Hi, ChocoHolic!
I made a similar subwoofer, but a very low cost version. Second hand Magnat 8 coll 150W speaker with 2 coils, 7 l box (resonance about 100 Hz), driven by a 100W amp. Obviously the SPL was not enough unless I stand it into the corner.
I like your design very much! But I think you have too much money!
You inspired me to build it with 7 pieces 6" Sony speaker with 1,5" coils. Xmax is about 6 mm. The amplifier is allready built, matter of fact there is even too much of Class D amp in my... (damn, I dont know the word...)
2 kW @ 2-2,66 ohm bridged, with MOSFETs:
http://www.hszk.bme.hu/~sp215/PWM_2kW/27.jpg
500W @ 4 ohm full bandwidth, 1000W low bandwidth, SE, with IGBTs:
http://www.hszk.bme.hu/~sp215/PWM_500W/311.jpg
http://www.hszk.bme.hu/~sp215/PWM_500W/316.jpg
Both have undestroyable overcurrent protection.
I saw You want to build one. Those Semicron MOSFET modules seems to be too large. If you use XXL MOSFETS, the conducting losses decreased, but switching losses and EMI increased, therefor I think you should compromisem, if you want an optimal solution! Could you write me a link to the datasheet of these MOSFETS? I would compare it to others! Where would you buy them? In hungary It's allmost impossible to buy any special parts.
I think doubled FDP2552 would be perfect to your bridged amp too! Best driver (IMHO): IR2010, best modulation method: DB (3-level PWM).
P.S.: I am chocoholic too! What is your daily dose?
I made a similar subwoofer, but a very low cost version. Second hand Magnat 8 coll 150W speaker with 2 coils, 7 l box (resonance about 100 Hz), driven by a 100W amp. Obviously the SPL was not enough unless I stand it into the corner.
I like your design very much! But I think you have too much money!
You inspired me to build it with 7 pieces 6" Sony speaker with 1,5" coils. Xmax is about 6 mm. The amplifier is allready built, matter of fact there is even too much of Class D amp in my... (damn, I dont know the word...)
2 kW @ 2-2,66 ohm bridged, with MOSFETs:
http://www.hszk.bme.hu/~sp215/PWM_2kW/27.jpg
500W @ 4 ohm full bandwidth, 1000W low bandwidth, SE, with IGBTs:
http://www.hszk.bme.hu/~sp215/PWM_500W/311.jpg
http://www.hszk.bme.hu/~sp215/PWM_500W/316.jpg
Both have undestroyable overcurrent protection.
I saw You want to build one. Those Semicron MOSFET modules seems to be too large. If you use XXL MOSFETS, the conducting losses decreased, but switching losses and EMI increased, therefor I think you should compromisem, if you want an optimal solution! Could you write me a link to the datasheet of these MOSFETS? I would compare it to others! Where would you buy them? In hungary It's allmost impossible to buy any special parts.
I think doubled FDP2552 would be perfect to your bridged amp too! Best driver (IMHO): IR2010, best modulation method: DB (3-level PWM).
P.S.: I am chocoholic too! What is your daily dose?
I've heard that someone has a patent on subwoofers operating entirely at frequencies below their (unusually high) system resonant frequency. If true, this proves once again that you can patent anything. When did you first design your system? I wonder if it constitutes prior art.
[Added a couple of minutes later] I did a quick google and came up with this: http://ldsg.snippets.org/boxes.php3
It says Bag End makes such a sub. Whether or not they claim patent rights, I do not know. All I know is what I read on the net.
[Added a couple of minutes later] I did a quick google and came up with this: http://ldsg.snippets.org/boxes.php3
It says Bag End makes such a sub. Whether or not they claim patent rights, I do not know. All I know is what I read on the net.
They don't own the ELF patent, they are a licencee. The patent of the ELF belongs to Long and Wickersham (one of them is the owner of Alembic).
It is perfectly possible to make low-end EQs that don't have the restrictions of LT (not possible to make dynamically tuneable) and ELF (intrinsic LPF).
BTW: The thing about "not introducing phase errors" with the ELF is nonsense. The ELF circuit does of course not introduce phase-inversion (like stated in their docu) but two times 90 degrees phase DELAY. Being a 2nd order closed box system it is still one of the better solutions regarding transient behaviour however.
Regards
Charles
phase_accurate mentioned:
Does anyone have experience with Dr. Leach's method described in a JAES paper from the late 80's/ early 90's? I have a copy somewhere around here, but it's faded and barely readable.
