Thanks Kevin, SY,
Will what you're suggesting, the lower value resistors for the first triode, change the gain? And you mention a grid stopper resistor? I reckon you mean a resistor in series with the input, right after the volume control pot? Furthermore, how do you calculate the cathode current, and how does that relate to the plots on the ECC82 datasheet?
Best regards,
Sander Sassen
http://www.hardwareanalysis.com
Will what you're suggesting, the lower value resistors for the first triode, change the gain? And you mention a grid stopper resistor? I reckon you mean a resistor in series with the input, right after the volume control pot? Furthermore, how do you calculate the cathode current, and how does that relate to the plots on the ECC82 datasheet?
Best regards,
Sander Sassen
http://www.hardwareanalysis.com
Hi SY,
This is on a PCB, but I can certainly work around that, no problem. I would however like to use the ECC82. The kit came supplied with two of the Electro Harmonix brand, but I am currently using two JJ ECC82's that I bought seperately as it was suggested to me that the EH's would be of much lesser quality. Would be a waste to not use these and buy ECC88's instead no?
Do you still want those voltages? If so I can open the preamp up again and do some measurements.
Best regards,
Sander Sassen
http://www.hardwareanalysis.com
This is on a PCB, but I can certainly work around that, no problem. I would however like to use the ECC82. The kit came supplied with two of the Electro Harmonix brand, but I am currently using two JJ ECC82's that I bought seperately as it was suggested to me that the EH's would be of much lesser quality. Would be a waste to not use these and buy ECC88's instead no?
Do you still want those voltages? If so I can open the preamp up again and do some measurements.
Best regards,
Sander Sassen
http://www.hardwareanalysis.com
Think of it as sunk cost.😉
It would be interesting to see how much this circuit strains Vhk(max), so sure, I'd love to know the plate voltage of the first stage.
Hi SY,
I might just do that, in the mean time please find the voltages listed below in the schematic, this is after having let the preamplifier settle in for about 10-mins, with no signal and the volume turned fully down.
Both channels were within a few mV from each other, hence I'm only listing voltages for one channel here. Obviously all voltages are referenced to ground (GND).
Best regards,
Sander Sassen
http://www.hardwareanalysis.com
I might just do that, in the mean time please find the voltages listed below in the schematic, this is after having let the preamplifier settle in for about 10-mins, with no signal and the volume turned fully down.
Both channels were within a few mV from each other, hence I'm only listing voltages for one channel here. Obviously all voltages are referenced to ground (GND).
Best regards,
Sander Sassen
http://www.hardwareanalysis.com
Hi Sander,
At less than 1mA cathode current 6.3V/7.2K (0.875mA) your ECC82 is operating nowhere close to its most linear operating point. Before you proceed too much farther you need to check the rating of the high voltage secondary on your power transformer, if it is at least 25mA -30mA you have adequate current available to boost it to at least 4mA.
As a quick approximation replace the cathode resistor with a 1.5K and replace the 100K decoupling resistor to the 47uF cap with a 2.2K, this should still provide some ripple rejection, but won't drop too much voltage. You should have (roughly) somewhere between 100 - 130V on the plate of the ECC82.
Kevin
At less than 1mA cathode current 6.3V/7.2K (0.875mA) your ECC82 is operating nowhere close to its most linear operating point. Before you proceed too much farther you need to check the rating of the high voltage secondary on your power transformer, if it is at least 25mA -30mA you have adequate current available to boost it to at least 4mA.
As a quick approximation replace the cathode resistor with a 1.5K and replace the 100K decoupling resistor to the 47uF cap with a 2.2K, this should still provide some ripple rejection, but won't drop too much voltage. You should have (roughly) somewhere between 100 - 130V on the plate of the ECC82.
Kevin
Thanks Kevin,
The transformer is labelled as 60mA at 220V secondary, and to be honest it never even gets warm to the touch, even after hours of use. The transformer rating, including the other secondary windings (see image posted previously) is 16.2VA.
