No idea because I don't know what that is. I'm faking my way through this stuff if you didn't notice that already.
I should note that my interest here is more academic than practical, in the sense that I typically won't be pushing my amp to clipping and I am perfectly happy with the way it sounds. I simply noticed the negative phase hit clipping well before the positive phase while playing with my new o-scope, and it made me very curious. I thought it was maybe the result of me building the amp wrong (cathode bias resistor too high or too low).
I'm surprised George doesn't have an opinion on this.
I'm surprised George doesn't have an opinion on this.
You can minimize distortion by chosing a favourable operating point and load. As the load increases, the load-line becomes more horizontal, so it intersects a more evenly spaced set of characteristic curves. Also, pick an operating point where the characteristic curves are most evenly spaced.
It is a balance of Ia, B+ and Pd.
When bias Vgk is too low, Ia will be low. Vgk is negative, something like -14V, lower means larger such as -16V. But SSE has Vg at 0V so what you can do to increase Ia is reduce Vk by using lower value cathode resistor.
But then if B+ is the same, lower Vk means higher Vak hence higher Pd. Tube service life will be significantly lowered if used at excessive Pd. So to balance the increase in Ia, it is better to reduce B+ to keep design intended Pd.
When bias Vgk is too low, Ia will be low. Vgk is negative, something like -14V, lower means larger such as -16V. But SSE has Vg at 0V so what you can do to increase Ia is reduce Vk by using lower value cathode resistor.
But then if B+ is the same, lower Vk means higher Vak hence higher Pd. Tube service life will be significantly lowered if used at excessive Pd. So to balance the increase in Ia, it is better to reduce B+ to keep design intended Pd.
First off, by putting the scope probe on the speaker terminals it is observed that one side clips before the other. It is assumed to be the bottom, but this may not be the case. Some OPT's invert the phase, some don't. The only way to know for sure is to probe the plate of the output tube. This is not recommended for someone who is not sure if their scope can deal with the nearly 1000 volts that's present on the plate when the amp is driven hard enough to see clipping.
The picture in post #1 talks about bias voltage. This refers to a fixed bias amp like the TSE. More (negative) voltage reduces the tube current. In an SSE you would install a larger valued resistor, to reduce the tube current. This will increase the positive voltage on the cathode.
The SSE, like many SE amps will usually clip asymmetrically. It may be possible to play with the cathode resistor and find a point where the amp has symmetrical clipping for a given tube. This is likely not the point of optimum sound. Why?
The clipping point of any amplifier, tube, solid state, chip amp or even an amp powered by magic crystals, is determined by the load it sees. Lower the load impedance and the load current goes up, the active devices must push this added current into the load or it will clip, distort, or blow up. Increase the load impedance and the load will draw less current, but need more voltage to make the same power. At some point a higher load impedance will need more voltage than the power supply has, and the amp will clip.
Push pull amps have two output devices, one pushes current into the load on one half cycle, and the other pushes current into the load on the next half cycle. Given equal devices in each side, the clipping will be symmetrical.
Single ended amps have a single output device. In a tube amp the tube will attempt to pull power supply current through the OPT (and speaker) on one half cycle. It will then release the energy stored in the OPT's magnetic field into the speaker on the next half cycle. Balance depends on the load AND the OPT. Perfect electrical balance does not account for magnetic losses.
This can be balanced so the the store and release half cycles are equal at exactly one load impedance, usually the rated load for the OPT.
Your speakers are NOT 8 ohms (or whatever it says on the back) at all frequencies. Good speakers should have an impedance VS frequency curve associated with them......guess what, this isn't accurate either! The curve for a given speaker is valid for a specific power level, usually 1 watt, and under static conditions, IE one sine wave tone at a time.
That curve will usually have a large impedance peak around the woofer's resonance where the impedance goes up, a lot. My Yamahas are about 30 ohms at 70 Hz. Where does clipping happen most of the time, big fat bass notes. Need to drive a big bass note into 30 ohms? remember, a lightly loaded amp will clip for lack of supply voltage. This is why the SSE runs such a high supply voltage compared to most SE tube amps.
Say some musical information is pulling the woofer cone inward and a big bass drum hit tries to instantly reverse it's direction and push that cone the other way. Does that speaker remain 30 ohms? No it doesn't. That speaker is a coil with weight and momentum in a magnetic field, and therefore generating electricity, which it stuffs back into the amp, fighting the amp's attempt to change it. That drops the instantaneous dynamic impedance to a very low, and possibly negative value. This is called counter EMF, and one of the reasons that some amp enthusiasts hate GNFB.
I spent about a year of my life a long time ago studying this phenomenon, only to find an ever deepening rabbit hole. I finally gave up and vowed never to go there again. Just understand that steady state amp measurements with a fixed resistor load are valuable, and useful in designing an amp. They do not always correlate with sound quality.
I have some scope pictures of an SSE putting 45 volts peak to peak (16 Vrms) of a clean bass guitar note across my "8 ohm" Yamahas. This does not mean that the SSE was making 32 watts of clean power.
The picture in post #1 talks about bias voltage. This refers to a fixed bias amp like the TSE. More (negative) voltage reduces the tube current. In an SSE you would install a larger valued resistor, to reduce the tube current. This will increase the positive voltage on the cathode.
