Greetings friends. Having recently figured out how to hook up an oscilloscope+signal generator to an amplifier, I'd like to analyze my daily listener. So I put the sig gen in the 3k-65k range and put the scope on the 4-ohm resistor across the output terminals. Get a wave that looks like this:
As you can see, there's major rounding of that leading edge. What gives? Amp has no feedback so there's no phase lead cap to adjust. Here's the schematic:
Anything I should look at?
Thanks!
As you can see, there's major rounding of that leading edge. What gives? Amp has no feedback so there's no phase lead cap to adjust. Here's the schematic:
Anything I should look at?
Thanks!
That is not so bad for 10K. No jingling at the edges. Maybe get a shot of a 1K square wave. Looking at that shot I would expect the amplifier to make it past 20K with a sine wave.
I fit is not the output transformer I would suspect you need lower impedance driving the power tube. You could reconfigure the 6sl7 for 6sn7 after changing plate and cathode values. Would have lower gain though.
I fit is not the output transformer I would suspect you need lower impedance driving the power tube. You could reconfigure the 6sl7 for 6sn7 after changing plate and cathode values. Would have lower gain though.
Sure it would mean rolling off of high frequency. My eyeball suggests somewhere around 30K.
You could run a frequency sweep past 20K to know for sure,
You could run a frequency sweep past 20K to know for sure,
Here is a 1kHz sq wave
And here's a wave somewhere 10-20kHz
And it loses that flat top and looks more like a shark fin as the freq's increase
And here's a wave somewhere 10-20kHz
And it loses that flat top and looks more like a shark fin as the freq's increase
Ya but what makes you think that? What makes the wave curve like that? Pretend I didn't study engineering. 😉
Imho, your sig gen may have too slow a rise time, at something that looks like 50us, which is not going to adequately excite output transformer resonances out past 100kHz.
Another way to get a view of any ringing characteristics is to use no-load, or a capacitance load - if the amp is still stable (hopefully) then those loads may well show up high frequency damped ringing - which may be severe ringing but if they are damped (and not outright oscillation) then that is something you may be able to tweak feedback on to improve the stability margin. Not all output transformers generate an overshoot with damped ring with nominal load.
Another way to get a view of any ringing characteristics is to use no-load, or a capacitance load - if the amp is still stable (hopefully) then those loads may well show up high frequency damped ringing - which may be severe ringing but if they are damped (and not outright oscillation) then that is something you may be able to tweak feedback on to improve the stability margin. Not all output transformers generate an overshoot with damped ring with nominal load.
It does not hurt to measure from your signal generator while connected to the amplifier.
Also measure after the 6sl7 driver at the power tube grid.
The driver tube may be limited by the input capacitance of the power tube.
Also measure after the 6sl7 driver at the power tube grid.
The driver tube may be limited by the input capacitance of the power tube.
Your square wave depicts a smooth frequency response roll off. This is an especially desired behaviors for stages relying on global feedback.
This roll-off can be either due to leakage inductance or shunt capacitance, but most manufacturers rely on shunt capacitance roll-off, as it is less dependent of the load nature. Leakage inductance roll-off output transformers can give very different frequency responses with changing speaker loads.
The smaller the rounding radius of the leading edge, the higher frequency response is (for the given frequency). Bit of a lumberjack explanation, but easy to understand.
In the end, your square wave is okay. Nothing to worry about.
It would be worrying if you got a shape resembling a sinewave instead of a 10kHz squarewave.
This roll-off can be either due to leakage inductance or shunt capacitance, but most manufacturers rely on shunt capacitance roll-off, as it is less dependent of the load nature. Leakage inductance roll-off output transformers can give very different frequency responses with changing speaker loads.
The smaller the rounding radius of the leading edge, the higher frequency response is (for the given frequency). Bit of a lumberjack explanation, but easy to understand.
In the end, your square wave is okay. Nothing to worry about.
It would be worrying if you got a shape resembling a sinewave instead of a 10kHz squarewave.
That's really not bad for an SE amp with no feedback. I would not obsess over it. ;-) How does it sound?
To produce a perfect square wave you need infinite frequency response. Any rounding of the leading edge of the square wave shows that you don't have infinite frequency response.
However, as others have said, the fact that it still resembles a square wave at 10kHz suggests that your frequency response is good past 20kHz and probably rolls off around 30kHz.
Some of our comrades can glance at a reproduced square wave and tell you all you need to know about the amplifier...
Cheers, Mike
A better test would be a sine wave sweep from 0 to 30KHz.
A 1KHz square wave is a mixture of all the odd harmonics up into the many MHz region.
While some audiophiles insist on MHz performance it really isnt needed with the average human ear.
A 1KHz square wave is a mixture of all the odd harmonics up into the many MHz region.
While some audiophiles insist on MHz performance it really isnt needed with the average human ear.
It sounds correct, but I'm well aware of confirmation bias. A friend had commented that the bass was a bit lacking, switching over to the 4-ohm tap took care of that.
That's quite good performance actually, I've designed and built dozens of SE amplifiers and until you get into very exotic output transformers 30 - 40kHz small signal bandwidth is about what you will see.
To get good square waves at 10kHz would require an output transformer flat to at least several hundred kHz with minimal (read no) phase shift and higher bandwidth circuitry.
That's a neat little oscilloscope. Do confirm that it can display undistorted 10kHz square waves.
To get good square waves at 10kHz would require an output transformer flat to at least several hundred kHz with minimal (read no) phase shift and higher bandwidth circuitry.
That's a neat little oscilloscope. Do confirm that it can display undistorted 10kHz square waves.
The DSO138 is a 200kHz scope so it is unlikely to display a perfect rendition of a 10kHz square wave, but it should do a reasonable facsimile. I think I have to buy the kit. 😀