Have I to mod any resistor or cap value to use 12AU7 / ECC82 & 12AX7 / ECC83?
Could you post heater wiring?
Hi,
The Stereophile reviews and measurements on CAT SL-1 have drawn my attention, so I’ve spent some time simulating different variants of the circuits shown in this thread. I think the AnalogMetric circuit works fine. Only the SRPP cathode cap should be 150pF, not 100n, as it creates non-linearities around 1kHz.
I’ve tried to replace the input ECC82 tube with ECC88, as in the newer versions of CAT SL-1. According to my simulations all the current/voltage values are very close to the ECC82 version. I’ve replaced the cathode resistor with 630 Ohms, and the 150pF in the SRPP stage with 330pF to keep the HF behaviour stable. The ECC88 version has lower distortion but much higher gain. To lower the gain the only thing I could think of was to put an input divider. All the other things I’ve tried have changed the overall ballance of the circuit.
The simulated distortion is negligible - for an 18V pulse at the output on 47k dummy load the second harmonic is only 150uV.
BTW some 6moon review pics show there’s a 0,22uF cap at the input of the original preamp - it is connected directly to the attenuator circuitry. So maybe some observations in this thread are correct.
I’d be glad if SO has any ideas how to improve the ECC88 version and lower the gain without changing the balance the circuit has.
Asen
The Stereophile reviews and measurements on CAT SL-1 have drawn my attention, so I’ve spent some time simulating different variants of the circuits shown in this thread. I think the AnalogMetric circuit works fine. Only the SRPP cathode cap should be 150pF, not 100n, as it creates non-linearities around 1kHz.
I’ve tried to replace the input ECC82 tube with ECC88, as in the newer versions of CAT SL-1. According to my simulations all the current/voltage values are very close to the ECC82 version. I’ve replaced the cathode resistor with 630 Ohms, and the 150pF in the SRPP stage with 330pF to keep the HF behaviour stable. The ECC88 version has lower distortion but much higher gain. To lower the gain the only thing I could think of was to put an input divider. All the other things I’ve tried have changed the overall ballance of the circuit.
The simulated distortion is negligible - for an 18V pulse at the output on 47k dummy load the second harmonic is only 150uV.
BTW some 6moon review pics show there’s a 0,22uF cap at the input of the original preamp - it is connected directly to the attenuator circuitry. So maybe some observations in this thread are correct.
I’d be glad if SO has any ideas how to improve the ECC88 version and lower the gain without changing the balance the circuit has.
Asen
Hi reading this thread, I get a quite good understanding. BUT: there is a resistor in the upper valve of the u-follower: R6, that is 1K. Discussion said, it must be 1M. That is my understanding. BUT: Guys here built it with 1K. So the upper valve is not modulated the correct way... Please clarify, why not 1M, but 1K should be used...
In the circuit linked from post 42 it is R2 which should be much greater than 1k. At 1k the upper valve does not operate quite as intended as an active load. For that the grid needs to be AC coupled to the lower anode but DC coupled to the bottom of the 3k cathode bias ressitor. 1k is almost a short circuit in this position so the anode load for the lower triode is an active load with lower impedance than intended (although still highish) in series with a parallel CR of 0.22uF and 47k. Note that this is not a mu-follower; that would have the signal taken from the upper cathode not the lower anode.
The gain of the second stage will be a bit lower than intended and distortion will be a bit higher than intended. Fortunately the high amount of global negative feedback will largely hide this problem. Better performance (with the 1k value for R2) might be obtained by simply removing the capacitor and 47k so the two 12AX7 triodes have a bit more voltage so can run more linearly.
This circuit shows how forgiving valves can be. The circuit contains several mistakes, yet it still (sort of) works.
The gain of the second stage will be a bit lower than intended and distortion will be a bit higher than intended. Fortunately the high amount of global negative feedback will largely hide this problem. Better performance (with the 1k value for R2) might be obtained by simply removing the capacitor and 47k so the two 12AX7 triodes have a bit more voltage so can run more linearly.
This circuit shows how forgiving valves can be. The circuit contains several mistakes, yet it still (sort of) works.
If you remove the capacitor and 47k then the voltage in the cathode of V4A will be higher and consequently the bias of V3A will be lower , so I do no consider it a good idea .
Removing the 47K and the cap means that you change the mu-follower into a SRPP (if you take the output from the upper cathode).
