Checked all the values and the connections. It's all correct. Yeah me!
Since I was there, I also noticed the amp schematic is incorrect, what I labeled +B2 in the power supply schematic goes straight to the plates of the 6S19P tubes. There is no 68 ohm resistor in between. In fact, there isn't even that capacitor (C3) near the 68 ohm resistor. So, 116V get delivered to the plates of the 6S19P tubes.
The same voltage is fed to the plate resistor of the 12AX7A.
I'll update the schematic to reflect that tomorrow.
Here is the amp circuit schematic updated with the power supply reversed engineered numbers:
I ran this .tran simulation and here are the AC voltages before the grid of the 12AX7 (green), after C1 (blue) and at the output (red):
What do you say?
Looks normal?
And this is with the signal source set at 1V amplitude at 1 kHz:
I ran this .tran simulation and here are the AC voltages before the grid of the 12AX7 (green), after C1 (blue) and at the output (red):
What do you say?
Looks normal?
And this is with the signal source set at 1V amplitude at 1 kHz:
On that last picture, the top half of the blue curve is fatter than the bottom part.
Also, if I increase the signal amplitude, the top part is first to clip!
I found that interesting.
Also, if I increase the signal amplitude, the top part is first to clip!
I found that interesting.
You can change the grid bias toward better symmetry.
Counting on thumbs, this amp clipping is FAR more power than I want in headphones. It does not need further "optimization".
So, I built this circuit using a partially built kit that went along with the schematics.
It turns out the power transformer runs on the low side of specs for the HT secondary.
So much so that 105V come out of supply section.
That plate voltage at the 12AX7 at barely 34V in actual measurements. The grid bias is roughly at -1V.
I verified it comes down to a valid operating point, but very much in the bottom left corner of the plate characteristics chart.
Plugging real life values into the simulation, it’s evident, the 12AX7 will clip if fed more than 0.8V of signal amplitude.
I do not understand this operating point choice…
Well, I goofed...I took my last measurements with a 12AU7 plugged in so of course...
The 12AX7 gives this:
which gives more headroom for the plate voltage swings between 12 and 93 V.
Still, not the most linear area of the chart.
The 12AX7 gives this:
which gives more headroom for the plate voltage swings between 12 and 93 V.
Still, not the most linear area of the chart.
You have 99V-30V or 70Vpp or 24Vrms swing at the plate without any serious kink. If you get that through the final stage into 300 Ohm load, that is 2 WATTS. Headphones "typically" need 1/1000th of the power of loudspeakers. Do you need 2,000 Watts in speakers?
The headroom is nice to have though - If I subbed in a 12AX7 in place of the 6F12P in my VAPIDR board, the loadline would look like this (B+ 600V, Ua 400V, Ia 2mA:
Red = DC, Green = AC
Red = DC, Green = AC
Last edited:
Headroom is very nice to have.
And then, with the headphones on your head, the RCA plug comes partially out, ground return is disconnected . . .
and you get HUMMM at Full Power right into your ears. OUCH!
And then, with the headphones on your head, the RCA plug comes partially out, ground return is disconnected . . .
and you get HUMMM at Full Power right into your ears. OUCH!
Had a taste of that with that loose headphones jack I will replace tomorrow. Surprising!
Is that why in the original design ground was llifted?
I've never experienced that here - I build everything with a common ground apparently... I found out I can send audio through my DSP EQ using only one wire. (RCA with only the tip inserted. Earth as return I think).
PE, also called "safety ground" or "third wire ground" is a good conductor, sometimes equal to or better than the intended signal "ground" return. If everything is kept local, meaning that all of the PE "grounds" are plugged into the same wall outlet, and no other earth grounds are connected remotely, this usually works just fine. But it's not ideal, to the extent that no professional installation is done the way we do home installations.
