Dear All,
as requested by Kaffimann, deef, Vincula and Perceval, I started measuring the Xizi Fast 5 in their TML enclosures.
The TSP have already been posted, but here again, together with first REW measurements:
Loudspeaker freestanding in the room, 0,3 m distance, on axis, with an uncalibrated Superlux 999 micro and a Steinberg URL soundcard, both should be OK up to 10 kHz and possibly further, as per measurements from others.
Let´s see: both units vary for approximately 1 dB, that´s OK. Both units have trouble at around 10 Khz, clearly visible with the impedance sweeps, with the frequency sweeps, and in the distortion plots as well (k3 at 3,3 kHz).
The TML and especially the damping method seems to work quite well, as the third frequency sweep (main driver lightly shielded from micro, loudspeaker upside down) indicates.
Next will be nearfield measurements at angles.
All the best
Mattes
as requested by Kaffimann, deef, Vincula and Perceval, I started measuring the Xizi Fast 5 in their TML enclosures.
The TSP have already been posted, but here again, together with first REW measurements:
Loudspeaker freestanding in the room, 0,3 m distance, on axis, with an uncalibrated Superlux 999 micro and a Steinberg URL soundcard, both should be OK up to 10 kHz and possibly further, as per measurements from others.
Let´s see: both units vary for approximately 1 dB, that´s OK. Both units have trouble at around 10 Khz, clearly visible with the impedance sweeps, with the frequency sweeps, and in the distortion plots as well (k3 at 3,3 kHz).
The TML and especially the damping method seems to work quite well, as the third frequency sweep (main driver lightly shielded from micro, loudspeaker upside down) indicates.
Next will be nearfield measurements at angles.
All the best
Mattes
Hi,
please find below the 0° to 60° sweeps in the nearfield.
It looks like the problem at 3.something kHz could be diffraction at the enclosures edges, but it is obvious that the drivers have a problem at nearly 10 kHz.
In how far this should bother the user and if there are any measures to control the problem area (I can hear all DSP users already, but I myself will stay passive...), will become more clear after farfield measurements in the listening room (that´s how I measured with the NEUTRIK frame 30 years ago, and how I´m measuring today as well...).
In the end, our ears shall be the final and most important judgement.
More to follow, but this evening old friends came to visit me, no further efforts today, need to have a glass of red wine and need to listen to Karl Seglem with the black flowers...
All the best
Mattes
please find below the 0° to 60° sweeps in the nearfield.
It looks like the problem at 3.something kHz could be diffraction at the enclosures edges, but it is obvious that the drivers have a problem at nearly 10 kHz.
In how far this should bother the user and if there are any measures to control the problem area (I can hear all DSP users already, but I myself will stay passive...), will become more clear after farfield measurements in the listening room (that´s how I measured with the NEUTRIK frame 30 years ago, and how I´m measuring today as well...).
In the end, our ears shall be the final and most important judgement.
More to follow, but this evening old friends came to visit me, no further efforts today, need to have a glass of red wine and need to listen to Karl Seglem with the black flowers...
All the best
Mattes
Attachments
Hi Perceval and 2stuffed,
don´t be afraid of these first measurements in the nearfield, these drivers sound better than measurements suggest. Furthermore, those measurements are nothing uncommon with other fullrange drivers of equal built - light paper membrane - look at several Fostex, Lowther or even original Coral Beta 8 measurements, they´re not looking any better.
Of course, in a state like this (unfiltered) and in an enclosure like this they will hurt Perceval´s ears and lead 2stuffed to the opinion those don´t sound good.
In my opinion, it is important to work out a correction filter for nearly every fullrange driver, I tried several times without filter and gave up this approach a few decades ago...
2stuffed, have you tried to build a BSC and contour filter? What types of enclosures have you realized? I will show the way to a nicely working filter later, and I think it sounds quite OK, albeight not perfect... but after all, a perfect speaker is very hard to achieve...
After all, this is a relatively cheap chinese driver, and of course there are alternatives available, the TB 5-inchers or, if you´re going the connoisseur route, the FERTIN 13 WLA (of which I have a pair, but haven´t yet done anything with them. Maybe I´ll make a small OB with 2 10-inch bass units... but not now).
Perceval, this would be my advise:
Ossicles | HIGHRESAUDIO
All the best
Mattes
don´t be afraid of these first measurements in the nearfield, these drivers sound better than measurements suggest. Furthermore, those measurements are nothing uncommon with other fullrange drivers of equal built - light paper membrane - look at several Fostex, Lowther or even original Coral Beta 8 measurements, they´re not looking any better.
Of course, in a state like this (unfiltered) and in an enclosure like this they will hurt Perceval´s ears and lead 2stuffed to the opinion those don´t sound good.
In my opinion, it is important to work out a correction filter for nearly every fullrange driver, I tried several times without filter and gave up this approach a few decades ago...
2stuffed, have you tried to build a BSC and contour filter? What types of enclosures have you realized? I will show the way to a nicely working filter later, and I think it sounds quite OK, albeight not perfect... but after all, a perfect speaker is very hard to achieve...