In principle I agree with the main point of DF96: the anode currents of the input 12AU7 and the 12AX7 mu-follower are a bit too low. Maybe, thanks to the feedback it measures well but I would not be sure about the sound....
I would change R3 and R13 from 3K to 1.5K. This way the 12AX7 should work more or less at 0.65 mA and anode of the lower valve should be around 112V instead of 120V. Not a big deal, really!
I would also tweak R14 and R8 to get a more linear behavior from the input valve.
Last edited:
Well it's not a good idea as formulated , but with further changes it will work fine , of course the circuit will turned to SRPP , I know it , and instead of changing R3 and R13 from 3K to 1.5K , you can change the value of R10 such a way so the voltage at the anode of V2A remain 120V or close ( and you can add a decoupling capacitor for more filtering ) then you have to change the value of R7 too , as you see there is a place for a lot of improvements .
Last edited:
Gentlemen.
Sorry, but there seems to be a lot of "voodoo" embedded in this thread. Many years ago, 1950 or so, I learned from my father, a high end Audio "guru" from the 1930s, that the way you design wonderfully sounding audio amplifiers is to design a good sounding amplifier with about 15dB more gain than you need and then put about 15dB of negative feedback around the amplifier to reduce the gain to what you really need.
This superior technique has always worked well for me.
For instance my College Senior EE power amplifier using transistors was designed using that technique. The Mackintosh Amplifier Clinic came on Campus to test our amplifiers and declared my design as the first "straight wire with gain" amplifier they had ever measured.
I am "preaching" this way because the various circuit configurations that you adhere to
have all sorts of internal negative feedback loops. "Cathode Ray" writing in the Wireless World in the 1950s did articles on feedback loops within feedback loops.
If you don't watch what you are doing carefully you may wind up with an amplifier that perform worse than if you used a simpler configuration with limited feedback.
I look at the various forms of tube hookups as an introduction of internal loops of positive and negative feedback. The tube has no idea which circuit it is connecter into, it just does what the potentials between the electrodes tell it to do. If the voltage conditions tell it to distort the signal it will, if not it won't.
Hans J Weedon.
Sorry, but there seems to be a lot of "voodoo" embedded in this thread. Many years ago, 1950 or so, I learned from my father, a high end Audio "guru" from the 1930s, that the way you design wonderfully sounding audio amplifiers is to design a good sounding amplifier with about 15dB more gain than you need and then put about 15dB of negative feedback around the amplifier to reduce the gain to what you really need.
This superior technique has always worked well for me.
For instance my College Senior EE power amplifier using transistors was designed using that technique. The Mackintosh Amplifier Clinic came on Campus to test our amplifiers and declared my design as the first "straight wire with gain" amplifier they had ever measured.
I am "preaching" this way because the various circuit configurations that you adhere to
have all sorts of internal negative feedback loops. "Cathode Ray" writing in the Wireless World in the 1950s did articles on feedback loops within feedback loops.
If you don't watch what you are doing carefully you may wind up with an amplifier that perform worse than if you used a simpler configuration with limited feedback.
I look at the various forms of tube hookups as an introduction of internal loops of positive and negative feedback. The tube has no idea which circuit it is connecter into, it just does what the potentials between the electrodes tell it to do. If the voltage conditions tell it to distort the signal it will, if not it won't.
Hans J Weedon.
Sorry but I cannot see your point in keeping 120V rather than 112V. Also have a look at the 12AU7 working at low current with just 20K load....it's basically a variable mu triode even for very small signal.
My experience is that feedback can work well only if the circuit is well designed and devices are used properly (i.e. as linearly as possible).
I never said it was a good idea. A good idea would be to redesign the stage so that it acts as a proper mu-follower, instead of merely looking superficially like a mu-follower while actually being a poorly designed active load stage.Dimitris AR said:
An even better idea might be to design a better preamp, or not use a preamp at all!
What ?12AU7 , there is no 12AU7 in the circuit .
And the term variable mu that you use , are referred to
variable Transconductance tube , do you want to say that 12AU7 are variable Transconductance tube ? first time to hear it , it's not linear at low currents ( as most tubes ) but not variable Transconductance at all .
In the posted file the bottom circuit has a 12AU7 at the input.
I cannot see a practical difference between non linear and variable mu in such case!
