Are we sure about that? H3 at 1.5k - 3k is about -60 dB, and at 4k - 6k it is about -70 dB. That is a change from 0.1% to 0.03% in H3. I would guess that it would take a quite exceptional listener to perceive that difference, especially at 105 db SPL.
I have no experience with current drive amps... Can you give me some examples of current drive amps?
This only would make sense when the midrange produces less H3! Which it doesn't. And we are talking about <0,1% at 105dBSpl ... the ear should not be able to hear that.
And be careful - the limiter sets in at exactly that frequency area! This will have influence to the result, so let's look how the response is at 100dBSpl without limiting.
H3 goes down from 2kHz. The tweeter is producing even less H3 as the 3" midrange. And we are talking of about 0,03% of H3 ... most setups can't even measure that
So in THIS case the tweeter is producing some more H2 and less H3 as the midrange at 2kHz. Both produce exceptional low THD!
But H3 is really low in this area? 10dB less as both 12" produce ... and they have 12x the membrane area ... (with about 70x the weight btw)I meant potentially all the peaks from 600Hz to 3.5kHz. An A/B comparison is not easy, but should allow you to hear the effect - or prove that I am wrong
The peak at 1kHz could be nicer - but it only shows up cause H2 is so low.
Let's see how the Beryllium version behaves here ...
I thought about putting a passive resistor in front of the midrange to do some dampening ... will check this in the next build. A/B-ing will be easy - with measurements. But pretty hard for a listening test.
It depends if you are a DIY constructor or not, which is after all what this forum is for. There have been commercial loudspeakers exploiting current drive - the now-dissolved, Celtic Audio, for example - and there is a lot of literature on practical implementations - David Birt and Malcolm Hawksford/Paul Mills, for examples. Many examples you will find described within this forum.
H2 can be largely benign, H3 due to magnetic linearities not so. Quite possibly because current dependent non-linearities are not precisely SPL related, and so could be discerned as a separate noise rather than part of the recorded material? But that is just me making up a theory to better account for its audibility.
And there is no telling if H3 due to the bass drivers is showing current dependent non-linearities too. My suspicion was raised only by the peaky nature of H3 evident in the mid range.
A narrow band "peak" of H3 is probably not enough to discern what is going on. Often we need to see what H2 is doing, as well as higher order harmonics. For instance, if H4/H5/H6 was also elevated, it might not be the driver, but a loose screw/cable/wiring etc.
In this case, the M74A series are well reviewed by @HiFiCompass :
https://hificompass.com/en/reviews/bliesma-m74a-6-m74b-6-m74p-6-and-m74s-6
I concur with @IamJF - H3 appears high because the H2 is low:
M74A with 8V drive
Compare that to the M74P with 8V drive
Note 8V is 9dB more than 2.83V volt, and thus matches up with 96dB(sensitivity of the M74A) + 9 = 105dB , the exact same SPL
Reference:
Hificompass.com
In this case, the M74A series are well reviewed by @HiFiCompass :
https://hificompass.com/en/reviews/bliesma-m74a-6-m74b-6-m74p-6-and-m74s-6
I concur with @IamJF - H3 appears high because the H2 is low:
M74A with 8V drive
Compare that to the M74P with 8V drive
Note 8V is 9dB more than 2.83V volt, and thus matches up with 96dB(sensitivity of the M74A) + 9 = 105dB , the exact same SPL
Reference:
Hificompass.com
What considerations have been made about the ear height to speaker placement and the upside down tweeter configuration? I see you guys are talking of high crossovers for the tweeter/mid... This suggest that the tweeter be at ear level or higher but never lower. As a floor stander you have the tweeter at ~37". From there looking at KA and CtC spacing, nothing works right and I see the flaw of not being able to match polars. The best worse case scenario is having a higher directive tweeter.... with strictly dynamic drivers, you have the worse worse case with the tweeter having less directivity than the lower driver. 0.5 Ka for the tweeter is ~2.1khz and 2 ka is ~1.8khz for the mid range driver, and no hopes of 1/4wl spacing in any case. Wouldn't it make sense to at least have a low KA on one of the drivers involved?
