AudioTechnology drivers are only available direct, as I recall.
Re the Super5, Harbeth out of necessity major on their history & many (not all) of the design principles Harwood et al created at the BBC, or in their own time outside. The Super5 is in a sense a loose descendent of the distantly related Spendor BC1. Back when the BC1 was introduced, the HF1300 tweeter couldn't make it much past 13KHz, hence the reason for the Coles supertweeter. Fast forward 3+ decades, there's no problem with the 1in dome tweeter getting to 20KHz & beyond, but tradition does count and some in the commercial market still like both the look and the idea of a smaller additional HF driver, so Spendor and Harbeth alike continue to make a product with that general layout. To be honest, it's inference on my part that Harbeth aren't running a separate XO for the the second HF unit, so I'm speaking under correction, but if there is, there isn't much sign of it in the measurements that I can make out, and what you certainly can see is a large amount of interference between the two HF drivers in the final octave because they're both covering the same BW: Harbeth Super HL5plus loudspeaker Measurements | Stereophile.com
Re the Super5, Harbeth out of necessity major on their history & many (not all) of the design principles Harwood et al created at the BBC, or in their own time outside. The Super5 is in a sense a loose descendent of the distantly related Spendor BC1. Back when the BC1 was introduced, the HF1300 tweeter couldn't make it much past 13KHz, hence the reason for the Coles supertweeter. Fast forward 3+ decades, there's no problem with the 1in dome tweeter getting to 20KHz & beyond, but tradition does count and some in the commercial market still like both the look and the idea of a smaller additional HF driver, so Spendor and Harbeth alike continue to make a product with that general layout. To be honest, it's inference on my part that Harbeth aren't running a separate XO for the the second HF unit, so I'm speaking under correction, but if there is, there isn't much sign of it in the measurements that I can make out, and what you certainly can see is a large amount of interference between the two HF drivers in the final octave because they're both covering the same BW: Harbeth Super HL5plus loudspeaker Measurements | Stereophile.com
Last edited:
jimbones, I wasn't trying to upset you! Just my humour. But I do hate threads that drag on for hundreds of pages. I forget what we are trying to do after a while.
But seriously, to imitate a Harbeth 30 is quite an undertaking.This is a mature and complicated design with, IMO, some unique design features that most people at this forum simply don't understand. I think the way forward here is the BBC LS5/9 design refined:
BBC and Rogers Loudspeakers › BBC LS5/9
Like the BBC LS3/5A, it's unique sound probably comes from the 90 degree Butterworth 3 crossover. Not a single person at this forum has ever agreed with me, but I am probably right. I have the advantage of having heard these things too. And can build BW3. And very nice it is.
I couldn't really care less exactly what woofer you use, they are all much the same in the end amongst the well-behaved variety. 1" tweeters are all much the same too.
You do end up with some fairly alarming power response and dispersion anomolies at the top end, but that is the sound, and it's OK.
But seriously, to imitate a Harbeth 30 is quite an undertaking.This is a mature and complicated design with, IMO, some unique design features that most people at this forum simply don't understand. I think the way forward here is the BBC LS5/9 design refined:
BBC and Rogers Loudspeakers › BBC LS5/9
Like the BBC LS3/5A, it's unique sound probably comes from the 90 degree Butterworth 3 crossover. Not a single person at this forum has ever agreed with me, but I am probably right. I have the advantage of having heard these things too. And can build BW3. And very nice it is.
I couldn't really care less exactly what woofer you use, they are all much the same in the end amongst the well-behaved variety. 1" tweeters are all much the same too.
