StigErik if that was your site I have to say it was very educational. I loved the design you created. Is there any chance you could put that info on a thread here or would that cause a conflict with Griffen? I used to have a lot of questions about that build but now the biggest questions I have revolve around the subwoofer. I just want to know what the exact driver was the the tuning you went with. I think that the subs would make great speaker stands for a pair of Gedlee speakers.
The g1 looks like an amazing speaker. The price is way to high though but when you are dealing with people setting up mastering studios I guess they could afford it. I bet you could build that whole system for 1/10th of the price utilizing the same drivers, your own woodwork, miniDSP, and quality amps.
The g1 looks like an amazing speaker. The price is way to high though but when you are dealing with people setting up mastering studios I guess they could afford it. I bet you could build that whole system for 1/10th of the price utilizing the same drivers, your own woodwork, miniDSP, and quality amps.
As mentioned, the TD15S will work quite well in a sealed enclosure. It is about 93.5dB 1W or 95.1dB 2.83V in 8ohm nominal. At 30hz it will be around 7dB down getting you 88dB at 30hz with 1W. This is in 7.5cf with typical box fill giving a Qtc of .6 as you prefer. F3 is 48hz and F10 is about 24.5Hz.
You could also look at the SBP15 which uses the same coil but has a dustcap and a little more mass. It is designed get a lower F3 in sealed boxes, but you go up to about 8.5cubic foot. It is 90.7dB 1W and 92.7dB 2.83v in the 8ohm nominal version. This gives you -5.4dB at 30hz for 87.1dB 1W at that point. The F3 is 39hz with F10 being 21Hz.
SBP15
Either the SBP15 or TD15S can be done in a 16ohm coil. We have a company in the UK that does recording studios that switched to the SBP15 for all their sealed box monitoring setups. They were previously using TAD1601's at over 2x the cost.
Something like the JBL 2242H is really not practical for sealed box use. While they are more efficient, you don't get much extension due to the laws of physics. The 2242H is 99dB 1W but by 30hz in a Qtc .6 enclosure the response is down over 18dB for only 82.4dB 1W. The 2245 is better at 95dB 1W but response is still down 9.3dB at 30hz for 87dB. In comparison to the TD15S or SBP15 though, it has only 9.5mm Xmax vs 14mm. It has over 4x the inductance at 1.4mh vs .3mH for the TD or SBP. It has no shorting rings to handle flux modulation or keep inductance linear either.
John
You could also look at the SBP15 which uses the same coil but has a dustcap and a little more mass. It is designed get a lower F3 in sealed boxes, but you go up to about 8.5cubic foot. It is 90.7dB 1W and 92.7dB 2.83v in the 8ohm nominal version. This gives you -5.4dB at 30hz for 87.1dB 1W at that point. The F3 is 39hz with F10 being 21Hz.
SBP15
Either the SBP15 or TD15S can be done in a 16ohm coil. We have a company in the UK that does recording studios that switched to the SBP15 for all their sealed box monitoring setups. They were previously using TAD1601's at over 2x the cost.
Something like the JBL 2242H is really not practical for sealed box use. While they are more efficient, you don't get much extension due to the laws of physics. The 2242H is 99dB 1W but by 30hz in a Qtc .6 enclosure the response is down over 18dB for only 82.4dB 1W. The 2245 is better at 95dB 1W but response is still down 9.3dB at 30hz for 87dB. In comparison to the TD15S or SBP15 though, it has only 9.5mm Xmax vs 14mm. It has over 4x the inductance at 1.4mh vs .3mH for the TD or SBP. It has no shorting rings to handle flux modulation or keep inductance linear either.
John
Hi John. Cheers. I already had problems the way you measure Xmax as a percentage of Xlim, and not as max. excursion (peak) with 10%distortion (didn't know there are different ways). I guess that is a trade thing, and is not measured as a standard parameter for all brands. My question, now, is a different one. Do you measure (voice coil inductance) Le@1KHz or Le@20KHz. Because your Le is one "almost impossible".
I think you said once that your voice coils are small diameter and specially wounded. That can be seen in the impedance curve anyway, as a reflex of inductance.
Last edited:
Hi John. Cheers. I already had problems the way you measure Xmax as a percentage of Xlim, and not as max. excursion (peak) with 10%distortion (didn't know there are different ways). I guess that is a trade thing, and is not measured as a standard parameter for all brands. My question, now, is a different one. Do you measure (voice coil inductance) Le@1KHz or Le@20KHz. Because your Le is one "almost impossible".
I'm not sure what you mean by a problem the way we measure Xmax. The industry standard for Xmax is determined by the lesser value of either Xmag and Xsus. This is defined by both Dumax testing since the mid 95's and also by Klippel more recently. Xsus is typically always a higher value and Xmag is almost always the determining value. Xmag is defined as the point where Bl drops to 70% of it's rest value. This can be verified with Klippel testing, or simply through FEA analysis to tell where this point is.
Other companies give general "guesses" as to where this happens. Many use (VC length - gap height)/2 and then add 50% of the gap height back in. Generally this is close, but depending on what the height of the gap plate is, it is not always accurate. For example, we have a 38mm coil and 18mm thick gap plate. If we were to use that formula we would get 10mm overhang plus 9mm for half gap and get 19mm. This is overstated. Others add only 25% of the gap. Some don't add any. However if you model and do FEA or have a Klippel test verify the driver, the 70% Bl value is at 14mm one way as we claim. The Bl curve is the best determining factor.
