Can be a can of worms, using a laser technically tells you when 10% distortion actually
happens. So as a manufacture you would tend to prefer laser.
Usually laser measured numbers are higher than the gap/coil height equation.
I tend to respect manufacturers that publish coil and gap height.
Think many are trying to stay away from ridicule.
If you published gap/coil equation and laser.
The laser number would be higher and people question it.
I guess people in the know how. Can also question
manufactures that do not publish coil height.
They obviously prefer to publish the higher laser measured spec.
Otherwise as far as limited 16 ohm drivers on the market.
One point I was interested in finding same thing. But it struck
my curiosity with using series drivers.
Also spun a interest in series crossovers to maybe do .5 alignment
in series. To work around low availability of 16 ohm drivers in
certain price ranges.
Came to a point where sometimes 2 drivers was more affordable
than a single driver. And possible a .5 experiment could be fun
even a advantage sonically without draining the wallet
Last edited:
Based on the spec sheet, this satisfies all of your requirements.
https://www.lorantz.com.au/loudspeaker-products/loud-speakers/15-inch-bass-mid/ac386t-pa-500w/
It is not local to you (postage would be a killer), but does show that the product you want still exists. Maybe you can find a small manufacturer near you that has something similar, or contact some repair places to see if they can recone a pair of dead/salvaged PA woofers to your specs.
---
A single 8ohm 18" option may be less hassle to source, and quite a few are around the 98dB mark. Conveniently for your budget, high(ish) Qts is usually found in the cheaper / lower power part of the pro market, e.g.:
https://usspeaker.com/beyma 18wrs600-1.htm (currently available new)
https://usspeaker.com/paudio c18-650el-1.htm (maybe available 2nd hand)
A quad of cheap 8ohm drivers would also be easy to pick up. If your have plenty of space, and if application is open baffle or infinite baffle, using a small array of really cheap drivers can work well.
There's the acoustic mutual loading at the lowest frequencies, but then again there is no free lunch because at the frequencies above that the array of two will become more directional. Actually doesn't the on-axis sensitivity go up + 6 dB? I also remember an AES paper saying this was all due to superposition, not loading, but remember only dimly because it was a long time ago.
Another factor to consider is for the same woofer, the efficiency and Q won't be the same with different coils. My recollection is that while some makers might specify them as the same, that's a lie, the magnetics don't work like that. Also remember the DC resistance is on the bottom of Small's efficiency equation, so it's harder for the high impedance design to be as efficient.
However the easier load on the amplifier I think is a very good thing. Amplifier power ratings are all tested with resistors, but speakers aren't resistors, they can run an amp out of its SOA (Safe Operating Area) more easily. And, amp heat losses are inversely proportional to the square of the resistance, and the increased static power into resistors only increases 1-2 dB into lower and lower impedances. So running amps into low complex impedances I believe does not actually gain more dynamic power at clipping.
Last edited:
im in the same boat. looking for a 100db 15" woofer. dont think its really possible. should have went with dual 15's myself to fully meet my mids.
You mean in the context of tube amplifiers, right? From the amplifier side? Because given the same magnet structure and cone, different impedance coils will give different efficiency* for the woofer. That was what I meant above, that you can't balance out the turns in the gap versus the impedance to get the same parameters.
*midband efficiency. Huh, now I want to play with a simulator and see what happens just at the low frequencies, which to my recollection that's just dominated by the box/woofer compliance and the moving mass.
No, +6dB would be the figure for a pair of parallel drivers and a solid-state amplifier (i.e. both driven by the same voltage), so double the power. Which is not the case here.
The drivers become directional where diameter approaches half the wavelength of interest, so about 500Hz for 15" drivers.
Small's equation also includes l (the overall length of conductor in the magnetic field) - which will vary with voicecoil resistance. The important things is for Bil to be invariant (i being applied current), which it will be, unless the manufacturer messes around with motor geometries (which is possible, but I'm guessing rare).
- "not the case here" = tube amp, where the output transformer would thus output the same power on the new tap?
- BiL...yes, mmm, wait but if you change the impedance AND the voltage...too off topic. My recollection is those things won't change at the same pace but I'll start another thread if I can't find my coil simulator.
