I'm not familiar with how Ebay works, but I know that RenoMark
is attending to his mother's funeral and is not available, and if
BillC has decided not to purchase the amplifier after all, there
are other buyers already standing by. Therefore I am assuming
it is some form of mistake or somebody's B.S.
is attending to his mother's funeral and is not available, and if
BillC has decided not to purchase the amplifier after all, there
are other buyers already standing by. Therefore I am assuming
it is some form of mistake or somebody's B.S.
Current drive with Tube amp
Ever since I read Nelson's current drive paper several weeks ago I've been fascinated by the concept, especially since I have Lowthers. I've ben working on an amp using BDTs (Beam Deflection Tubes) which are inhearantly good transconductance devices, but I wasn't getting the same effects Nelson was talking about in the paper.
Then it hit me that the problem was the output transformers, when we design a tube amp the OPT is set to deliver the optimal voltage swing/current value from the tube to 8 ohms at the speaker. But with the speaker going much higher in impedance at some frequencies the transconductance amp was running out of voltage swing. Since I was using a transformer with a lot of different taps I changed things around so the voltage ratio was half what it was before, WOW this radically changed the sound. I was now getting deep bass out of a Lowther!
After listening for a while I decided to see if different settings on the Decware Gizmos would work with this arrangement. It turns out the sound was the same no matter WHAT the gizmos were set to! The sound was actually better with them out of the circuit altogeher. That was very surprising since the gizmos had made a very significant improvement in the sound with other amps. I guess its not surprizing since the gizmo is in series with the speaker and will just get compensated for by the transconductance amp.
Every thing seemed fuller and smoother, deeper soundstage and even better subtle detail resolution than before. I was stunned and all by changing a couple wires!! But the more I listened the more I realized it was now too hot on the treble side. So this morning I tried adding C1 and R1. Looking at the article I couldn't decide which to use, I'm using Lowther EX3, in a dual quarter wave pipe, but with a front horm for the midrange, it didn't seem to match any of the examples in the paper. Since I had some 10 ohm resistors and some 8uf oil caps I decided to try them. Well that did wonders, the sound is now incredibly balanced, smooth, open, detailed, lively and just plain beautiful.
I've been listening now for the last 9 hours and not getting tired of it at all, I'm still enthralled by the music this makes. (I finally had to eat somthing and decided to share this with you)
Thank you Nelson for publishing your paper, I know its initiated a fair amount of controversy, but I can say, empahtically even, IT WORKS!! (even with tubes)
John S.
Ever since I read Nelson's current drive paper several weeks ago I've been fascinated by the concept, especially since I have Lowthers. I've ben working on an amp using BDTs (Beam Deflection Tubes) which are inhearantly good transconductance devices, but I wasn't getting the same effects Nelson was talking about in the paper.
Then it hit me that the problem was the output transformers, when we design a tube amp the OPT is set to deliver the optimal voltage swing/current value from the tube to 8 ohms at the speaker. But with the speaker going much higher in impedance at some frequencies the transconductance amp was running out of voltage swing. Since I was using a transformer with a lot of different taps I changed things around so the voltage ratio was half what it was before, WOW this radically changed the sound. I was now getting deep bass out of a Lowther!
After listening for a while I decided to see if different settings on the Decware Gizmos would work with this arrangement. It turns out the sound was the same no matter WHAT the gizmos were set to! The sound was actually better with them out of the circuit altogeher. That was very surprising since the gizmos had made a very significant improvement in the sound with other amps. I guess its not surprizing since the gizmo is in series with the speaker and will just get compensated for by the transconductance amp.
Every thing seemed fuller and smoother, deeper soundstage and even better subtle detail resolution than before. I was stunned and all by changing a couple wires!! But the more I listened the more I realized it was now too hot on the treble side. So this morning I tried adding C1 and R1. Looking at the article I couldn't decide which to use, I'm using Lowther EX3, in a dual quarter wave pipe, but with a front horm for the midrange, it didn't seem to match any of the examples in the paper. Since I had some 10 ohm resistors and some 8uf oil caps I decided to try them. Well that did wonders, the sound is now incredibly balanced, smooth, open, detailed, lively and just plain beautiful.
I've been listening now for the last 9 hours and not getting tired of it at all, I'm still enthralled by the music this makes. (I finally had to eat somthing and decided to share this with you)
Thank you Nelson for publishing your paper, I know its initiated a fair amount of controversy, but I can say, empahtically even, IT WORKS!! (even with tubes)
John S.
I was surprised at how well it works also. I just finished a
pair of .5 cf bookshelves with DX55s and they sound very good,
although being Lowthers, they still can't play very loudly on the
bottom end.
The Decware networks are based on the output impedance of
his tube amps, which are low compared to a current source,
so of course they are not optimal.
Those who want to play with this without building a
transconductance amp can simply use a higher power amplifier
and place R0 in series with the output instead of in parallel.
