i built this amp on my way to the existing circlotron amp that i'm using now.
It measured as advertised.
One thing that will help you, is to figure out whether the gross distortion is coming from a ground loop.
I'm not sure if your meter can help you identify this, failing which a balanced input soundcard and FFT software may be the way forward.
It measured as advertised.
One thing that will help you, is to figure out whether the gross distortion is coming from a ground loop.
I'm not sure if your meter can help you identify this, failing which a balanced input soundcard and FFT software may be the way forward.
Could you elaborate on what you have in mind? I don't see any possibility of ground loops in my setup, but I'd be happy if you could point out something I might have overlooked.
Chris
it may well be something else other than a grounding issue in your case. But when I first started messing with circs, I spent some time chasing down wild-geese due to grounding setups.
If your analyser shows that there are large amplitude harmonics at frequencies much lower than your signal, then that's a dead giveaway. I don't know anything about the Keithley analyser you're using so afraid I can't offer much help with setting up the measurement for this.
Out of curiosity though, why are you using an external siggen and not the one which is built into your analyser ?
Partly a matter of convenience (easier to set the desired frequency and amplitude, and more accessible), and also because my external generator produces a somewhat lower distortion sinewave than the built-in one.
I also looked at FFT curves, and there is no sign of anything untoward other than the high level of the harmonics of the signal generator fundamental. Principally, it is second harmonic.
Chris
Well, it's just a small battery-operated device, fully floating. So it could be used as a single-ended or balanced source, depending on what it's hooked up to.
Chris
Last edited:
Yes, quite sure. To quote from the users' manual, "The maximum common-mode voltage (voltage between INPUT LO and the chassis ground) is 500V peak." So both the LO and HI inputs (- and +) are floating wrt ground.
And just to make absolutely sure (since I agree that the distortion results could have been suggestive of something in the measuring setup that was inducing an asymmetry), I also verified that nothing changes if I temporarily run the Keithley meter through a "cheater plug" that removes its grounding, so that it is unambiguously floating.
And likewise, when I look at the FFT plots, using a Velleman USB scope hooked to a laptop, the setup is floating.
Chris
Sorry Chris, I should have asked if both inputs were substantially isolated from ground, not just floating. For example, your amp's output terminals are floating but, being tethered to ground through about 100 ohms each, are not much isolated from it.
For accurate balanced measurement, both input terminals on the distortion analyzer should measure very high and equal resistance to its own ground and to the amplifier signal ground. The factors you mention don't necessarily mean that the Keithley is set up this way. Maybe it is, but if so, I'd expect to see it called out in the product lit.
For accurate balanced measurement, both input terminals on the distortion analyzer should measure very high and equal resistance to its own ground and to the amplifier signal ground. The factors you mention don't necessarily mean that the Keithley is set up this way. Maybe it is, but if so, I'd expect to see it called out in the product lit.
Last edited:
I'm not sure the Keithley is suitable for this type of balanced measurement. I think Jan has a little familiarity with this meter, maybe we could get his thoughts.
One potentially cheap and easy verification is hooking a cheap speaker up and listening to the output. 2% is fairly audible and any hums, etc. would show up as well. Just thinking out loud.
One potentially cheap and easy verification is hooking a cheap speaker up and listening to the output. 2% is fairly audible and any hums, etc. would show up as well. Just thinking out loud.
I know what you mean, but the Keithley is designed to allow its input terminals (both + and -) to be anything up to 500 V away from ground (chassis) potential. And indeed, I verified that the resistance between - terminal and ground, and between + terminal and ground, exceeds 20 M ohms. And, as I mentioned, the high distortion readings I get remain unaltered if I isolate the Keithley meter from ground altogether, using a cheater plug.
And, furthermore, I see high harmonic distortion in FFT plots, using a USB oscilloscope connected to a laptop with no ground connection.
Chris
Hum measures at about 0.5mV into 8 ohms. It is essentially inaudible on my high-efficiency Lowther speakers. I never really get the chance to hear the high distortion, since for practical listening it probably never even approaches 1W. (I should add that the distortion drops a lot below 1W. For example, about 0.25% at 100mW, rising to about 1% at 1W.)
Just looking at the waveform into an 8 ohm dummy load, I can actually see the sinewave peaks starting to round off as I crank up the power beyond about 12W or so.
Chris
Last edited:
Well, assuming you've also verified 0 VDC between your analyzer ground and your amp ground, I think it should work. An offset between the two grounds could conceivably cause an artificially high even-order distortion reading. (Perhaps this is the "ground loop" issue that kasey197 referred to earlier?)
FWIW, I've found that cheater plugs don't always provide enough isolation for low distortion measurement. What has worked well is the other solution you describe, of using a laptop and a PC-based test instrument (in my case, a USB-powered Picoscope). But I do find it works best to run the laptop on battery only (no wall charger) and let the amp outputs establish the balanced connection to ground for the two analyzer input terminals.
If you see excessive distortion with such a setup, I agree that you likely have an amp problem. Still, it seems odd that it's similar in both channels.
That's notable for sure, though I'm not quite sure what the cause might be yet.
Like Joe, I keep going back to a measurement error somewhere, since both channels have the same problem. But that certainly doesn't preclude other stuff. How is the balzen stage connected to the circlotron?
This is getting to the heart of the matter, I think! Prompted by your question, I've been making a few measurements. First, I checked the signal levels of the two supposedly balanced outputs from the Zen preamp. They are in fact a long way from being balanced; 1.00V on one output corresponds to 1.31V on the other output.
Having discovered this, I then tried disconnecting the preamp altogether and applying my floating signal generator directly between the balanced inputs of the circlotron. At the same power levels as before, the distortion is now very much smaller. For example, now at 1W into 8 ohms it is less than 0.1%, compared with about 1% previously. The distortion that remains still seems to be predominantly 2nd harmonic. I suppose this could be due to a small mismatch of the 2SJ74 drivers (or the output MOSFETS).
I can try playing around with a potentiometer to balance up the signal levels on the two outputs from the preamp. In fact, I suppose that could also assist with compensating for the residual mismatch in the circlotron amp itself.
In the longer term, I would be quite happy to do away with the Zen preamp stage altogether, if I could find a better way to convert from an unbalanced source to get balanced outputs for driving the circlotron. Any suggestions would be welcome!
Anyway, the underlying cause of the problem seems now to be identified. Many thanks to all who helped!
Chris
Hi Chris,
It is a good news!
I think at the same time, it is worth it to investigate, why this is happened!
One of my first Pass projects was a BOSOZ preamp, I'm using it more than 15 years.
(Maybe I should just finished it already.)
And I'm not really searching for a better one.
It is a great and simple, so ZEN preamp.
Regards,
Gyuri
It is a good news!
I think at the same time, it is worth it to investigate, why this is happened!
One of my first Pass projects was a BOSOZ preamp, I'm using it more than 15 years.
(Maybe I should just finished it already.)
And I'm not really searching for a better one.
It is a great and simple, so ZEN preamp.
Regards,
Gyuri
Last edited: