Hello all,
I've been trying to find a suitable driver for use with my Alan Eaton 45 SET amp, and have been all around the world for several weeks, first trying to go the fullrange route, ordering a pair of Liisong's new PT-10 drivers. I cancelled that order when I discovered the Radian series of coax drivers, and settled on the 5210, a 10" cone with a 1" aluminum alloy compression driver. Here is the page for it: https://cdn.shopify.com/s/files/1/0111/0324/0254/files/5210C.pdf?v=1617035289</ul>
I decided to go with the Radian because of the success some people have had with them, particularly Frank Fazzalari of Coherent Audio. His speakers win high praise and not from the usual commercial circuit. There's also a company in Germany making beautiful speakers with these same drivers.
So. I've got lots of experience building and DIY'ing including amp mod stuff, but this is my first foray into a from-scratch design/build. I think I'm in the ballpark but "the more I learn, the less I know" has never been more true. Night after night I read until it makes no sense anymore. Recently read Martin King's work on MLTL/QW theory, and learned a lot although the math was about 40 years late for me. I'm needing assistance in two areas, the crossover and the boxes.
For the boxes, I've got a pretty simple idea for a BR enclosure, 2.2cu.ft/62l, or, if I can find a way to design it, a 3.3cu.ft/93l. Both of these were generated with free online software and compared to existing enclosures. I also got a version using Eminence software through US Speaker, where I purchased the drivers.
First - I don't know whether this driver is more at home in the TQWT or the BR, but I feel that a QW pipe with a full terminus would be a better choice. I don't know all the correct terms, so by 'full terminus' I mean where the throat opening is not restricted and is a continuation of the pipe cross section as in true Voigt pipes. I'd prefer the throat to be bottom front rather than an open base, as in Troels Graveson's designs. Try as I might, I cannot find a reliable resource to design this box, and would really appreciate some help. I will even pay someone to generate the design, as I'm way behind on my work and getting behinder all the time right now. But first I'd like to know if a simple bass reflex might be as good or better, though I doubt it. I prefer the bass response of the horn, not driving a narrow band slug at the terminus then falling away.
Second - the crossovers! I've plugged the T/S numbers in every free software I can find, and the a friend and coconspirator who has a Windows rig used Jeff Bagby's crossover designer as well, and I think I'm in the general ballpark but still not sure. I will post the FR of the drivers below, as well as the tweeter response with 5uF, 10uF, and 15uF caps on the tweeter (first order). Here are my questions about the crossover:
1. Frank declined to sell me his crossovers, and wouldn't give me the values, but we had a great conversation anyway, very good guy. Although he uses first order on both drivers, he said I might want to go second order on the tweeter. I would rather avoid an inductor on the tweeter mostly due to phase concerns, which is really high on my list and part of the reason I was originally going to use full-rangers. So first gate: first or second order?
2. With first order Butterworth, I've got pretty typical values at 1200hz, 1350hz, and 1500hz: 16.56uF/1.06mH(woofer), 14.72uF/.94mH, and 13.25uF/.85mH. Second order on the tweeter only lowers the cap values a bit and adds a small inductor.
As you will see in the frequency/impedance graph for these drivers, there's a really knotty issue due to a sharp impedance spike from the tweeter at just below 800hz, and this is really interfering with the falloff of the tweeeter and pushing up the FR almost all the way to the intended xover point (I'm using 1350 as a general target). I don't know how to remedy this, and am beyond reluctant to even go to 2nd order, let alone 4th. I'd rather start from scratch with a different driver, BUT this doesn't seem to bother Frank, he's got some way of dealing with the impedance that I don't understand. Zu audio is another example of a popular speaker using cheap Eminence drivers and squeezing out very good performance. I'm guessing these guys are manipulating the enclosures to make it all work?
Below I will post the following: The Radian EQ/Imp graph, and the tweeter response for first order using 4.7, 10, and 15uF values plugged into Jeff's program. Any and all guidance is much appreciated, and like I said earlier I'm willing to pay to learn here, I wish I had a mentor in my area but there is nobody doing this. Cheers, Bryan
I APOLOGIZE FOR THE LAST TWO IMAGES BEING ROTATED, I DON'T KNOW WHY THEY LOADED THAT WAY!