The SSE, like many SE amps will usually clip asymmetrically. It may be possible to play with the cathode resistor and find a point where the amp has symmetrical clipping for a given tube. This is likely not the point of optimum sound. Why?
The clipping point of any amplifier, tube, solid state, chip amp or even an amp powered by magic crystals, is determined by the load it sees. Lower the load impedance and the load current goes up, the active devices must push this added current into the load or it will clip, distort, or blow up. Increase the load impedance and the load will draw less current, but need more voltage to make the same power. At some point a higher load impedance will need more voltage than the power supply has, and the amp will clip.
Push pull amps have two output devices, one pushes current into the load on one half cycle, and the other pushes current into the load on the next half cycle. Given equal devices in each side, the clipping will be symmetrical.
Single ended amps have a single output device. In a tube amp the tube will attempt to pull power supply current through the OPT (and speaker) on one half cycle. It will then release the energy stored in the OPT's magnetic field into the speaker on the next half cycle. Balance depends on the load AND the OPT. Perfect electrical balance does not account for magnetic losses.
This can be balanced so the the store and release half cycles are equal at exactly one load impedance, usually the rated load for the OPT.
Your speakers are NOT 8 ohms (or whatever it says on the back) at all frequencies. Good speakers should have an impedance VS frequency curve associated with them......guess what, this isn't accurate either! The curve for a given speaker is valid for a specific power level, usually 1 watt, and under static conditions, IE one sine wave tone at a time.
That curve will usually have a large impedance peak around the woofer's resonance where the impedance goes up, a lot. My Yamahas are about 30 ohms at 70 Hz. Where does clipping happen most of the time, big fat bass notes. Need to drive a big bass note into 30 ohms? remember, a lightly loaded amp will clip for lack of supply voltage. This is why the SSE runs such a high supply voltage compared to most SE tube amps.
Say some musical information is pulling the woofer cone inward and a big bass drum hit tries to instantly reverse it's direction and push that cone the other way. Does that speaker remain 30 ohms? No it doesn't. That speaker is a coil with weight and momentum in a magnetic field, and therefore generating electricity, which it stuffs back into the amp, fighting the amp's attempt to change it. That drops the instantaneous dynamic impedance to a very low, and possibly negative value. This is called counter EMF, and one of the reasons that some amp enthusiasts hate GNFB.
I spent about a year of my life a long time ago studying this phenomenon, only to find an ever deepening rabbit hole. I finally gave up and vowed never to go there again. Just understand that steady state amp measurements with a fixed resistor load are valuable, and useful in designing an amp. They do not always correlate with sound quality.
I have some scope pictures of an SSE putting 45 volts peak to peak (16 Vrms) of a clean bass guitar note across my "8 ohm" Yamahas. This does not mean that the SSE was making 32 watts of clean power.
I guess I got lost in the explanation and forgot to explain what I did to my own SSE. Like others I put a switch in the cathode circuit to choose several different resistors at the twist of a knob. The best way to do this is to put a high value 5 watt resistor in the amp. I used 750 ohms. Wire a switch across it that can put 4 to 6 different resistors in parallel with the 750 ohm resistor to give values from 750 ohms (no parallel resistor) to somewhere in the 450 ohm range. These numbers may need adjustment for your particular tubes, or care in use. 450 ohms is too low for a 6L6GC, but just a little over spec for a 6550.
I then put my scope across the speaker leads and cranked up some real music, since nobody listens to pure sine waves. I found a track with nearly sinusoidal bass guitar notes, cranked it into mild clipping and picked the setting that looked the most symmetrical. Having a digital storage scope makes this a bit easier since I can keep pressing the capture button until I capture some notes.
My Yamaha NS-10M Studio monitors like a fairly high resistor value (don't remember which one now, 10 years later) but my open baffle Hawthorne Silver Iris speakers with a 15 inch woofer need as much current that the 6550 would stand. I ran them in the 450 ohm range where they were eating over 100 mA. The old Electro Harmonix tubes didn't seem to mind, and are still alive after 10 years. The Allied 6K7VG power transformer got far too hot to touch. Sucking 230 mA from a 150 mA rated transformer will do that. 10 years later that transformer is still good, but the clear coating over the black paint is blistered and there is a bit of rust on the laminations.
I then put my scope across the speaker leads and cranked up some real music, since nobody listens to pure sine waves. I found a track with nearly sinusoidal bass guitar notes, cranked it into mild clipping and picked the setting that looked the most symmetrical. Having a digital storage scope makes this a bit easier since I can keep pressing the capture button until I capture some notes.
My Yamaha NS-10M Studio monitors like a fairly high resistor value (don't remember which one now, 10 years later) but my open baffle Hawthorne Silver Iris speakers with a 15 inch woofer need as much current that the 6550 would stand. I ran them in the 450 ohm range where they were eating over 100 mA. The old Electro Harmonix tubes didn't seem to mind, and are still alive after 10 years. The Allied 6K7VG power transformer got far too hot to touch. Sucking 230 mA from a 150 mA rated transformer will do that. 10 years later that transformer is still good, but the clear coating over the black paint is blistered and there is a bit of rust on the laminations.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.