Hello,
I have been studying tube amp since some few weeks now, and there's still something I dont quite grasp.
Seen from the moon, a tube is a voltage controlled current source, when the signal oscillate, it induces changes in the current going through the transformer (and it's inductance). When this current lower, the transformer will increase the voltage on the primary, and this tension will add up to the power supply's one.
It's quite a big inductance, we are talking about several henry's.
I understood that the voltage on the tube followed loosely the load line of the impedance ?
Why is this phenomenon never taken into account when designing an amp ?
Can it harm the tube in extreme cases ? if the quintessence point of the tube is already the maximum anode voltage, some 100V+ from the transformer can be harmful ?
Geek things :
what is the laplace transform for such a system ? (a tension source, a transformer (or a simple inductance ?) and a tube.
Thank you so much !
I have been studying tube amp since some few weeks now, and there's still something I dont quite grasp.
Seen from the moon, a tube is a voltage controlled current source, when the signal oscillate, it induces changes in the current going through the transformer (and it's inductance). When this current lower, the transformer will increase the voltage on the primary, and this tension will add up to the power supply's one.
It's quite a big inductance, we are talking about several henry's.
I understood that the voltage on the tube followed loosely the load line of the impedance ?
Why is this phenomenon never taken into account when designing an amp ?
Can it harm the tube in extreme cases ? if the quintessence point of the tube is already the maximum anode voltage, some 100V+ from the transformer can be harmful ?
Geek things :
what is the laplace transform for such a system ? (a tension source, a transformer (or a simple inductance ?) and a tube.
Thank you so much !
> this tension will add up to the power supply's one.
Yes, the plate voltage will "kick" above the supply voltage. Yes, a 6V6 at 300V will pull-down to under 100V and "kick" to over 500V. This is part of plotting a load-line.
It is a lot to digest in "some few weeks".
Yes, the plate voltage will "kick" above the supply voltage. Yes, a 6V6 at 300V will pull-down to under 100V and "kick" to over 500V. This is part of plotting a load-line.
It is a lot to digest in "some few weeks".
Because the inductance of the output transformer is never taken into account. The impedance is the only considered parameter.
Other question, how strict is the maximum tension rating on the anode. Like is it ok to reach 500v for a tube given for 300v maximum?
But the value of the inductance of output transformers is often taken into account, mainly for the frequency response at the low side. For drawing loadlines, the value of the inductance is not so important.
For power tubes, the maximum anode voltage is given while conducting a reasonable amount of current (in a normal working point). When the anode swings more positive, the anode current goes down. Power tubes are made to take these swings up (and down), so they can take the 500 V if they are rated for 300 V.
For power tubes, the maximum anode voltage is given while conducting a reasonable amount of current (in a normal working point). When the anode swings more positive, the anode current goes down. Power tubes are made to take these swings up (and down), so they can take the 500 V if they are rated for 300 V.
Last edited:
On the contrary, it is ALWAYS taken into account by competent designers. In fact, even by many less than competent. Spice models of OPTs include it.
Not sure where you get your information, but some of it is less than accurate.
C’est pas par là, c’est par ici! 😀
When the tube draws the plate voltage down while the grid is positive, the current in the transformer primary stores magnetic energy in the iron core of the OPT. As the voltage on the grid goes negative the energy stored in the iron core maintains the plate current. But only until the energy is dissipated, determined by the inductance of the primary winding.
So for an amplifier more inductance is better. All determined by the core size & the magnetic material used.🙂
So for an amplifier more inductance is better. All determined by the core size & the magnetic material used.🙂
In design triodes are usually treated a voltage controlled voltage generators in series with their plate resistance. OTOH, pentodes are modeled as a voltage controlled current source in parallel with their plate resistance.
But can also be taken the other way & sometimes are. In both cases these are assumed to be the small signal parameter.
Triodes are OK with a broad range of load impedance's, usually at least 2rp & more. For a power amp the LF rolloff is determined primarily by the rp of the tube in parallel with the load Rl reflected by the load back thru the OPT.
OTOH, The pentode rp is very high & does little to keep the load on the tube much less than the reflected load of Rl. And for max power output the load is critical.
So the triode power amp is able to deliver power to the load thru a given OPT to a lower frequency than a pentode. The 3db point is reached when the inductive reactance equals the tube rp in parallel with the reflected load.
More primary inductance is better. So is a low rp amplifier such as a triode. Do the Math.
There are several other factors the OPT designer needs to consider. To name a few, leakage reactance, winding capacity, skin effect & proximity effect.😀
But can also be taken the other way & sometimes are. In both cases these are assumed to be the small signal parameter.
Triodes are OK with a broad range of load impedance's, usually at least 2rp & more. For a power amp the LF rolloff is determined primarily by the rp of the tube in parallel with the load Rl reflected by the load back thru the OPT.
OTOH, The pentode rp is very high & does little to keep the load on the tube much less than the reflected load of Rl. And for max power output the load is critical.
So the triode power amp is able to deliver power to the load thru a given OPT to a lower frequency than a pentode. The 3db point is reached when the inductive reactance equals the tube rp in parallel with the reflected load.
