Coming from this thread which unfortunately only had a couple of replies, I pass on the question to you tube amplifier cracks: what would the perfect multi-way speaker for a tube amp look like?
Long story short: I am looking into building a multi-way speaker which makes life as easy as possible for a tube amplifier. Since I know about speakers but don't have too much of a clue about what a tube amp really would love to see at its output terminals, I hope that you guys can help me gain that knowledge. There is no specific amp I have in mind, so let's just note that we are talking about an amplifier with at least a couple of watts per channel. As for the speaker, I am thinking of a standard floorstander design, likely 3-way, with a maximum bass driver size of 12". So basically, I want to make a "normal" floorstander with the exception of optimising it for tube amps.
What I am aware of are the basics such as a linear and preferably not-too-low impedance and high efficiency. What I couldn't wrap my head around yet are things like coils and the inevitable back electromotive force (BEMF) they send to the amplifier.
Do inductors in the crossover play a role similar to the voice coils of the drivers? Should I optimise the drivers for low Le (voice coil inductance)? Or rather avoid voice coils at all if possible (magnetostatic ribbon tweeter)? Does the moving mass of the drivers (Mms) play a role in storing energy and returning it as BEMF?? Does a high magnetic power of the motor (BxL) raise the BEMF and thus should be avoided? Would 96dB/1W/1m and 4 ohms (two 8-ohm bass drivers in parallel) be better or worse than 90dB/1W/1m and 16 ohms (two 8-ohm bass drivers in series)?
Questions and more questions, but no definitive answers yet. Can you help? 😕
Long story short: I am looking into building a multi-way speaker which makes life as easy as possible for a tube amplifier. Since I know about speakers but don't have too much of a clue about what a tube amp really would love to see at its output terminals, I hope that you guys can help me gain that knowledge. There is no specific amp I have in mind, so let's just note that we are talking about an amplifier with at least a couple of watts per channel. As for the speaker, I am thinking of a standard floorstander design, likely 3-way, with a maximum bass driver size of 12". So basically, I want to make a "normal" floorstander with the exception of optimising it for tube amps.
What I am aware of are the basics such as a linear and preferably not-too-low impedance and high efficiency. What I couldn't wrap my head around yet are things like coils and the inevitable back electromotive force (BEMF) they send to the amplifier.
Do inductors in the crossover play a role similar to the voice coils of the drivers? Should I optimise the drivers for low Le (voice coil inductance)? Or rather avoid voice coils at all if possible (magnetostatic ribbon tweeter)? Does the moving mass of the drivers (Mms) play a role in storing energy and returning it as BEMF?? Does a high magnetic power of the motor (BxL) raise the BEMF and thus should be avoided? Would 96dB/1W/1m and 4 ohms (two 8-ohm bass drivers in parallel) be better or worse than 90dB/1W/1m and 16 ohms (two 8-ohm bass drivers in series)?
Questions and more questions, but no definitive answers yet. Can you help? 😕
Last edited:
Good question.....I've been wondering the same thing.......I suppose you may have to narrow the scope a little. For example, amps without feedback will have high output impedance (on the order of a couple of ohms) and would be happier from a damping factor standpoint with higher impedance speakers.
So series wired doesn't do anything for increasing sensitivity, but gets better damping factor and parallel wiring gets the opposite. Then, 4 drivers in series-parallel gets some of both but does driving 4 voice coils with a flea powered amp make sense? Seems like the tail will be wagging the dog..
So series wired doesn't do anything for increasing sensitivity, but gets better damping factor and parallel wiring gets the opposite. Then, 4 drivers in series-parallel gets some of both but does driving 4 voice coils with a flea powered amp make sense? Seems like the tail will be wagging the dog..
Last edited:
That's the problem exactly. I kind of understand that all of these things matter, but I don't know which ones dominate others and therefore should be regarded the most.
Design a speaker for a reasonable source impedance of 1 - 2 ohms, don't assume a perfect voltage source as most speaker designers and software do.
Note that a complex X-O can be difficult for small SE amps to drive, so relative simplicity is important.
This also means that drivers should have reasonably flat impedances or ones that are easily corrected with zobel networks.
Load impedance should not drop appreciably below 8 ohms at any frequency. (Say 6 ohms)
Efficiency should be reasonably high. (Say >94dBSpl)
All this is quite arbitrary, but are similar to the sort of criteria I have used in designing my speaker systems.
There is no one speaker that is ideal for all situations or amplifiers, you will have to make some decisions about what will be compatible and what will not.
Note that a complex X-O can be difficult for small SE amps to drive, so relative simplicity is important.
This also means that drivers should have reasonably flat impedances or ones that are easily corrected with zobel networks.
Load impedance should not drop appreciably below 8 ohms at any frequency. (Say 6 ohms)
Efficiency should be reasonably high. (Say >94dBSpl)
All this is quite arbitrary, but are similar to the sort of criteria I have used in designing my speaker systems.
There is no one speaker that is ideal for all situations or amplifiers, you will have to make some decisions about what will be compatible and what will not.
See, this is one of the questions I ask myself.
Think 8" woofer. Let's assume I have two choices: a proper woofer and a fullrange which is sturdy enough to be used as a woofer. The fullrange one has a smaller voice coil and a weaker motor, which means that the coil has a much smaller inductance and the driver will be controlled by the amp much easier. Sounds nice. But to make the driver work with a midrange, I need to use a large coil in the crossover to linearise the frequency response. The other driver, the proper woofer, would produce the same result with a smaller coil. So is it better to use the fullrange and cope with 4,7mH, or does the amp like a proper woofer with 2.2mH inbetween more? I have no clue! 🙁
Unfortunately I'm not deep enough into tube amps to give a list of specific amps the speaker should work with. And anyway, it should work fine with a range of amps, so let's just say it should be optimised to be as "friendly" as possible to any amplifier.
The zenith console I have that has a tube amp in it has a 10 inch and 2 3 inch speakers one acting as mid and one as tweeter.
I would pretty much stay away from ribbons due to their tendency to have impedance changes with frequency they are playing.
That's all I know.
Cool.
Srinath.
I would pretty much stay away from ribbons due to their tendency to have impedance changes with frequency they are playing.
That's all I know.
Cool.
Srinath.
nelson pass once said that its very hard to better(if at all) the sound of a current output valve amp with full range speakers.............
hav'ing said that i would go for 2x 6" per channel for the bass, angled so the centre of the outputs(sound)of the speakers meets at about 7' from the drives. with say a 3" and tweeter dome, keeping all the drives as close as possible on the front of the cabinet which would need to be an odd shape.
also a 3way 2nd order x over should sound ok
hav'ing said that i would go for 2x 6" per channel for the bass, angled so the centre of the outputs(sound)of the speakers meets at about 7' from the drives. with say a 3" and tweeter dome, keeping all the drives as close as possible on the front of the cabinet which would need to be an odd shape.
also a 3way 2nd order x over should sound ok
Attachments
My take off the top of my head is that if you are not going to biamp then you should go for efficient midwoofs with a broad smooth frequency response and a smooth impedance curve. The goal is to allow it to run up relatively high and use the simplest low pass possible (preferably first order). If you can get a full range with solid low end performance so much the better. In some cases no low pass is needed at all and this is ideal IMO.
Now you will need a tweeter (or helper tweeter) of high efficiency that can be run low enough to blend smoothly with the woofer with a simple high pass filter. Again first order is preferred. A nice compression horn might be a good choice here.
For enclosure I would suggest either open baffle or something that tends to flatten the impedance curve such as aperiodic, back loaded horn or 1/4 wave (TL).
The more appropriate the drivers chosen are the less complicated the crossover will need to be which is desirable.
Now you will need a tweeter (or helper tweeter) of high efficiency that can be run low enough to blend smoothly with the woofer with a simple high pass filter. Again first order is preferred. A nice compression horn might be a good choice here.
For enclosure I would suggest either open baffle or something that tends to flatten the impedance curve such as aperiodic, back loaded horn or 1/4 wave (TL).
The more appropriate the drivers chosen are the less complicated the crossover will need to be which is desirable.
Thank you for your inputs srinath, pointy & mashaffer! 🙂
The approach I want to take however is a bit of a different one. The idea is to build a capable three-way floorstander system which any average listener would accept to use in his living room, but which has been optimised to be as gentle as possible on the amplifier while offering the merits of a normal 3-way design like a reasonably deep and tight bass, crisp mids and a clear treble. So the idea is not so much to design a speaker which is perfect for a tube amp (which would be a horn with a fullrange driver) rather than asking how one can optimize a 3-way floorstander to work best with a delicate tube amp.
Here's what I have in mind right now, just a prototype design growing in my head:
Bass: Ciare HX 201
Actually an 8" full range driver, but quite capable of deep bass, too. Very light diaphragm, high efficiency, low voice coil inductance, fairly soft suspension, surprisingly resilient in terms of input power
Mids: Fostex FE126En
A 4" fullrange as well as a great mid driver. High efficiency, very light diaphragm, very soft suspension, very low voice coil inductance
Tweeter: a random magnetostatic ribbon tweeter like the Swans RT2E-A
No voice coil means no inductance, no inbuilt transformer (as opposed to "true" ribbon tweeters), virtually no mass and thus no BEMF
This should make a well-proportioned floorstander with a real world efficiency of 90-92 dB/1W/1m and 8 ohms which should be very easy to propel for any amplifier, given my understanding of the matter is correct...
EDIT: Oh, and there's the enclosure type question as well: would a tube amp benefit from a bass speaker working in a sealed enclosure as opposed to a vented enclosure? In a sealed enclosure, the driver's excursion will be mechanically limited by the enclosed air for all frequencies. Also, the impulse response is known to be better, which that the diaphragm will stop moving faster and easier after the impulse as well. In a bass reflex enclosure, the driver typically behaves like in a free air setup below the fb of the enclosure. Also, the impulse response typically is worse. In my understanding, a bass driver therefore will produce more BEMF in a vented enclosure, which is why a sealed enclosure might be a better choice.
The approach I want to take however is a bit of a different one. The idea is to build a capable three-way floorstander system which any average listener would accept to use in his living room, but which has been optimised to be as gentle as possible on the amplifier while offering the merits of a normal 3-way design like a reasonably deep and tight bass, crisp mids and a clear treble. So the idea is not so much to design a speaker which is perfect for a tube amp (which would be a horn with a fullrange driver) rather than asking how one can optimize a 3-way floorstander to work best with a delicate tube amp.
Here's what I have in mind right now, just a prototype design growing in my head:
Bass: Ciare HX 201
Actually an 8" full range driver, but quite capable of deep bass, too. Very light diaphragm, high efficiency, low voice coil inductance, fairly soft suspension, surprisingly resilient in terms of input power
Mids: Fostex FE126En
A 4" fullrange as well as a great mid driver. High efficiency, very light diaphragm, very soft suspension, very low voice coil inductance
Tweeter: a random magnetostatic ribbon tweeter like the Swans RT2E-A
No voice coil means no inductance, no inbuilt transformer (as opposed to "true" ribbon tweeters), virtually no mass and thus no BEMF
This should make a well-proportioned floorstander with a real world efficiency of 90-92 dB/1W/1m and 8 ohms which should be very easy to propel for any amplifier, given my understanding of the matter is correct...
EDIT: Oh, and there's the enclosure type question as well: would a tube amp benefit from a bass speaker working in a sealed enclosure as opposed to a vented enclosure? In a sealed enclosure, the driver's excursion will be mechanically limited by the enclosed air for all frequencies. Also, the impulse response is known to be better, which that the diaphragm will stop moving faster and easier after the impulse as well. In a bass reflex enclosure, the driver typically behaves like in a free air setup below the fb of the enclosure. Also, the impulse response typically is worse. In my understanding, a bass driver therefore will produce more BEMF in a vented enclosure, which is why a sealed enclosure might be a better choice.
Last edited:
This may serve as a starting point. Mostly general concepts, but well written and good info.
Symphony Sound - Tube Friendly Speakers
Symphony Sound - Tube Friendly Speakers
Well, it already cleared up that I used the term efficiency when I actually meant sensitivity. 😀
Will absorb that article, thank you!
Will absorb that article, thank you!
Hi Interceptor
I'm in the process of building a 4way where each driver is driven buy its own SET amp .The OPT's are wound to suit each drivers impedence .
Heres a pic of the top section's cabinet , drivers will be Raven1 & PHL's .
3D sketch of 300B SETs
Cheers
I'm in the process of building a 4way where each driver is driven buy its own SET amp .The OPT's are wound to suit each drivers impedence .
Heres a pic of the top section's cabinet , drivers will be Raven1 & PHL's .
An externally hosted image should be here but it was not working when we last tested it.
3D sketch of 300B SETs
An externally hosted image should be here but it was not working when we last tested it.
Cheers
Yes, personally I would avoid normal ported enclosures for your application. So as I understand it you are looking to build the tall skinny style towers and not large enough to offend postmodern sensibilities. Looking at your drivers I see why you are thinking three way.
I found other documentation on that tweeter that indicates that a 3K crossover is as low as you want to go. It didn't say what order but I assume they are thinking at least 2nd order so if you wanted to go first order then it would have to be higher yet. That is a neat looking tweeter and is very intriguing.
It looks like the woofer might have some breakup starting at about 600 Hz so it may not be too fond of running up very high. I like the fostex driver as a mid/full range but it seems to me that the woofer is not going to provide as much extension and power as you could get by going just a little bit larger. With that high fs and low xmax the added grunt looks to be minimal. I think I would be looking for a 10" with a little more low end squirt if I was going to go three way for more extension and power on the bottom with less overlap in functionality.
In a 1/4 wave tower you still might be able to eliminate the woofer if you used an 8" FR. A third option that comes to mind is to use a plate amp in the base of each speaker to drive an 8 or 10 inch subwoofer in the lower part of the tower and let the tube amp drive the fostex and tweeter only to ease the load on it. The only thing with that approach is of course the need to plug the speaker into a wall outlet but the resulting power an clarity ought to be stunning.
I found other documentation on that tweeter that indicates that a 3K crossover is as low as you want to go. It didn't say what order but I assume they are thinking at least 2nd order so if you wanted to go first order then it would have to be higher yet. That is a neat looking tweeter and is very intriguing.
It looks like the woofer might have some breakup starting at about 600 Hz so it may not be too fond of running up very high. I like the fostex driver as a mid/full range but it seems to me that the woofer is not going to provide as much extension and power as you could get by going just a little bit larger. With that high fs and low xmax the added grunt looks to be minimal. I think I would be looking for a 10" with a little more low end squirt if I was going to go three way for more extension and power on the bottom with less overlap in functionality.
In a 1/4 wave tower you still might be able to eliminate the woofer if you used an 8" FR. A third option that comes to mind is to use a plate amp in the base of each speaker to drive an 8 or 10 inch subwoofer in the lower part of the tower and let the tube amp drive the fostex and tweeter only to ease the load on it. The only thing with that approach is of course the need to plug the speaker into a wall outlet but the resulting power an clarity ought to be stunning.
If you are going to design a speaker and an amp then design the two to work together. This means having more than one amp. One per drive unit in fact. In other words do not think about 'amp design' or 'speaker design', but instead think about 'system design'.
Design the complete system as one coherent whole.
Design the complete system as one coherent whole.
Gazumped!
Damn, Gordy said everything I was going to say!
Oh, wait a minute, he didn't say "solid state". Solid state makes a lot of sense for bass. Keep the valves for 200Hz upwards.
Damn, Gordy said everything I was going to say!
Oh, wait a minute, he didn't say "solid state". Solid state makes a lot of sense for bass. Keep the valves for 200Hz upwards.
To me the perfect speaker for a tube amp is a 16 or 8 ohms fullrange driver, in second place a speaker like the Klipsch RF83, two-way, 100dB SPL, 8ohms compatible and it is not too expensive.
RF-83 Floorstanding Speaker - RF-83 --- Floorstanding Stereo Speaker - Klipsch Reference Series - Overview
RF-83 Floorstanding Speaker - RF-83 --- Floorstanding Stereo Speaker - Klipsch Reference Series - Overview
@56oval: wow, that will be an awesome setup. Would love to listen to that! 🙂
@EC8010: thank you for your input as well, but I need it all passive and versatile.
Sensitivity (SPL at 1 meter with 2.8 volts of pink-noise input): 96 dB
Impedance (minimum/nominal): 2.8 (@180 Hz)/11 ohms
Tall and skinny is not exactly what I am looking for, I'd accept a 12" bass driver on the front of the speaker. That's the absolute max however, I'd rather go with an 8" or a 10" (or two of each for that matter). But I want to design what one would call a classic 3-way floorstander.
The tweeter was just an example. Most of the magnetostatic tweeters of that kind are very much alike. And since I plan to use a fullrange as a mid, a high crossover frequency isn't a problem anyway.
I'd happily accept a 10" woofer if I were to find one which fulfils the needs. Unfortunately, I'm not even sure whether the stats I am looking for are the ones I need.
It would, but I'd love to go all passive if I can.
It would be a perfect solution, but it's not an option. I need a speaker which will work with the majority of average tube amps as it is. Thanks for the input though! 🙂
@EC8010: thank you for your input as well, but I need it all passive and versatile.
Unfortunately, the Klipsch RF-83 actually meets neither of those figures. 100dB/1W/1m, 8 ohms and -3dB at 29Hz simply is not possible with this setup. The figures measured in this review are much more believeable:
Sensitivity (SPL at 1 meter with 2.8 volts of pink-noise input): 96 dB
Impedance (minimum/nominal): 2.8 (@180 Hz)/11 ohms
Hi,
Do you think the Klipsch factory datasheet is lying?? Or there is explanation for this larges differences??
Klipsch always seems a good manufacturer to me. This SoundVision test inform the freq.range graphic and SPL level was made on a home room, factory tests may not.
I read a thread where a RF83 owner say he drive the RF83 with 1 Watt with fantastics results and the low impedance looks does not affect the tube amp.
Regards, Gustavo
Given your need for sensitivity you might have to go with a ported or passive radiator system for the part under 200 Hz. Maybe using a pro driver like this one or similar.
Parts-Express.com:Eminence Kappa Pro-12A 12" Cast Frame Driver | kappa pro-12a 12" pa driver 12" bass speaker pa woofer cast frame driver
Just another angle to consider. I haven't run any sims on this driver but I bet you could find a suitable one amongst the many out there.
Parts-Express.com:Eminence Kappa Pro-12A 12" Cast Frame Driver | kappa pro-12a 12" pa driver 12" bass speaker pa woofer cast frame driver
Just another angle to consider. I haven't run any sims on this driver but I bet you could find a suitable one amongst the many out there.
- Status
- Not open for further replies.