With 4.7uF cap, lower green line:
With 10uF cap:
15uF:
I've been trying to find a suitable driver for use with my Alan Eaton 45 SET amp, and have been all around the world for several weeks, first trying to go the fullrange route, ordering a pair of Liisong's new PT-10 drivers. I cancelled that order when I discovered the Radian series of coax drivers, and settled on the 5210, a 10" cone with a 1" aluminum alloy compression driver. Here is the page for it: https://cdn.shopify.com/s/files/1/0111/0324/0254/files/5210C.pdf?v=1617035289</ul>
I decided to go with the Radian because of the success some people have had with them, particularly Frank Fazzalari of Coherent Audio. His speakers win high praise and not from the usual commercial circuit. There's also a company in Germany making beautiful speakers with these same drivers.
So. I've got lots of experience building and DIY'ing including amp mod stuff, but this is my first foray into a from-scratch design/build. I think I'm in the ballpark but "the more I learn, the less I know" has never been more true. Night after night I read until it makes no sense anymore. Recently read Martin King's work on MLTL/QW theory, and learned a lot although the math was about 40 years late for me. I'm needing assistance in two areas, the crossover and the boxes.
For the boxes, I've got a pretty simple idea for a BR enclosure, 2.2cu.ft/62l, or, if I can find a way to design it, a 3.3cu.ft/93l. Both of these were generated with free online software and compared to existing enclosures. I also got a version using Eminence software through US Speaker, where I purchased the drivers.
First - I don't know whether this driver is more at home in the TQWT or the BR, but I feel that a QW pipe with a full terminus would be a better choice. I don't know all the correct terms, so by 'full terminus' I mean where the throat opening is not restricted and is a continuation of the pipe cross section as in true Voigt pipes. I'd prefer the throat to be bottom front rather than an open base, as in Troels Graveson's designs. Try as I might, I cannot find a reliable resource to design this box, and would really appreciate some help. I will even pay someone to generate the design, as I'm way behind on my work and getting behinder all the time right now. But first I'd like to know if a simple bass reflex might be as good or better, though I doubt it. I prefer the bass response of the horn, not driving a narrow band slug at the terminus then falling away.
Second - the crossovers! I've plugged the T/S numbers in every free software I can find, and the a friend and coconspirator who has a Windows rig used Jeff Bagby's crossover designer as well, and I think I'm in the general ballpark but still not sure. I will post the FR of the drivers below, as well as the tweeter response with 5uF, 10uF, and 15uF caps on the tweeter (first order). Here are my questions about the crossover:
1. Frank declined to sell me his crossovers, and wouldn't give me the values, but we had a great conversation anyway, very good guy. Although he uses first order on both drivers, he said I might want to go second order on the tweeter. I would rather avoid an inductor on the tweeter mostly due to phase concerns, which is really high on my list and part of the reason I was originally going to use full-rangers. So first gate: first or second order?
2. With first order Butterworth, I've got pretty typical values at 1200hz, 1350hz, and 1500hz: 16.56uF/1.06mH(woofer), 14.72uF/.94mH, and 13.25uF/.85mH. Second order on the tweeter only lowers the cap values a bit and adds a small inductor.
As you will see in the frequency/impedance graph for these drivers, there's a really knotty issue due to a sharp impedance spike from the tweeter at just below 800hz, and this is really interfering with the falloff of the tweeeter and pushing up the FR almost all the way to the intended xover point (I'm using 1350 as a general target). I don't know how to remedy this, and am beyond reluctant to even go to 2nd order, let alone 4th. I'd rather start from scratch with a different driver, BUT this doesn't seem to bother Frank, he's got some way of dealing with the impedance that I don't understand. Zu audio is another example of a popular speaker using cheap Eminence drivers and squeezing out very good performance. I'm guessing these guys are manipulating the enclosures to make it all work?