You'll have the losses, but the effect will be similar. In the
case of most Lowthers in sealed boxes, the R0 value would
be 20-30 ohms, and if you rear horn load them, you are looking
at lower values, more like 8 ohms or so.
pair of .5 cf bookshelves with DX55s and they sound very good,
although being Lowthers, they still can't play very loudly on the
bottom end.
The Decware networks are based on the output impedance of
his tube amps, which are low compared to a current source,
so of course they are not optimal.
Those who want to play with this without building a
transconductance amp can simply use a higher power amplifier
and place R0 in series with the output instead of in parallel.
You'll have the losses, but the effect will be similar. In the
case of most Lowthers in sealed boxes, the R0 value would
be 20-30 ohms, and if you rear horn load them, you are looking
at lower values, more like 8 ohms or so.
I'm also amazed with this article.Thanks Nelson for this one.It is excellent!
I have read it three or four times trying to understand everything.
I was also reading Martin J. King excellent site ( http://www.quarter-wave.com/ ) and i can see there are some similarities.
I happend to have a pair of Fostex FE 204 drivers and found old speaker boxes about 60 litres capacity,then I exchanged the front panel with the new one cut to fit Fe 204 and played with the port size (i went for bass reflex).I started with series resistance and found that for given resistance there is a proper capacity of the box and size of the port.I found the speaker sounded best with lower bass boosted about 1-2db which gave tuning at 47Hz for my box and 10 ohm resistor.Then I added 4mH inductor to attenuate midrange a little bit and 2uF Siemens MKV cap to boost the treble, both in parallel with resistor.Perfect balance for me.
Now I was all the time playing with this using copy of Alpeh P preamp and excellent sounding Gain Clone built with Black Gates ,Caddocks and Rikens as a power amp.I must admit I haven't used Alphs for a quite long time becouse in most cases (speakers) they sounded worse than Gain Clone.
The big surprice came when i hooked up ZEN v4 amp today and it sounded little better than Gain Clone with the set up listed above.So I went to other room and brought Aleph 3 that was unused for a couple of months and when I hooked it up I was amazed.With these speakers Aleph sounded much better than Gain Clone.The sound became full,rich and natural,kind of full bodied.Also very clean and transparent with the nice silence between the notes.Awsome.I was listening for a couple of hours.
This Aleph is built with quite cheap parts and that is a big surprice for me!
Now there are some things I don't quite fully understand.
Nelson, if there is a voltage amp and only R1,C1 and L1 (in parallel) in series with the speaker,is R1 still acting as R0 or not?
I'm using 10 ohm in parallel with 2uF cap and 4mH coil.Does it qualify as current driving?
Then reading your article I can understand what R0 does but then is R1/C1/L1 only a kind of notch filter?
What exactly R1 ,C1 and L1 does?
Bartek
I have read it three or four times trying to understand everything.
I was also reading Martin J. King excellent site ( http://www.quarter-wave.com/ ) and i can see there are some similarities.
I happend to have a pair of Fostex FE 204 drivers and found old speaker boxes about 60 litres capacity,then I exchanged the front panel with the new one cut to fit Fe 204 and played with the port size (i went for bass reflex).I started with series resistance and found that for given resistance there is a proper capacity of the box and size of the port.I found the speaker sounded best with lower bass boosted about 1-2db which gave tuning at 47Hz for my box and 10 ohm resistor.Then I added 4mH inductor to attenuate midrange a little bit and 2uF Siemens MKV cap to boost the treble, both in parallel with resistor.Perfect balance for me.
Now I was all the time playing with this using copy of Alpeh P preamp and excellent sounding Gain Clone built with Black Gates ,Caddocks and Rikens as a power amp.I must admit I haven't used Alphs for a quite long time becouse in most cases (speakers) they sounded worse than Gain Clone.
The big surprice came when i hooked up ZEN v4 amp today and it sounded little better than Gain Clone with the set up listed above.So I went to other room and brought Aleph 3 that was unused for a couple of months and when I hooked it up I was amazed.With these speakers Aleph sounded much better than Gain Clone.The sound became full,rich and natural,kind of full bodied.Also very clean and transparent with the nice silence between the notes.Awsome.I was listening for a couple of hours.
This Aleph is built with quite cheap parts and that is a big surprice for me!
Now there are some things I don't quite fully understand.
Nelson, if there is a voltage amp and only R1,C1 and L1 (in parallel) in series with the speaker,is R1 still acting as R0 or not?
I'm using 10 ohm in parallel with 2uF cap and 4mH coil.Does it qualify as current driving?
Then reading your article I can understand what R0 does but then is R1/C1/L1 only a kind of notch filter?
What exactly R1 ,C1 and L1 does?
Bartek
R1, C1 and L1 are a very low Q notch, usually coming in at
mid-band and taking down the upper midrange a bit and then
in cases where the top end of the driver is falling off, L1 gives it
a little kick up at the top. You can see some examples of these
curves at the very end (appendix) of the article on
www.firstwatt.com
Typically this notch is only a couple of decibels deep, depending
on the driver and so on.
I find myself using L1 only rarely. Often it will flatten the
response on the top a bit, but doesn't always sound better,
which is why you only want to use the networks I show as a
starting point.
