To drive the 9 woofers (8 ohm each) in my line array I would like to build a current/transconductance tube amplifier, as previous experience was quite nice. About 10W into 60-70 ohm would be nice.
A first attempt with a pentode push pull output stage was not successful, learned some new things from 6A3summer in this thread https://www.diyaudio.com/forums/tubes-valves/362804-current-drive-valve-amplifier-6.html
There is the possibility of applying current feedback in a global loop, some amplifiers with variable damping factor out there. As I have a 60-70 ohms load and do not need a lot of power, I think an OTL could be an option, but they do not always have a high output impedance.
So, curious about your ideas! Thanks, erik
A first attempt with a pentode push pull output stage was not successful, learned some new things from 6A3summer in this thread https://www.diyaudio.com/forums/tubes-valves/362804-current-drive-valve-amplifier-6.html
There is the possibility of applying current feedback in a global loop, some amplifiers with variable damping factor out there. As I have a 60-70 ohms load and do not need a lot of power, I think an OTL could be an option, but they do not always have a high output impedance.
So, curious about your ideas! Thanks, erik
An OTL would be just fine here. Only 2 PL519 or similar would be sufficiant here.
Even a 6AS7 / 6080 amp would be nice with less power.
Even a 6AS7 / 6080 amp would be nice with less power.
For 10W 60ohm, sqr(10/60) = 408mA. V=0.408*60 = 5.8V. That’s going to be spicy.
My current “complication” is an mini 8 triode/ch PP class A OTL into 32-55ohm. I run each PP pair as a separate output stage that combines on the output side.
I assume you’re going class AB?
As soon as you get into AB/B the impedance changes when the triodes switch on/off (given each triode is not identical). More so with AB or mosfet follower.
You could’ve run a tube master and then run a cathode mosfet follower hybrid for current.
Also an 6as7 on paper does 120mA 13W limit.. 4 tubes in PP pairs perhaps.
Circlotron perhaps?
My current “complication” is an mini 8 triode/ch PP class A OTL into 32-55ohm. I run each PP pair as a separate output stage that combines on the output side.
I assume you’re going class AB?
As soon as you get into AB/B the impedance changes when the triodes switch on/off (given each triode is not identical). More so with AB or mosfet follower.
You could’ve run a tube master and then run a cathode mosfet follower hybrid for current.
Also an 6as7 on paper does 120mA 13W limit.. 4 tubes in PP pairs perhaps.
Circlotron perhaps?
Last edited:
ErikdeBest,
You have nine 8 Ohm drivers in series, Right?
Each driver DCR is perhaps 6 Ohms, that is 54 Ohms DCR when in series.
The next question is, what is the impedance curve of the drivers in the line array, versus frequency?
Nine times the driver's impedance curve from the data sheet (and modified by the type of enclosure).
What about the 'enclosure'?
Is it a closed box?
Is it an open back?
What kind of enclosure?
Will the drivers have a flat frequency response when driven by a large impedance amplifier?
Will they be dampened by the enclosure and driver, without any electrical dampening?
Why not try a simple experiment?
1. Put an 8 Ohm power resistor across an amplifier output.
Connect a 60 Ohm, or 90 Ohm power resistor in series from the amplifier output to the line array of series connected drivers.
Your damping factor is approximately Unity (~ 1).
How does that arrangement sound?
2. Then connect the series line array directly across the amplifier output (that still has the 8 Ohm resistor across it).
How does that sound?
Compare the two different sounds of the line array.
Now, consider that using a true high impedance current source amplifier will sound more like #1 than #2 above.
if you doubt that, than repeat #1 above, but with a 300 Ohm series resistor (well, any amplifier that does not put out Lots of Watts is going to be very quiet with a 300 Ohm series resistor).
I used to use 5, 15, and 30 Ohm series resistors from a solid state amplifier to a nominal 8 Ohm 2 way loudspeaker. Fun listening.
Good luck experimenting and listening.
Have fun, that is what it is all about.
You have nine 8 Ohm drivers in series, Right?
Each driver DCR is perhaps 6 Ohms, that is 54 Ohms DCR when in series.
The next question is, what is the impedance curve of the drivers in the line array, versus frequency?
Nine times the driver's impedance curve from the data sheet (and modified by the type of enclosure).
What about the 'enclosure'?
Is it a closed box?
Is it an open back?
What kind of enclosure?
Will the drivers have a flat frequency response when driven by a large impedance amplifier?
Will they be dampened by the enclosure and driver, without any electrical dampening?
Why not try a simple experiment?
1. Put an 8 Ohm power resistor across an amplifier output.
Connect a 60 Ohm, or 90 Ohm power resistor in series from the amplifier output to the line array of series connected drivers.
Your damping factor is approximately Unity (~ 1).
How does that arrangement sound?
2. Then connect the series line array directly across the amplifier output (that still has the 8 Ohm resistor across it).
How does that sound?
Compare the two different sounds of the line array.