You do end up with some fairly alarming power response and dispersion anomolies at the top end, but that is the sound, and it's OK.
what is the difference between "energy frequency response" and "total amplitude"
I got the others. Thanks
I got the others. Thanks
That is about the quintessence of the remark I made earlier. The energy response (actually the power response) determines the in-room sound you will experience. That is because you not only hear the direct sound from the speaker but also the reflected sound from boundaries around it (the room). With systems like these Harbeths, there generally will be a dip in the power response at about 1 to 2kHz. A lot of us like this characteristic, probably because it sounds well at lower volumes. Which again can be explained from the sensitivity of our ears, that vary with frequency and with level. Long story short: it’s the acoustic flavor one likes (and not some magical cone). Steve probably is right about the 3d order contributing to it all, but most of this can be explained by simple acoustics.
I measured a pair of Harbeth SLH5 and there is a HUGE dip at 3.2 KHZ. I do not believe it is not a error in my measurements because I measure my speaker in the same exact location and mine measure flat.
Thanks for the clarification Mark B
Thanks for the clarification Mark B
I should ask this. My own DIY speakers do not have the dip at 1-2 khz as you mention. I guess I could do a simulation to give me that dip but then isn't it like throwing darts? sim it, build it, listen, repeat?
I would have thought Alan Shaw gets things right, but there was a thing called the BBC dip (at crossover) that wasn't too offensive to the ears:
Harbeth Super HL5plus loudspeaker Measurements | Stereophile.com
This one is a phase aligned fourth order LR4. It falls away very badly around 3kHz off-axis because the woofer is beaming. And in fact LR4 has a power hole at crossover just by its -6dB crossover nature, though it measures flat on axis.
The BBC idea is to use a BW3 response. The phase at crossover is aligned 90 degrees, which flattens what would otherwise be a 3dB peak at crossover. But it fills the power hole by 3dB. Which I think is a good thing. At this point people usually mention asymmetric lobing, but I don't care about that. BW3 tends to sound better all around the room, not just in some small sweet spot.
Harbeth Super HL5plus loudspeaker Measurements | Stereophile.com
This one is a phase aligned fourth order LR4. It falls away very badly around 3kHz off-axis because the woofer is beaming. And in fact LR4 has a power hole at crossover just by its -6dB crossover nature, though it measures flat on axis.
The BBC idea is to use a BW3 response. The phase at crossover is aligned 90 degrees, which flattens what would otherwise be a 3dB peak at crossover. But it fills the power hole by 3dB. Which I think is a good thing. At this point people usually mention asymmetric lobing, but I don't care about that. BW3 tends to sound better all around the room, not just in some small sweet spot.
As long as you can’t reproduce the directivity pattern, no. But you could alter the room acoustics by adding midband absorption. Which in fact is the way to go.
@systems7
Sounds like an interesting design. Where can I find more information about it?
Guess the morel CAT378 could work aswell.
It's possible, though not especially likely, they had the tweeters wired out of phase. Production errors like that are easy to make and certainly not unknown, although fortunately quite rare.
What I suspect is more possible is simply that, since the SuperHL5 on available evidence runs twin tweeters off the same filter, its response is inherently somewhat more variable in the vertical plane than you might expect with a single HF unit, even relatively small variations from the optimal reference axis (viz. on axis with the larger tweeter) having an obvious effect. Note the suckout measured by JA of the HL5plus (figure 6) lurking in the crossover region as you move away from the optimal: Harbeth Super HL5plus loudspeaker Measurements | Stereophile.com
Last edited:
mirroring the classics
Gotta chime in here... I own the Harbeth 30.2 and bought them to "replace" the LS5/9 I sold (and miss very much) back in 2000. In spite of the known tonal anomalies of the speaker, somehow it was just an enjoyable speaker to listen to... and as my back-up to ESL-63s and LS3/5a, the 5/9 stood between the 2 legends, sharing attributes of both.
I (largely) bought the 30.2 due to a number of the glowing comments from reviewers and the Harbeth reputation. So far, I have been on the fence... it is a good speaker, I struggle saying a great speaker, like the LS5/9. I think that the 30.2 IS a "better" more modern design speaker, however I still like the 5/9 ~ a lot ~ but not sure if I could live with them today.
I think that the 30.2 "weakness" is in the woofer and that like the 5/9 going with a larger diameter tweeter, gives the greater power in mid band response x at the expense of extension in the top end. Driven with very good OTL and MOSFET based amps, the Harbeth woofer clearly distorts before bottoming out (which I have honestly never done) above certain listening levels.
I thought that the speaker tends to require "a lot" of power, but does not take that much power, a real conundrum for me. Harbeth "says" an easy load to drive. My amps get hot driving them... which supports a complex crossover, as a portion of reduced sensitivity. My other comment or complaint, they are lacking in terms of air, they are dry sounding... Compared to Maggies (ribbon tweeter) or the Silverline SR17S (Dynaudio Esotar T-330), as direct comparisons.
When I got the 30.2 home, I pulled down the Spica T60 in one system and put the Harbeths in place and thought (!!!!!) these sound much like the Spicas, for a lot less money. Tonally, imaging and sense of space, they are along the lines of a few other great speakers...
So, yes YOU CAN create a fine speaker as defined within this thread (I made notes) with the many great suggestions, will follow this thread. I like what you are working towards. Some aspects of stated finish production products are way overstated, marketing hyperbole, maybe just fluff...
Gotta chime in here... I own the Harbeth 30.2 and bought them to "replace" the LS5/9 I sold (and miss very much) back in 2000. In spite of the known tonal anomalies of the speaker, somehow it was just an enjoyable speaker to listen to... and as my back-up to ESL-63s and LS3/5a, the 5/9 stood between the 2 legends, sharing attributes of both.
I (largely) bought the 30.2 due to a number of the glowing comments from reviewers and the Harbeth reputation. So far, I have been on the fence... it is a good speaker, I struggle saying a great speaker, like the LS5/9. I think that the 30.2 IS a "better" more modern design speaker, however I still like the 5/9 ~ a lot ~ but not sure if I could live with them today.
I think that the 30.2 "weakness" is in the woofer and that like the 5/9 going with a larger diameter tweeter, gives the greater power in mid band response x at the expense of extension in the top end. Driven with very good OTL and MOSFET based amps, the Harbeth woofer clearly distorts before bottoming out (which I have honestly never done) above certain listening levels.
I thought that the speaker tends to require "a lot" of power, but does not take that much power, a real conundrum for me. Harbeth "says" an easy load to drive. My amps get hot driving them... which supports a complex crossover, as a portion of reduced sensitivity. My other comment or complaint, they are lacking in terms of air, they are dry sounding... Compared to Maggies (ribbon tweeter) or the Silverline SR17S (Dynaudio Esotar T-330), as direct comparisons.
When I got the 30.2 home, I pulled down the Spica T60 in one system and put the Harbeths in place and thought (!!!!!) these sound much like the Spicas, for a lot less money. Tonally, imaging and sense of space, they are along the lines of a few other great speakers...
So, yes YOU CAN create a fine speaker as defined within this thread (I made notes) with the many great suggestions, will follow this thread. I like what you are working towards. Some aspects of stated finish production products are way overstated, marketing hyperbole, maybe just fluff...
wel, my ex has the M40.2 and the HL5+ and both are indeed not neutral (a bit dark sounding) but do sound very good and polite. I never heared them distort altough like Ozark says. They were driven by an audio research Ref 75 (the M40.2) and a Prima Luna Dialogue HP (the HL5+) set in triode modus.
How are the Volt brand polycone drivers for such a Harbeth clone task ?
@Systeme 7 : not sure I understood what you said about the recess and the 90° phase to lower the XO in relation to a third order filter ?
I assume we are talking of horizontal spacing. Should not the bass driver be advanced (and not recessed) for time phase alignement in the crossover overlap, i.e. time center adjustment ? Or are we not talking about delay but frequency range cut off (pardon my ignorance 😱). Is the impulse response alignement of the two drivers affected by the filter order phase delay(90° advance per order?)
