A straight transconductance output has Mongo-sized Zout. Not an advantage. You need to pull that back into the real world or your woofer's damping factor ceases to exist. The approach I find most intriguing (perhaps predictably) is to servo the woofer; in other words, include the driver itself in the feedback loop. I will have more to report on this at a later date.
You can use feedback to lower the Zout, but you pretty much have to have some sort of "real" resistance to ground at the output so as to have something to work with. (In the next few days, there'll be a preamp adaptation of this idea on Enjoy the Music.) Without the resistance, you get rather unpredictable results. The gain is undefined and the DC offset starts wandering all over the place.
You can use a low-ish resistance to ground to stabilize things and get away without using any feedback at all. See the front end of the GR-25 for an example. It wastes amplifier power if you put it at the output, though.
I was looking at an amplifier schematic the other day that used lotsa bipolar output devices--half followers (emitter-to-emitter), half transconductance (collector-to-collector)--not your father's output stage. to be sure. I'm still thinking about that one...and, yes, the driver stage is a little more complicated than usual.
In the abstract, the advantage to a transconductance amplifier is that the heating of the voice coil (a nasty distortion mechanism, incidentally) doesn't degrade the output of the amp as much. It just does the current source thing and keeps ramping up the voltage until it manages to push the desired amount of current through. If it weren't for the damping problem, you'd be home free. Prod at it long enough and you can get the damping factor under control.
Grey
You can use feedback to lower the Zout, but you pretty much have to have some sort of "real" resistance to ground at the output so as to have something to work with. (In the next few days, there'll be a preamp adaptation of this idea on Enjoy the Music.) Without the resistance, you get rather unpredictable results. The gain is undefined and the DC offset starts wandering all over the place.
You can use a low-ish resistance to ground to stabilize things and get away without using any feedback at all. See the front end of the GR-25 for an example. It wastes amplifier power if you put it at the output, though.
I was looking at an amplifier schematic the other day that used lotsa bipolar output devices--half followers (emitter-to-emitter), half transconductance (collector-to-collector)--not your father's output stage. to be sure. I'm still thinking about that one...and, yes, the driver stage is a little more complicated than usual.
In the abstract, the advantage to a transconductance amplifier is that the heating of the voice coil (a nasty distortion mechanism, incidentally) doesn't degrade the output of the amp as much. It just does the current source thing and keeps ramping up the voltage until it manages to push the desired amount of current through. If it weren't for the damping problem, you'd be home free. Prod at it long enough and you can get the damping factor under control.
Grey
Variac said:
I written about all that of course, but current-source amps are not going to
be a popular general choice. I think of them as novelty amplifiers that sound
quite good in a limited range of applications.
I think the most interesting thing about them is that they deliver more uniform
acceleration independent of motion induced impedance.
Nelson Pass said:I written about all that of course, but current-source amps are not going to
be a popular general choice. I think of them as novelty amplifiers that sound
quite good in a limited range of applications...
And how about driving somthing like an R1 ribbon? Would the close to resistive load be a good combination?
It's straightforward enough. One way is to get a big power
amp with about +/-80V swing (400 watts into 8) and place
the R1 ribbon in series with a big 8 ohm resistor.
The other way is to build an active current source capable of
about +/- 5 volts at 10 amp peaks. As a push-pull Class A,
this would need about +/-10V rails at 5A bias, for about 100
watts idle dissipation.
You could do this complementary Common Drain and with a
feedback loop for current source on the front end, or you could
do it complementary Common Source without feedback.
The latter would be the most interesting. If you guys are that
hot to have one, I do have a schematic and it could be a project.
amp with about +/-80V swing (400 watts into 8) and place
the R1 ribbon in series with a big 8 ohm resistor.
The other way is to build an active current source capable of
about +/- 5 volts at 10 amp peaks. As a push-pull Class A,
this would need about +/-10V rails at 5A bias, for about 100
watts idle dissipation.
You could do this complementary Common Drain and with a
feedback loop for current source on the front end, or you could
do it complementary Common Source without feedback.
The latter would be the most interesting. If you guys are that
hot to have one, I do have a schematic and it could be a project.
I think that Burning Amps has begun to define itself- After the first couple, which will be pretty conventional, it might become among other things, amps too specialized to be done even by First Watt!
Specialized or not, I just happen to use a ribbon tweeter!
Boy, a direct drive ribbon tweeter amp and a current amp for the bass..that's exotic!
Mark
Specialized or not, I just happen to use a ribbon tweeter!
Boy, a direct drive ribbon tweeter amp and a current amp for the bass..that's exotic!
Mark
For a ribbon amp the post #6 by ZenMod in the Choke Susy thread has posibilities. As it is posted the value of the 2 inductors
needs to be large for a fullrange 8 ohm load. But the value of
the inductor needed for a tweeter amp is a lot less and it also
falls when the load resistance is lower. For a .2ohm ribbon with
a low end cutoff of 2k a pair of .1mh chokes might work for
a low power ribbon amp. I am taking the liberty of reposting
the circuite ZenMod posted earlier.
needs to be large for a fullrange 8 ohm load. But the value of
the inductor needed for a tweeter amp is a lot less and it also
falls when the load resistance is lower. For a .2ohm ribbon with
a low end cutoff of 2k a pair of .1mh chokes might work for
a low power ribbon amp. I am taking the liberty of reposting
the circuite ZenMod posted earlier.
Attachments
The most interestingNelson Pass said:It's straightforward enough. One way...
The other way...
The latter would be the most interesting. If you guys are that
hot to have one, I do have a schematic and it could be a project.
That's the one...So, Zen Mod say's you probably have something in the drawer???
Something like that maybeNelson Pass said:...an active current source capable of
about +/- 5 volts at 10 amp peaks. As a push-pull Class A,
this would need about +/-10V rails at 5A bias, for about 100
watts idle dissipation.
You could do this complementary Common Drain and with a
feedback loop for current source on the front end, or you could
do it complementary Common Source without feedback.
The latter would be the most interesting. If you guys are that
hot to have one, I do have a schematic ...
It always doesNelson Pass said:By the way, the delay in the Burning Amp #1 is simply that the
decision was made to have it coincide with the release of the new
site software. That has taken a little more time than originally
estimated.
When a real engineer says 5Min... Count on at least an hour! Or he aint thinking enough - Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.