More primary inductance is better. So is a low rp amplifier such as a triode. Do the Math.
There are several other factors the OPT designer needs to consider. To name a few, leakage reactance, winding capacity, skin effect & proximity effect.😀
Impedance is nominal if too low inductance loads it down, what happens at the lower end of the spectrum.
If inductance were irrelevant, transformers would be matchbox size, light and cheap.
Inductance is not indicated in schematics , simply because it is ASSUMED you will use a COMPETENT output transformer, and not a piece of crap.
I can´t believe some of the comments I read here, by supposedly knowledgeable Members.
Get around this: transformer *isolation* is not written either on schematics , maybe because they are *supposed* to stand mains or +V voltages? 😎
Try loading your output valve with a 5k wirewound resistor instead of an OPT and tell us how well that works out for you 😀
the transformer has wire resistance as well anyway.
at dc it is just the r of primary.
at ac it is r of primary plus reflected impedance from secondary.
as for a 5k resistor, that wont sound good !
at dc it is just the r of primary.
at ac it is r of primary plus reflected impedance from secondary.
as for a 5k resistor, that wont sound good !
Some designers use a power resistor across the OPT primary, secondary open to determine the optimum load resistance. That almost eliminates guessing what the OPT is doing. Very easy to do & provides immediate results.🙂
Yes it provides immediate results. Personally I will take correct over immediate anytime 🙂
A slightly better idea and no more difficult is to load the secondary with various resistors.
A slightly better idea and no more difficult is to load the secondary with various resistors.
“Almost eliminates guessing what the OPT is doing” - read: don’t even need to know it’s turns ratio. The trafo’s primary becomes a choke load, and you can vary Ra-a directly. Test at a sufficiently high (midband) frequency And the low end roll off doesn’t come into play.
Charming. Would you be so kind to be more precise? Whom are you refering to? When or how did they fake their knowledge?
That also eliminates a bunch of the available power from the output stage too unless the resistor is so large that it is only really defining a maximum load impedance.
That also eliminates a bunch of the available power from the output stage too unless the resistor is so large that it is only really defining a maximum load impedance.
------------------------
For a triode, theoretical max power occurs when the load equals 2rp. In practice it is more like 2.5rp. But at that loading D% is high. So good design at a load of 3-4 rp results are better. No need to do the unloaded OPT thing as long as the triode rp is known. Unless you are curious to try the test.😀 Worth trying for your own experience.
Load for max power in a pentode stage is a lot more critical. The objective is to get the loadline to run up into the knee of the plate characteristic curves. Any deviation from that & less power is available. A speaker load will reduce the power available simply because the loadline resulting from the reactive load is not a straight line. And at any moment it is a rather complicated ellipse.
I design a PP Pentode stage so that the load Rl is less than optimum. One reason is that speaker published impedance is a target. For example an 8R speaker measures more like 6.5R at the mid-band. But could be anything at 400 Hz & 10 KHz, usually higher. Another reason, a steeper loadine results in lower 3H harmonics.🙂
------------------------
For a triode, theoretical max power occurs when the load equals 2rp. In practice it is more like 2.5rp. But at that loading D% is high. So good design at a load of 3-4 rp results are better. No need to do the unloaded OPT thing as long as the triode rp is known. Unless you are curious to try the test.😀 Worth trying for your own experience.
Load for max power in a pentode stage is a lot more critical. The objective is to get the loadline to run up into the knee of the plate characteristic curves. Any deviation from that & less power is available. A speaker load will reduce the power available simply because the loadline resulting from the reactive load is not a straight line. And at any moment it is a rather complicated ellipse.
I design a PP Pentode stage so that the load Rl is less than optimum. One reason is that speaker published impedance is a target. For example an 8R speaker measures more like 6.5R at the mid-band. But could be anything at 400 Hz & 10 KHz, usually higher. Another reason, a steeper loadine results in lower 3H harmonics.🙂
Attachments
All I'm saying is that if you have the primary of the output transformer in parallel with a loading resistor, then AC current being produced by the finals will go through the OT and the loading resistor, and whatever current goes through the loading resistor isn't going to hit the voice coil.
If, for example, we decide to operate a 2A3 at the book operating point for a 2.5K load but we use a 5K output transformer and a 5K resistor, we will only get about half the power we otherwise would(assuming an ideal OT and resistive load). If the 5K transformer is allowed to run on its own, we get a little more power than with the resistor, more damping, and less THD.
For the pentode scenario I would wonder if you're throwing away enough power that you're back at triode strapped numbers in an effort to present more of a resistive load line.
If you're designing a guitar amp, this could be a way to generate extra THD at lower levels. That could have some merit.
If, for example, we decide to operate a 2A3 at the book operating point for a 2.5K load but we use a 5K output transformer and a 5K resistor, we will only get about half the power we otherwise would(assuming an ideal OT and resistive load). If the 5K transformer is allowed to run on its own, we get a little more power than with the resistor, more damping, and less THD.
For the pentode scenario I would wonder if you're throwing away enough power that you're back at triode strapped numbers in an effort to present more of a resistive load line.
If you're designing a guitar amp, this could be a way to generate extra THD at lower levels. That could have some merit.