Now, consider that using a true high impedance current source amplifier will sound more like #1 than #2 above.
if you doubt that, than repeat #1 above, but with a 300 Ohm series resistor (well, any amplifier that does not put out Lots of Watts is going to be very quiet with a 300 Ohm series resistor).
I used to use 5, 15, and 30 Ohm series resistors from a solid state amplifier to a nominal 8 Ohm 2 way loudspeaker. Fun listening.
Good luck experimenting and listening.
Have fun, that is what it is all about.
Last edited:
A while back Tube Cad had a small OTL designed for a 32 ohm speaker that might be what you are looking for and no I didn't bookmark the issue it was in.
I think most OTL amplifiers have Voltage Negative Feedback, in order to Lower their output impedance.
Those are not 'current source' amplifiers.
Those are not 'current source' amplifiers.
https://www.diyaudio.com/forums/tubes-valves/373722-simple-otl-amp-driver-ef86-6sn7.html#post6695945
The amp does not use NFB instead use PF to achieve constant power.
I just have a briefly tested with 8 ohms speaker the frequency response, it is not flat as posted, it has top and bottom end raised, esp more on top gradually from 1Khz on. Now I know why my other OTL amp (NFB based) sound flat, less bass and less high. So you can design two type of OTL amp one is Constant Power source (Power Paradigm) and the other is Constant Voltage source.
Edit: to avoid apply too much power to the tweeter (when impedance goes high), keep the output low or shunt with resistor.
The amp does not use NFB instead use PF to achieve constant power.
I just have a briefly tested with 8 ohms speaker the frequency response, it is not flat as posted, it has top and bottom end raised, esp more on top gradually from 1Khz on. Now I know why my other OTL amp (NFB based) sound flat, less bass and less high. So you can design two type of OTL amp one is Constant Power source (Power Paradigm) and the other is Constant Voltage source.
Edit: to avoid apply too much power to the tweeter (when impedance goes high), keep the output low or shunt with resistor.
Last edited:
Although this is widely accepted and repeated, it may be overstated. For an excellent graphical example see :
Part 4: Of Loadlines, Power Output and Distortion
All good fortune,
Chris
They do. Plus my V=IR mu have had a mistype on the calc. 0.4*60= 24V.
Tubecad certainly have a multiple driver in series. The same article then followed by an OTL discussion.
Small OTL Amplifiers and Cathode-Coupled Amplifiers
More SE OTL & 32-Ohm Speakers
And possibly some local experience: OTL for dummies.
Also V-to-I PP OTL down the page here: V-to-I Amplifers
More SE OTL & 32-Ohm Speakers
And possibly some local experience: OTL for dummies.
Also V-to-I PP OTL down the page here: V-to-I Amplifers
Last edited:
I remember reading Broskie (tubecad) when researching OTLs. He made a very definite point about this in his OTL pages - I mistakenly had this associated with impedance in my head. It's transconductance.
Source: Brazilian OTL
Wow, many thanks for all the replies. I read a lot of Broskie's blogs and indeed saw the OTL for the 50 ohm headphone, and his comments on the advantages of a 72 ohm speaker when using OTL. I understand the theory and a simple Futterman or circlotron would drive these to crazy levels.
As noted by 6A3sUMMER, I would like to have a current output amplifier. Taking these triode based Futterman and Circlotrons and skipping the feedback may be a first approach, but the distortion increases and it is still not a true current amplifier. So, as commented by Lampie, my thoughts are indeed towards the pentodes, natural high impedance devices, possibly aided with some current feedback that reduces distortion and further increases the impedance. Unfortunately this does not go well with OPTs, so therefore OTL.
I collected some thoughts and will make some sketches later on, show them here!
As noted by 6A3sUMMER, I would like to have a current output amplifier. Taking these triode based Futterman and Circlotrons and skipping the feedback may be a first approach, but the distortion increases and it is still not a true current amplifier. So, as commented by Lampie, my thoughts are indeed towards the pentodes, natural high impedance devices, possibly aided with some current feedback that reduces distortion and further increases the impedance. Unfortunately this does not go well with OPTs, so therefore OTL.
I collected some thoughts and will make some sketches later on, show them here!
Thanks for the extensive answer 6A3sUMMER,
I did indeed have this idea of adding some power resistors in series with the woofer array as a first test for the effects of higher output impedance.
Currently the array is still configured in a 8 ohm fashion (three drivers in series, the the three string in parallel) and fed from a conventional amplifier, so this makes it easier to add some series resistance and still get nice SPL (I am using a 7 channel amplifier, 3 channels are still idle. If I want more power I could include a phase inverter on the amps input I could put another 2 channels to use for a balanced amplifier, that would still be happy feeding eg 16 ohms, 8 for the woofers and 8 for the series resistance)
I agree with you, it is indeed about having fun. Maybe I do not always sound like that, but I enjoy these endevours very much!
I did indeed have this idea of adding some power resistors in series with the woofer array as a first test for the effects of higher output impedance.