Below I will post the following: The Radian EQ/Imp graph, and the tweeter response for first order using 4.7, 10, and 15uF values plugged into Jeff's program. Any and all guidance is much appreciated, and like I said earlier I'm willing to pay to learn here, I wish I had a mentor in my area but there is nobody doing this. Cheers, Bryan
I APOLOGIZE FOR THE LAST TWO IMAGES BEING ROTATED, I DON'T KNOW WHY THEY LOADED THAT WAY!
With 4.7uF cap, lower green line:
With 10uF cap:
15uF:
Attachments
Hmm, since this amp isn't made by a big company I don't have any printed data, but I do know it's designed to really want to see an 8ohm load like a lot of small single ended triode amps, if that's any help. I think I understand what you're asking, but I find it odd that the Fs is 67 hz but that falls roughly between the two impedance peaks at 42hz and 90hz, is this because they're being summed somehow?
Another odd thing that keeps happening when my buddy is doing the modeling, is that if he sets the xover point at say 1350 in Jeff's software, the plotted crossover point shows up around 800hz. I'm new to this end of things so excuse my ignorance.
For example, the graph below is for second order on the tweeter, first on the woofer, for a target crossover at 1400:
He had to move the target to 2500 to get it to plot the crossover at 1400:
Another odd thing that keeps happening when my buddy is doing the modeling, is that if he sets the xover point at say 1350 in Jeff's software, the plotted crossover point shows up around 800hz. I'm new to this end of things so excuse my ignorance.
For example, the graph below is for second order on the tweeter, first on the woofer, for a target crossover at 1400:
He had to move the target to 2500 to get it to plot the crossover at 1400:
In free air does it actually measure 67 Hz since the mean between 42 - 90 = sqrt(42*90) = 61.48 Hz?
Note that manufacturer data can be its design spec or the mean between 'x' quantity samples or something else, though the pioneer's did quality assurance (QA) at +/- 10% of its design efficiency spec (n0).
Not really; my spreadsheet to accurately calculate it is long gone, but needs some of the amp's design, components data, though the manufacturer should know, yet too often in 'modern' times they just list its 8 ohm/whatever taps.
Anyway, here's some help if the manufacturer won't provide it or if willing to accept a potentially too large a box, then we can just use 8 ohms of series resistance + any for wiring, XO components, which increases its Qts': (Qts) + any added series resistance (Rs): http://www.mh-audio.nl/Calculators/newqts.html
(Rs) = 0.5 ohm minimum for wiring, XO components, so may be higher if a super small gauge is used as a series resistor and/or there's other series resistance.
Note that manufacturer data can be its design spec or the mean between 'x' quantity samples or something else, though the pioneer's did quality assurance (QA) at +/- 10% of its design efficiency spec (n0).
Thiele Small parameters equations – How each one affects the others
Anyway, those two peaks are the ones you ideally want a measured equal amplitude.Not really; my spreadsheet to accurately calculate it is long gone, but needs some of the amp's design, components data, though the manufacturer should know, yet too often in 'modern' times they just list its 8 ohm/whatever taps.
Anyway, here's some help if the manufacturer won't provide it or if willing to accept a potentially too large a box, then we can just use 8 ohms of series resistance + any for wiring, XO components, which increases its Qts': (Qts) + any added series resistance (Rs): http://www.mh-audio.nl/Calculators/newqts.html
(Rs) = 0.5 ohm minimum for wiring, XO components, so may be higher if a super small gauge is used as a series resistor and/or there's other series resistance.
Honestly GM, you lost me a bit here. I’m going to have to take Radian’s T/S numbers at face value and presume the Fs really is 67hz, and work with the values they’ve provided. Later on when I’m mic’ing and measuring, I’ll make physical adjustments to the components to bring things in line. I’ve looked at the links you provided, thank you, but it’s a lot to take in for a novice. Wouldn’t the output impedance of a small fleawatt tube SET amp be in the .5-1 ohm range?
Can this process proceed without knowing all the parameters of my system?
Can this process proceed without knowing all the parameters of my system?
Just sharing in case it helps with crossover design
https://www.mtg-designs.com/diy-speaker-plans/radian-5210
https://www.mtg-designs.com/diy-speaker-plans/radian-5210
Opening the PDF of the Radian speaker I read: Designed for High SPL applications...750W...3" voice coil.
So...what to Say? I'd Just Copy the available crossover schematic, and seeing a 3 mH coil on the woofer(2 on the tweeter) I would expect a narrow baffle, because of the BSC phenomenon.
Seeing the attenuation net on the tweeter, it'll play loud...together with the big attenuation of the woofer in the mid-treble range( probably).
A lot Better than the predicted 6dB slopes which intersect for a too Wide range
So...what to Say? I'd Just Copy the available crossover schematic, and seeing a 3 mH coil on the woofer(2 on the tweeter) I would expect a narrow baffle, because of the BSC phenomenon.
Seeing the attenuation net on the tweeter, it'll play loud...together with the big attenuation of the woofer in the mid-treble range( probably).
A lot Better than the predicted 6dB slopes which intersect for a too Wide range
Thanks pico. Since this is a pro audio speaker the crossover is designed mostly for the purpose of achieving a decent FR without regards to the same needs of the home audio community, ie dynamics, speed, phase coherence, etc, but more for protecting the drivers at high levels. The people who've succeeded in using these drivers for home audio scrap the stock xo design and always seem use very simple first and maybe second order (on the tweeter). I know it's a bit of a challenge but I'm trying not to step all over these drivers because the purpose of the project is to create a highly responsive speaker in the vein of big Altecs and other full range and coax/horn speakers. I could easily copy the Radian crossover or the one linked by vinee above, but it's taking me in the opposite direction. I know this can be done because I've seen it done, but don't have access to the designs used unfortunately. It may take a lot of fiddling with the enclosure once I get relatively close with the crossover, but I'm confident it can be done. The tweeter is very challenging though with the impedance spike at 800 hz.
I think the 12" and 15" versions are less problematic, but I really want to solve this as much as possible!
I think the 12" and 15" versions are less problematic, but I really want to solve this as much as possible!
Well, 2nd order on the tweeter is advisable. 3rd order with treble tilt (1uF//6Ω) might be too much. And the woofer might suffer ( !!) from that big coil...I mean, it kills most of the midrange so how is the tweeter asked to come into play? A balanced response from both drivers, as the norm, is advisable ( again). Otherwise it makes no sense to use a 300 W driver and attenuate it where major efficiency shines. But that are the skills for good crossover...and so this thread belongs to Multy-Way !
I done some reading and Coherent Audio model 10 have modified Radian coax driver. Some differences are visible on the pictures, blue cone and accordion suspension, while stock 5210 shows black cone and double half-roll suspension.
Even if they sell you the crossovers or give you the schematics its not the same thing beacuse these are not the same drivers.
Also, with stock 5210, average SPL of HF section is something like 7-8 dB above LF section, so you need some HF attenuation. If they only have a capacitor for HF, then it seems like HF section is also somehow modified, so it can match LF level in such way, just with single cap.
My point is, you need more complex crossover for stock unit, like the one mentioned in post #6.
Btw, judging by the component values in that crossover, it looks like its made for 5210 driver variant with 16 Ohm HF section.
The second image in your post #3 looks promising.
Even if they sell you the crossovers or give you the schematics its not the same thing beacuse these are not the same drivers.
Also, with stock 5210, average SPL of HF section is something like 7-8 dB above LF section, so you need some HF attenuation. If they only have a capacitor for HF, then it seems like HF section is also somehow modified, so it can match LF level in such way, just with single cap.
My point is, you need more complex crossover for stock unit, like the one mentioned in post #6.
Btw, judging by the component values in that crossover, it looks like its made for 5210 driver variant with 16 Ohm HF section.
The second image in your post #3 looks promising.
They just cut the dust cap off on the drivers. They are the same.
