addressed, point 3 is going to be a wrinkle. My point is, moving your mic a few cm, left/right or up/down will undo all the work you doing.

i have been puzzling about multidriver alignment and how to get consistent results over more than one listening location. physical offset seems to be the only reliable way to do this on the horizontal plane. or of course a point source.

let me know your thoughts.

After somewhat settling on the freq response in post #87 i wanted to work on the time domain aspect of the speaker. The step response of this FR plot is as below. Those who have been following my post know that i am a measurement freak and i trust my mic more than my ears. Unless i can objectively co-relate my subjective findings, i wont know where to look for a tweak. It will be more like shooting arrows in the dark with trial and errors.

View attachment 44191

They look a bit disconnected at the midrange and the woofer with the midrange hovering a little longer and the woofer unable to keep pace with that. Though subjectively they sounded quite ok to me, it had a feeling like the singer and the musicians wanted to perform quickly and leave the scene. The bass punch was getting masked by this stronger midrange. I measured the delay in the impulse response and noted that it was around 18usec. I had a tough time figuring out which driver was faster ( or ahead) of the other. Physically the mids and the woofer were ahead of the tweeter and i had taken some effort to align them in time, but yet there was a 18usec (-6mm) delay between their individual arrivals.

This is the freq response with adding only the delay on the midrange driver -
View attachment 44192


I then figured out that it was the midrange that require to slow down and needed some kind of a delay or a speed breaker to be added. I added passive components (inductor and resistor) on the midrange and noted that the delay has gone from 18usec(-6mm) to 12usec(-4mm). Getting aggressive on the components was causing the FR to get affected and i had to stop there. But some delay was required to slow down the midrange.

During my conversation with @prem couple of months ago, i recalled him mentioning about silver wires not be used for speaker wires blindly as they are more used as a band-aid for response issues. By using these words of wisdom, i thought why not increase the speed of the tweeter by adding silver wire on them and retaining copper on the midrange and the woofer. I did and measured again but again found the same delay of 12usec with nothing happening to the impulse response.

I later researched on speed of electrons in various metals and discovered that the drift velocity of electrons varies in different metals and the cross-section area, density of the material and molecular weight has significant effect on its speed. I also discovered that drift velocity is inversely proportional to the cross-section area. The mateial which came to my mind was aluminium and i did have some 4 sqmm thick solid aluminium wires with me for experiment. I added initially one meter of wire to the midrange speaker in the pretext of adding delay and measured the impulse response and the delay moved from -4mm to +19mm. So the trick worked. I later reduced the wire length from 1 meter to around 1 feet and measured again to get a perfect delay of 0mm or 0sec.

I was amazed with my measurement of how much a simple material type can add or subtract velocity of different frequencies. So blindly using any cables or wires for your amplifier power supply / speaker cables wont make sense and it has to be for the specific speed that you require. Sometime even a simple cheap aluminium wire could be good enough in some case.

This is the FR and the step response after both the delay ckt and the aluminium wire on the midrange, silver wire on the tweeter and copper wire on the woofer,
View attachment 44193

View attachment 44194

In this step response each driver picks-up from where the other leaves. The woofer pick-up is where the midrange leaves.

So how do they sound subjectively. The singers and musicians sounded much relaxed and not in a hurry and wanted to stay back for more.
Well its for others to listen and decide if they agree with the objective measurements.

 

[Hari Iyer]

Hi Hari,

the maths of this escapes me. (and the physics, which, of course is derived maths )

you reduced the length of the wire by 2.28ft (1m to 1 ft) and your delay went for +19mm to 0mm. At 346m/s (speed of sound at 25c) - thats 0.00005sec.

1st question - how can you be sure that this is not a measurement artifact?
2nd question- given that temperature will impact conductance, what are you doing to ensure that this repeatable at different atmospheric conditions?
3rd question - i am assuming that this measurement is on design axis. as you go off axis,(either horizontally or vertically) time of flight from each driver to listener changes, and therefore so how do you plan to compensate for that? so even if point 1 and 2 are demonstrably addressed, point 3 is going to be a wrinkle. My point is, moving your mic a few cm, left/right or up/down will undo all the work you doing.

i have been puzzling about multidriver alignment and how to get consistent results over more than one listening location. physical offset seems to be the only reliable way to do this on the horizontal plane. or of course a point source.

let me know your thoughts.

1- To keep measurement artifact to bare minimum I did all these measurement after 11:30PM to 2:30pm to keep the mic and speakers at the same position without moving them even an um. So any artifacts would be because of other reasons beyond my control. I took many measurements ( more than 4) for each change so that it can be averaged out.

2-the wires are inside the boxes where the temperature changes will be minimum due to the isolation from the ambient.

3-ime most drivers integrate quite well in the horizontal axis SPL wise, that's why you don't get too much variation in level when you move in the horizontal axis. But most drivers where there are timing issues the SPL changes when you move in the vertical plane. If the level when you sit lower to the listening axis or in in the listening axis or when you stand up changes then it's an indication of timing issues or integration between drivers. The trick is to get them same at all listening positions in both the horizontal and vertical planes.
Time shifts causes phase errors which will increase sharpness in the highs and make the vocals very shrill asking you to reduce the volume to reduce the overlaps between drivers.

I have kept the crossover frequency in multiple of wavelength of each drivers so that any overlaps between them will add in-phase atleast at the crossover frequency.

One major point noted was earlier I needed to toe-in the speaker around 5 Deg to 10 deg to get perfect center imaging in my room. After applying the appropriate delay the toe-in is not required and the speakers are now parallel to the wall. Also the sweet spot has widened and I get the same center image in all positions in my room

 

[Hari Iyer]

I am still confused Hari.

Now is this new tweak / optimization will be temperature dependent, say when you are listening to the speakers in cold to hot conditions ? (consider extreme climate swings in some Indian states when even the ambient temperature differs a lot)

Further, please explain once again, what if the speakers are made to play in a different room (other than yours), will your final crossover tweaks still hold good ? (Or will they have to be tweaked again for optimization ? Consider new room reflections too)

Yogi Bear, I have not considered any temperature variations in my calculations because the wires are inside the boxes and not external. So I am not expecting any major temperature changes on the wire to impact any velocity.
What you are exploring is a theoretical concept which will require funding and the means for me to explore further. As it's said the proof of the pudding is in the eating no matter how perfect your objective measurements are.

 

[keith_correa]

I was amazed with my measurement of how much a simple material type can add or subtract velocity of different frequencies. So blindly using any cables or wires for your amplifier power supply / speaker cables wont make sense and it has to be for the specific speed that you require. Sometime even a simple cheap aluminium wire could be good enough in some case.

Hari can you explain how one determines what specific "speed" one wants or should want for power supplies and speaker cables?

Ultimately, it all boils down to conductance which is the reciprocal of resistance. So isn't a simple answer for using a conductor (for the areas aforementioned), the highest conductance aka the lowest resistance? Or am I missing something here?

 

[Hari Iyer]

Hari can you explain how one determines what specific "speed" one wants or should want for power supplies and speaker cables?

Ultimately, it all boils down to conductance which is the reciprocal of resistance. So isn't a simple answer for using a conductor (for the areas aforementioned), the highest conductance aka the lowest resistance? Or am I missing something here?

Keith, I am not aware if conductance or velocity can be used interchangeably in this case. The velocity is also inversely proportional to the cross-section area. ie. more the surface area less the velocity. Thinner wires have more velocity. I did not do any math to arrive at the correct length just cut the wire and measured again and again to come at this length. I am not sure if this can be duplicated in other setups as it was more of a trail and error method using measurements.