Coax cables and twisted pairs both only give noise immunity because the same signal current flows in two nearby conductors in opposite polarity. Immunity to external noise and contamination depends on that external contamination being equal in both conductors so it can be cancelled at the receiver. In home installations, we violate that with every single-ended stereo interconnection; each of the interconnection shields (signal returns) carries half of its channels return current and half of the other channel's return current. And this is the best possible case, without including the possible path through PE. It's usually good enough for home use, where things can be kept fairly local and interfering noise fairly low.
If we wanted to do better we could separate signal grounds from PE and/or make inter-chassis connections with twisted pair plus shield (floated at one end). A step up from that would be a change to the receiving circuit that slightly lifts received return current from internal signal ground. A big step up from that is a differential receiver, where only the difference between received hot and return is recognized. This can then be optimized by making impedances on both ends equal for signal hot and return, to maximize the equality of contaminating signal voltages, for best rejection. This is a professional interconnection. Interestingly it does not require "balanced" equal voltages on signal hot and return, only equal impedances.
All good fortune,
Chris
Coax cables and twisted pairs both only give noise immunity because the same signal current flows in two nearby conductors in opposite polarity. Immunity to external noise and contamination depends on that external contamination being equal in both conductors so it can be cancelled at the receiver. In home installations, we violate that with every single-ended stereo interconnection; each of the interconnection shields (signal returns) carries half of its channels return current and half of the other channel's return current. And this is the best possible case, without including the possible path through PE. It's usually good enough for home use, where things can be kept fairly local and interfering noise fairly low.
If we wanted to do better we could separate signal grounds from PE and/or make inter-chassis connections with twisted pair plus shield (floated at one end). A step up from that would be a change to the receiving circuit that slightly lifts received return current from internal signal ground. A big step up from that is a differential receiver, where only the difference between received hot and return is recognized. This can then be optimized by making impedances on both ends equal for signal hot and return, to maximize the equality of contaminating signal voltages, for best rejection. This is a professional interconnection. Interestingly it does not require "balanced" equal voltages on signal hot and return, only equal impedances.
All good fortune,
Chris
I was referring to the RCA phono plug at the input of the headphone amplifier, like from a CD player with 2-wire power cord, that is plugged into 110V or 220V, (and uses a floating "ground" return on the RCA output plug).
A non 3 wire power cord and plug on a US 3 wire outlet with ground, neutral, and hot . . . If your living room stereo system RCA signal plug comes partly out (disconnecting the ground return, before the center signal pin), your speakers fill the room with 60Hz at max volume.
I did not mean the 1/4 Inch plug from the headphones to the amp output.
A non 3 wire power cord and plug on a US 3 wire outlet with ground, neutral, and hot . . . If your living room stereo system RCA signal plug comes partly out (disconnecting the ground return, before the center signal pin), your speakers fill the room with 60Hz at max volume.
I did not mean the 1/4 Inch plug from the headphones to the amp output.
Well, I think I have learned quite a bit with this circuit.
I built the circuit using a scrapped kit, and I am quite surprised by how good it sounds and its power output.
I should conclude this exercise by measuring the output from an objective standpoint, I guess. I already know I get 8 Vpp before clipping in my test headphones. But, I should figure out what a headphones dummy load should be and how to go from there.
Thanks so much to all of you who helped in my journey through this circuit. I do not take your help for granted and I am in fact very thankful.
I built the circuit using a scrapped kit, and I am quite surprised by how good it sounds and its power output.
I should conclude this exercise by measuring the output from an objective standpoint, I guess. I already know I get 8 Vpp before clipping in my test headphones. But, I should figure out what a headphones dummy load should be and how to go from there.
Thanks so much to all of you who helped in my journey through this circuit. I do not take your help for granted and I am in fact very thankful.
B+ 600V and 400V on the anode of a 12AX7 or equivalent seems a bit high. Would be great if the real tube could cope with such voltages but where do we find such a gem?
300V on the anode is normal operation max, some SQ type can take 330V, but i would not try 400V, did you try that In real life?
I remember some McIntosh amp where reported as having troubles because of running this tube above specs.