What you are suggesting is I omit both 3K6 (R13 and R17) resistors and replace them with a single 1K5 resistor right? Then reduce the value of the 100K series resistor (R14) to 2K2.
Best regards,
Sander Sassen
http://www.hardwareanalysis.com
The transformer is labelled as 60mA at 220V secondary, and to be honest it never even gets warm to the touch, even after hours of use. The transformer rating, including the other secondary windings (see image posted previously) is 16.2VA.
What you are suggesting is I omit both 3K6 (R13 and R17) resistors and replace them with a single 1K5 resistor right? Then reduce the value of the 100K series resistor (R14) to 2K2.
Best regards,
Sander Sassen
http://www.hardwareanalysis.com
Hi Sanders,
Yes, you got it.. 😀 Some tinkering may be required to get the best overall sound, but this seems like a good starting point.
Transformer secondary sounds like it should not have a problem with the additional current, but the 16.2Va rating doesn't sound quite right, just the 220V winding accounts of 13.2 of the total.. The fact that it does not get warm does seem to indicate that there is plenty of reserve.
Yes, you got it.. 😀 Some tinkering may be required to get the best overall sound, but this seems like a good starting point.
Transformer secondary sounds like it should not have a problem with the additional current, but the 16.2Va rating doesn't sound quite right, just the 220V winding accounts of 13.2 of the total.. The fact that it does not get warm does seem to indicate that there is plenty of reserve.
Hi Kevin, et al,
I've made the proposed changes, here's a updated schematic, again with all the voltages listed. I'm about to put it back into my system to see whether it sounds any different.
As you can see I used a 100-ohm grid stopper resistor as I did not have any other values below 1K available. I'm quessing this isn't a problem?
As before any feedback or suggestions are most welcome, the main question being whether the ECC82 now operating properly?
Best regards,
Sander Sassen
http://www.hardwareanalysis.com
I've made the proposed changes, here's a updated schematic, again with all the voltages listed. I'm about to put it back into my system to see whether it sounds any different.
As you can see I used a 100-ohm grid stopper resistor as I did not have any other values below 1K available. I'm quessing this isn't a problem?
As before any feedback or suggestions are most welcome, the main question being whether the ECC82 now operating properly?
Best regards,
Sander Sassen
http://www.hardwareanalysis.com
A quick test shows that it works, overall gain however is a lot higher. This is a bit of a downside, as I'm only able to use about 1/5 of the potmeter's range, beyond that it gets too loud.
Any suggestions?
Best regards,
Sander Sassen
http://www.hardwareanalysis.com
Any suggestions?
Best regards,
Sander Sassen
http://www.hardwareanalysis.com
Alright, I've calculated the gain for the common cathode amplifier, which is how the first triode is operated.
ECC82:
gm = 2200 microohmos (transconductance)
ra = 7.7Kohm (plate resistance)
mu = 17 (amplification factor)
For the gain we need to calculate the internal plate resistance first, as this will have increased due to the fact that we use a non-bypassed cathode resistor (Rk).
ra'(unbypassed Rk) = ra + (mu + 1)*Rk
= 7.7K + (18)*1.5K
= 34.7K
Now the gain is given by the following equation.
Av(unbypassed Rk) = (mu*Rp)/(Rp+ra')
= (17*33K)/(33K+34.7K)
= 8.2
Hence this is twice that of the original design, which had either a 2x or 4x gain. However the gain can thus be changed by changing Rp. With a gain of 2x in mind the following equation must be solved.
Av(unbypassed Rk) = (mu*Rp)/(Rp+ra')
2 = (17*x)/(x+34.7K)
2 = 17x/x+34.7K
2x+69.4K = 17x
69.4K = 15x
x = 69.4K/15
x = 4.7K
I'll be substituting the 33K resistor with a 4.7K resistor and adjust the series resistor in the power supply accordingly and report back whether this is accurate.
Best regards,
Sander Sassen
http://www.hardwareanalysis.com
ECC82:
gm = 2200 microohmos (transconductance)
ra = 7.7Kohm (plate resistance)
mu = 17 (amplification factor)
For the gain we need to calculate the internal plate resistance first, as this will have increased due to the fact that we use a non-bypassed cathode resistor (Rk).
ra'(unbypassed Rk) = ra + (mu + 1)*Rk
= 7.7K + (18)*1.5K
= 34.7K
Now the gain is given by the following equation.
Av(unbypassed Rk) = (mu*Rp)/(Rp+ra')
= (17*33K)/(33K+34.7K)
= 8.2
Hence this is twice that of the original design, which had either a 2x or 4x gain. However the gain can thus be changed by changing Rp. With a gain of 2x in mind the following equation must be solved.
Av(unbypassed Rk) = (mu*Rp)/(Rp+ra')
2 = (17*x)/(x+34.7K)
2 = 17x/x+34.7K
2x+69.4K = 17x
69.4K = 15x
x = 69.4K/15
x = 4.7K
I'll be substituting the 33K resistor with a 4.7K resistor and adjust the series resistor in the power supply accordingly and report back whether this is accurate.
Best regards,
Sander Sassen
http://www.hardwareanalysis.com
You'll have an impressive amount of distortion from this arrangement; the load line will be practically vertical.
You really have a fundamental problem: to linearize that tube (such as it can be!), you'll end up with more gain than you want. You can either build in attenuation so you're knocking the signal down, then amplifying it, or try running just the CF and see if unity gain is enough for your system. It may well be.
You really have a fundamental problem: to linearize that tube (such as it can be!), you'll end up with more gain than you want. You can either build in attenuation so you're knocking the signal down, then amplifying it, or try running just the CF and see if unity gain is enough for your system. It may well be.
Hi SY,
You'll need to explain that to me if you don't mind, as I don't know what you mean by that. I've put in the changes as I proposed, which resulted in the below schematic with the noted voltages.
Best regards,
Sander Sassen
http://www.hardwareanalysis.com
You'll need to explain that to me if you don't mind, as I don't know what you mean by that. I've put in the changes as I proposed, which resulted in the below schematic with the noted voltages.
Best regards,
Sander Sassen
http://www.hardwareanalysis.com
As for the gain, I need something with about 2 ~ 4x gain. This would give all my sources enough gain to fully utilize the potential of my amplifier. Unity gain is not enough, and for that I can use my passive preamplifier, which uses two stepped attenuators.
Best regards,
Sander Sassen
http://www.hardwareanalysis.com
Best regards,
Sander Sassen
http://www.hardwareanalysis.com
Rather than analysing the circuit algebraically, you'd do better to analyse it graphically. If you do that, you will discover that you have drawn a very steep loadline and that high distortion is inevitable. Now you know why people like to use 12B4 for line stages - it has mu = 6.
Hi EC8010,
Care to explain to me how that works? As is evident from my posts I have a basic knowledge of tubes and to be honest haven't used them in any of my audio projects for quite some time (since '93 to be exact), I could use a few pointers.
For your information, the below link has detailed information about the ECC82 tubes I'm using. This datasheet also has a number of graphs that I can't seem to grasp at the moment, I guess that's what you are referring to?
JJ ECC82
http://www.drtube.nl/datasheets/ecc82jj2003.pdf
Thanks and best regards,
Sander Sassen
http://www.hardwareanalysis.com
Care to explain to me how that works? As is evident from my posts I have a basic knowledge of tubes and to be honest haven't used them in any of my audio projects for quite some time (since '93 to be exact), I could use a few pointers.
For your information, the below link has detailed information about the ECC82 tubes I'm using. This datasheet also has a number of graphs that I can't seem to grasp at the moment, I guess that's what you are referring to?
JJ ECC82
http://www.drtube.nl/datasheets/ecc82jj2003.pdf
Thanks and best regards,
Sander Sassen
http://www.hardwareanalysis.com
The graph you need to look at is colloquially known as the anode characteristics and is a plot of Ia against Va for a variety of grid voltages.
Let's suppose that you wanted to analyse your circuit with a 4k7 anode load resistor and 280V HT. If there was zero current flowing through the resistor, there would be zero voltage drop across it, which would mean that anode would be at HT voltage. We can plot that point as (x, y) (Va = HT, 0mA).
