Here you go:
An externally hosted image should be here but it was not working when we last tested it.
http://dcoward.best.vwh.net/analog/Aug00S&V.pdf
Vacuum-Tubes Operational Amplifiers
They used to made real ICs in glass bulbs, back then in Germany! Only coils and capacitors were usually needed to build a cheap radio receiver for masses.
K2-W opamp used 4 triodes (2 per bulb) and a pair of neon VR bulbs to shift zero. The design was classical, exactly like modern opamps: diffstage, VAS, output follower.
Hi SY,
While your circuit looks intersting, I have little clue as how to re-calculate the (gyrator) values in case a different crossover frequency is needed. For example, if I need a 220 Hz hi-pass with a Q of 0.707, how do I recalculate the crossover values. I have been reading some info about gyrators here and there, but this approach is not as widely used as SK topology, so there is still some mud present...
Thanks
Vix
While your circuit looks intersting, I have little clue as how to re-calculate the (gyrator) values in case a different crossover frequency is needed. For example, if I need a 220 Hz hi-pass with a Q of 0.707, how do I recalculate the crossover values. I have been reading some info about gyrators here and there, but this approach is not as widely used as SK topology, so there is still some mud present...
Thanks
Vix
Think of the circuit as looking exactly like a high pass 2nd order Butterworth passive crossover for a tweeter, but with a really high tweeter impedance.
We start by picking a load resistance. We want one big enough that it doesn't load down the buffer and that will be large compared to the gyrator's equivalent DC resistance. Let's say 47k, a nice easy value (this will correspond to R2 on the gyrator diagram in the Acheron article).
The C and L are calculated exactly like a 47k tweeter- using the calculator-ready formulas from Dickason, C = 0u011, L = 48.1H. Then use the gyrator equations to get the target L, keeping the equivalent DC resistance (corresponding to R1 ibid) below 5k or so (easy).
We start by picking a load resistance. We want one big enough that it doesn't load down the buffer and that will be large compared to the gyrator's equivalent DC resistance. Let's say 47k, a nice easy value (this will correspond to R2 on the gyrator diagram in the Acheron article).
The C and L are calculated exactly like a 47k tweeter- using the calculator-ready formulas from Dickason, C = 0u011, L = 48.1H. Then use the gyrator equations to get the target L, keeping the equivalent DC resistance (corresponding to R1 ibid) below 5k or so (easy).
SY, thank you for the answer, but I am still battling with some calculations. Excuse my ignorance and basic math questions, but, in order to be more specific, I will let you which part I understood and which not, so maybe it's going to be easier to go from there. I hope it will also motivate other people to try this kind of circuit. So,
In your case, its looks like R1 is chosen to be 5k1. C is chosen to be 10 nF. In order to get to 2nd order 110 Hz, butterworth hi-pass, R2 needs to be 100K, so the simulated inductor equals to 204 H. I hope this is correct. But what confuses me is the actual schematic: http://syclotron.com/acheron_signal_schematic.gif?phpMyAdmin=CayBDFTmCDnn1lPC7FR2Is66kHd
Let's go in order: C2=10 nF, C4=10nF, R7=5k1 (this is clear)
But, I don't get how R6=120 K and R8=200 K correlate here?
If I was to build it without schematic, I would omit R6 and R8 would be 100K.
In your answer above, it turns out that to get to the 220 Hz, R8 should be 47K?
I am sorry for perhaps stupid questions, but since there are no online calculators for this kind of circuit, it goes a bit harder...
In your case, its looks like R1 is chosen to be 5k1. C is chosen to be 10 nF. In order to get to 2nd order 110 Hz, butterworth hi-pass, R2 needs to be 100K, so the simulated inductor equals to 204 H. I hope this is correct. But what confuses me is the actual schematic: http://syclotron.com/acheron_signal_schematic.gif?phpMyAdmin=CayBDFTmCDnn1lPC7FR2Is66kHd
Let's go in order: C2=10 nF, C4=10nF, R7=5k1 (this is clear)
But, I don't get how R6=120 K and R8=200 K correlate here?
If I was to build it without schematic, I would omit R6 and R8 would be 100K.
In your answer above, it turns out that to get to the 220 Hz, R8 should be 47K?
I am sorry for perhaps stupid questions, but since there are no online calculators for this kind of circuit, it goes a bit harder...
PMILLET VARIABLE ACTIVE @ WAY TUBE CROSSOVER
Dear Mr: Millet,
I just Purchased your 2 Way Active Variable Tube Crossover PCB on The 29/04/16,
I Plan to Use it With my KLIPSCH K33 15" Bass Speaker and My JBL 2386*B Horn Flares With a 2" JBL Driver,
I would like to know as What Parts I will Need to Have the BASS Working from 15 or 20 HZ to around 500HZ , And My Midrange To Take over From 500Hz to 5000 Khz, and From There onwards I Will Connect To The Passive Klipshorn Corner Horn ORIGINAL Crossover and From 5000Khz Onwards .
Using My JBL SUPER BULLET 2402 TWEETERS.
I need to know The Parts List for The Crossover Frequenciers, Please Keep me Informed.
Kind Regards
Francis Jansz
Whereas... Pete has never really been happy with his Sallen-Key crossover.
Whereas... I have been more than delighted with my Acheron, a gyrator-based design.
Whereas... I have recently joined Pete in the Great State of Texas.
Whereas... Pete and I are looking for an excuse to get together and do some drinking,
Be it now announced that, if there's sufficient interest, we will be collaborating on a project. Pete has left the grunt design work to me; he'll do the PCB design. We're strongly leaning toward replacing the op-amp gyrators with pentodes, but using the basic cathode follower and servo circuitry from my original design.
Here's the options: this will be a second-order circuit. Most likely to get done quickly is a two way, with the bass section being handled by solid state and the upper section being done with tubes. Sections could be cascaded to give 4th order. If people want the bass to be done with tubes and/or a three-way, that could take a bit longer (my design and bench time are sadly limited these days), but if that's what it will take, that's what it will take. regulated supplies, of course. And single ended, though the enterprising could certainly add input and/or output transformers.
So... is there interest enough to inspire us to put the time in, start the breadboarding, and pop a few Belgian ales?[/QUOTE]
Dear Mr: Millet,
I just Purchased your 2 Way Active Variable Tube Crossover PCB on The 29/04/16,
I Plan to Use it With my KLIPSCH K33 15" Bass Speaker and My JBL 2386*B Horn Flares With a 2" JBL Driver,
I would like to know as What Parts I will Need to Have the BASS Working from 15 or 20 HZ to around 500HZ , And My Midrange To Take over From 500Hz to 5000 Khz, and From There onwards I Will Connect To The Passive Klipshorn Corner Horn ORIGINAL Crossover and From 5000Khz Onwards .
Using My JBL SUPER BULLET 2402 TWEETERS.
I need to know The Parts List for The Crossover Frequenciers, Please Keep me Informed.
Kind Regards
Francis Jansz
Whereas... Pete has never really been happy with his Sallen-Key crossover.
Whereas... I have been more than delighted with my Acheron, a gyrator-based design.
Whereas... I have recently joined Pete in the Great State of Texas.
Whereas... Pete and I are looking for an excuse to get together and do some drinking,
Be it now announced that, if there's sufficient interest, we will be collaborating on a project. Pete has left the grunt design work to me; he'll do the PCB design. We're strongly leaning toward replacing the op-amp gyrators with pentodes, but using the basic cathode follower and servo circuitry from my original design.
Here's the options: this will be a second-order circuit. Most likely to get done quickly is a two way, with the bass section being handled by solid state and the upper section being done with tubes. Sections could be cascaded to give 4th order. If people want the bass to be done with tubes and/or a three-way, that could take a bit longer (my design and bench time are sadly limited these days), but if that's what it will take, that's what it will take. regulated supplies, of course. And single ended, though the enterprising could certainly add input and/or output transformers.
So... is there interest enough to inspire us to put the time in, start the breadboarding, and pop a few Belgian ales?[/QUOTE]
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.