What amplifier do you use for taking your measurements when designing a system?
I tried researching what to use but didn't find much. I am wondering if some amplifiers color the sound enough to actually affect the measurements. I couldn't find anything definitive.
I will probably try a couple different amplifiers next time I am taking measurements just to see if there is any different in response at all. My guess is there probably isn't. If there is, I would hazard to guess its either at the far high or far low end of the frequency spectrum.
Currently I use this cheap class D mono amp. Its actually very clean sounding. I use one of these in my bedroom with better sound than I think you'd get anywhere else for $30.
I found a couple places and also heard from A4e that I maybe should be using a class AB. I do not currently have one of those. I had one in my car but it died years ago. Class AB and hot summers don't mix well. I am wondering if I should pick one of these up or a cheaper car audio 2 channel class AB.
The logical part of my brain says to use the same amplifier that I plan on using with the system. My current project I will be running this amplifier as I already own it and have upgraded the tubes. BTW, if you have one of these, upgrade the tubes and it will bring a world of clarify to the high end.
I tried researching what to use but didn't find much. I am wondering if some amplifiers color the sound enough to actually affect the measurements. I couldn't find anything definitive.
I will probably try a couple different amplifiers next time I am taking measurements just to see if there is any different in response at all. My guess is there probably isn't. If there is, I would hazard to guess its either at the far high or far low end of the frequency spectrum.
Currently I use this cheap class D mono amp. Its actually very clean sounding. I use one of these in my bedroom with better sound than I think you'd get anywhere else for $30.
I found a couple places and also heard from A4e that I maybe should be using a class AB. I do not currently have one of those. I had one in my car but it died years ago. Class AB and hot summers don't mix well. I am wondering if I should pick one of these up or a cheaper car audio 2 channel class AB.
The logical part of my brain says to use the same amplifier that I plan on using with the system. My current project I will be running this amplifier as I already own it and have upgraded the tubes. BTW, if you have one of these, upgrade the tubes and it will bring a world of clarify to the high end.
What we measure is part of a chain- PC to the DAC to the amp to the driver to the mic to the pre-amp to the ADC back into the PC.
If we want to measure something it’s important that our chain is transparent.
Ideally we want something that has higher voltage swing, wider frequency bandwidth and lower distortion that what we are trying to measure.
Many class D amplifier had variations in their HF response, depending on the resistance of the load. This becomes problematic because speakers have frequency dependant resistance (impedance) and the measured frequency response will fluctuate. Further, many of them had distortion at low or high frequencies.
However, models from Hypex, Purifi and the later TI TPA3251/5 with PFFB do not have this problem of HF variation dependant on the load (aka load invariant)
Class AB amps were deemed suitable for measurements because they are generally ruler flat between 20Hz to 20KHz, and if they display some load variation, often it is far beyond 20KHz.
I had been using the Thomann t. Amp PM40c, that is mentioned in the ARTA Handbook (page 61), as an affordable mono amp to test speakers. Unfortunately, it a less than ideal implementation of the National LM3886 amplifier, with distortion only in the -60dB (0.1) THD+N range at 1W. If you want to test distortion, this will be a limiting factor.
A more suitable one would perhaps be @tomchr ’s Modulus’ modules.
@3eaudio also has a selection of TPA3251/5 amplifiers that would make a better test amp.
If we want to measure something it’s important that our chain is transparent.
Ideally we want something that has higher voltage swing, wider frequency bandwidth and lower distortion that what we are trying to measure.
Many class D amplifier had variations in their HF response, depending on the resistance of the load. This becomes problematic because speakers have frequency dependant resistance (impedance) and the measured frequency response will fluctuate. Further, many of them had distortion at low or high frequencies.
However, models from Hypex, Purifi and the later TI TPA3251/5 with PFFB do not have this problem of HF variation dependant on the load (aka load invariant)
Class AB amps were deemed suitable for measurements because they are generally ruler flat between 20Hz to 20KHz, and if they display some load variation, often it is far beyond 20KHz.
I had been using the Thomann t. Amp PM40c, that is mentioned in the ARTA Handbook (page 61), as an affordable mono amp to test speakers. Unfortunately, it a less than ideal implementation of the National LM3886 amplifier, with distortion only in the -60dB (0.1) THD+N range at 1W. If you want to test distortion, this will be a limiting factor.
A more suitable one would perhaps be @tomchr ’s Modulus’ modules.
@3eaudio also has a selection of TPA3251/5 amplifiers that would make a better test amp.
Last edited:
I use a smal TDA7265 class AB (power supply +/-23V), which is my PC amplifier, so it's already hooked up to be used by REW.
It's pretty flat in the range I need to measure speakers, reasonable distortion at low power levels and there is no high frequency noise normally found in class D.
I don't like to see in the oscilloscope the high frequency (200kHz, 300kHz) some class D left unfiltered.
It's pretty flat in the range I need to measure speakers, reasonable distortion at low power levels and there is no high frequency noise normally found in class D.
