Wow took a while to catch up... some good info. I think it may be worthwhile to try out an FR soon. As someone mentioned, the basic physics are hard to overcome. Still, there is hope, and it seems I should be checking out something other than the poor examples I've heard thus far...
I can only recall two times I thought subs were well integrated. both were multiple subs and both were carefully tuned x-overs. They were also high quality subs, not typical home theater boomboxes.
this has been my fear of FR. the stuff I'd seen in the past looked pretty nasty. but I've not paid much attention since then...
) I think there are some good designs around today. It is, as this thread highlights, debatable if an FR could replace a very good 2-way (given the same LF ability) But I think they are competitive. To really replace 2-way, they would need to overcome the inherent shortcomings of resonances and HF directivity and such. (even if they are tameable...)
They are usually hard to integrate. I decided it was not worthwhile. My current main speakers can go low enough...
I can only recall two times I thought subs were well integrated. both were multiple subs and both were carefully tuned x-overs. They were also high quality subs, not typical home theater boomboxes.
Yes. The bottom one or two octaves are not critical to musical enjoyment, though they are sometimes nice to have. I decided long ago accuracy was better than sloppy extension
this sounds promising. I've been meaning to invest in a good DSP that could be used to digitally x-over a 2-way. Now I have some new ideas
this has been my fear of FR. the stuff I'd seen in the past looked pretty nasty. but I've not paid much attention since then...
as someone else pointed out why would the original mfg not do this? is it really just cost? or is this simply the end listener doing some personal tailoring? It seems if such damping is required, the FR is either not so good to begin with or is not really FR?
this is the good part! (and maybe the bad and the ugly are manageable after all
) I think there are some good designs around today. It is, as this thread highlights, debatable if an FR could replace a very good 2-way (given the same LF ability) But I think they are competitive. To really replace 2-way, they would need to overcome the inherent shortcomings of resonances and HF directivity and such. (even if they are tameable...)I may yet
It is cost, and momentum. For instance, to do what i do would probably triple the build cost of the driver (an estimate based on direct conversation with a manufacturer). Given the multiplying factor of getting drivers into people's hands drivers would be far too expensive to make the business viable.
dave
I don't either. but only EQ for good, not for evil
...sort of like Andy's dislike for extra components - well it's not like we add them arbitrarily
OTOH, I've seen passive x-overs with an incredible number of components - each presumably solving some problem, but you have to stop and think that if there were that many problems, maybe the design was flawed from the start
while slightly OT, mind my asking what sort of EQ/DSP hardware you use? There are some new kits and products that look promising, and of course the seemingly ubiquitous Behringer stuff...
thx
On the subject of adding extra components like a passive crossover, I realise that my thinking is based on designing and building DHT amplifiers. So I'm kind of "extending" the way I'd design the amp into the crossover - trying to minimise and optimise parts.
Since this is a speaker forum, though, it's obvious that speaker designers will equally want to optimise the speaker design! So I recognise I'm thinking in a different framework.
Andy
Since this is a speaker forum, though, it's obvious that speaker designers will equally want to optimise the speaker design! So I recognise I'm thinking in a different framework.
Andy
It is critical when using a high output impedance amplifier like your SET that any affect of an XO on the impedance of the speaker does not cause it to get "bumpy" in respect to amplitude or phase. Complex XOs that create a wild ride impedance are the enemy of many amplifiers.
This can be used to advantage in dealing with baffle step.
dave
This can be used to advantage in dealing with baffle step.
dave
I don't think it's irrelevant to extend some degree of "amplifier think" to passive crossovers. In a SET you would go to the pains of direct coupling stages just to eliminate a 0.1uf cap. In speaker crossovers you'd almost take a 47uF cap for granted. So I guess I'm trying to "direct couple" to the speaker unit to put it one way. If this avoids problems with complex crossovers as you say, so much the better.
So are you saying that a typical 300b SET would very much like to see a full-range speaker unit?
Andy
So are you saying that a typical 300b SET would very much like to see a full-range speaker unit?
Andy
Umm...