I'm wonder also if Harbeth is using BS compensation in their filter ?
Edit : also forget : the poly from SB Acoustic have very low distorsion while perhaps being close to what "plastic" cone do (polite=a little less details?) ? The Rolls-Royce as already said being Audiotechnology drivers while expensive.
@Systeme 7 : not sure I understood what you said about the recess and the 90° phase to lower the XO in relation to a third order filter ?
I assume we are talking of horizontal spacing. Should not the bass driver be advanced (and not recessed) for time phase alignement in the crossover overlap, i.e. time center adjustment ? Or are we not talking about delay but frequency range cut off (pardon my ignorance 😱). Is the impulse response alignement of the two drivers affected by the filter order phase delay(90° advance per order?)
I'm wonder also if Harbeth is using BS compensation in their filter ?
Edit : also forget : the poly from SB Acoustic have very low distorsion while perhaps being close to what "plastic" cone do (polite=a little less details?) ? The Rolls-Royce as already said being Audiotechnology drivers while expensive.
Last edited:
Also worth a read for the op : Heissmann Acoustics | Compact speaker | Heissmann-Acoustics
Wavecore good drivers and already tiped Seas horn tweeter (Systeme7 post)
Wavecore good drivers and already tiped Seas horn tweeter (Systeme7 post)
Harbeth "sound"
Concur, with the nature of the Harbeth "family" sound. *as my words.
I listen mostly to instrumental music, very little to vocals, and believe the Harbeth "are voiced to be tonally balanced" to be smooth in the mid band / vocal range... perhaps at the expense of a more fleshed out overall character.
As a "one man show" with (only a) singular designer and person voicing the(ir) products. Not sure how I feel about this?
Concur, with the nature of the Harbeth "family" sound. *as my words.
I listen mostly to instrumental music, very little to vocals, and believe the Harbeth "are voiced to be tonally balanced" to be smooth in the mid band / vocal range... perhaps at the expense of a more fleshed out overall character.
As a "one man show" with (only a) singular designer and person voicing the(ir) products. Not sure how I feel about this?
Good call re the Volt units -how I forgot them I don't know. The BM228.8 for e.g. would (on the basis of the published data) be a good option for something along the lines of a 30.2. SB do not currently produce an 8in unit with a polypropylene cone, and typically their units are not ideal for the higher crossover frequencies Harbeth tend to use as they start to break up quite early. The same applies to the Wavecor units -decent drivers of their type, but not particularly well-suited to this application.
Yes, of course Harbeth compensate for step-loss in their filters, as you can see from the Stereophile measurements -with the arguable exception of the P3ESR, which has a slightly different set of priorities / requirements.
Yes, of course Harbeth compensate for step-loss in their filters, as you can see from the Stereophile measurements -with the arguable exception of the P3ESR, which has a slightly different set of priorities / requirements.
They are tuned to be smooth in general, but their final test is spoken word said Allan Shaw in a video on youtube. If it doesn't present the voice like it should, it doesn't pass.
And my ex mainly listen to jazz and electronic music, what is also mostly instrumental. No singer songwriters there... I played quiet a few dub and (oldskool) hiphop trough that system, and it still sounds good, but lacks a bit low extention. But if you want a bright "in your face" or a neutral sound, this is not the speaker you want.
Last edited:
This has to do with the way you tune a 3d order BW in general, as opposed to a LR. The LR adds up nicely on axis and the resulting SPL is +6dB. Off axis it’s another story and in general the sound power curve of a LR is worse than that of an BW uneven order. So the in room response of the latter can be better.
There in fact lies your challenge with an 8”-1” system. Optimizing it on-axis isn’t the way to go by default. Some smart partly correction of the baffle step could work out better. Bottom line: since the acoustics aren’t perfect, tweaking the crossover to the actual listening conditions is inevitable.
Last edited: