I wouls NOT recommend changing to solid state rectifiers. That amp looks a lot like one I was playing with and it ran the tubes not far from the max dissipation ratings. Replace tube rectifiers with semis and you will be way over max ratings. It will work probably but tube life will be very short.
To remove nfb, just clip the feedback resistor at the transformer output. Sound likely won't change too much other than getting louder, the transformer doesn't allow much feedback (and measured distortion was.... awful at best even with the nfb). Didn't sound too bad, though.😉
To remove nfb, just clip the feedback resistor at the transformer output. Sound likely won't change too much other than getting louder, the transformer doesn't allow much feedback (and measured distortion was.... awful at best even with the nfb). Didn't sound too bad, though.😉
I don't think so.
I have active Studio Monitors which are perfectly flat, and a SS amplifier with my hi-fi speakers which sound practically the same.
This goal you've posted is accomplished almost in every modern stereo system.
The reason I bought a tube hi-fi amp is to experience the sound of tubes interacting with the speakers impedance for music (current drive) which people claim to be more alive and musical.
That is why I put so much emphasis on the lack of NFB to have minimum damping factor and maximum speaker impedance interaction.
As I understand, in Pentode mode without NFB the frequencies of the speaker impedance peaks will distort first as the tube power will be used more there.
I agree, this is not what we want from a hi-fi amp, but this is an essential part of the unique sound of tube amps.
That is why I like having a range of output modes to choose from that naturally have different feedback response.
Triode mode may sound too "in control" just like SS to me... I shall listen very soon.
If this is the case, Tubes and this hobby are altogether obsolete. 😀DF96 said:
EDIT:
Thanks bwaslo.
Care to elaborate on "the transformer doesn't allow much feedback".
What did you hear/measure?
Last edited:
5% was at around 4W, and not all 2nd or 3rd harmonic either.
The limited bandwidth of the transformer would cause the amp to oscillate if too much fb were used.
The limited bandwidth of the transformer would cause the amp to oscillate if too much fb were used.
You assume you can't have a high output impedance and high distortion with solid state. Which is wrong.
You assume tubes can't have low output Z and low distortion. Which is also wrong.
You assume tubes can't have low output Z and low distortion. Which is also wrong.
Hmm....... A little knowledge is a Dangerous Thing....
I would recommend the O/P read through this--
http://www.tubebooks.org/Books/Atwood/Langham 1950 High-Fidelity Techniques.pdf
It May give him some better understanding as to what he's on about!
I would recommend the O/P read through this--
http://www.tubebooks.org/Books/Atwood/Langham 1950 High-Fidelity Techniques.pdf
It May give him some better understanding as to what he's on about!
Last edited:
Thanks for the informative book.
Mind you, referring to a person right beside you in third person is not polite.
Yes, I'm new at this and may talk nonsense at first. 😱 But I expect to be corrected, that's why I'm here.
I guess I wasn't born an engineer like many of you...
Briefly going through the book you posted, it seems like in the 50's 0.5W with 1% harmonic distortion was considered Hi-Fi and enough power for domestic music listening (well at least to the author).
It also appears that the author much prefers Triode to Beam Tetrode when distorted, and power tubes have distortion anyway at any usable power.
This means that peaks music signal will always be distorted first, and the speaker impedance will further suck power in the its peaks unless a negative feedback is used.
So the output tube configuration is significant for tube amplifier sound because it is always distorted to some degree.
OK I get it, Tubes and Transformers distort and limit bandwidth so NFB is essential in tube amps if Hi-Fi is the goal of the pursuit.
In the end, I will be the judge of what sounds good to me even if it turns out not so hi-fi even by 50's standards.
No, this is not what I assume. I perfectly aware that anything can be done and probably is done.famousmockingbird said:
My goal is NOT dead flat Hi-Fi tube amp, I got my SS for that! Furthermore, I don't think 150W with only 1W of undistorted audio power is efficient.
My goal it to get better understanding of how tubes sound and interact with the speakers in different output configurations and find the sweet spot for me.