After all, this is a relatively cheap chinese driver, and of course there are alternatives available, the TB 5-inchers or, if you´re going the connoisseur route, the FERTIN 13 WLA (of which I have a pair, but haven´t yet done anything with them. Maybe I´ll make a small OB with 2 10-inch bass units... but not now).
Perceval, this would be my advise:
Ossicles | HIGHRESAUDIO
All the best
Mattes
That sounds well recorded... I will investigate. Thanks!
All my FR drivers have needed some work to sound great... and the TB W8-1772 have needed the most! They were hurting my ears when I first heard them!
But, when tamed in all the right places, they sound great.
It's funny, but the best FR I have heard that sounded good without corrections, were quite cheap... for example, the TC9FD, and another cheap bamboo paper looking driver that was almost flat from 300Hz to 12kHz. and only cost $20. Then again, my Visaton B80 were not cheap and sounded pretty good, and only needed a small tweak.
The Fertins do look the part! I'll see if a trip to France is in the works ... it would be cool to pick up a pair!
Right now, I'm having fun with the new TB Coax, and I have some SBAcoustics drivers on the way that should be fun as well!
Exciting times!
Hi Perceval,
should you really someday be interested in the FERTIN drivers, I might help, as I´m in friendly contact with the artist, errrr, manufacturer.
You´re right, all FR drivers need some time.
I have the large TB as well, but the W-1808. This is, in my humble opinion, the best driver in it´s price range, and if you want to surpass it (or make a sidestep...), you´ll have to pay a lot more.
Of course I´m following your coax adventures...
Have fun!
Mattes
should you really someday be interested in the FERTIN drivers, I might help, as I´m in friendly contact with the artist, errrr, manufacturer.
You´re right, all FR drivers need some time.
I have the large TB as well, but the W-1808. This is, in my humble opinion, the best driver in it´s price range, and if you want to surpass it (or make a sidestep...), you´ll have to pay a lot more.
Of course I´m following your coax adventures...
Have fun!
Mattes
Dear All,
let´s continue with some work in order to get the Fast 5 in a state where it´s listenable.
First the BSC, after some attempts, 1,2 mH and 5,6 Ohm seems to be about right.
Normally, a small cap is also needed to turn this into a contour filter and to avoid too much treble attenuation, but it seems this is not needed here, we´ll see later.
All the best
Mattes
let´s continue with some work in order to get the Fast 5 in a state where it´s listenable.
First the BSC, after some attempts, 1,2 mH and 5,6 Ohm seems to be about right.
Normally, a small cap is also needed to turn this into a contour filter and to avoid too much treble attenuation, but it seems this is not needed here, we´ll see later.
All the best
Mattes
Perceval said... It's funny, but the best FR I have heard that sounded good without corrections, were quite cheap... for example, the TC9FD...
Agreed. One of my favorite full range drivers was the cheap Pioneer B20. As for the TC9, I’ve used it similarly to the B20 - by simply adding bass and treble support.
Agreed. One of my favorite full range drivers was the cheap Pioneer B20. As for the TC9, I’ve used it similarly to the B20 - by simply adding bass and treble support.
Dear All,
next step: a second parallel notch takes care about the 9 kHz spike. Values are 1,8 Ohm, 2,0 uF and 0,19 mH. The region around the spike is also slightly attenuated, that´s how it is...
Some people prefer serial LCRs for shaping, but as long as we have significant R in series with the speaker, we can use parallel LCRs.
Later impedance measurements will show if we´ve run into trouble with this configuration.
One important hint: all parallel Rs are total R, that means R of the coil and R of the resistor together.
All the best
Mattes
next step: a second parallel notch takes care about the 9 kHz spike. Values are 1,8 Ohm, 2,0 uF and 0,19 mH. The region around the spike is also slightly attenuated, that´s how it is...
Some people prefer serial LCRs for shaping, but as long as we have significant R in series with the speaker, we can use parallel LCRs.
Later impedance measurements will show if we´ve run into trouble with this configuration.
One important hint: all parallel Rs are total R, that means R of the coil and R of the resistor together.
All the best
Mattes
Hi Zilla,
of course, there are many ways to be happy.
I just want to show what can be done with a speaker which looks a bit problematic from the start, by throwing a handful of crossover elements at it.
In no way I´m claiming this to be the best or only way...
Also I´m not regarding the Fast 5 to be a top driver... but as before, with the correction this driver sounds quite nice, despite obvious trouble in the treble!
All the best
Mattes
Dear All,
after the 2 notches, we have slightly falling response and not too much treble.
Experience says that due to directivity in the treble, a slightly rising response is favorable with FR drivers.
So, the 2,7 uF cap which was omitted a few hours ago found it´s way back into the filter.
I think that the result looks quite OK, we have +- 3 dB in the room between 50 Hz and app. 13 kHz, distortions are OK, waterfall doesn´t show anything to worry in the mids and highs (for a paper FR driver).
That was the easy part, now we have a base for further finetuning and matching ones room and taste.
I will stop here, I won´t have the time for tuning by ear before april, and even if I had it, my final results will probably be irrelevant for somebody with another enclosure, room and taste.
