A ribbon (meaning low impedance capable) version of Aleph-X would be my dream too!
But there are things to remember. First off, I remember Pass talking about a minimum voltage over the FETS is something like 10V, and to get good sound you should try to avoid lower than 15V, or was that 20V?
To do things like this in practice you need to monitor the current given out...... aka current feedback and high output impedance.
And the possibility to skip the transformer certainly seems like a good idea to me.
The idea to use longer and narrower strips of aluminum is certainly possible, but that means you have to use a binding material. In my ribbons I have used the kind of tape used to keep bandages in place. That white fiber type. This has really good damping of resonances. But it has a high mass, and probably has a roll of in the highs because of this. The ideal thing to do would be to make one small tweeter and have only aluminum as radiator and an amp capable of driving it directly.
The really nice thing about ribbons is that it must be about the easiest driver to make totally DIY (and at the same time sound really good too). And now even the neodymium magnets are starting to get cheap enough to use, and that means a decent sensitivity on the drivers is attainable too!
But there are things to remember. First off, I remember Pass talking about a minimum voltage over the FETS is something like 10V, and to get good sound you should try to avoid lower than 15V, or was that 20V?
To do things like this in practice you need to monitor the current given out...... aka current feedback and high output impedance.
And the possibility to skip the transformer certainly seems like a good idea to me.
The idea to use longer and narrower strips of aluminum is certainly possible, but that means you have to use a binding material. In my ribbons I have used the kind of tape used to keep bandages in place. That white fiber type. This has really good damping of resonances. But it has a high mass, and probably has a roll of in the highs because of this. The ideal thing to do would be to make one small tweeter and have only aluminum as radiator and an amp capable of driving it directly.
The really nice thing about ribbons is that it must be about the easiest driver to make totally DIY (and at the same time sound really good too). And now even the neodymium magnets are starting to get cheap enough to use, and that means a decent sensitivity on the drivers is attainable too!
Re: Re: Re: What about current amplifiers??
Thanks for the reply. I greatly appreciate your practical experience on these matters. May I ask for a clarification, though--I can think of two ways of interpreting your comment: Are you saying that the modulation distortion from the mid and bass SPL moving the ribbon around in the gap when driven by a low impedance amp is not material or that you can't achieve enough electrical damping with a ribbon to have a material impact, so don't bother?
I found this sort of modulation distortion to be highly significant in my current speakers (which use compression drivers) when I was using a third-order high pass. Since then I go to extra lengths to keep the out of band look-back impedance as low as possible--no more odd-order high passes for me. In fact, one of the reasons I am now using a fully active crossover is to avoid this problem.
I should mention, though, that my speakers have the midrange and tweeter mounted in the same horn (see pic in the following thread: http://www.diyaudio.com/forums/showthread.php?s=&threadid=17642&highlight= ). I am probably therefore more sensitive to this distortion than most.
I imagine that a ribbon has even worse problems in this regard since it has such high mechanical compliance and was working out how best to achieve a high level of electrical damping. Maybe I'm worrying about nothing, though.
Thanks, hadn't though of that. I suppose those parasitics would make it hard to achieve an amp with a low enough output impedance to reach my goals--high electrical damping, no transformer.
Ahhhh, my misunderstanding. Yes, you are of course absolutely right on that point. Since I started using DSP I'm afraid I don't even give linear distortion a second thought becuase it is so easy to fix with DSP. I am so fixated on non-linear distortion these days. A current source amp addresses a lot of non-linear distortion mechanisms that I think aren't that important for a ribbon, but as Nelson pointed out above, there are linear distortion issues in a ribbon that a current source does address.
Sorry, I should have been more clear. I'm not worried about damping of mechanical resonances in the ribbon, I'm worried about SPL from the room moving the ribbon about--especially the SPL from the mid and bass which will lead to large movements in the ribbon. This is what I was refering to when I was talking abut "flapping in the wind." This I imagine would be a big source of distortion in a ribbon driven by a current source. I could be wrong, though.
The transformer does lower the output impedance of the amp, but the problem is that the BL of the ribbon is also almost nothing and you need an output impedance that matches the BL. Not easy to achieve even with a step-down transformer for two reasons, ribbons tend to have low B and the transformer itself adds resistance.
Yes, sounds yummy. Since I don't need any gain, though--or even negative gain, what you think of a single stage p channel follower? Are there any devices that could be biased at something like 10A and be reasonably linear? I'm afraid I am lost when it comes to reading transistor transfer function charts. I speak fluent tubes, but it's not always easy to translate into transistorese.
thanks everyone for the helpful insights.