I'm absolutely considering this! Will do more detailed off axis measurements outside with the final version of the speaker, this one needs to go out beginning of next week.
I made a few quick and very dirty horizontal off axis measurements in the room and it looked promising for mid and tweeter and less pronaunced for the 12"s as expected. But the frequency response is rising of the 12", 3.5 would be easily possible.
Tweeter position would be to low for usual use, this is just a setup to check what level is possible and if one 3" is strong enough for the 2 12".
I have a MTM configuration in my listening room and prefer to just use one M of them. And also prefer to use the upper one as I'm sitting high an a chair during mixing/listening. Also in my home cinema I prefer a slightly higher center of the music as too low. So I would put the tweeter at LEAST to 1m or higher. In my reference room it's 113cm. (here it's only 90cm, that's needed for the measurement configuration)
The concept of the high sitting 12" is to excite the room more regularly, idealy they would sit at about 0,6m and 1.8m for normal sized rooms in Europe - but that would be a very large speaker - need to decide for a compromise. So there is enough room and space to be flexible with tweeter placement.
What's KA spacing? I am definitely not in the CtC boat! You produce a huge chancellation to "design" a better ceiling reflection for usual listening positions. As my speakers are also used for close listening and should be able to produce an even sound in the whole room this will produce major problems. In my listening trials I compare to KEF LS50 and KH120ii and always prefered close drivers to get a more "coax like" point source reproduction.
That's also the reason to have a LOW crossover frequency - speakers blend to a close to coax reproduction.
Tweeter and Midrange are now 10cm apart - will lower that with a changed mid frontplate to less then 9cm, that's the best I can achieve.
The 3" has no beaming at 2kHz - that's not a 17cm driver ;-). So no jump in directivity but it would make a jump at higher crossover frequencies, that for sure.
Also the use of this speaker is in dedicated home cinemas and serious listening environments. In both ceiling treatment is recommended. I would concern more about floor reflection.
This is an important topic no question!
Ok, just trying to play devils advocate for you. On the topic of directivity KA is a spec I learned from Dr. Geddes paper on directivity. A driver isn't truly omnidirectional until Ka gets low like 1 or 0.5. The 3" has more directivity than the 1" even at 2khz but it may not be a noticeable amount. Being just a little higher than ka=2, and 2 being an expectation for close monitoring. A big issue for me where my ctc is much larger so viewing angle from the lower driver must be considered.
These issues increase and decrease with frequency, where exactly are you crossing over at?
These issues increase and decrease with frequency, where exactly are you crossing over at?
I cannot agree with this statement. The nature of third harmonic distortion due to eddy currents (more specifically due to magnetic non-linearties in the pole pieces in which they flow) is generally peaky. Such distortion generally presents differently to typical displacement-dependent second harmonic distortion, for example (where higher harmonics can also give useful information).
This is one of the most interesting aspects of this thread. The M74A has enough (more than enough) low distortion high SPL capability to keep up with the two very capable 12" woofers.
j.
That's OK by me if you disagree.
Though perhaps I should have emphasised
“probably” in my post, via bold or italics.
Did you get a chance to review the evidence I posted?
What I suggested, by perhaps failed to convey, was that had the M74P been used, few people would have noticed it to be high.
I proposed that it’s only because the alu M74A model, as used my @IamJF , was used, that it allowed H2 to be low enough (-65 to -70dB) that H3 could be seen to poke its head up (-55dB) above H2.
When the M74P and M74A are both played at 105 dB & 1m; H3 is -50dB (good) or -55dB (very good) respectively. @IamJF ’s dats is consistent with @HiFiCompass ’s data. At 105dB.
AFAIK, the current state of the art for a cone/dome direct radiator is about ~-75 H3 (0.02%) at 106 dB for 1Khz @ 1m (not horn loaded)
I’m was not proposing the mechanism of H3 (or H2), they are complex and involve a combination of Le(x), Sd(x), BL(x), Kms(x), as well as current(A)