Hello.
I want to bias the Long-Tailed PI for the lowest distortion, so it should be biased as symmetrical as possible.
I have a 360V supply and I allow 90V for the tail. I am left with 270V for 100K RLoad, 2.7mA. Now the loadlines :
Is is right to bias at 177.5v since it's the middle point of the voltage swing, for the lowest distortion (and IMD) ? Or Am I doing it wrong ?
Thank you.
I want to bias the Long-Tailed PI for the lowest distortion, so it should be biased as symmetrical as possible.
I have a 360V supply and I allow 90V for the tail. I am left with 270V for 100K RLoad, 2.7mA. Now the loadlines :
An externally hosted image should be here but it was not working when we last tested it.
Is is right to bias at 177.5v since it's the middle point of the voltage swing, for the lowest distortion (and IMD) ? Or Am I doing it wrong ?
Thank you.
Last edited:
You have the AC swinging up to the 0v bias line, in reality you can't do this as you'll get grid current etc, swinging up to -1v is the usual, however the valve will only conduct over the range of the IP signal, so something like +0.5v, - 0.5v, so OP will be a lot less than you've drawn.
To get the least distortion at a glance, your load line wants to cross the grid lines at 90 degrees and you want to bias where the grid lines are evenly spaced. That said look at the Philips datasheet for the ECC83, it gives the THD for a wide range of values of Va, Rk etc, etc and the conditions for very low THD are all using big Ra's and high HT with the load line more or less horizontal and in the curved area of the grid lines.
Biasing for excellent THD is more complex and nuanced though, you have to take into account frequency, IP Z, IP C, OP Z, OP Z and god knows what else, the gain of the stage effects every parameter. Download and have a read of Merlins excellent article here - The Valve Wizard then apply what you've read on the bench, you need both to fully understand the subject.
Having said all this the THD of a stage might not be the most important parameter because if you have a PP OP stage some of this distortion will be rejected, also a bit of 2nd harmonic distortion can be a good thing, it's the main reason we use valves, that and the way they clip.
Andy.
To get the least distortion at a glance, your load line wants to cross the grid lines at 90 degrees and you want to bias where the grid lines are evenly spaced. That said look at the Philips datasheet for the ECC83, it gives the THD for a wide range of values of Va, Rk etc, etc and the conditions for very low THD are all using big Ra's and high HT with the load line more or less horizontal and in the curved area of the grid lines.
Biasing for excellent THD is more complex and nuanced though, you have to take into account frequency, IP Z, IP C, OP Z, OP Z and god knows what else, the gain of the stage effects every parameter. Download and have a read of Merlins excellent article here - The Valve Wizard then apply what you've read on the bench, you need both to fully understand the subject.
Having said all this the THD of a stage might not be the most important parameter because if you have a PP OP stage some of this distortion will be rejected, also a bit of 2nd harmonic distortion can be a good thing, it's the main reason we use valves, that and the way they clip.
Andy.
Thank you guys. I am learning hifi, prior I only built mic preamp, so it's the first time I have to design a Push pull power section from scratch.
The easiest route would be to copy an existing amp, but I will not learn. Yet if you have an example of "perfect/low distortion" hifi design Long tail Push pull schematic, I would like to take a look !
Hi rayma,
Input : I have a volume control, I "need" 2.5 or 3Vpp it would be perfect.
Output : I want to have enough for 2 EL34 push pull, with an Ultra linear OT (Hammond 1650N) without adding driver stage.
Diabolical Artificer,
I figured out it's impossible to bias completely symmetrical, as tube will always produce non linear distortion, no matter the bias (it's part of the tone). But if I am right there is 2 distortions, the input signal distortion (clipping) and output signal distortion (harmonic distortion). The perfect bias would be to be as close to the center for both input and output signals.
What bias point would you choose in my place ? The closet to 90° would be 1.5V.
The philips datasheet gives for 250Vb/100K/0.86mA=3.9% distortion.
The easiest route would be to copy an existing amp, but I will not learn. Yet if you have an example of "perfect/low distortion" hifi design Long tail Push pull schematic, I would like to take a look !
Hi rayma,
Input : I have a volume control, I "need" 2.5 or 3Vpp it would be perfect.
Output : I want to have enough for 2 EL34 push pull, with an Ultra linear OT (Hammond 1650N) without adding driver stage.
Diabolical Artificer,
I figured out it's impossible to bias completely symmetrical, as tube will always produce non linear distortion, no matter the bias (it's part of the tone). But if I am right there is 2 distortions, the input signal distortion (clipping) and output signal distortion (harmonic distortion). The perfect bias would be to be as close to the center for both input and output signals.
What bias point would you choose in my place ? The closet to 90° would be 1.5V.
The philips datasheet gives for 250Vb/100K/0.86mA=3.9% distortion.
"What bias point would you choose in my place ?" I tend to do some quick calculations then knock the circuit up on the bench, then use a pot and fixed resistor for the cathode resistor, then "tune" it in. I mainly use a LTP with a CCS that has this adjustability too, power up, give it time to warm up then adjust for best distortion and OP.