Now suppose that the valve had switched on so hard that there was no voltage drop across it and all the voltage was across the anode load resistor. The current through that resistor would be determined by Ohm's law, 280V/4k7 = 59.6mA. We can plot that point as (0V, 59.6mA).
The voltage across the resistor is linear (V= IR), so if we have two points on the line, we can join them together. So we draw a line between the two points we just plotted. We've just drawn a loadline, and the loadline shows all possible combinations of anode voltage and anode current for that particular HT voltage and anode load resistor. Not only that, but the loadline tells us the Vgk that is required to set a particular anode voltage. We just look to see which grid line intercepts our chosen operating point. Even better, once we've chosen an operating point, we can investigate distortion. Suppose we were at Vgk = -6V. We could look to see the swing in anode voltage by moving to Vgk = -5V, then see if it's the same as the swing moving to -7V. If it isn't, then we have distortion. Try different loadlines, and you will find that shallower loadlines generated by large value anode load resistors produce lower distortion.
Let's suppose that you wanted to analyse your circuit with a 4k7 anode load resistor and 280V HT. If there was zero current flowing through the resistor, there would be zero voltage drop across it, which would mean that anode would be at HT voltage. We can plot that point as (x, y) (Va = HT, 0mA).
Now suppose that the valve had switched on so hard that there was no voltage drop across it and all the voltage was across the anode load resistor. The current through that resistor would be determined by Ohm's law, 280V/4k7 = 59.6mA. We can plot that point as (0V, 59.6mA).
The voltage across the resistor is linear (V= IR), so if we have two points on the line, we can join them together. So we draw a line between the two points we just plotted. We've just drawn a loadline, and the loadline shows all possible combinations of anode voltage and anode current for that particular HT voltage and anode load resistor. Not only that, but the loadline tells us the Vgk that is required to set a particular anode voltage. We just look to see which grid line intercepts our chosen operating point. Even better, once we've chosen an operating point, we can investigate distortion. Suppose we were at Vgk = -6V. We could look to see the swing in anode voltage by moving to Vgk = -5V, then see if it's the same as the swing moving to -7V. If it isn't, then we have distortion. Try different loadlines, and you will find that shallower loadlines generated by large value anode load resistors produce lower distortion.
EC8010 said:
Hi EC8010,
Thanks for that suggestion. I’m just looking for such a low Mu tube for a Mu follower. Do you have any idea where these are available in Europe?
Cheers 😉
<Edit>
Found! But uhhh, 12.50 eu a piece 🙁
I can't remember where I got mine from, but if I needed some more I'd just do an Internet search of American vendors and order them from there. With the disgraceful performance of the dollar it's really quite cheap.
Hi EC8010,
Just edited my post. Yes maybe ordering from the US is advantageous and is Sander also interested in?
Cheers 😉
Just edited my post. Yes maybe ordering from the US is advantageous and is Sander also interested in?
Cheers 😉
Hi Peter,
Yes, I'd be interested. I've been optimizing this design for the past two days now, simply by trying out different values for the components and measuring its performance with RightMark Audio Analyzer. I've now reached a point where linearity is flawless up to about 80KHz and THD as low as 0.05% at 6Vpp output. I've spent half a day trying to get better results, but that was a fruitless exercise. Hence that's the best I can get out of this particular design, can't see how I can significantly improve upon it.
Best regards,
Sander Sassen
http://www.hardwareanalysis.com
Yes, I'd be interested. I've been optimizing this design for the past two days now, simply by trying out different values for the components and measuring its performance with RightMark Audio Analyzer. I've now reached a point where linearity is flawless up to about 80KHz and THD as low as 0.05% at 6Vpp output. I've spent half a day trying to get better results, but that was a fruitless exercise. Hence that's the best I can get out of this particular design, can't see how I can significantly improve upon it.
Best regards,
Sander Sassen
http://www.hardwareanalysis.com
- Status
- Not open for further replies.