John
Nelson Pass said:
Thanks for the reply. I greatly appreciate your practical experience on these matters. May I ask for a clarification, though--I can think of two ways of interpreting your comment: Are you saying that the modulation distortion from the mid and bass SPL moving the ribbon around in the gap when driven by a low impedance amp is not material or that you can't achieve enough electrical damping with a ribbon to have a material impact, so don't bother?
I found this sort of modulation distortion to be highly significant in my current speakers (which use compression drivers) when I was using a third-order high pass. Since then I go to extra lengths to keep the out of band look-back impedance as low as possible--no more odd-order high passes for me. In fact, one of the reasons I am now using a fully active crossover is to avoid this problem.
I should mention, though, that my speakers have the midrange and tweeter mounted in the same horn (see pic in the following thread: http://www.diyaudio.com/forums/showthread.php?s=&threadid=17642&highlight= ). I am probably therefore more sensitive to this distortion than most.
I imagine that a ribbon has even worse problems in this regard since it has such high mechanical compliance and was working out how best to achieve a high level of electrical damping. Maybe I'm worrying about nothing, though.
Nelson Pass said:
Thanks, hadn't though of that. I suppose those parasitics would make it hard to achieve an amp with a low enough output impedance to reach my goals--high electrical damping, no transformer.
akb1212 said:
Ahhhh, my misunderstanding. Yes, you are of course absolutely right on that point. Since I started using DSP I'm afraid I don't even give linear distortion a second thought becuase it is so easy to fix with DSP. I am so fixated on non-linear distortion these days. A current source amp addresses a lot of non-linear distortion mechanisms that I think aren't that important for a ribbon, but as Nelson pointed out above, there are linear distortion issues in a ribbon that a current source does address.
akb1212 said:
Sorry, I should have been more clear. I'm not worried about damping of mechanical resonances in the ribbon, I'm worried about SPL from the room moving the ribbon about--especially the SPL from the mid and bass which will lead to large movements in the ribbon. This is what I was refering to when I was talking abut "flapping in the wind." This I imagine would be a big source of distortion in a ribbon driven by a current source. I could be wrong, though.
Brian Donaldson said:
The transformer does lower the output impedance of the amp, but the problem is that the BL of the ribbon is also almost nothing and you need an output impedance that matches the BL. Not easy to achieve even with a step-down transformer for two reasons, ribbons tend to have low B and the transformer itself adds resistance.
Brian Donaldson said:
Yes, sounds yummy. Since I don't need any gain, though--or even negative gain, what you think of a single stage p channel follower? Are there any devices that could be biased at something like 10A and be reasonably linear? I'm afraid I am lost when it comes to reading transistor transfer function charts. I speak fluent tubes, but it's not always easy to translate into transistorese.
thanks everyone for the helpful insights.
John
Okay, let's see...amps for ribbons and ribbons for amps.
First, let's note that the Magneplanar ribbon is direct drive and has a resistance of around 2.5 ohms. It is 60" long. The company has loopholes through which you may buy the drivers outright, but I think you'll find that it's cheaper to pick up a used pair of Maggies on ebay and pull out the tweeters. Besides, you could do worse than a pair of 3.6s.
Second, be aware that there's a great deal of confusion about the use of the term ribbon. Many--actually most--of the drivers being sold as 'ribbons' are actually planar magnetics. For instance, the Bohlender-Graebner drivers are planar magnetics, not ribbons. There's a lot of cache in the term ribbon, so companies use the term rather freely. Add to that the fact that a lot of consumers are pretty hazy about the differences between the different drivers and, well, the waters get murky.
Suffice it to say that there's no such thing as a wide-range ribbon. Period. The lowest you're going to get out of a true ribbon is perhaps 2kHz, plus or minus a bit, and at that you're likely to have reliability problems. Best to keep ribbons restricted to tweeter use.
Planar magnetics, on the other hand, are also purely resistive, and can go much deeper. You can get pretty reasonable resistances from planar magnetics--enough so that you don't have to worry about exotic amp topologies...unless you just feel like doing something for the fun of it.
Not to push either project, but the Mini-A and the Aleph-X were both conceived originally as dedicated tweeter amps. Nelson's SOZ is a current amp, something that people tend to overlook. Then he went back and fiddled with it some more sometime later in the Zen series. I did a SOZ w/current source thread once upon a time, and so on and so forth. The idea being that ideas are lying about here and there...pick one up and mess with it. You don't have to wait until the F-1 schematics become available. You can build either a heavily biased voltage amp, or a current amp in the here and now.
Incidentally, a MOSFET amp will work with 8-10V rails, but the distortion will be much higher than the same circuit at, say, 20-25V.
Grey
First, let's note that the Magneplanar ribbon is direct drive and has a resistance of around 2.5 ohms. It is 60" long. The company has loopholes through which you may buy the drivers outright, but I think you'll find that it's cheaper to pick up a used pair of Maggies on ebay and pull out the tweeters. Besides, you could do worse than a pair of 3.6s.
Second, be aware that there's a great deal of confusion about the use of the term ribbon. Many--actually most--of the drivers being sold as 'ribbons' are actually planar magnetics. For instance, the Bohlender-Graebner drivers are planar magnetics, not ribbons. There's a lot of cache in the term ribbon, so companies use the term rather freely. Add to that the fact that a lot of consumers are pretty hazy about the differences between the different drivers and, well, the waters get murky.
Suffice it to say that there's no such thing as a wide-range ribbon. Period. The lowest you're going to get out of a true ribbon is perhaps 2kHz, plus or minus a bit, and at that you're likely to have reliability problems. Best to keep ribbons restricted to tweeter use.
Planar magnetics, on the other hand, are also purely resistive, and can go much deeper. You can get pretty reasonable resistances from planar magnetics--enough so that you don't have to worry about exotic amp topologies...unless you just feel like doing something for the fun of it.
Not to push either project, but the Mini-A and the Aleph-X were both conceived originally as dedicated tweeter amps. Nelson's SOZ is a current amp, something that people tend to overlook. Then he went back and fiddled with it some more sometime later in the Zen series. I did a SOZ w/current source thread once upon a time, and so on and so forth. The idea being that ideas are lying about here and there...pick one up and mess with it. You don't have to wait until the F-1 schematics become available. You can build either a heavily biased voltage amp, or a current amp in the here and now.
Incidentally, a MOSFET amp will work with 8-10V rails, but the distortion will be much higher than the same circuit at, say, 20-25V.
Grey
I must agree there really is no such thing as a true wide range
ribbon. i did assemble and use a 46 inch long pair years ago
and was quite happy with them.
but overall i agree that they are most happy when built and used as a tweeter.
I have also used planars which can be very good indeed.
but allways yearned for the ribbons later.
the 46 inchers seemed happy to run on a rather conventional power amp running on a lower voltage power supply as their d.c. resistance measrured around around .8 - .9 ohms if memory serves.
a lower voltage SOZ might hold the best promise.
not having built any of the zen power amps
any comments from those who have?
ribbon. i did assemble and use a 46 inch long pair years ago
and was quite happy with them.
but overall i agree that they are most happy when built and used as a tweeter.
I have also used planars which can be very good indeed.
but allways yearned for the ribbons later.
the 46 inchers seemed happy to run on a rather conventional power amp running on a lower voltage power supply as their d.c. resistance measrured around around .8 - .9 ohms if memory serves.
a lower voltage SOZ might hold the best promise.
not having built any of the zen power amps
any comments from those who have?
Elektor had published sometime ago a bipolar output amp using MAT02 and MAT03 input pair with +-15 volt rails and a huge output current capability which was specifically meant for Ribbons below 1 ohm impedance.
Someone around here can help with the schematics and PCB layouts as already published.
Someone around here can help with the schematics and PCB layouts as already published.
"Cache" in my previous post should have had a "t" on the end. Sigh.
For those who are building their own ribbons, don't forget that thinner conductors are also good for increasing resistance. People keep slicing the driver element narrower and narrower--and there are advantages to be had, such as a more even magnetic field due to closer magnets--but thinness is a virtue, too.
You can never be too rich or too thin.
(I'm slender enough, but I'm still waiting on the rich part...)
That doesn't mean that it's easy to find thin aluminum, especially in long lengths. There's a website somewhere out there where the guy went to the trouble to separate the foil from gum wrappers. Ultra thin, but a little on the short side. Still, DIY folk have to take what they can get.
Grey
For those who are building their own ribbons, don't forget that thinner conductors are also good for increasing resistance. People keep slicing the driver element narrower and narrower--and there are advantages to be had, such as a more even magnetic field due to closer magnets--but thinness is a virtue, too.
You can never be too rich or too thin.
(I'm slender enough, but I'm still waiting on the rich part...)
That doesn't mean that it's easy to find thin aluminum, especially in long lengths. There's a website somewhere out there where the guy went to the trouble to separate the foil from gum wrappers. Ultra thin, but a little on the short side. Still, DIY folk have to take what they can get.
Grey
Hi eLarson,
The E is an abbreviation for EMF a physical quantity which describes the ability or "potential" of an electrical source to deliver energy. You can also think of it as the property of the source which creates current in a circuit. Derived from the nineteenth century term "electromotive force", but it is not a force as we define force now, but rather an electron potential difference. The emf of a battery is responsible for producing a potential difference between the battery's terminals. If the battery is not connected to anything else, that potential difference is equal to the emf. The SI unit is the volt, symbol V.
[Not to be confused with electromagnetic field which the popular press sometimes refers to as EMF.]
---------------------------------------------
SIMPLE USEFUL RIBBON EFFICIENCY EQUATIONS FOR AMP DESIGNERS:
For a ribbon tweeter, magnetic fields of 0.5 to 1 Tesla are possible at modest cost.
A=ribbon area meter^2
B=magnetic field in Telsa
%= percent of ribbon which conducts current(100=100%)
m= mass of ribbon in grams
Efficiency = A^2 * B^2 * % / m
spl 112.2 + 10*(LOG(wfficiency)/LOG(10))
senS effeciency + LOG(8/resistance)/LOG(10)
Prolonged exposure to > 85 db SPL can cause hearing loss. 94 db at 1m = 85 db at 8m, which is a good size room. A 110 db @ 1m design would seem to have enough margin. For a 2m long ribbon, a displacement of +/- 0.1" should be capable of 200 - 20,000 Hz at 110 db
The E is an abbreviation for EMF a physical quantity which describes the ability or "potential" of an electrical source to deliver energy. You can also think of it as the property of the source which creates current in a circuit. Derived from the nineteenth century term "electromotive force", but it is not a force as we define force now, but rather an electron potential difference. The emf of a battery is responsible for producing a potential difference between the battery's terminals. If the battery is not connected to anything else, that potential difference is equal to the emf. The SI unit is the volt, symbol V.
[Not to be confused with electromagnetic field which the popular press sometimes refers to as EMF.]
---------------------------------------------
SIMPLE USEFUL RIBBON EFFICIENCY EQUATIONS FOR AMP DESIGNERS:
For a ribbon tweeter, magnetic fields of 0.5 to 1 Tesla are possible at modest cost.
A=ribbon area meter^2
B=magnetic field in Telsa
%= percent of ribbon which conducts current(100=100%)
m= mass of ribbon in grams
Efficiency = A^2 * B^2 * % / m
spl 112.2 + 10*(LOG(wfficiency)/LOG(10))
senS effeciency + LOG(8/resistance)/LOG(10)
Prolonged exposure to > 85 db SPL can cause hearing loss. 94 db at 1m = 85 db at 8m, which is a good size room. A 110 db @ 1m design would seem to have enough margin. For a 2m long ribbon, a displacement of +/- 0.1" should be capable of 200 - 20,000 Hz at 110 db
Thanks, LS. I think it makes more sense to me now with E referring to e.m.f. 
Out of curiosity, how does the motion of the ribbon relate to a drive current? It seems to me that the low inherent impedance of the ribbon would be well-suited to current mode drive.
It's easy to get a higher impedance... more costly to get beneath the ribbons low impedance for voltage drive, though.
Out of curiosity, how does the motion of the ribbon relate to a drive current? It seems to me that the low inherent impedance of the ribbon would be well-suited to current mode drive.
It's easy to get a higher impedance... more costly to get beneath the ribbons low impedance for voltage drive, though.
eLarson...here are a few simplified equations for ribbon motion. 3D modeling software is freely available.
The ribbon speaker is the only audio transducer where a purely resistive electrified element immersed in a uniform field directly generates sound with no physical obstructions in the direct listening path. The principle of operation is quite simple and based on the Lorentz force. The thin ribbon of metal, just a few microns thick, is suspended in a powerful transverse magnetic field. The alternating signal current i passes through the ribbon length L, perpendicular to the magnetic flux B, and creates a forward-backward force F = B*L*i. which uniformly vibrates the entire ribbon area creating sound. This electromagnetic circuit has complete control over ribbon motion. The sound is projected by the ribbon itself, because the ribbon serves both as both voice coil and diaphragm. Every part of the ribbon is driven directly, uniformly, and simultaneously with minimal energy storage. As the current reverses, the foil moves in the opposite direction with corresponding mechanical oscillations at the frequency of the electrical signal.
F=force
B=transverse magnetic field
L=ribbon length
i=current through ribbon
R=ribbon resistance
x=acceleration
m=ribbon mass
A=radiating area
P=pressure
d=density of air
r=radial distance from ribbon
E=voltage across ribbon
Force F = B*L*i = B*L*(V/R)
In the mass controlled region the mechanical loss resistance and stiffness are negligible compared to the ribbon mass, and allows some simplified equations for acceleration and pressure
Acceleration x can be approximated by x = B*L*E / m*R
Farfield pressure on axis P = d*A*x / 2*pi*r
P = d*A*B*L*V / 2*pi*r*m*R
The pressure response in this range is seen to be independent of frequency.
For a ribbon of 0.047 ohms yielding an input resistance of 2.3 ohms, with a 7 to 1 transformer, the pressure equation, yields a level of 94 dB at 1 meter for 2.83 volt reference.
The ribbon speaker is the only audio transducer where a purely resistive electrified element immersed in a uniform field directly generates sound with no physical obstructions in the direct listening path. The principle of operation is quite simple and based on the Lorentz force. The thin ribbon of metal, just a few microns thick, is suspended in a powerful transverse magnetic field. The alternating signal current i passes through the ribbon length L, perpendicular to the magnetic flux B, and creates a forward-backward force F = B*L*i. which uniformly vibrates the entire ribbon area creating sound. This electromagnetic circuit has complete control over ribbon motion. The sound is projected by the ribbon itself, because the ribbon serves both as both voice coil and diaphragm. Every part of the ribbon is driven directly, uniformly, and simultaneously with minimal energy storage. As the current reverses, the foil moves in the opposite direction with corresponding mechanical oscillations at the frequency of the electrical signal.
F=force
B=transverse magnetic field
L=ribbon length
i=current through ribbon
R=ribbon resistance
x=acceleration
m=ribbon mass
A=radiating area
P=pressure
d=density of air
r=radial distance from ribbon
E=voltage across ribbon
Force F = B*L*i = B*L*(V/R)
In the mass controlled region the mechanical loss resistance and stiffness are negligible compared to the ribbon mass, and allows some simplified equations for acceleration and pressure
Acceleration x can be approximated by x = B*L*E / m*R
Farfield pressure on axis P = d*A*x / 2*pi*r
P = d*A*B*L*V / 2*pi*r*m*R
The pressure response in this range is seen to be independent of frequency.
For a ribbon of 0.047 ohms yielding an input resistance of 2.3 ohms, with a 7 to 1 transformer, the pressure equation, yields a level of 94 dB at 1 meter for 2.83 volt reference.
amp-guy....
take a look at post 29, a simple Class A symmetrical topology that can drive 0.5-1 ohm. Perhaps lowering the output power supply voltage would make it efficient and safe on lower resistance ribbons.
I'd like to see a digital amp design that has the output stage adjusted for < 1 ohm. Perhaps using the TI chips.
take a look at post 29, a simple Class A symmetrical topology that can drive 0.5-1 ohm. Perhaps lowering the output power supply voltage would make it efficient and safe on lower resistance ribbons.
I'd like to see a digital amp design that has the output stage adjusted for < 1 ohm. Perhaps using the TI chips.
thanks linesource ;
the schematic you attached in post 29
looks interesting lots of ccs and a cascode. and whats up with the opamp looks like maybe a d.c. servo.
there are some obscure semis thou.hmmm
when i had ribbons about 10 years ago i used a bgw 250
with the rail voltage dropped to +-20 volts. (i think)
it ran whithout smoke escaping but didnt sound that great.
the schematic you attached in post 29
looks interesting lots of ccs and a cascode. and whats up with the opamp looks like maybe a d.c. servo.
there are some obscure semis thou.hmmm
when i had ribbons about 10 years ago i used a bgw 250
with the rail voltage dropped to +-20 volts. (i think)
it ran whithout smoke escaping but didnt sound that great.
Is this mean that lack of flux density can be compensated with more amp power with same impulse characterisitics of the ribbon motion/acceleration (sound quality) ?
I want to figure out one thing - is the sound quality of ribbon speaker depends of how strong (and more expensive) the magnets used, or it is only power efficience that suffers.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.