I’m BShaws bud that’s helping him, but I’m a novice as well.
I’ve been using winPCD and I’ve come up with a pretty good solution I think.
1st order on the LF and second order HF.
Crossing at 1300. Any issues/problems/suggestions from you pros?
I’m BShaws bud that’s helping him, but I’m a novice as well.
I’ve been using winPCD and I’ve come up with a pretty good solution I think.
1st order on the LF and second order HF.
Crossing at 1300. Any issues/problems/suggestions from you pros?
Attachments
Thanks Davor, I hadn't noticed the suspension, you're absolutely right. That change alone changes the T/S numbers quite a bit I'm sure. What Cam says is true, Frank is ordering the drivers without the dustcap, I think in earlier times he cut them out if I am not mistaken.
That second image you refer to is where we had the issue of the crossover shifting to the left on the graph, Cam had to push the input values to 2500 hz to get it to behave like that! But he realized there was some volume matching issues causing that, so reworked things and came up with the latest scheme, which looks fantastic and crosses at 1300, using a 5.6uF cap and .64mH inductor on the tweeter, and a really small .25 inductor on the woofer. I was surprised especially by the woofer value, but the impedance is rising rapidly through the crossover area so maybe that's making it easier to choke the woofer off?
I'm learning fast here, but still in the dark forest for the most part. Loudspeaker Cookbook will be here tomorrow, wish I'd bought it two years ago.
An L-Pad on the tweeter for attenuation may also help flatten the impedance.
If that's not enough, another approach would be an LCR series notch filter across the driver terminals. Something around 37 µF, 1 mH, and 8 ohms should be in the ballpark.
Last edited:
Yeah, lots to know to do an optimized tube system in general and especially a low damping factor (DF), no negative feedback (-fb) audio system and why the pioneers marketed only optimized systems. That said, fast forward to renovating vintage gear and modern SET amps 'craze', many folks are so 'enamored' with its 'syrupy' sound that it's plenty good enough as is, at least till they hear an optimized high power one for 'x' mass quantities of $$ more and forced to either learn how to DIY or accept what one has.
Regardless, when I posted I didn't see that the double impedance hump was actually a cab's tuning, which is around 60 Hz with your goal being to tune to actual Fs wherever it is, but you'll know it when the these two peaks are equal amplitude, though as you noted, it's best to start with published Fs if not tested and move tuning up/down as required.
No, that would be a OTL, PP or even early SS amp values; otherwise (much) higher with most of the ones I once owned being ~matching impedance for cheapness, max bass speaker output using 'FR' drivers. Anyway, remembered I have a lot of links if you want to go deeper into it and/or measuring yours if interested and there's forums here and elsewhere devoted to tube systems.
Again, not really due to its high output impedance, i.e. no -fb, so as I previously noted this added series resistance will effectively increase the driver's total Qts'.
For example, use these basic max flat vented alignment's formulas (note Qts' dominates!) to calculate with no series resistance and then using the link to calculate it with 8 ohms added to see how much bigger, lower tuned it will be and how mis-tuned both will be compared to one tuned to actual Fs:
Vented net volume (Vb) (L) = 20*Vas*Qts'^3.3 (Ft^3 = (Vb)*~28.31685)
Vented box tuning (Fb) (Hz) = 0.42*Fs*Qts'^-0.96
F3 (Hz) = Fs*0.28*Qts^-1.4
So I think I"m getting close. Thanks mattstat you were right, I just can't get these to play together without an l-pad. So right now it's second order on both drivers, plus the l-pad on the tweeter, crossed at 1500. Will start with a pretty good sized ported enclosure for testing and work from there.
I've been furiously reading the loudspeaker design cookbook, wow what a great resource! The only thing I stumble on a bit is his explanations of phase responses with the different types/orders of crossovers, some of the language isn't intuitive to me.
I've been furiously reading the loudspeaker design cookbook, wow what a great resource! The only thing I stumble on a bit is his explanations of phase responses with the different types/orders of crossovers, some of the language isn't intuitive to me.