After spending a lot of time and effort on an amp project to get it to look presentable and more importantly, make some sound that at least passes as music, how do you or could you quantify or evaluate its performance?
I know you could take it along to a friends place, or cheekily take it for comparison when auditioning some 'new' replacement, but are there any tests that would, in theory, be telling of its abilities?
I would think that it would also help to have some sort of reference product for comparison also.
I know you could take it along to a friends place, or cheekily take it for comparison when auditioning some 'new' replacement, but are there any tests that would, in theory, be telling of its abilities?
I would think that it would also help to have some sort of reference product for comparison also.
Obviously I listen to it, but I also do a suite of measurements. Large and small signal response, response and wave shape into various loads, THD and a spectral plot on the PC. I like to see music on the output and very little music on the supply rails. I want to see no sticking to the rails when it clips. I want the bias to be stable with temperature and have no chance of thermal runaway. I want full output at 20 kHz with no tendency to short the supplies together. I don't normally include protection because I figure if I built it, I can fix it if I'm stupid enough to blow it up. But, if it has protection, I want to verify that it works as intended without working on the music when it shouldn't. I also check very carefully for RF from the supply and look at how HF or RF can affect the input.
CH
CH
Member
Joined 2009
Paid Member
I have an o'scope and a freq generator , if i can't see it and I like what I hear then it's good to go. Frequently you will hear someone say they heard an amp they liked despite distortion measurements.
Keep listening and reading,Find what you like and stick with it.
Happy listening(holidays), Elwood
Keep listening and reading,Find what you like and stick with it.
Happy listening(holidays), Elwood
Build one channel and test it for THD, stability, square wave response, clean clipping, bias stability etc. You can't really do all this with spice 'cos the real world is a very different place!
If all is well then I build the other channel and listening can begin...
Some amps that have measured great can fail this last test by a long way!
If all is well then I build the other channel and listening can begin...
Some amps that have measured great can fail this last test by a long way!
The first priority is to make sure it is working properly, and you can't determine that by listening (or from spice 
). Check for oscillation - an oscillating amp can sound fine so a scope will help out there. Then the tests that jez recommended - square wave, clipping performance, frequency response.
When you do these tests and everything looks good have a listen and I'll bet it sounds good too.
). Check for oscillation - an oscillating amp can sound fine so a scope will help out there. Then the tests that jez recommended - square wave, clipping performance, frequency response.When you do these tests and everything looks good have a listen and I'll bet it sounds good too.
even a cheap pc motherboard audio chipset today is going to gve better than 90 dB s/n and often 96 or 192 K sample rate
dedicated pci or usb soundcards can be noticably better - ESI Juli@ is good for the price
with some resistor dividers to scale the power amp's output to fit the soundcard line inputs you can do pretty comprehensive noise, distortion, frequency response and crosstalk measurements
free RMAA software rolls the measurements into a easy to use test suite
Download. Audio Rightmark
not a replacement for a fast 'scope but can make higher resolution measurements in the audio frequency band
for a more "manual" interface http://audacity.sourceforge.net/ lets you control the soundcard and edit/play/record multiple tracks
there are also some other power amp test setups such as the Hafler difference test
http://www.diyaudio.com/forums/everything-else/13415-null-difference-testing.html
and there is output "interface distortion"/complex load simulation by connecting your power R test load across 2 channels and driving with different phase or frequency signals on each end
both of these could use the soundcard to generate the test signals and view audio frequency differences and IMD distortion
I haven't used this software yet but it looks useful:
http://libinst.com/Audio DiffMaker.htm
note that continuous signal tests at full power/frequency and complex load impedance simulation can lead to excessive power dissapation - higher than musical signals or reasonable loudspeaker loads
dedicated pci or usb soundcards can be noticably better - ESI Juli@ is good for the price
with some resistor dividers to scale the power amp's output to fit the soundcard line inputs you can do pretty comprehensive noise, distortion, frequency response and crosstalk measurements
free RMAA software rolls the measurements into a easy to use test suite
Download. Audio Rightmark
not a replacement for a fast 'scope but can make higher resolution measurements in the audio frequency band
for a more "manual" interface http://audacity.sourceforge.net/ lets you control the soundcard and edit/play/record multiple tracks
there are also some other power amp test setups such as the Hafler difference test
http://www.diyaudio.com/forums/everything-else/13415-null-difference-testing.html
and there is output "interface distortion"/complex load simulation by connecting your power R test load across 2 channels and driving with different phase or frequency signals on each end
both of these could use the soundcard to generate the test signals and view audio frequency differences and IMD distortion
I haven't used this software yet but it looks useful:
http://libinst.com/Audio DiffMaker.htm
note that continuous signal tests at full power/frequency and complex load impedance simulation can lead to excessive power dissapation - higher than musical signals or reasonable loudspeaker loads
Last edited:
The key is "foot tapping"
If the result is foot tapping, I make this equation:
(foot tapping time / total listening session time) * WAF = satisfaction factor.
No foot tapping at all= obvious fail
Sounds like a married Linnie
Married to his Naim amps of course.The music i like seldom involves any hint of foot tapping so i guess it's a fail by default.
An interesting question. One i have never considered seriously.
When i was 11 i had a turntable with an inbuilt amp. 4 germanium wonders per channel (those outputs must be worth a fortune today) and full range OB speakers. Really basic stuff. So, when i turned 12 and was officially granted permission to use a soldering iron, my first priority was to build an amp which sounded better than what i already had. Not counting some teenage obsession with science and numbers (read in a book that all good amps sound the same, provided THD is < 1%) i've been doing pretty much the same ever since.
Measurements don't even give a hint of what an amp may sound like. Stiil, everything i build gets thoroughly tested. Harmonic distortion may not be that important but i'd rather have it low and square waves better look square.
nearly any pc built in sound chipset would let you make the measurements with the free software I pointed to
there are also software for sound card signal generator and 'scope like functionality
It really is worthwhile to make at least the RMAA measurements since its so easy and you could tell if somethings are grossly wrong from the results
all the test hardware you need is to solder a few resistors to a RCA or 1/8" phono TRS connector
there are also software for sound card signal generator and 'scope like functionality
It really is worthwhile to make at least the RMAA measurements since its so easy and you could tell if somethings are grossly wrong from the results
all the test hardware you need is to solder a few resistors to a RCA or 1/8" phono TRS connector
Last edited:
Member
Joined 2009
Paid Member
Perhaps not everyone will have access to a scope, but I agree, it's close to being the 'Silver Bullet' for sorting out problems.
I have read somewhere that in some amplifiers, if you have excessive oscillation you may be able to detect it without a scope, although less reliably, by:
a) put a radio next to your amp - you'll pick up any rf emissions (most wayward hf oscillations will be up in the rf)
b) feel for a 'warm' zobel resistor due to it working to dissipate unwanted hf
c) ask Santa's reindeer if they can hear the hf
p.s. maybe the best way to evaluate your DIY amplifier has to be go grab some buddies and some beer and have a listening session
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.