Hi,
On the subject of changing the feedback circuit responsible for RIAA, I am actually thinking about increasing gain of my phono stage by altering some of the resistors in this circuit.
Can I alter any of these feedback resistors here and just change the gain, not spoiling the RIAA equalisation?
On the subject of changing the feedback circuit responsible for RIAA, I am actually thinking about increasing gain of my phono stage by altering some of the resistors in this circuit.
Can I alter any of these feedback resistors here and just change the gain, not spoiling the RIAA equalisation?
Original gain of V1 (1/2 ECC83) is ( (ra//R3) / (rk+R4) ) 52k / 2.4k ≈ 22x, so there is some margin.
A small dc current (85μA) is flowing from the cathode V3 to the cathode of V1, so #2 is best solution.
A small dc current (85μA) is flowing from the cathode V3 to the cathode of V1, so #2 is best solution.
Changing R4 or shunting it will affect the RIAA EQ. Changing R6 or R7 won't, but it will upset the operating point of V3 which is direct coupled.
R6 and R7 are inside the feedback loop, so changing their values will barely change the closed-loop gain.
And I am also wrong about R4. It is Lipshitz's R0, and participates in gain but not in EQ.
Not a good start to the day :-(
Not a good start to the day :-(
Thanks for this, so regarding post 2, what is the reason to also add a capacitor - if the change in bias from the reduction in cathode resistance by adding a parallel cap is increasing the gain, why do you need a bias capacitor when there was not one there before?
R4 is 1.8K that set a current thru valve. If you change the value of this resistor you will change the static operational point of this tube. To not alter the static operational point, the new resistor must be connected in series with an capacitor to have effect only in AC and to not change DC.
ahh ok, I thought the point of the suggestion was to change the operating point in order to change the gain !
So the new resistor will sit parallel to R4 and change the gain by altering the feedback loop: I am guessing it will take some of the feedback and share this to ground via the resistor and capacitor, so reducing the level fed to the grid of V1?
Any idea of how to imagine how i might say get an additional 6dB of gain or so, with the choice of resistor value here?
So the new resistor will sit parallel to R4 and change the gain by altering the feedback loop: I am guessing it will take some of the feedback and share this to ground via the resistor and capacitor, so reducing the level fed to the grid of V1?
Any idea of how to imagine how i might say get an additional 6dB of gain or so, with the choice of resistor value here?
Just put an adjustable resistor, set at a level requested, measure it and replaced with a fix resistor.
...to change the operating point in order to change the gain...
That's the unwanted way to change gain, as it changes the dc setting to: #9 is right.
...and change the gain by altering the feedback loop: I am guessing it will take some of the feedback and share this to ground...
Not to 'take something of the feedback', the feedback is defined by ( ( Z-riaa-network + R4 ) / R4 ), in short (Zn + R4) / R4.
By changing R4, the gain is changed. Decreasing R4 increases the gain. As far as the open loop gain permits.
...so reducing the level fed to the grid of V1...
The level to the cathode is reduced, resulting in an increase of output level to maintain the same differential signal between grid and cathode.
So I have finally got my phono circuit working (see other thread here as I had problems with DC leaking through the C40 coupling capacitor!!).
I am going to try a 6dB gain with a parallel resistor of 1.8K in combination with R4 - can I just check that this is all I need to do, no need for adding a capacitor as well to start with?
Thanks,
Rich
Ok- so I added a 3.1K resistor in parallel with R4 (1.8K) and measured a resulting resistance of about 1.1K.
Following the provided maths this increased the gain by 4dB. I did this on one channel first and noticed the gain, and drive difference and then did the same on both channels.
Resulting voltages on the valve were changed quite a lot.
Anode v1 140V [from 175V] cathode 1.05V [from 1.4V]
It sounds a lot better now, more power and dynamics and actually a lot warmer and richer [The 3K was a 2W Allen Bradley] - is there any way of logically determining the limit of what is sensible within the valves operating boundaries?
Thanks again fro all the advice!!
Following the provided maths this increased the gain by 4dB. I did this on one channel first and noticed the gain, and drive difference and then did the same on both channels.
Resulting voltages on the valve were changed quite a lot.
Anode v1 140V [from 175V] cathode 1.05V [from 1.4V]
It sounds a lot better now, more power and dynamics and actually a lot warmer and richer [The 3K was a 2W Allen Bradley] - is there any way of logically determining the limit of what is sensible within the valves operating boundaries?
Thanks again fro all the advice!!
Ik =Ia = ( Vps - Va ) / Ra = ( 390 - 170 ) / 300k ≈ 0.67mA (original); Pa ≈ 220 * 0.67m ≈ 175mW.
Now the current is ( 390 - 140 ) / 300k ≈ 0.83mA (+24%); Pa ≈ 250 * 0.83m ≈ 208mW (+19%).
You can find the limits is the databooks, the ECC83 is well documented.
With the setpoint changed, the distortion level is also changed.
Now the current is ( 390 - 140 ) / 300k ≈ 0.83mA (+24%); Pa ≈ 250 * 0.83m ≈ 208mW (+19%).
You can find the limits is the databooks, the ECC83 is well documented.
With the setpoint changed, the distortion level is also changed.
Well I should have known better, as I had just spent a lot of my free time implementing the Lipshitz paper in Java code. All of it, including the limited-gain stuff, which led me down the path of solving cubics. And then using all that to implement transfer function classes for several dozen vintage preamps. By that time I knew quite a lot of the Lipshitz formulae from memory, so I really shouldn't have made that simple mistake. R4 (Lipshitz R0) appears in the non-inverting gain equation but not in the EQ equation.
I just had another listen to the revised circuit and it sounds great so I will either leave it alone with the old 1.82K and the new 3.2K in parallel on R4 or swap them both out for a single 1K resistor. The previous owner had also put in some Kimber cable between the RCA socket and the circuit board on the input (where these resistors are placed) and used solder like icing a cake so I may tidy the whole thing up...perhaps
Thanks
Rich
Thanks
Rich