Keep in mind that even your excellent QuantAsylum equipment has own distortion and noise comparable to this amplifier at 1W. I had to purchase a new measurement gear.Full day ahead, but I may have some time to do a few measurements.
IAIMH,
Your fresh build has motivated me to bump up the bias a bit.
After an hour of playing time, checked the voltage at R17 and it’s 0V252, exactly the same as the day the lid was installed 3 months ago. Now both channels are set to 0V300.
The Kill-A-Watt meter was showing 105W now showing 116W from the wall.
👍🏻
Edit:
Temperature increased by 3°C, now sinks are 39°C
Your fresh build has motivated me to bump up the bias a bit.
After an hour of playing time, checked the voltage at R17 and it’s 0V252, exactly the same as the day the lid was installed 3 months ago. Now both channels are set to 0V300.
The Kill-A-Watt meter was showing 105W now showing 116W from the wall.
👍🏻
Edit:
Temperature increased by 3°C, now sinks are 39°C
Last edited:
I had some time free up. I was listening to the amp in the background most of morning. I also did some quick and dirty measurements. I am still learning how to measure properly, and this was far from ideal. Mainly, I wanted to make sure that it was at the very least built without any major flaw that would show up in measurements, but not through listening.
tl;dr - it all seems just fine to me on both accounts.
Quick Power vs. THD - Both channels behave identically and to expectations for all practical purposes.
Representative channel at 1W 4R and 1W 8R - channels behave the same for all practical purposes.
PSU is clean enough. It's not to the extremes of my "wretched excess" external PSU, but it's quite good and about 1/10th the weight... and it's in the same chassis.
15W 8R and 25W 4R - Practical / Rounded power output limits.
Clipping behavior into 8R - Nice and even. I could have sworn I did 4R also... but I may have forgotten to save the graph.
Noise and A-weighted noise. Looks a mess, but it's all below ~95dBV. What really matters is that it's dead quiet...
What I built this amp for and what it definitely excels at is low-level listening. However, what I also wanted was a "bullet-proof" system for when kids come over and want to turn it up to "11". With an Iron Pre at 6dB of gain + the 12dB of gain with the amp, it can be safely turned ALL the way up. It plays to pleasant if not loud volumes without sounding strained at all.
More to come, but lovely, lovely amplifier.
If anyone sees anything of concern, let me know, but for now... it's just getting settled into the system.
tl;dr - it all seems just fine to me on both accounts.
Quick Power vs. THD - Both channels behave identically and to expectations for all practical purposes.
Representative channel at 1W 4R and 1W 8R - channels behave the same for all practical purposes.
PSU is clean enough. It's not to the extremes of my "wretched excess" external PSU, but it's quite good and about 1/10th the weight... and it's in the same chassis.
15W 8R and 25W 4R - Practical / Rounded power output limits.
Clipping behavior into 8R - Nice and even. I could have sworn I did 4R also... but I may have forgotten to save the graph.
Noise and A-weighted noise. Looks a mess, but it's all below ~95dBV. What really matters is that it's dead quiet...
What I built this amp for and what it definitely excels at is low-level listening. However, what I also wanted was a "bullet-proof" system for when kids come over and want to turn it up to "11". With an Iron Pre at 6dB of gain + the 12dB of gain with the amp, it can be safely turned ALL the way up. It plays to pleasant if not loud volumes without sounding strained at all.
More to come, but lovely, lovely amplifier.
If anyone sees anything of concern, let me know, but for now... it's just getting settled into the system.
All looks perfectly well. As expected, QuantAsylum own distortion prevents proper amplifier distortion measurement. Still, results are very useful and confirm successful build.
For noise measurement, did you put short circuit on inputs?. Otherwise, you are measuring noise generated by 47 K resistor at input + some EMI.
For noise measurement, did you put short circuit on inputs?. Otherwise, you are measuring noise generated by 47 K resistor at input + some EMI.
^ Definitely. There are some tips and tricks for lowering the amount of distortion added by the measurement system, but I generally don't go to extremes.
I've never tried shorting the inputs for the noise measurements. Wonderful suggestion!
Here's a before and after... That's now part of my process. Very interesting. Also, for full transparency, the amplifier was measured sitting in the rack with all sorts of other nonsense spewing nasty EMI toward it, so no doubt I'm also measuring things not associated directly with the amplifier, but it's measured in it's "natural environment.
Also, a totally irrelevant side note, but the amp does have a slight turn-on thump. Nothing concerning, but it's there.
I've never tried shorting the inputs for the noise measurements. Wonderful suggestion!
Here's a before and after... That's now part of my process. Very interesting. Also, for full transparency, the amplifier was measured sitting in the rack with all sorts of other nonsense spewing nasty EMI toward it, so no doubt I'm also measuring things not associated directly with the amplifier, but it's measured in it's "natural environment.
Also, a totally irrelevant side note, but the amp does have a slight turn-on thump. Nothing concerning, but it's there.
Shorting inputs is a standard procedure for noise measurements. Otherwise, high input impedance allows for EMI incursion. While measuring single uV levels and below, there is always something “in the air” around.I've never tried shorting the inputs for the noise measurements.
Your noise measurement is now close to mine(post #4) with noise peaks at 1.6 uV. It is nice to observe nothing at 120 Hz. Confirms PSRR performance of both amplifier and voltage regulators.
Distortion measurements for high power are spot on, but at 1W, measured distortion is actually from QuantAsylum. In example, measured distortion for 1W and 15W into 8Ω are almost the same: 0.00068% and 0.00076%. Result for 1W is not right.
I’ve used measurement equipment with own total distortion at 0.000001% level.
https://www.diyaudio.com/community/...na-70-to-outperform-apx555b-for-30000.386001/
Binning lateral MOSFETs by Vgs is useful for paralleling them in a powerful output stage. With this amplifier, it can’t bring any benefits.Is there value in picking a specific "select color code" for the laterals from Profusion?
- Home
- Amplifiers
- Solid State
- Building an ultimate low power class A amplifier – my way