Currently the array is still configured in a 8 ohm fashion (three drivers in series, the the three string in parallel) and fed from a conventional amplifier, so this makes it easier to add some series resistance and still get nice SPL (I am using a 7 channel amplifier, 3 channels are still idle. If I want more power I could include a phase inverter on the amps input I could put another 2 channels to use for a balanced amplifier, that would still be happy feeding eg 16 ohms, 8 for the woofers and 8 for the series resistance)
I agree with you, it is indeed about having fun. Maybe I do not always sound like that, but I enjoy these endevours very much!
True, OTLs typically follow a low output impedance.
I'll be watching this thread, it seems current follows voltage and resistance. Surely you'd need some sensing feedback (either direct or through balance) given what you're after is a constant voltage, variable current with a variable impedance..
Maybe I'm missing something? (which is normally the case 😀)
Last edited:
I have a stash of 22JF6 and an idea to use these in enhanced triode mode (G1 at gnd, drive signal at G2). The operating point in this example would be about 100V at the plate and 200mA, when it needs around 60V on G2, 0 on G1. The drawn loadline is 100R. The load is provided by a mosfet gyrator, with about 50V over it, to allow for enough swing and still be in its linear region (couple volts from drain to source).
In real life I would probably use 3 valves in parallel, with about 150-160mA per valve. Or PP, with 2 valves per side, and a lot less 2nd order harmonics 🙂 Also, some previous curve tracing with added cathode resistance and G2 drive has shown more linear curves and less gain (higher voltage and swing needed on G2), both consequences of the current feedback and good for my intention. Dual drive from SmokingAmp, or cathode driven from a P mosfet are also possibilities...
It is of course class A and a lot of heat dissipation!
Also, adding some current feedback to a futtermann OTL should not be that hard, just saw this thread Transconductance converted Aleph J
But first get some power resistors and experiment with those 🙂
In real life I would probably use 3 valves in parallel, with about 150-160mA per valve. Or PP, with 2 valves per side, and a lot less 2nd order harmonics 🙂 Also, some previous curve tracing with added cathode resistance and G2 drive has shown more linear curves and less gain (higher voltage and swing needed on G2), both consequences of the current feedback and good for my intention. Dual drive from SmokingAmp, or cathode driven from a P mosfet are also possibilities...
It is of course class A and a lot of heat dissipation!
Also, adding some current feedback to a futtermann OTL should not be that hard, just saw this thread Transconductance converted Aleph J
But first get some power resistors and experiment with those 🙂
Attachments
I treated my simple PP pairs as separate slices only being connected on the output side of each slice’s output cap. This allows easier biasing and matching. Otherwise you’ll end up with a very complex system to tune as they’re all interdependent.
That steepness if vertical load line looks familiar!
A Google showed you can constant voltage load it by variable transconductance you’ll get the most out of the tube. A voltage governing triode Cathode tied amp with a variable current triode providing both would work to vary transconductance. Would need to experiment with LTSpice..
That steepness if vertical load line looks familiar!
A Google showed you can constant voltage load it by variable transconductance you’ll get the most out of the tube. A voltage governing triode Cathode tied amp with a variable current triode providing both would work to vary transconductance. Would need to experiment with LTSpice..
Last edited:
Note the point at the bottom about capacitor size and current.
(This is where the wisened will shake their heads)
I have four PP pairs (slices). Each with 4x330uF+0.22uF, so in total 5,200uF per channel to reduce impedance from the bipolar 330uF caps with 0.22uF film cap. I could reduce that number of 330uF but for now it remains until I hear the amp running. In the grand scheme £5 difference isn't going to break the bank.
Many OTL amplifiers have very steep load lines on the output tubes.
That makes for lots of distortion.
How to fix that . . .
In order for OTL to have low distortion, negative feedback is needed:
Voltage sensed (low output impedance). Not what ErikdeBest wants
Current sensed (high output impedance). More like what he is looking for.
The challenge of you OTL designers is to design an OTL that can drive 10 Watts into 60 - 90 Ohms of series connected drivers, and such amplifier having current feedback and output impedance of 900, 9,000 Ohms or more, and yet that is stable with that kind of load (series drivers and enclosure effects that makes the driver impedance versus frequency vary all over the map: resistive, capacitive, inductive, and combination RLC).
That is way beyond my knowledge.
. . . Come on OTL designers, come up with something for ErikdeBest.
That makes for lots of distortion.
How to fix that . . .
In order for OTL to have low distortion, negative feedback is needed:
Voltage sensed (low output impedance). Not what ErikdeBest wants
Current sensed (high output impedance). More like what he is looking for.
The challenge of you OTL designers is to design an OTL that can drive 10 Watts into 60 - 90 Ohms of series connected drivers, and such amplifier having current feedback and output impedance of 900, 9,000 Ohms or more, and yet that is stable with that kind of load (series drivers and enclosure effects that makes the driver impedance versus frequency vary all over the map: resistive, capacitive, inductive, and combination RLC).
That is way beyond my knowledge.
. . . Come on OTL designers, come up with something for ErikdeBest.
Last edited: