I am planning to build a loudspeaker that complements my MPP. I have build a preamp and Brianco is doing a poweramp so in the next month a system will be made of Phonoamp, Preamp, Poweramp and Speaker for DIY.
The speaker is based on a wideband unit in a baffle.
Under 80 Hz it will be complimented by a dipol woofer and over 7kHz there will be a horn loaded Ribbon.
The wideband driver got developped over a period of one year.
It has a classical paper cone and Alnico magnet.
I got the finished samples some days ago and did some measurements.
The driver is designed for very low distortion.
It is under 0.5% up to 100dB in 1m second harmonic. Thirs id under 0.1%.
Today i will post some measurements.
The driver is in a 90cm by 125cm baffle placed asymmetric due to golden cut.
I will pulish photos and drawings later.
Impedance is 6 Ohm Re and sensitivity is 96dB for 1W/ 1m.
The speaker is based on a wideband unit in a baffle.
Under 80 Hz it will be complimented by a dipol woofer and over 7kHz there will be a horn loaded Ribbon.
The wideband driver got developped over a period of one year.
It has a classical paper cone and Alnico magnet.
I got the finished samples some days ago and did some measurements.
The driver is designed for very low distortion.
It is under 0.5% up to 100dB in 1m second harmonic. Thirs id under 0.1%.
Today i will post some measurements.
The driver is in a 90cm by 125cm baffle placed asymmetric due to golden cut.
I will pulish photos and drawings later.
Impedance is 6 Ohm Re and sensitivity is 96dB for 1W/ 1m.
Hello!
I can see typical nasty peak at around 3 kHz
Isn't there really any way to get rid of it? Not by means of ex post facto passive or active correction but by preventive measures? Are such peaks really unavoidable?
interesting what Hartley wrote in 1958:
Audio Design Handbook, page 12
Audio Design Handbook, page 153
best!
graaf
I can see typical nasty peak at around 3 kHz
Isn't there really any way to get rid of it? Not by means of ex post facto passive or active correction but by preventive measures? Are such peaks really unavoidable?
interesting what Hartley wrote in 1958:
Audio Design Handbook, page 12
Audio Design Handbook, page 153
best!
graaf
Last edited:
You will not beleave it but 140 protos have being build over the years.Hello!
I can see typical nasty peak at around 3 kHz
Isn't there really any way to get rid of it? Not by means of ex post facto passive or active correction but by preventive measures? Are such peaks really unavoidable?
interesting what Hartley wrote in 1958:
Audio Design Handbook, page 12
Audio Design Handbook, page 153
best!
graaf
A version can be build without that peak. The version i have shown has the finished magnet system and membrane.
I also think the peak should go away because it is in the most sensitive area of human hearing. Other versions can be build that have more extended treble but for my profect i decided on an aditional tweeter so i will go with a variety of the shown driver.
then why leave it? if "a version can be build without" it?
prohibitive costs? design-for-manufacturability considerations?
ribbon at 7 kHz and up would not help, even crossing at 3 kHz would not help
and we have been already sailing away from "Full Range" land into the waters of "Multi Way"
50 years have passed, we presumably have all this hi-tech (from new materials to new measurements) unavailable to Hartley and His contemporaries and yet we can't fix those same old problems?
this is so disappointing
what is interesting speakers designed according to Hartley ideas are allegedly free from them
here is an example - Hartley speaker paired with a horn loaded ribbon filtered "6dB down at 10kHz and 12dB down at 5kHz":
Stereophile: Listening #37
yes, this example is outrageously overpriced, nevertheless the driver seems to be working according to the design principles 50 years old
so... then
why leave the peak?
all Hartley's patents expired long time ago, why not use it?
or ideas from Stewart Hegeman and others of "non-uniform asymmetric diaphragm treatment" or "differential wave impedance cone":
http://www.diyaudio.com/forums/full-range/126234-8-inches-paper-cone-treatments-4.html#post2088939
or Weber Rehde's patents and so on
why not?
best regards!
graaf
Last edited:
50 years have passed, we presumably have all this hi-tech (from new materials to new measurements) unavailable to Hartley and His contemporaries and yet we can't fix those same old problems?
this is so disappointing
I couldn't agree more. And I don't understand this mystery any more than you do. I was simply amazed at how awful the expensive and inexpensive fullrange drivers sounded (Lowther, AER, Fostex, etc). I have no other word for it. Amazed. If a cone can't be driven above 90db without eardrum killing distortion around 1-4khz what good is it? No good at all!
Since I don't like the vague sound of contemporary multiway drivers and designs, I've decided there are only 2 possible avenues from here: DIY cone treatments to make the FR behave or buy a set of the large Bohlender Graebener units; both of which I plan to do.
Good Widebands are posible
We can make a wideband that is flat from 60Hz to 10kHz + - 2dB with 95dB efficiency at 8 Ohm.