This tube amp is not going to replace my SS as the main listening system, but will serve more like a learning project for pure enjoyment and knowledge, and it looks nice as a decorative item with the wood, chrome and glowing tubes.
Last edited:
Now we have established that you are not interested in hi-fi from your tube amp (despite your initial question) we can answer appropriately.
Great!
I will shoot some thoughts and questions so please correct me if I'm wrong.
BTW, the amp is not here yet, it is still shipping.
So the questions a sort of preparations without any actual measurements.
Now for some OPT primary impedance vs Plate voltage and Power dissipation questions.
bwaslo said I may exceed the tubes power dissipation with solid state rectifier in place of the tube one.
The OPT transformers of the amp I bought have 3.5k ohm primary impedance.
According to valve wizard (and ohms law):
Z = Va^2 / Pa
Pa = Va^2 / Z
Va = sqrt (Z*Pa)
Lets say the Va is 300V,
Pa = 300^2/3500 = 25.7W.
This automatically exceeds EL34 max dissipation Power... hmm.
So to get the most power from a tube, the primary impedance (reflected from the secondary at correct load) of the OPT should match but not exceeded?
So the less the impedance the tube "sees" the more power it uses.
A short at the transformer secondary will look like zero impedance to the tube which will require endless power (current) from the tube which will destroy it, or a tube has current limit?.
But it will keep the transformer safe from flyback voltage. Better a dead tube than a dead transformer?
How does this OPT impedance Power calculation related to the Power dissipation when biasing a tube?
If I understand correctly, a Class A single ended always conducts at maximum power even when no signal goes through it? If so, that is extremely inefficient.
How then the measured current across the cathode resistor to calculate the power dissipation it idle is related to the current that goes through the OPT primaries when amplifying?
If I'm not mistaken the reflected impedance is only "visible" to the tube through an AC waveform, this should mean that the tube is not conducting when there is no AC signal through it. But it is contradictory to what I know about Class A amps..... anyone cares to explain?
Thanks.
I think you've misunderstood the formula OR mixed up between output power and plate dissipation OR voltage swing and plate voltage.
Plate dissipation, Pdiss, is simply Vak * Ia. Vak is not necessarily Va since the "k" part (which is the cathode) can be grounded (fixed bias) or elevated (cathode bias). For example, in your 2nd schematic, the cathode is at 20V while plate is at 320V. This means Vak = 300v. Also, Ia can be simply calculated since we know Rk = 300R, which means Ia = 20V/300R = 66mA.
Finally, Pdiss is 20W.
So basically, on Class AB push pull, usually you bias the tube at 70% of max dissipation. Since you know the Va, you then bias the tube to flow current Ia to get this target Pdiss.
Plate dissipation, Pdiss, is simply Vak * Ia. Vak is not necessarily Va since the "k" part (which is the cathode) can be grounded (fixed bias) or elevated (cathode bias). For example, in your 2nd schematic, the cathode is at 20V while plate is at 320V. This means Vak = 300v. Also, Ia can be simply calculated since we know Rk = 300R, which means Ia = 20V/300R = 66mA.
Finally, Pdiss is 20W.
So basically, on Class AB push pull, usually you bias the tube at 70% of max dissipation. Since you know the Va, you then bias the tube to flow current Ia to get this target Pdiss.
Last edited:
ballpencil,
I understand.
But what about OPT primary in relation to biasing??
What if the OPT primary impedance is too low, how then we'll bias the amp?
What if I plugged my 8ohm speaker to the 4ohm output? this will result in half the apparent impedance to the tubes.
I understand.
But what about OPT primary in relation to biasing??
What if the OPT primary impedance is too low, how then we'll bias the amp?
What if I plugged my 8ohm speaker to the 4ohm output? this will result in half the apparent impedance to the tubes.
OPT primary impedance does not come into factor when determining the bias point. What matters is the primary resistance because this tells you how much voltage will be dropped in the primary before reaching the plate.
Let's say the primary resistance from the B+ entry to the plate connection is 100 ohms. With 66mA, this means there is a drop of 6.6V on the primary winding. With B+ of 320V, the plate will be at 320-6.6=313.4V. Not much difference but it might matter in other cases.
Let's say the primary resistance from the B+ entry to the plate connection is 100 ohms. With 66mA, this means there is a drop of 6.6V on the primary winding. With B+ of 320V, the plate will be at 320-6.6=313.4V. Not much difference but it might matter in other cases.
Page 37 of that book onwards I linked to will explain Load Impedance and the differences between Triode and Pentode, and how these two modes affects the operation of the output stage.
As Ballpencil said, Bias has nothing much to do with the output Transformer Impedance Ratio.
Bias is the amount of Current flowing under quiescent conditions at the normal supply voltages and no signal. Its a DC condition.
For SE Class-A, then this Bias will usually be set just under the maximum P-Diss of the output valve. Sometimes, if you dont care about the valve life, you can re-bias a bit more and get just a little more output, This also can affect the distortion the stage generates.
Its been my experience, these cheap Chinese amps tend to bias rather on the Hot side, and the supplied valves are generally not the highest quality, and have a short --but Hot life....
As Ballpencil said, Bias has nothing much to do with the output Transformer Impedance Ratio.
Bias is the amount of Current flowing under quiescent conditions at the normal supply voltages and no signal. Its a DC condition.
For SE Class-A, then this Bias will usually be set just under the maximum P-Diss of the output valve. Sometimes, if you dont care about the valve life, you can re-bias a bit more and get just a little more output, This also can affect the distortion the stage generates.
Its been my experience, these cheap Chinese amps tend to bias rather on the Hot side, and the supplied valves are generally not the highest quality, and have a short --but Hot life....
Last edited:
You can't get a 300V peak swing from a 300V supply rail, even without cathode bias. The valve needs some voltage across it in order to conduct. Then for Class A output power can at most be half the input power.
What you say contradicts this:
Source: The Valve Wizard -Single EndedValve Wizard said:
According to this, Class A will swing twice the HT voltage.
The plate will dissipate full power of 25W (DC) when no signal at the input of the power tube, and use LESS power when actually playing a sinewave.
Since the OPT is an AC device, it will only see the power of the sinewave which is always less than the DC biasing (25W).
In other words, the tube dissipates most power when idle (no AC), and the transformer least when idle.
When an AC signal is played through the tranformer, it comes alive and does its reflection magic from side to side (again, ONLY with AC signal!).
This means that when no AC signal is at the transformers primary windings, we should not worry about having no load at the secondary, unless the tube oscillates of course.
Last edited:
Hi James. Take the time to read the book and think about what DF96 is saying. I would trust what he says over valvewizzard, which also tends to have gitaramp schematics mixed in with the lot. Its not a bad site, but its not the be-all and end-all..
If you want, you can easily similate how a PSU can behave by using Duncan Amp's PSU Designer II software. Download it and configure it for your circuit with the quiescent 'nominal' current draw.
Then change that load and see what happens to your power supply. How will the current draw change you might ask? It will change with the swing in your driving signal. You can plot values you get from PSU II and discover what DF96 is saying. PSU Designer II yields quite reasonable values, and it is much easier than actually building a circuit and making measurements. Of course you can also calculate values using plate curves and Ohm's law if you think your power supply is on the closer side of perfect...
Depending on the quality of the power supply you will see that lower current drawn results in quite a quick swing in the DC voltage upwards. If this swing is likely with your input signal, but does not violate the operating limits of the Tube/Valve then that is a good thing.
But there is a limits to this wonderful swing, and they can be easily seen on the plate curves as well. The Voltage also has to cleanly swing down when the AC signal changes accordingly. So you need to look at your plate curves and make sure that your amplified signal does not get into the knees of the curves...
The single ended chinese amps that many of us have encoutered tend to be biased HOT. Why is that? Design weakness or is it on purpose? What does HOT bias mean? To me, it means the tubes are deep into the range of Class A... the cathode voltages are too low relative to the grid - the difference between grid and cathode is too small... This means the amplifier will go into those knees of the plate curves quite easily too, which is not good for Hi-Fi.
Can you fix these (typically older model) chinese amps? Yes, but you will likely discover that the power supply needs to be re-designed or perhaps replaced.. which means a different mains transformer.. which is costly..
Please trust us when we suggest that you really do want Hi-Fi for true sound reproduction. This does not necessarily mean you can not have power as well. One trusted way to get excellent Hi-Fi with lots of power is to build push-pull, and there are many good schematics you can copy from.
If you want to start with Single Ended, and get gobs of power too, then you will likely discover that the power supply will not be trivial. You won't get more power by strapping a pentode into triode mode (I think you know that though).
I guess what I am saying is this: after reading everything on this thread it is not clear to me what you want. It seems to be this:
1. More power from your pentode
2. But in Triode mode
3. Without Feedback (although judicious application of this can actually reduce distortion in the correct topology).
I can't say how you will manage this.
If you want, you can easily similate how a PSU can behave by using Duncan Amp's PSU Designer II software. Download it and configure it for your circuit with the quiescent 'nominal' current draw.
Then change that load and see what happens to your power supply. How will the current draw change you might ask? It will change with the swing in your driving signal. You can plot values you get from PSU II and discover what DF96 is saying. PSU Designer II yields quite reasonable values, and it is much easier than actually building a circuit and making measurements. Of course you can also calculate values using plate curves and Ohm's law if you think your power supply is on the closer side of perfect...
Depending on the quality of the power supply you will see that lower current drawn results in quite a quick swing in the DC voltage upwards. If this swing is likely with your input signal, but does not violate the operating limits of the Tube/Valve then that is a good thing.
But there is a limits to this wonderful swing, and they can be easily seen on the plate curves as well. The Voltage also has to cleanly swing down when the AC signal changes accordingly. So you need to look at your plate curves and make sure that your amplified signal does not get into the knees of the curves...
The single ended chinese amps that many of us have encoutered tend to be biased HOT. Why is that? Design weakness or is it on purpose? What does HOT bias mean? To me, it means the tubes are deep into the range of Class A... the cathode voltages are too low relative to the grid - the difference between grid and cathode is too small... This means the amplifier will go into those knees of the plate curves quite easily too, which is not good for Hi-Fi.
Can you fix these (typically older model) chinese amps? Yes, but you will likely discover that the power supply needs to be re-designed or perhaps replaced.. which means a different mains transformer.. which is costly..
Please trust us when we suggest that you really do want Hi-Fi for true sound reproduction. This does not necessarily mean you can not have power as well. One trusted way to get excellent Hi-Fi with lots of power is to build push-pull, and there are many good schematics you can copy from.
If you want to start with Single Ended, and get gobs of power too, then you will likely discover that the power supply will not be trivial. You won't get more power by strapping a pentode into triode mode (I think you know that though).
I guess what I am saying is this: after reading everything on this thread it is not clear to me what you want. It seems to be this:
1. More power from your pentode
2. But in Triode mode
3. Without Feedback (although judicious application of this can actually reduce distortion in the correct topology).
I can't say how you will manage this.
Last edited:
Is the 6n8p in some kind of SRPP? The schematics provided are not likely...
This appears to be the amplifier pictured:
http://www.aliexpress.com/store/pro...ng-Amplifier-Finished/924212_32270738481.html
This appears to be the amplifier pictured:
http://www.aliexpress.com/store/pro...ng-Amplifier-Finished/924212_32270738481.html
Thanks soulmerchant.
The 6n8p is the Chinese version of the common 6SN7 (GT).
I have bought this one: Rivals hifi exquis EL34 tube amps Single ended Class A handmade Scaffolding amplifier finished product-in Amplifier from Consumer Electronics on Aliexpress.com | Alibaba Group
I requested a schematic directly from the chinese seller: http://kfdown.a.aliimg.com/kf/UT8RI1YXzFXXXagOFbXr.jpg
Although it seems to be missing the Chok so I don't trust it's the correct one.
The 6N8P (6SN7) have their both halves in parallel.
A Pentode connection is illustrated here with somewhat big 1K G2 resistor.
This particular amp is designed to be used with the 6P3P which is equivalent to 6L6GC.
An EL34 may over dissipate and have short life in this amp but a 6L6GC will be just fine, is it?
As I understand it, the PSU in tube amps should provide stable DC and enough current without sagging, which is unrelated to the AC swing at the transformer primaries or tubes plate.
If the PSU can provide enough current the HT should not sag, correct?
Lets say the HT/B+ is solid 300V from an ideal PSU:
As I understand it, The Plate swings 600V AC peak-to-peak sinewave on a 300V DC, so from 0V to 600V;
I think we should make distinction between DC from the PSU which should be stable as a rock, and AC which swings and should not have any effect on the DC.
What you are saying is that there IS some interaction between the AC swing and the DC PSU, I understand that, nothing is ideal.
Correct me if I'm wrong.
Biasing colder which results in more negative voltage between the grid and cathode will reach cutoff (voltage limit) first.
Biasing hotter which results in less negative voltage between the grid and cathode will reach saturation (0V grid) first.
When the Kinks come into play? When the valve is closer to saturation (0V grid)?
The 6n8p is the Chinese version of the common 6SN7 (GT).
I have bought this one: Rivals hifi exquis EL34 tube amps Single ended Class A handmade Scaffolding amplifier finished product-in Amplifier from Consumer Electronics on Aliexpress.com | Alibaba Group
I requested a schematic directly from the chinese seller: http://kfdown.a.aliimg.com/kf/UT8RI1YXzFXXXagOFbXr.jpg
Although it seems to be missing the Chok so I don't trust it's the correct one.
The 6N8P (6SN7) have their both halves in parallel.
A Pentode connection is illustrated here with somewhat big 1K G2 resistor.
This particular amp is designed to be used with the 6P3P which is equivalent to 6L6GC.
An EL34 may over dissipate and have short life in this amp but a 6L6GC will be just fine, is it?
soulmerchant, what I understood from DF96 post that he is referring to the HT as a rail voltage (like in SS amps) which can not be exceeded without distorting.DF96 said:
As I understand it, the PSU in tube amps should provide stable DC and enough current without sagging, which is unrelated to the AC swing at the transformer primaries or tubes plate.
If the PSU can provide enough current the HT should not sag, correct?
Lets say the HT/B+ is solid 300V from an ideal PSU:
As I understand it, The Plate swings 600V AC peak-to-peak sinewave on a 300V DC, so from 0V to 600V;
I think we should make distinction between DC from the PSU which should be stable as a rock, and AC which swings and should not have any effect on the DC.
What you are saying is that there IS some interaction between the AC swing and the DC PSU, I understand that, nothing is ideal.
Correct me if I'm wrong.
EDIT Correction.soulmerchant said:
Biasing colder which results in more negative voltage between the grid and cathode will reach cutoff (voltage limit) first.
Biasing hotter which results in less negative voltage between the grid and cathode will reach saturation (0V grid) first.
When the Kinks come into play? When the valve is closer to saturation (0V grid)?
Simple, I want understanding which leads to predictability without to much trial and error. 🙂soulmerchant said:
Last edited:
No, this simply can't be with real tubes and transformer.
First, as i mentioned above, the primary transformer winding will drop several volts before so the plate itself will not get 300VDC. But okay let's assume that Va = B+ for simplification. There is the second problem:
As DF96 stated, all tubes need some Vak to conduct. Taken from valvewizard's page, here is a loadline for SE EL34 at Va = 300V.
An externally hosted image should be here but it was not working when we last tested it.
Imagine the purple loadline crossing Vg1=0v, it can be seen that Va does not reach 0V. By eye-approximation it is still around 10-15V which means the peak swing on that side is 285-290V.. this should conclude that Va=300VDC does not yield Vpp swing of 600V.
Last edited:
- Status
- Not open for further replies.