The reason 70% is taken is this is considered a standard value to get to that 10% Xmax. In some drivers with high flux modulation you have much higher odd order harmonics and distortion can be much higher than 10% at that value. In a driver where proper shorting rings are used to drastically reduce the odd order harmonics, those Xmax figures will be valid. Klippel can even go a step farther in testing and just give an excursion where you reach 10% distortion based not only on the Bl curve but also physical noises in the driver, suspension stiffness increase, etc.
As far as inductance goes, keep in mind that a VC is simply a coil of wire. It is no different than any other inductor. An iron core inductor has much higher inductance. In a typical driver the coil functions as an iron core as long as the windings are next to the pole. At rest, part of the windings may be past the tip of the pole and you have an Le value based on that "core". As the coil goes inward now the entire coil has an iron "core" and inductance goes up. As it goes forward, there is less iron "core" and the inductance goes down. This is why typical drivers have an S shaped inductance curve. This means more upper frequency output on the outward stroke than the inward. This changing response also changes the phase of the driver. The changing phase and impedance curve changes how the driver interacts with a passive crossover network. This is a huge problem.
The solution is to either take the iron out of the core or make it seem like it isn't there so it becomes an air core at all points. Alloys like Hiperco 50 are magnetically permeable but not electrically conductive. The same goes for powdered iron. Either of these will make the coil act like an air core. The problem is Hiperco 50 is nearly impossible to machine and EXPENSIVE. I was looking to source a 2" diameter rod to make poles for a special TC woofer. To do the nearly 4" long pole would be over $800 just for the raw Hiperco50. Then it has to be OD ground to tolerance instead of machined. It would cost around $1500 just to make one pole piece. The powdered iron isn't so expensive, but it has it's own cost. While not electrically conductive, it is also about 20-25% less magnetically permeable than low carbon steel. This means a much larger motor would need to be used to get back to the same motor strength. It would nearly double the cost of the drivers.
The other option is to sleeve the pole with copper. Now all eddy currents that try to flow through the pole are quickly shorted out and the coil appears to be an air core inductor. No matter what position the coil is in, the Le does not change. Brandon Thill measured several of the TD woofers and after awhile even quit measuring Le(x). He commented: "Well this was the last TD driver I tested Le(x) for after seeing all the others what's the point? Put a well designed full copper sleeve on the polepiece and Le(x) will be excellent. In the case of ALL the TD drivers it's flawless. "
You can see the results from the TD15X here:
http://sites.google.com/site/drivervault/driver-measurements/15/ae-speakers-td15x/le-x
If you go back you can see that none of the other drivers from any manufacturer have this kind of Le(x) linearity.
As far as measuring the Le values, most measurement systems automatically measure at 1KHz. Here is an impedance curve from a TD15 though. You can see how flat it is at all points.
John
Last edited:
Yes ... 1/10 is very close to the actual production cost of any commercial speaker system.
The original G1 was using an RCF L18P300 woofer, but they have switched to an other unit now (B&C I think).
I sold all my rights regarding Griffin, and I cannot share any details with you. Sorry....
But anyhow - these days I would recommend a dipole subwoofer over anything else.
Sb, Fostex and acknowledgement
Dear Inductor, it feels wierd addressing a piece of copper wire wound around an iron bar, but anyway I am sorry.
Jokes apart - the dip in frequency at 500 Hz seen on the measurement does not correspond to any anormality in the impedance graph. I think it is a diffraction or measurement artifact.
Kris
I've posted SB Acoustics SB34NRX75-6 12" on post#6. http://www.diyaudio.com/forums/mult...-drivers-sealed-bass-channel.html#post2326193
...But you never thank-you-me.
Pay attention for graph of response curve (dB) measurement with "mic" at 31.6cm (or 1/3M).
Also look at this driver http://eminence.com/pdf/kappalite-3012lf.pdf
Dear Inductor, it feels wierd addressing a piece of copper wire wound around an iron bar, but anyway I am sorry.
Jokes apart - the dip in frequency at 500 Hz seen on the measurement does not correspond to any anormality in the impedance graph. I think it is a diffraction or measurement artifact.
Kris
AE woofers
Hi John, your drivers are always extremely appealing and the TD 15S as well as the SBP15 seems to yield a truly heroic bass response. However, the enclosure required is quite large - I am currently seeking something a wee bit smaller. I drive the 150 - 800 range with a 12" midrange unit 100 db sensitivity, so I do not need a woofer that can do 1 kHz clean. If I needed 1 K from a 12" or 15" I'd buy a AE speaker in a heartbeat.
Best
Kris
Hi John, your drivers are always extremely appealing and the TD 15S as well as the SBP15 seems to yield a truly heroic bass response. However, the enclosure required is quite large - I am currently seeking something a wee bit smaller. I drive the 150 - 800 range with a 12" midrange unit 100 db sensitivity, so I do not need a woofer that can do 1 kHz clean. If I needed 1 K from a 12" or 15" I'd buy a AE speaker in a heartbeat.
Best
Kris
LAT 250
Hello Doug,
LAT 250 - How do you think they handle the 100 - 300 Hz region? I will most likely cross actively 1. order or 2. order at 140 or 150 Hz so they have to be well behaved at least an octave above this point.
Kris
Hello Doug,
LAT 250 - How do you think they handle the 100 - 300 Hz region? I will most likely cross actively 1. order or 2. order at 140 or 150 Hz so they have to be well behaved at least an octave above this point.
Kris
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.