A friend of mine built one for a car system back then with excellent results . His only comment was "It was a mathematical nightmare..."
Tim
Does anyone have experience with Dr. Leach's method described in a JAES paper from the late 80's/ early 90's? I have a copy somewhere around here, but it's faded and barely readable.
A friend of mine built one for a car system back then with excellent results . His only comment was "It was a mathematical nightmare..."
Tim
What is its working principle ?
Edit: Is it like the first circuit on this page ? :
http://members.hometown.aol.com/_ht_a/lmdmkm/mini-sub/mini-sub.htm
Regards
Charles
Edit: Is it like the first circuit on this page ? :
http://members.hometown.aol.com/_ht_a/lmdmkm/mini-sub/mini-sub.htm
Regards
Charles
O.K,. but the manufacturer lets one believe that the circuit only generates phase-inversion, which is not correct. Furthermore the driver & ELF combination is a 4th order lowpass. It's transient behaviour is the same as a closed box with a 2nd order lowpass as x-over, neither better nor worse.
Regards
Charles
Regards
Charles
Hi Dave
LT= Linkwitz transform
Advantage of LTF:
Very wide range of possible driver/tuning combinations. No intrinsic HF rolloff.
Disadvantages:
Hard to make tuneable (the implementation via state-variable filter is a possible workaround) let alone dynamically tuneable for overexcursion protection. Needs 6 fairly accurate capacitor values.
ELF (extended low frequency)
Advantages:
Simple. Lower cutoff easily made tuneable, even electronically for dynamic excursion/amp-overdrive control. Basically only two capacitors needed.
Disadvantage:
Intrinsic lowpass function, depending on driver fc.
Regards
Charles
LT= Linkwitz transform
Advantage of LTF:
Very wide range of possible driver/tuning combinations. No intrinsic HF rolloff.
Disadvantages:
Hard to make tuneable (the implementation via state-variable filter is a possible workaround) let alone dynamically tuneable for overexcursion protection. Needs 6 fairly accurate capacitor values.
ELF (extended low frequency)
Advantages:
Simple. Lower cutoff easily made tuneable, even electronically for dynamic excursion/amp-overdrive control. Basically only two capacitors needed.
Disadvantage:
Intrinsic lowpass function, depending on driver fc.
Regards
Charles
phase_accurate asked:
I'm not sure. The page you referenced mentions Dr Leach's name, but IIRC he followed his original publication with another in the 80's.
When I read about the LT on Linkwitz' site, it really sounded like them principle Leach used in the paper I read. However, I don't know if their implementation is equivalent.
I guess it's time to dredge up the old, faded thermal copy and see if I can still read it...
Tim
I'm not sure. The page you referenced mentions Dr Leach's name, but IIRC he followed his original publication with another in the 80's.
When I read about the LT on Linkwitz' site, it really sounded like them principle Leach used in the paper I read. However, I don't know if their implementation is equivalent.
I guess it's time to dredge up the old, faded thermal copy and see if I can still read it...
Tim
Hi Pafi!
...was on holidays since June 4th....
Great, really nice to see that my sub inspired you
The link to the DataSheet of that MosFet-modules is:
http://www.semikron.com/databook/05mosfet/m120b020.pdf
I also feel that they are oversize and their high gate charge pushed me think about that heavy FET-drivers UC3710T from TI/Unitrode.
http://focus.ti.com/docs/prod/folders/print/uc3710.html
I have received the two SADHARA drivers and since had them in my hands in real life, I decided to make two subwoofer with one drive each. One driver already has a weigth of approx. 20kg.
A box with two drivers would result in about 60kg....
To heavy, to difficult to find a proper position....
Alright, so the required power reduces.
Also I am now planning to integrate the amp into the box, so I will not need a short circuit protection anymore. This will also relax the power requirements for the MOSFETs.
Which types would you recommend?
If I will keep the possibility to go for a rail voltage of about 130V, then I will need at least 150V types, better 200V.
If I calculate the max. value of the sine wave current by dividing the
rail by the DC resistance of the speaker (2.3 Ohms) I will end up short above 50A .
As the frequency is vey low and the speakers are driven much below the resonance frequency, they will show an impedance close to their DC-value. And as the frequencies are very low the SOA of the MOSFETS will not allow much more current than their DC-rating...