I think I'd just avoid SETs, for their high output impedance. You can model the effect on the bass response in winISD (under the "signal" tab, play with the series resistance) - the result for one ported design was 7dB gain (from where the response was) at 80Hz. You could tell when listening to the result, too.
Mind you, it was a 12" co-ax playing. 5kHz crossover on the tweeter.
For smaller drivers, the lift in bass may be welcome, but the ability of the system to replicate exactly what went in will be hindered somewhat.
Modelling something fairly typical... FE127 in a ~13L cabinet, tuned to 64Hz (flat to a -3dB of 60Hz). A 4 ohm output impedance would yield 3dB gain at ~100Hz (giving a broad bump in response). This might sound pleasing, giving a fuller sound, perhaps. But once accustomed to a low output impedance, I'm sorry, but SETs just don't do bass as well.
For me, there is no way a 4-5" driver will ever reproduce <40Hz with decent volume. As soon as the cone starts flapping around, you start running into non-linearities - the midrange/treble reproduction has to suffer because of this.
So, instead of adding a tweeter, I've added subwoofers. Leave the bass to something that can hack it, properly. That way, there's no need to say "well, it's good [bass extension] for what it is". This is just good.
Chris
I think I'd just avoid SETs, for their high output impedance. You can model the effect on the bass response in winISD (under the "signal" tab, play with the series resistance) - the result for one ported design was 7dB gain (from where the response was) at 80Hz. You could tell when listening to the result, too.
Mind you, it was a 12" co-ax playing. 5kHz crossover on the tweeter.
For smaller drivers, the lift in bass may be welcome, but the ability of the system to replicate exactly what went in will be hindered somewhat.
Modelling something fairly typical... FE127 in a ~13L cabinet, tuned to 64Hz (flat to a -3dB of 60Hz). A 4 ohm output impedance would yield 3dB gain at ~100Hz (giving a broad bump in response). This might sound pleasing, giving a fuller sound, perhaps. But once accustomed to a low output impedance, I'm sorry, but SETs just don't do bass as well.
For me, there is no way a 4-5" driver will ever reproduce <40Hz with decent volume. As soon as the cone starts flapping around, you start running into non-linearities - the midrange/treble reproduction has to suffer because of this.
So, instead of adding a tweeter, I've added subwoofers. Leave the bass to something that can hack it, properly. That way, there's no need to say "well, it's good [bass extension] for what it is". This is just good.
Chris
I quite see your point, Chris, but I'm addicted! I'm getting good results with full-range units anyway.
I tried my SET on stacked Quad 57s and the bass was indeed lumpy. Mind, the mid and treble was simply glorious.
I take your point about subwoofers, and friends use them. I try and screw as much bass as I can out of a 5" or 6" unit, and I listen at lowish volumes in a smallish room so it's not a big deal. In a bigger room at higher volumes it would be a real issue.
Andy
Last edited:
This is called efficient engineering! If you study the sound system designs of the pioneers of audio you will find that thanks to having no preconceived notions about what a sound system's design should be beyond the known electro-mechanical-acoustical physical limitations, the need for the highest practical system efficiency and the 'deep pockets' of Bell Labs, RCA, MGM plus a few others to fund the research you'll be hard pressed to improve on them in any substantial way beyond what becomes available due to advances in materials, manufacturing technology.
Once you move past these initial designs you'll see that later 'improvements' other than strictly efficiency related, are all predominently profit/market driven with the insertion of negative feedback [-fb] being the first and most deleterious to accurate sound reproduction AFAIK.
WRT passive XOs, the more you separate the amp's design from the speaker's, the more complex the XO must be to properly bridge the electro-mechanical-acoustical gap between them, always trading efficiency for nominally flat BW.
The $64 k question in my mind then became 'how much sonic accuracy is potentially sacrificed using high power designs'. Until the Babb Lorelei, my experience had been 'way too much' hence my need for high electrical efficiency systems at minimum and multi-way ultra high efficiency for any critical listening. For everything else, a good set of headphones.
The Babb wasn't particularly budget friendly, especially since it couldn't be repaired if damaged, and still needed some fine tuning development, but its core design was solid, i.e. 16-20+ kHz usable BW, nominally flat sounding ~32-20+ kHz at high peak power [to offset its low efficiency] with no obviously audible thermal or mechanical compression to me and the few other DIYers in attendance.
Bottom line, you're in the right 'frame of mind', just apparently either not educated enough yet to design optimized sound systems and/or lack the relatively high budget required, be it size/WAF/$$$/whatever related.
GM
I also found, interestingly enough, that on the sims at least, a FE126E can be an almost perfect drop-in for a FE127E with 6.1ohms output impedance. (13L box tuned to 64Hz for both, the deviation was <0.2dB)
Perhaps, with the right additions, a FE126 can be a replacement for the FE127s? Variable Amplifier Impedance
I suspect amplifier choice plays a big part in which speakers we choose: multi-way cabinets are perhaps more suited to SS amplification, whereas the simpler load of a single driver allows the use of valve amplifiers, which are glowy.
Chris
Perhaps, with the right additions, a FE126 can be a replacement for the FE127s? Variable Amplifier Impedance
I suspect amplifier choice plays a big part in which speakers we choose: multi-way cabinets are perhaps more suited to SS amplification, whereas the simpler load of a single driver allows the use of valve amplifiers, which are glowy.
Chris
I'm educated enough to design and optimise SETs with DHTs - that's not an issue - but I'm nowhere near educated enough in speaker design, which clearly is quite an advanced black art. Since all I need to complement my SET is a good full-ranger, I happily don't have to venture far into the black arts (though I do try and read the theory and I've built several cabinets in my time). Hence my interest in how good full-rangers can get.
Best
Andy
Last edited:
OK, but I said sound systems [amp+speaker+wiring+XO [if any]]. 'Black art' it's not, just known physics and my 'take' on your remarks is that you're trying to find a 'FR' driver to best 'fit' your idea of an optimized SET whereas I'm saying its design must complement the driver and vice versa to create an optimized system.
GM
GM
I think you'd find in practical application, that such sims wouldn't tell the whole story. At least in the one case I tried it a few years back with the Brynn enclosure - even with "appropriate" series R ( somewhere under 4 ohms IIRC), the 126 didn't quite equal the midbass weight or dynamics of the 127. I'm sure there was something missing from my primitive efforts, but at the time I wasn't sufficiently motivated to further pursue the matter.
glowy is nice, but I'm not so sure all valve amps need to see the simpler load of a well damped SD speaker - certainly a 30-50W P/P pentode or UL KT88, etc with judiciously tailored feedback would have no issue muscling through the reactance of many a multi-way - how coherent and musically delicate that would sound to the ear of a listener "atuned" to a 2A3/BLH system is another question
Last edited:
Whether one calls it a "black art" or not, there's a very large degree of complex technology in speaker design, even if it's all known physics. It can seem a black art to those who haven't delved deep into the equations and theory. But you're right - my amp is already pretty optimised, so I'm looking for the closest match in a FR. That's a different way of looking at optimisation in in my particular case it suits my listening priorities. Wouldn't work for most others, and frankly I could count the number of people I know who have all-DHT SETs on the fingers of one hand so it's hard to make comparisons.
Andy
The article I posted above has some circuitry that allows the user to play with the output impedance. I might try it with my little tripath amp. Will report back if I do, but ~6ohms looks somewhere near optimum for the bass.
I'd agree that a high power (relatively speaking) valve amp would probably be fine with a large multiway design, but many of the full-range aficionados here have flea-watt SETs, so it was those I was referring to.
High output impedance can be a good thing. Current amplifiers for instance have the potential to dramatically reduce distortion. Get thee a copy of ETM's book, or dig out some of the papers on the subject.
Also given that most of a SET's distortion is 2nd order, careful system setup can have that canceled by the primarily 2nd order distortion in the loudspeaker.
One analytical study (wish i could find the URL) compared FE208s horn driven by a SET and an NAD with very low distortion measured into a resistor. The SET + speakers had dramatically lower combined distortion than the NAD + speakers.
dave
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.