Hi Deerhunt,
Nice work! You are fast. Make sure you don’t have air leaks. I have found that asymmetric cats ears are sometimes due to an air leak. Try adding some very loose stuffing in exit vent to bring the height of ears down. A chunk of high porosity fish tank filter foam might work too. Add a little more stuffing in the closed end of the chamber. Cutting the first divider down would probably increase the tuning frequency but not sure it is needed. How would you describe the sound as it is?
I think the 100Hz is normal part of the response of the TL.
I think a BSC to the Purifi might equalize and make the 40Hz bass match the upper end more.
On Harsch XO I have 4th order and do not notice a 2.7kHz peak.
Nice work! You are fast. Make sure you don’t have air leaks. I have found that asymmetric cats ears are sometimes due to an air leak. Try adding some very loose stuffing in exit vent to bring the height of ears down. A chunk of high porosity fish tank filter foam might work too. Add a little more stuffing in the closed end of the chamber. Cutting the first divider down would probably increase the tuning frequency but not sure it is needed. How would you describe the sound as it is?
I think the 100Hz is normal part of the response of the TL.
I think a BSC to the Purifi might equalize and make the 40Hz bass match the upper end more.
On Harsch XO I have 4th order and do not notice a 2.7kHz peak.
Last edited:
While waiting for the new cabinets to be built, I decided to do some XO work using the PTT6.5 and the RAAL 70-20xr tweeter. I installed the tweeter into the spare baffle I have with the PTT6.5. I added some MakerBeam to give it some support so that I could take some measurements of it as an open baffle. The XO is going to be around 3.5kHz anyway so I don't think I need to see what the response is going to be like below 500Hz anyway.
Here is the setup - I have a large piece of leftover foam from a giant Pelican case that I am using to reduce the effect of the back reflected wave off the wall.
Here is a closeup showing how the drivers are mounted. I am using thw 12ohm tap on the RAAL and using a 2.2uF cap to give it about a 3.5kHz crossover point. Don't want to take raw data without a cap even.
I then proceeded to take data with the tweeter only, woofer only, and connecting the woofer and tweeter in parallel. The amp is my TPA3255 Reference Class D amp. Mic is UMIK-1 calibrated by Cross Spectrum Labs. Measuring at 0.5m away and on tweeter axis. Will probably redo this because the absolute SPL should be coming out to be closer to 88dB, not 94dB.
Here are three measurements shown on same graph, and amazingly, this is the first time I have seen an automatic crossover work itself out this nicely. The PTT6.5 is running fullrange, the RALL is via a 2.2uF film cap with built in transformer set to the 12ohm tap to level match the woofer better. The combined response looks like a real crossover has been developed for this, but purely accidental (and the XO frequency landed right about at 3.5kHz):
Here is the measured acoustical phase:
Here is the measured distortion (not sure what those peaks are but they are coming from the RAAL):
Here is the measured impulse step response:
All of these actually don't look that bad. The frequency response is automaticlally within +/-2.5dB. Very strange and lucky. What is weird is that there is no cancellation dip at the crossover frequency despite both drivers in same +ve phase???
Listening to it, it sounds very good actually. Anyhow, will measure the ZMA curves later and then load all of this into Xsim and see how much better we can make it. But have you guys ever seen such a good accidental XO?
Here is the setup - I have a large piece of leftover foam from a giant Pelican case that I am using to reduce the effect of the back reflected wave off the wall.
Here is a closeup showing how the drivers are mounted. I am using thw 12ohm tap on the RAAL and using a 2.2uF cap to give it about a 3.5kHz crossover point. Don't want to take raw data without a cap even.
I then proceeded to take data with the tweeter only, woofer only, and connecting the woofer and tweeter in parallel. The amp is my TPA3255 Reference Class D amp. Mic is UMIK-1 calibrated by Cross Spectrum Labs. Measuring at 0.5m away and on tweeter axis. Will probably redo this because the absolute SPL should be coming out to be closer to 88dB, not 94dB.
Here are three measurements shown on same graph, and amazingly, this is the first time I have seen an automatic crossover work itself out this nicely. The PTT6.5 is running fullrange, the RALL is via a 2.2uF film cap with built in transformer set to the 12ohm tap to level match the woofer better. The combined response looks like a real crossover has been developed for this, but purely accidental (and the XO frequency landed right about at 3.5kHz):
Here is the measured acoustical phase:
Here is the measured distortion (not sure what those peaks are but they are coming from the RAAL):
Here is the measured impulse step response:
All of these actually don't look that bad. The frequency response is automaticlally within +/-2.5dB. Very strange and lucky. What is weird is that there is no cancellation dip at the crossover frequency despite both drivers in same +ve phase???
Listening to it, it sounds very good actually. Anyhow, will measure the ZMA curves later and then load all of this into Xsim and see how much better we can make it. But have you guys ever seen such a good accidental XO?
Attachments
That’s not even a first try XO - it’s basically no XO. PTT6.5 running full range and 2.2uF cap on RAAL to avoid 0.6ohm short due to transformer coupling.
The RAAL protrudes 3.62in behind the 3/4in thick baffle. The TL is 4in deep so there should be enough room to fit inside.
RAAL has its own built in rear chamber that is sealed from main enclosure. Are you saying a new ribbon design that is OB or dipole sort of like some AMTs?
The RAAL protrudes 3.62in behind the 3/4in thick baffle. The TL is 4in deep so there should be enough room to fit inside.
RAAL has its own built in rear chamber that is sealed from main enclosure. Are you saying a new ribbon design that is OB or dipole sort of like some AMTs?
Looks good for not using a crossover for the Purifi! 
Seems like you don't have that 3,7k notch that I see. Perhaps I need to redesign my baffle?
I did some testing of damping and you're right that some loose damping in the open end makes a big difference on impedance peaks (both).
What evens out the peaks most is by putting the 80g stuffing more towards the closed end. Lets say from magnet and up. Actually it the magnet keeping the damping from moving down.
Green - evenly spread damping in the whole back chamber
Red - damping closed end
Purple - 10g damping open end
I measured the other box and the impedance looks almost identical. So the 270Hz notch can be seen in both.
Seems like you don't have that 3,7k notch that I see. Perhaps I need to redesign my baffle?
I did some testing of damping and you're right that some loose damping in the open end makes a big difference on impedance peaks (both).
What evens out the peaks most is by putting the 80g stuffing more towards the closed end. Lets say from magnet and up. Actually it the magnet keeping the damping from moving down.
Green - evenly spread damping in the whole back chamber
Red - damping closed end
Purple - 10g damping open end
I measured the other box and the impedance looks almost identical. So the 270Hz notch can be seen in both.
Last edited:
I don’t show my data below 500Hz since open baffle losses are great. Are you taking about the tiny wiggle in impedance curve at 270Hz?
That’s nothing to worry about. Very small. I don’t see anything in frequency. 270Hz would not be caused by baffle. It’s a 31cm 1/4wave characteristic length. That might be reflection from driver membrane to closed end wall. Maybe add more stuffing in closed end. Or it maybe reflection from vent mouth to first turn. Add more foam in corner. Take ruler and look for 31cm or 62cm 1/2wave distance.
Good work in effect of damping on impedance curve. As it gets lower in peaks the bass should sound drier and more articulate.
Baffle diffraction effects are of order 1kHz to 8kHz. You do have a stepped baffle there and that might show up somewhere as non smooth response.
So 3.7kHz peak is probably baffle effect. That’s a 4.6cm characteristic distance for 1/2 wave reflection. Looks like from your woofer to the raised step edge is about that distance? You also 4 raised equidistant edges on that stepped baffle. Whereas on mine it is flush and baffle edges are not all equidistant. Just two side ones and those may be the 3kHz peak. 5.7cm is about distance from midpoint of membrane from center to surround and this to the side?
Your box and baffle look great though. Very nice work. Is that a Bliesema tweeter? Everyone is using those with the Purifi it seems.
That’s nothing to worry about. Very small. I don’t see anything in frequency. 270Hz would not be caused by baffle. It’s a 31cm 1/4wave characteristic length. That might be reflection from driver membrane to closed end wall. Maybe add more stuffing in closed end. Or it maybe reflection from vent mouth to first turn. Add more foam in corner. Take ruler and look for 31cm or 62cm 1/2wave distance.
Good work in effect of damping on impedance curve. As it gets lower in peaks the bass should sound drier and more articulate.
Baffle diffraction effects are of order 1kHz to 8kHz. You do have a stepped baffle there and that might show up somewhere as non smooth response.
So 3.7kHz peak is probably baffle effect. That’s a 4.6cm characteristic distance for 1/2 wave reflection. Looks like from your woofer to the raised step edge is about that distance? You also 4 raised equidistant edges on that stepped baffle. Whereas on mine it is flush and baffle edges are not all equidistant. Just two side ones and those may be the 3kHz peak. 5.7cm is about distance from midpoint of membrane from center to surround and this to the side?
Your box and baffle look great though. Very nice work. Is that a Bliesema tweeter? Everyone is using those with the Purifi it seems.
Last edited:
Thanks for the useful tips!
My open end has some edges and that might be the resonance we're seeing. Giving reflections. Measuring from edge to internal "last corner" is 31cm!!!
I will cut the last away the edge (15mm on the sides and 5mm low end).
You're right about the 4,6cm also. It's the distance between membrane and edge of baffle. Hats off!
I will make a better baffle in the final version, giving it smoother transition hopefully.
Yes, it the BlieSMa T34A. It seems to integrate well with the Purifi. It's only a shame to throw away 10dB of gain...
My open end has some edges and that might be the resonance we're seeing. Giving reflections. Measuring from edge to internal "last corner" is 31cm!!!
I will cut the last away the edge (15mm on the sides and 5mm low end).
You're right about the 4,6cm also. It's the distance between membrane and edge of baffle. Hats off!
I will make a better baffle in the final version, giving it smoother transition hopefully.
Yes, it the BlieSMa T34A. It seems to integrate well with the Purifi. It's only a shame to throw away 10dB of gain...
Glad that the physics works out well. I used to use the same trick when setting up optics and lasers to find inadvertent reflections that causes dips.. Frequencies much higher but same concept. Light travels about 11in per nano second vs sound at 342m per second.
The nice thing with RAAL is it has 3 transformer taps to vary sensitivity from 89dB (12ohms) to 93dB (5ohms). So far, not a single resistor in the circuit.
The nice thing with RAAL is it has 3 transformer taps to vary sensitivity from 89dB (12ohms) to 93dB (5ohms). So far, not a single resistor in the circuit.
So here is my first trial crossover design for the RAAL and PTT6.5. I tried to reduce the stuff above 3.5kHz on the PTT6.5 and added a small BSC to level the signals. There is something off with my SPL meter - it seems to be registering +5.0dB higher than it should so ignore the absolute SPLs for the moment.
It is all positive phase crossover still and seems quite time aligned. I'll prototype this pretty quickly with my XO dev kit and Wago connectors. I think I have all of these values on hand. I had to apply a manual 1.98in time delay on woofer to get the raw signals to predict the combined trace for the interferometry step.
Schematic (not shown is inherent 2.2uF series cap and built in transformer on RAAL):
Predicted Freq Response:
Predicted Impedance Response (note this is OB so only single peak from woofer):
Predicted Step Response:
My feeling is that if I could make a 2in deep waveguide for the RAAL, I could also apply the Harsch XO here for a transient perfect XO without the phase wraps.
It is all positive phase crossover still and seems quite time aligned. I'll prototype this pretty quickly with my XO dev kit and Wago connectors. I think I have all of these values on hand. I had to apply a manual 1.98in time delay on woofer to get the raw signals to predict the combined trace for the interferometry step.
Schematic (not shown is inherent 2.2uF series cap and built in transformer on RAAL):
Predicted Freq Response:
Predicted Impedance Response (note this is OB so only single peak from woofer):
Predicted Step Response:
My feeling is that if I could make a 2in deep waveguide for the RAAL, I could also apply the Harsch XO here for a transient perfect XO without the phase wraps.
Attachments
Last edited:
Here is the data with the temporary XO now implemented with Wago connectors. The predicted response matches the measured response almost spot-on. XO simulators are pretty cool aren't they?
Values as implemented in schematic (note that there is 2.2uF cap in series after resistor for the tweeter that is not shown as it is part of the raw data):
Comparison of Predicted (blue) with Measured (orange):
Measured Freq response:
Measured Acoustical Phase:
Measured Step Response:
Listening to it now - very very nice! RAALs really have a nice smooth top end that is so clear.
Temporary developmental XO implementation:
@Danny:
I'll play around with the XO some more in Xsim and try moving it lower - I think the 3kHz peak is the sharp baffle edge diffraction though.
Values as implemented in schematic (note that there is 2.2uF cap in series after resistor for the tweeter that is not shown as it is part of the raw data):
Comparison of Predicted (blue) with Measured (orange):
Measured Freq response:
Measured Acoustical Phase:
Measured Step Response:
Listening to it now - very very nice! RAALs really have a nice smooth top end that is so clear.
Temporary developmental XO implementation:
@Danny:
I'll play around with the XO some more in Xsim and try moving it lower - I think the 3kHz peak is the sharp baffle edge diffraction though.
Attachments
Last edited:
I have been doing some listening with well known test tracks. This combo of the RAAL 70-20xr and the PTT6.5 with this XO sounds superb! Very lifelike realism and clarity and yet not fatiguing. It conveys the emotion of the performance very well. I hear things that I have not heard before - I know that's a common thing to say but it really does happen with the RAAL, and always surprising when you think you know a track because you have heard it hundreds of times before.
I have my 10F/RS225 FAST TL's playing on left channel to round out the bass lacking in the right channel open baffle at present.
Yuri Honing Trio "Walking on the Moon"
Suzanne Vega "Rosemary"
Natalie Merchant "Maggie Said"
Ahmad Jamal "Autumn Rain"
Omer Avital "Maroc"
Circuit Bent "Orchestrality"
Eagles "Hotel California"
Steely Dan "Hey Nineteen"
Metallica "Enter Sandman"
Anne Bisson "September in Montreal"
Chris Brubeck "In Your Own Sweet Way"
The Who "See Mee Feel Me Listening to You"
Gordon Goodwin's Big Phat Band "Garaje Gato"
Norah Jones "Burn"
It's all excellent and very enjoyable. You know you have a special combination when it makes you want to re-listen to your collection. Can't wait to hear it with the TL on the PTT6.5. I think a lot of the credit goes to the PTT6.5 for handling such a large portion of the bandwidth up to 3.5kHz is basically all of the mid range and it is so clear and low distortion. I should add that the PTT6.5 does indeed sound excellent as a mid range in an OB. I am definitely not using it for bass here and there is a nice effect of the driver being open on the back. Maybe I should make a 3-way with a bass TL, an OB mid, and a ribbon top.
I have my 10F/RS225 FAST TL's playing on left channel to round out the bass lacking in the right channel open baffle at present.
Yuri Honing Trio "Walking on the Moon"
Suzanne Vega "Rosemary"
Natalie Merchant "Maggie Said"
Ahmad Jamal "Autumn Rain"
Omer Avital "Maroc"
Circuit Bent "Orchestrality"
Eagles "Hotel California"
Steely Dan "Hey Nineteen"
Metallica "Enter Sandman"
Anne Bisson "September in Montreal"
Chris Brubeck "In Your Own Sweet Way"
The Who "See Mee Feel Me Listening to You"
Gordon Goodwin's Big Phat Band "Garaje Gato"
Norah Jones "Burn"
It's all excellent and very enjoyable. You know you have a special combination when it makes you want to re-listen to your collection. Can't wait to hear it with the TL on the PTT6.5. I think a lot of the credit goes to the PTT6.5 for handling such a large portion of the bandwidth up to 3.5kHz is basically all of the mid range and it is so clear and low distortion. I should add that the PTT6.5 does indeed sound excellent as a mid range in an OB. I am definitely not using it for bass here and there is a nice effect of the driver being open on the back. Maybe I should make a 3-way with a bass TL, an OB mid, and a ribbon top.
Last edited:
Great that you have such a perfect correlation between simulation and measurements !
Yes, those RAALs are great, they sound natural and detailed without asking for attention.
I've crossed them at 3kHz or a little higher with my Accutons, they blend perfectly.
I've tried lower than 3kHz but that didn't sound good.
Enjoy your new speaker !
Yes, those RAALs are great, they sound natural and detailed without asking for attention.
I've crossed them at 3kHz or a little higher with my Accutons, they blend perfectly.
I've tried lower than 3kHz but that didn't sound good.
Enjoy your new speaker !