I don't think that I have to much money, I simply optimized my financial range:
Which thing is sucking really a lot of a man's money, but in fact is not really necessary? -- NO! It is not his wife!!!
It's the car.
@ Yuri:
Why should 80Hz be the definite limit?
I see to major reasons for distorsions:
1. ..speakers cone may bend or break up into partial modes....
2. ..non linearity of the air suspension...
(Air suspension of a sealed box is stiffer for compreshion and less
stiffer for decompression)
Point 1 is related to the selected speakers.
Due to this I decided to pick that crazy expensive ESOTECs, which
drive a 110mm cone with a 75mm voice coil.
Point 2 is important if the ratio volume of enclosure / volume displacement is getting small (for my taste: less than 20 at full excursion). You overcome Point 2 by an active excursion
regulation and skip the LT. I think some of the VELODYNE's top level Subs are also UnderResonancePrincipleSubwoofers (URPS) with active excursion regulation...
Of course I agree: The higher you choose the resonance frequency the more power you require. A hell of lot of more power, because the speakers output transfer function drops with second order....
On the other hand your subwoofer should not have a resonance frequency which matches to one of your 3 major resonances of the
room were you want to listen to music.
In normal rooms the highest main resonance is about 71 Hz, which corelates to the commeon 2.4m height of the room....
So I would prefer to stay above that.
Cheers to all
Markus
P.S.
I started that subwoofer in 1999 and I use it since 2000.
...have no idea, if URPS was new at that time or not.
I simply had the idea, tried it, and it was working really fine.
..didn't care about patenting it, because I don't like our completely derailed patent system...
...was on holidays since June 4th....
Great, really nice to see that my sub inspired you
The link to the DataSheet of that MosFet-modules is:
http://www.semikron.com/databook/05mosfet/m120b020.pdf
I also feel that they are oversize and their high gate charge pushed me think about that heavy FET-drivers UC3710T from TI/Unitrode.
http://focus.ti.com/docs/prod/folders/print/uc3710.html
I have received the two SADHARA drivers and since had them in my hands in real life, I decided to make two subwoofer with one drive each. One driver already has a weigth of approx. 20kg.
A box with two drivers would result in about 60kg....
To heavy, to difficult to find a proper position....
Alright, so the required power reduces.
Also I am now planning to integrate the amp into the box, so I will not need a short circuit protection anymore. This will also relax the power requirements for the MOSFETs.
Which types would you recommend?
If I will keep the possibility to go for a rail voltage of about 130V, then I will need at least 150V types, better 200V.
If I calculate the max. value of the sine wave current by dividing the
rail by the DC resistance of the speaker (2.3 Ohms) I will end up short above 50A .
As the frequency is vey low and the speakers are driven much below the resonance frequency, they will show an impedance close to their DC-value. And as the frequencies are very low the SOA of the MOSFETS will not allow much more current than their DC-rating...
I don't think that I have to much money, I simply optimized my financial range:
Which thing is sucking really a lot of a man's money, but in fact is not really necessary? -- NO! It is not his wife!!!
It's the car.
@ Yuri:
Why should 80Hz be the definite limit?
I see to major reasons for distorsions:
1. ..speakers cone may bend or break up into partial modes....
2. ..non linearity of the air suspension...
(Air suspension of a sealed box is stiffer for compreshion and less
stiffer for decompression)
Point 1 is related to the selected speakers.
Due to this I decided to pick that crazy expensive ESOTECs, which
drive a 110mm cone with a 75mm voice coil.
Point 2 is important if the ratio volume of enclosure / volume displacement is getting small (for my taste: less than 20 at full excursion). You overcome Point 2 by an active excursion
regulation and skip the LT. I think some of the VELODYNE's top level Subs are also UnderResonancePrincipleSubwoofers (URPS) with active excursion regulation...
Of course I agree: The higher you choose the resonance frequency the more power you require. A hell of lot of more power, because the speakers output transfer function drops with second order....
On the other hand your subwoofer should not have a resonance frequency which matches to one of your 3 major resonances of the
room were you want to listen to music.
In normal rooms the highest main resonance is about 71 Hz, which corelates to the commeon 2.4m height of the room....
So I would prefer to stay above that.
Cheers to all
Markus
P.S.
I started that subwoofer in 1999 and I use it since 2000.
...have no idea, if URPS was new at that time or not.
I simply had the idea, tried it, and it was working really fine.
..didn't care about patenting it, because I don't like our completely derailed patent system...
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.