Another thought to run by y'all. I really like OB so I had an idea to do horn tweeter, highish efficiency MR or FR, and an Alpha 15 or similar all OB and then have a sealed woofer below that. The though was that I could set it up so that the sealed woofer could be biamped or alternately paralleled with the Alpha.
The basic idea is not only to add a little extra sensitivity to the bass but also to compensate some for the OB rolloff. The basic response of each is shown below (sim in Hornresp).
Alpha OB:
GRS 18PT Sealed:
The box on the 18 is sized so that the final rolloff is at about the same frequency as the Alpha. If the box is made a little bigger it will flatten out and give about another 10Hz before rolloff. In either case I can adjust the slope of the sealed response to more closely mirror the Alpha like this.
Does this seem like feasible thing to do?
The basic idea is not only to add a little extra sensitivity to the bass but also to compensate some for the OB rolloff. The basic response of each is shown below (sim in Hornresp).
Alpha OB:
GRS 18PT Sealed:
The box on the 18 is sized so that the final rolloff is at about the same frequency as the Alpha. If the box is made a little bigger it will flatten out and give about another 10Hz before rolloff. In either case I can adjust the slope of the sealed response to more closely mirror the Alpha like this.
Does this seem like feasible thing to do?
First point - correct. You only get 3dB for free when you double up drivers - the other 3dB is down to doubling the power.
Second point - you can take a 4 ohm voicecoil for example), double the wire length by halving its cross-sectional area. You end up with 4x the Re (i.e. 16 ohms), but this is balanced (in the equation) by the l-squared term, which has likewise quadrupled. So, efficiency remains the same.
OK that seems logical, though my post-COVID-booster addled brain is feeling "halving its cross-sectional area" ends up with a different length coil because the diameter won't be half (??). I started getting off my butt and rewriting my coil design spreadsheet which I'll start another thread for.
Efficiency no=(rho*B^2*L^2*Sd^2)/(2*pi*c*Mms^2*Re)*100%
I'm getting that from Wikipedia which does not define L (they show lower case l but then it looks like I=current so I put L). This is the length of wire intercepting B, right? My copy of Small's thesis is too buried to dig out right now; sun is going down and gotta walk the dog before coyotes come out.
So for these purposes changing only the coil, it's trading the L^2/Re.
Then it comes down to what you said. Maybe the "no free lunch" that I recall is you can't end up with the same coil overhang...gonna walk the doggie and look at my coil spreadsheet after that.
Voicecoil details don't change the fundamental principles - for the purposes of analysis, imagine wire with a rectangular cross-section (which maximises available space, as would the more popular hexagonal).
l is the accepted symbol for voicecoil length - as in Bl.
l is the accepted symbol for voicecoil length - as in Bl.
Can I ask you cognocienti a question: If I take a driver and rewind the voice coil with thinner wire and more windings so the impedance goes up, is there any change to other driver parameters like the Q's and Fs? Assuming I fill the voice coil former as it was before so it cuts the same number of magnetic flux lines.
Jan
Jan
Shouldn't do, though in practice there may be differing amounts of wasted voicecoil volume due to the wire's insulation. Did you have anything specific in mind?
I came across some designs with the vintage Philips 800 ohms speaker, that could be driven from a tube amp without an output xformer. This was in the 1960-ies, OTL avant la lettre.
Wondering how different such a speaker would be from an 8 ohms, apart from the coil of course.
Jan
Wondering how different such a speaker would be from an 8 ohms, apart from the coil of course.
Jan
Wow, that's an oddball! Is it intended for normal push/pull operation, or is it meant to be capacitively connected to a single-ended output stage? Actually, 800 ohms is quite low, implies quite high power.
Given that the OPT is the Achilles Heel of the valve amplifier (cost/performance), it sounds like a great idea (though potentially somewhat dangerous!)...
Given that the OPT is the Achilles Heel of the valve amplifier (cost/performance), it sounds like a great idea (though potentially somewhat dangerous!)...
Yes. There are some commercial builds kinda like this.
My living room build has been similar for a few years:
18" OB (80-700Hz)
sealed 2*15"s (<80Hz)
In a fairly big room, the 18" midbass is far more sensitive than the LF section. I used DSP to get level matching / flatter FR.
I've salvaged parts from an audio console, that had source, amp and stereo speakers built into a mid-century cabinet. The speakers were hundreds of ohms.
My impression was that this was not rare (i.e. not just a Philips thing).