Mind you, this doesn't work with a voltage amp. For that you
need to create a somewhat different network.
mid-band and taking down the upper midrange a bit and then
in cases where the top end of the driver is falling off, L1 gives it
a little kick up at the top. You can see some examples of these
curves at the very end (appendix) of the article on
www.firstwatt.com
Typically this notch is only a couple of decibels deep, depending
on the driver and so on.
I find myself using L1 only rarely. Often it will flatten the
response on the top a bit, but doesn't always sound better,
which is why you only want to use the networks I show as a
starting point.
Mind you, this doesn't work with a voltage amp. For that you
need to create a somewhat different network.
You can make a current source amp out of any "op amp"
power amp arrangement simply by taking your feedback
off a current-sensing resistor, most usually located between
the speaker's negative terminal and ground.
Rod Elliot has an article on his site showing how to create
variable output impedances in this manner.
This works as advertised, but in my experience does not have
the quite the same sound as a current source which does not
have the speaker in a feedback loop. Don't ask me why.
power amp arrangement simply by taking your feedback
off a current-sensing resistor, most usually located between
the speaker's negative terminal and ground.
Rod Elliot has an article on his site showing how to create
variable output impedances in this manner.
This works as advertised, but in my experience does not have
the quite the same sound as a current source which does not
have the speaker in a feedback loop. Don't ask me why.
CSA can be done from one opamp and 2 output transistors (pos. and neg. halves). 2 sensing resistors. Summed output. Then you can put speaker between the output and ground. Have done this in the eighties for measuring purposes as a voltage controlled current output amp for AC signals with grounded load.
Mr. Pass,
I am in the proocess of building a 5 to 8 watt transconductance amplifier and I have a few of questions.
First off I should say that I have reread all of the Zen articles with an eye toward designing a transconductance amp and was surprised at the abundance of ideas and alternatives to be found there. It made them all new again. My final design ended up looking not like a SOZ but more like a stripped down Aleph P built with IRFP240 and 9240 transistors. It has also taken every ounce of will power I have not to X the thing in the first iteration. I chose to use two CCs's on the Q sources because of mechanical layout issues and more flexibility in making future modifications. There is no gain adjusting resistor between the transistor sources.
My questions are:
1) In the Zen line stage article you give the method for calculating gain. How did you derive the 12 ohm "apparent source resistance" value for the 610 gain transistors?
2) Does changing from a dual supply (plus and minus) to a single supply improve efficiency with the SOZ? This is implied in a couple of places because this change has been lumped in with other efficiency improvements. If so then why?
3) Why is the Aleph P not considered a transconductance amplifier? There are CCS's on the sources and the drains of the gain transistors. After removing the input and ouput coupling circuitry, the only parts that would appear to make the difference are the drain load resistors to ground.
4) In the F1 functional schematic the implication is that the CCS's balance out - top and bottom. I feel uneasy about this. It seems to be an invitation for low level motorboating or worse. I had thought to make one set of CCS's somewhat larger in current value and then put in load resistors to soak up the difference and keeping everything nice and stiff. This is part of the reason for question 3. Am I worrying over nothing here?
I apologize for so many questions at once but they've been gathering up. Forgive me too if you have answered any of them previously. I have searched the site thoruoghly and haven't found anything. Thank you.
Graeme
I am in the proocess of building a 5 to 8 watt transconductance amplifier and I have a few of questions.
First off I should say that I have reread all of the Zen articles with an eye toward designing a transconductance amp and was surprised at the abundance of ideas and alternatives to be found there. It made them all new again. My final design ended up looking not like a SOZ but more like a stripped down Aleph P built with IRFP240 and 9240 transistors. It has also taken every ounce of will power I have not to X the thing in the first iteration. I chose to use two CCs's on the Q sources because of mechanical layout issues and more flexibility in making future modifications. There is no gain adjusting resistor between the transistor sources.
My questions are:
1) In the Zen line stage article you give the method for calculating gain. How did you derive the 12 ohm "apparent source resistance" value for the 610 gain transistors?
2) Does changing from a dual supply (plus and minus) to a single supply improve efficiency with the SOZ? This is implied in a couple of places because this change has been lumped in with other efficiency improvements. If so then why?
3) Why is the Aleph P not considered a transconductance amplifier? There are CCS's on the sources and the drains of the gain transistors. After removing the input and ouput coupling circuitry, the only parts that would appear to make the difference are the drain load resistors to ground.
4) In the F1 functional schematic the implication is that the CCS's balance out - top and bottom. I feel uneasy about this. It seems to be an invitation for low level motorboating or worse. I had thought to make one set of CCS's somewhat larger in current value and then put in load resistors to soak up the difference and keeping everything nice and stiff. This is part of the reason for question 3. Am I worrying over nothing here?
I apologize for so many questions at once but they've been gathering up. Forgive me too if you have answered any of them previously. I have searched the site thoruoghly and haven't found anything. Thank you.
Graeme
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.