A LTP with a CCS has excellent AC balance but it's always a good idea to have a means of adjusting AC and DC balance in the OP stage, see - YouTube ( I think it's that one) also this might be worth a read - Practical Phase Inverters. bit here too - Vacuum Tubes Part 4 Differential: A more elegant solution. Article By Grey Rollins . You'll find lot's of articles on the subject online, just takes some time to find them. There's a few good thread on here about the LTP too discussing how the LTP is prone to higher 3H distortion than a cathodyne due to the different IP Z's of the two grids/triodes.
Hope that helps, Andy.
A LTP with a CCS has excellent AC balance but it's always a good idea to have a means of adjusting AC and DC balance in the OP stage, see - YouTube ( I think it's that one) also this might be worth a read - Practical Phase Inverters. bit here too - Vacuum Tubes Part 4 Differential: A more elegant solution. Article By Grey Rollins . You'll find lot's of articles on the subject online, just takes some time to find them. There's a few good thread on here about the LTP too discussing how the LTP is prone to higher 3H distortion than a cathodyne due to the different IP Z's of the two grids/triodes.
Hope that helps, Andy.
Why the use of 83 for driving EL34?
In my opinion is always better to use the double triode with a low Rp and the best linearity possible as 6H30 or 6N6. Or ECC99, also 12BH7.
They have normally a low gain bu it can be compensated by the previous stage.
In a right configuration you can get a great swing with low distortion and, very important, a low Zout of the splitter so you can drive better the grid of 34.
I don't like ccs but using them the balance is fine.
I trim the anode resistors to get the right values.
And, last but not least, the deep selection of tube for this purpose is mandatory, the sections must be similar.
Walter
In my opinion is always better to use the double triode with a low Rp and the best linearity possible as 6H30 or 6N6. Or ECC99, also 12BH7.
They have normally a low gain bu it can be compensated by the previous stage.
In a right configuration you can get a great swing with low distortion and, very important, a low Zout of the splitter so you can drive better the grid of 34.
I don't like ccs but using them the balance is fine.
I trim the anode resistors to get the right values.
And, last but not least, the deep selection of tube for this purpose is mandatory, the sections must be similar.
Walter
Last edited:
The ecc83 is for low distortion at low mA. Most of the high current tubes are higher distortion with low mA, the waveform is less evenly spaced.
Also for the gain because i want the pi to have enough signal gain for a pp of el34. I might change the tube later if it don't fit the need. But you are right the ecc83 is far from ideal as a Phase splitter i read a lot about it.
I will definitely look at the ccs tail. But it seem that waveform are only approximation to get close to an ideal point, yet for me it seems i need to bias between -0.5/120v and -3/250v, at -1.75v at 190v. Does it sound good to you guys ?
Also for the gain because i want the pi to have enough signal gain for a pp of el34. I might change the tube later if it don't fit the need. But you are right the ecc83 is far from ideal as a Phase splitter i read a lot about it.
I will definitely look at the ccs tail. But it seem that waveform are only approximation to get close to an ideal point, yet for me it seems i need to bias between -0.5/120v and -3/250v, at -1.75v at 190v. Does it sound good to you guys ?
Do you think that if it was possible to drive a p-p of EL34 directly from splitter someone in the world didn't built it?
The problem is the right swing at the ends of the OT trafo to get the power you want.
In addition, which configuration of output you are looking for?
Triode, UL or pentode; there different aspect to be considered.
And there are just some calculation to understand the right gain of each stage.
Walter
The best splitter I made was with 6H30 for my monos, but the gain, of course , is low.
Coupled with a first stage with 6922 the THD is very low.
The problem is the right swing at the ends of the OT trafo to get the power you want.
In addition, which configuration of output you are looking for?
Triode, UL or pentode; there different aspect to be considered.
And there are just some calculation to understand the right gain of each stage.
Walter
The best splitter I made was with 6H30 for my monos, but the gain, of course , is low.
Coupled with a first stage with 6922 the THD is very low.
Member smoking-amp has said some very interesting things about biasing and the gm curves.
Here are the search results for "constant gm" by smoking-amp.
By the way, does anyone have a pdf of
"System Design Factors for Audio Amplifiers" by M.V. Kiebert and P.R. Mallory Co."?
Smoking-amp posted a link but the file is bad.
Here are the search results for "constant gm" by smoking-amp.
By the way, does anyone have a pdf of
"System Design Factors for Audio Amplifiers" by M.V. Kiebert and P.R. Mallory Co."?
Smoking-amp posted a link but the file is bad.
Oh yeah, weird. I guess with the new site setup that times out. Anyway it was "constant gm" search with smoking-amp in the user name.
Still hoping somebody has a file they can post of
"System Design Factors for Audio Amplifiers" by M.V. Kiebert and P.R. Mallory Co.
Thanks
Still hoping somebody has a file they can post of
"System Design Factors for Audio Amplifiers" by M.V. Kiebert and P.R. Mallory Co.
Thanks