All the best
Mattes
after the 2 notches, we have slightly falling response and not too much treble.
Experience says that due to directivity in the treble, a slightly rising response is favorable with FR drivers.
So, the 2,7 uF cap which was omitted a few hours ago found it´s way back into the filter.
I think that the result looks quite OK, we have +- 3 dB in the room between 50 Hz and app. 13 kHz, distortions are OK, waterfall doesn´t show anything to worry in the mids and highs (for a paper FR driver).
That was the easy part, now we have a base for further finetuning and matching ones room and taste.
I will stop here, I won´t have the time for tuning by ear before april, and even if I had it, my final results will probably be irrelevant for somebody with another enclosure, room and taste.
All the best
Mattes
Attachments
Dear All,
****!!! I finally measured the loudspeaker´s impedance with this filter, and have to give a warning: do not build this filter!
I´ll have to do it again some day, sorry, the impedance drops to 2 Ohms in the treble, not good.
This approach to FR filters has worked many times, but not this time, sorry.
All the best
Mattes
****!!! I finally measured the loudspeaker´s impedance with this filter, and have to give a warning: do not build this filter!
I´ll have to do it again some day, sorry, the impedance drops to 2 Ohms in the treble, not good.
This approach to FR filters has worked many times, but not this time, sorry.
All the best
Mattes
Dear All,
please disregard my last post, I have mixed up my cables.... shame on me!
I´ve thought 2 minutes about the problem, then it became clear that I´ve attached the cables for the impedance measurement AFTER the BSC and not before it, that couldn´t work...
So, also the final impedance check is OK, as attached, impedance drops to 6.something Ohms in the treble, that should be OK for every amp, and I´m happy as well, as I didn´t have to worry until april what went wrong...
Now, time again for a glass of wine and some northern jazz: Bass Alone from Ole Amund Gjersvik.
All the best
Mattes
please disregard my last post, I have mixed up my cables.... shame on me!
I´ve thought 2 minutes about the problem, then it became clear that I´ve attached the cables for the impedance measurement AFTER the BSC and not before it, that couldn´t work...
So, also the final impedance check is OK, as attached, impedance drops to 6.something Ohms in the treble, that should be OK for every amp, and I´m happy as well, as I didn´t have to worry until april what went wrong...
Now, time again for a glass of wine and some northern jazz: Bass Alone from Ole Amund Gjersvik.
All the best
Mattes
Thanks for taking the time and sharing the details, Mattes.
It's very appreciated. And cool, another jazz band to look up!
How do you design your XO?
I've been using XSim, and it has been a sanity, if not life, saver.
XSim free crossover designer
It's very appreciated. And cool, another jazz band to look up!
How do you design your XO?
I've been using XSim, and it has been a sanity, if not life, saver.
XSim free crossover designer
Hi Perceval,
I started making loudspeakers in a time faaaaar away from any simulation programs...
I use AJHorn for some simulations, but for crossovers, I only rely first on rough measurements with DATS and REW, second on my ears. Of course, final checks with REW are also done, but voicing by ear is the longest part.
So, even a simple XO design like above starts with several boxes of coils, caps and resistors, and a lot of measurements.
For the XO from yesterday, I have for sure made a total of 60 or 70 measurements, but experience helps insofar that I have some feeling for which values might be needed, so I don´t start at point Zero.
All the best
Mattes
I started making loudspeakers in a time faaaaar away from any simulation programs...
I use AJHorn for some simulations, but for crossovers, I only rely first on rough measurements with DATS and REW, second on my ears. Of course, final checks with REW are also done, but voicing by ear is the longest part.
So, even a simple XO design like above starts with several boxes of coils, caps and resistors, and a lot of measurements.
For the XO from yesterday, I have for sure made a total of 60 or 70 measurements, but experience helps insofar that I have some feeling for which values might be needed, so I don´t start at point Zero.
All the best
Mattes
I think you would love to play with XSim.
It would at least save you 50 measurements!
You start with the first one, with the driver on the baffle. Import that into XSim, along with the impedance curve from DATS.
From there, you add parts in the software and see the effect on the whole system, including FR, Impulse, Step Response, and resulting Impedance. Every part change is accurately displayed.
It would save you a lot of time.
Once you find something that looks good, you can use your ears to finalise the design according to your taste, by changing some values in the network. That's where your experience comes in.
Using XSim will take a lot out of initial guessing.
Give it a try during your next trip.
It would at least save you 50 measurements!
You start with the first one, with the driver on the baffle. Import that into XSim, along with the impedance curve from DATS.
From there, you add parts in the software and see the effect on the whole system, including FR, Impulse, Step Response, and resulting Impedance. Every part change is accurately displayed.
It would save you a lot of time.
Once you find something that looks good, you can use your ears to finalise the design according to your taste, by changing some values in the network. That's where your experience comes in.
Using XSim will take a lot out of initial guessing.
Give it a try during your next trip.
I'd love to incorporate your "flower" with that driver, but I have a hard time coming up with something like that, and a sealed enclosure behind!
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.