Distortion in that driver is very low and linear over the whole response.
See some measurements i did with multitone.
The driver has a dynamic range of 70dB. Marker 1 and 2 show hum from the poweramp.
The best old drivers i measured had distortion much higher and not more then 60dB dynamic range. I also did an inverserse FFT from the hidden spectrum.
Noise in that driver is only 250nP. See again the 50 Hz component that does not come from the driver but is hum from the measurement amp.
I used an Alesis professional amp for this.
We can make a wideband that is flat from 60Hz to 10kHz + - 2dB with 95dB efficiency at 8 Ohm.
Distortion in that driver is very low and linear over the whole response.
See some measurements i did with multitone.
The driver has a dynamic range of 70dB. Marker 1 and 2 show hum from the poweramp.
The best old drivers i measured had distortion much higher and not more then 60dB dynamic range. I also did an inverserse FFT from the hidden spectrum.
Noise in that driver is only 250nP. See again the 50 Hz component that does not come from the driver but is hum from the measurement amp.
I used an Alesis professional amp for this.
it certainly looks nicer
but it is hard to compare those two graphs (the previously posted and the second) and to draw any conclusions because of different scales, averaging etc. although I can see general similarity of both FR curves
+/- 2 dB certainly is impressive
in fact there are widebanders for OB, pretty flat and with low distortions (according to manufacturer's proclamations), more or less expensive, like PHY-HP or Visaton B200
there is bigger problem with fullrangers
I am looking for something with a decent headroom (which means 8 inches at least), free from any shouts (those "eardrum killing distortion around 1-4khz" as put by InclinedPlane few posts above) that would work nicely without a tweeter and be capable of providing at least some real, non-resonant bass, not necessarily down to 20-ies - some 40-ies in room would suffice
but without horns or any other refrigerator-sized cabinets, rather in a moderate-sized acoustic suspension
oh well ported could be also ok provided it is small like Rehdeko 115, no bigger
I do like OBs but I think that they are just impractical for most music lovers, it' s mostly question of their size and placement requirements
and multi-way OBs are complicated from DIY perspective, very difficult to tune all the elements in
I think that it is not coincidence that AR-1 - the first compact bookshelf speaker - was such an epoch-making product
Is there any chance for a fullranger without shouts that could work nicely in a compact box? and that would not cost a small fortune? Alnico not really necessary
currently I am tweaking Fostex FE206E, nice affordable driver with great tweaking potential, a widely known fact for sure
, but it is strictly a tweaker's speaker, it is just horrible out of the box and not really suited to the use in a compact enclosureFE206E can be succesfully tweaked at home (using ideas similar to Hegeman's or Rehde's) but it would be great to have simply a driver with Fostex's problems fixed by means of all professional design instruments, measurements, standardization and so on
all measured and clearly audible, unlike in case of various "phase plugs" or "EnABLs" etc. the effects of which can hardly be measured not to mention to be consistently audible or at least rationally explained...
I had a dream ...
best regards!
graaf
Last edited:
but BG, even those largest, are just very big midtweeters - manufacturer's "network recommendations": 150Hz, minimum 12 dB/oct. electrical minimum"
You end up with multi-way speaker anyway
therefore "DIY cone treatments to make the FR behave" is the only way IMO
and more precisely - "8 inches paper cone treatments" as You already know
best,
graaf
To me one of the drawbacks of conventional cone
fullrangers and widebanders is, that with enough
tweaking they can be made reasonably flat on axis,
but this cannot go together with a flat power response.
Angular dispersion narrows with increasing frequency
reasonably continouus in the best case.
In the worst case there are several discontinuities in
angular dispersion dependent from frequency,
which will make the result very room dependent -
and prone to "sweet spot listening".
I'll give conventional full/wide range technology about a
decade to survive, at least in the high quality segment.
Arrays of integratively manufactured micro transducers,
developing DML and related designs will take over.
I do not say this to "downplay" the good results
presented here, but the more tweaking is done,
the more the limitations get clear also.
The popularity of vintage drivers, which are in many
cases competitive to recent products, is an indicator for
that too.
Just my opinion.
fullrangers and widebanders is, that with enough
tweaking they can be made reasonably flat on axis,
but this cannot go together with a flat power response.
Angular dispersion narrows with increasing frequency
reasonably continouus in the best case.
In the worst case there are several discontinuities in
angular dispersion dependent from frequency,
which will make the result very room dependent -
and prone to "sweet spot listening".
I'll give conventional full/wide range technology about a
decade to survive, at least in the high quality segment.
Arrays of integratively manufactured micro transducers,
developing DML and related designs will take over.
I do not say this to "downplay" the good results
presented here, but the more tweaking is done,
the more the limitations get clear also.
The popularity of vintage drivers, which are in many
cases competitive to recent products, is an indicator for
that too.
Just my opinion.
Last edited: