Thank you!
While I agree the PT2C is an overall better performer, I’m looking at this as a budget experiment with less expensive drivers all around. Building one, not much lost as I could always repurpose the TC9’s and sell off the PT Minis at a loss.
My most recent exposure to some of the new Magnepan models such as the 1.7 have me questioning even bothering with a home use array as the performance of the 1.7 is simply amazing.......but being addicted to DIY which gives me purpose to ask questions keeps me in the game. I’m confident the SB in your array is a real gem of a driver and the new 3” has peaked my interest as well.
While I agree the PT2C is an overall better performer, I’m looking at this as a budget experiment with less expensive drivers all around. Building one, not much lost as I could always repurpose the TC9’s and sell off the PT Minis at a loss.
My most recent exposure to some of the new Magnepan models such as the 1.7 have me questioning even bothering with a home use array as the performance of the 1.7 is simply amazing.......but being addicted to DIY which gives me purpose to ask questions keeps me in the game. I’m confident the SB in your array is a real gem of a driver and the new 3” has peaked my interest as well.
In an attempt to put this to rest, I did a progression of diffraction simulations in which I kept the pitch of the drivers the same 95mm but gradually increased the height of each driver's radiating surface from the 60 mm original up to 95mm at which point there is a continuous line source. At 86 mm where the gap between drivers is about 10% of the spacing, things started looking pretty good.
Given this progression showing continuous improvement with decreasing gap, I am inclined to believe that the simulation accurately predicts where one would see combing effects in the array. (I've also asked Kimmo, the author of Vituix, to comment). That doesn't mean the array wouldn't sound good because those comb lines disappear with smoothing.
Perhaps you should consider a short line of 4 minis next to the line of TC9s.
There has been a lot of discussion re' a really good tweeter in the middle of a line array and how to keep its sound in balance with the line at more than one distance. Frequency tapering the line array would be how to do it. All TC9s would contribute at low frequencies. Each group of 4 or so TC9s would get a lower cutoff frequency the further away from array center it was located.
How many PT-minis can be used before destructive combing sets in?
Here is a picture with 12
That is encouraging. You can use more drivers but then you need EQ in the top end. Notice its the floor reflection that is problematic. Do you have carpets. Looks to me like non full height PT mini array would be the way to go. Plenty of opportunity for fun experiments.
(Have to run now so might be overlooking something...)
Given this progression showing continuous improvement with decreasing gap, I am inclined to believe that the simulation accurately predicts where one would see combing effects in the array. (I've also asked Kimmo, the author of Vituix, to comment). That doesn't mean the array wouldn't sound good because those comb lines disappear with smoothing.
Perhaps you should consider a short line of 4 minis next to the line of TC9s.
There has been a lot of discussion re' a really good tweeter in the middle of a line array and how to keep its sound in balance with the line at more than one distance. Frequency tapering the line array would be how to do it. All TC9s would contribute at low frequencies. Each group of 4 or so TC9s would get a lower cutoff frequency the further away from array center it was located.
How many PT-minis can be used before destructive combing sets in?
Here is a picture with 12
That is encouraging. You can use more drivers but then you need EQ in the top end. Notice its the floor reflection that is problematic. Do you have carpets. Looks to me like non full height PT mini array would be the way to go. Plenty of opportunity for fun experiments.
(Have to run now so might be overlooking something...)
Attachments
Got a suggestion from Kimmo as to how to get more accurate sim of PTmini - model it as multiple smaller drivers that touch. If its diaphragm is 12x60 mm, it can be modeled as 5 12x12s. This will take a little bit of time but not all that much. Meanwhile, the short array approach still looks encouraging to me.
Well, not that much time at all. I placed 64 of these 12x12 mm drivers then deleted 3 out of every 8 leaving a gap of 36mm between groups of 5. Directivity for the 12x12s was computed at 1 degree steps.
The figure shows both this sim and results from the diffraction tool. Vituix limits me to 65 drivers total so I can't do a bigger array this way but the small differences gives confidence in the diffraction tool results.
The figure shows both this sim and results from the diffraction tool. Vituix limits me to 65 drivers total so I can't do a bigger array this way but the small differences gives confidence in the diffraction tool results.
Attachments
Interesting results!
Your mention of a central tweeter and a tapered line harkens to my favorite alignment which is a WWMTMWW.......which can very much reduce artifacts from floor/ceiling reflections while maintains near point source behavior in the critical hearing range.......works well in almost any listening room. Never tried such a design with controlled horizontal directivity though.....waveguide loaded tweeter and maybe OB/Dipole mids?......the high QTS TC9 would work really well as a mid in this configuration........maybe a horn loaded AMT/Ribbon in between?........better be staying put when listening though unless......unless 1st order high crossover point?
Your mention of a central tweeter and a tapered line harkens to my favorite alignment which is a WWMTMWW.......which can very much reduce artifacts from floor/ceiling reflections while maintains near point source behavior in the critical hearing range.......works well in almost any listening room. Never tried such a design with controlled horizontal directivity though.....waveguide loaded tweeter and maybe OB/Dipole mids?......the high QTS TC9 would work really well as a mid in this configuration........maybe a horn loaded AMT/Ribbon in between?........better be staying put when listening though unless......unless 1st order high crossover point?
Last edited:
WWMTMWM is another approach addressing the floor and ceiling reflection problem. I'm liking the original concept more as I go along and this version is certainly much cheaper than a high end WWMTMWW would be.
I think now I've done enough simulation to believe that it is indeed worth trying It appears that the amount of combing increases with the length of the array, as does the vertical coverage. The next question is how long should the array of PTminis next to the full height array of TC9s be.
If I were building this, I would go with 12 PTminis spanning sitting through standing ear height. With this length PTmini array, I don't see a need to frequency taper the TC9 array.
I think now I've done enough simulation to believe that it is indeed worth trying It appears that the amount of combing increases with the length of the array, as does the vertical coverage. The next question is how long should the array of PTminis next to the full height array of TC9s be.
If I were building this, I would go with 12 PTminis spanning sitting through standing ear height. With this length PTmini array, I don't see a need to frequency taper the TC9 array.
Attachments
I would say the length of the vertical line of PT Minis could be limited to two factors.....the desired vertical listening window size and power handling......which is the bottleneck here. Power compression is REALLY evident with planar drivers as there’s little to dissipate heat which contributes to power compression. I originally considered a length of 16 units for an overall HF line length of 60”.
A third factor is when you get combing instead of just ripples in the frequency response. There is a just a hint of combing at 16 drivers (top right of previous graph) at around 18 khz and starts at about 12 khz with 20 drivers (bottom right in previous graph).
Well as far as power goes, the array both giveth and taketh away. You get more devices in parallel for more efficiency and more power handling but you also get a 3db per octave roll off. There are only ~2.5 octaves between 4 khz and 20 Khz so that is only 7.5 db but you have 9 db of array gain with just 8 drivers and 9 db more power handling. Power compression shouldn't be an issue for the tweeter array even limited to 12 drivers.
The graphs suggest you can go to a longer PTmini array and avoid combing so long as you frequency taper it. Probably smarter to keep it simple.
Well as far as power goes, the array both giveth and taketh away. You get more devices in parallel for more efficiency and more power handling but you also get a 3db per octave roll off. There are only ~2.5 octaves between 4 khz and 20 Khz so that is only 7.5 db but you have 9 db of array gain with just 8 drivers and 9 db more power handling. Power compression shouldn't be an issue for the tweeter array even limited to 12 drivers.
The graphs suggest you can go to a longer PTmini array and avoid combing so long as you frequency taper it. Probably smarter to keep it simple.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.