I don't like to see in the oscilloscope the high frequency (200kHz, 300kHz) some class D left unfiltered.
I used Dayton Audio APA150 for the last 5 years.
I used an AudioSource Amp 100 for the last 6 months or so.
I just got a Fosi V3 monoblock that I haven't used yet.
The Fosi V3 has PFFB which is one of the things you want if using class D. Not sure if there are other issues to look out for, hopefully others will chime in on the class D issues.
I used an AudioSource Amp 100 for the last 6 months or so.
I just got a Fosi V3 monoblock that I haven't used yet.
The Fosi V3 has PFFB which is one of the things you want if using class D. Not sure if there are other issues to look out for, hopefully others will chime in on the class D issues.
I mostly use an older AudioSource Amp Two. It's 80 watts per channel and survived and sounded good in years of testing with extreme electrostatic speaker loads, so I don't have any worries about it handling whatever else I throw at it. It also has front panel gains for left and right channels, which come in handy when prototyping multiway speakers. Last time I looked a few years ago, their newer stuff didn't seem as interesting though.
There's also an NAD C316 on my test cart. It's only 40 watts per channel, and doesn't get used as much. It's kind of extra in general, since it was originally bought as a fill-in when something else failed.
If you're testing lower sensitivity speakers, something around 80 watts tends to be necessary if you want to be able to play music at higher levels. To me, 40 watts is marginal and I don't listen that loud. But if you only need the amp for frequency response testing you can get by with less power.
There's also an NAD C316 on my test cart. It's only 40 watts per channel, and doesn't get used as much. It's kind of extra in general, since it was originally bought as a fill-in when something else failed.
If you're testing lower sensitivity speakers, something around 80 watts tends to be necessary if you want to be able to play music at higher levels. To me, 40 watts is marginal and I don't listen that loud. But if you only need the amp for frequency response testing you can get by with less power.
At home I use a small dual LM1875 board hanging off a 15-0-15Vac transformer, in a case, with no level control, Clio takes care of that and it means I know exactly what I am feeding the driver.
At the workshop it's a Hypex nc252mp in a chassis, again no level control for the same reasons.
At the workshop it's a Hypex nc252mp in a chassis, again no level control for the same reasons.
It depends, but it's mostly a first gen SMSL SA-36A Pro (3118 chip) or an Audiophonics MPA-S125NC RCA (NCore NC122MP board), both very clean amplifiers. I use whatever is the easiest (closst) to me at that moment.
But in general, use clean amps with high damping factor, not coloured amps or with low damping factor as they colour the sound.
But in general, use clean amps with high damping factor, not coloured amps or with low damping factor as they colour the sound.
Watch out for DC on the output, if you wanna probe things, as you probably do.
I'm also using this "cheap" thomann AB amp mentioned in ARTA manual https://www.thomann.de/fi/the_tamp_pm40c_endstufenmodul.htm
"Cheap" in quotes, as it has doubled in price since I bought mine, so it's cheap only if your wage has also doubled since, unlike mine.
I'm also using this "cheap" thomann AB amp mentioned in ARTA manual https://www.thomann.de/fi/the_tamp_pm40c_endstufenmodul.htm
"Cheap" in quotes, as it has doubled in price since I bought mine, so it's cheap only if your wage has also doubled since, unlike mine.
Buckeye Hypex Ncore based Class D because it has plenty of power for any reasonable build, almost no frequency dependence and distortion is super low. It's basically a transparent amplifier in most cases so it stays out of the way of the measurement. Plus, I can get 6 channels for super cheap and do an active 3 way build before turning it into a passive build.
https://www.buckeyeamp.com/shop/amplifiers/hypex/nc252mp/6_channel
https://www.buckeyeamp.com/shop/amplifiers/hypex/nc252mp/6_channel
The graph reads about -60dB 2nd harmonic . I doubt if anyone can recognize that. Higher order goes down below -70dB in the range we are anything but insensitive for harmonic distortions. I wouldn’t worry about this implementation for speaker measurements.
I apologies, my mistake. 0.01% is -80dB. Transparent for most transducers, yes.
Now what was the reason I switched out this amp in favour of another I can’t recall. I should put it back into service and see.
Having said that, these days we are getting an order of magnitude better at 0.001 %(-100dB), I can’t in good conscience recommend someone buy that amp today, as a test instrument.
On the other hand, here’s a well designed LM3886:
Reference:
https://neurochrome.com/collections/power-amplifiers/products/modulus-86
Now what was the reason I switched out this amp in favour of another I can’t recall. I should put it back into service and see.
Having said that, these days we are getting an order of magnitude better at 0.001 %(-100dB), I can’t in good conscience recommend someone buy that amp today, as a test instrument.
On the other hand, here’s a well designed LM3886:
Reference:
https://neurochrome.com/collections/power-amplifiers/products/modulus-86
Last edited:
Yes, the composite builds on LM3886 chips can perform miracles. Not only for testing environments.
Helpful discussion regarding our test instruments/tools. I’ll need to use a better amp when I do more serious testing. So far, just to get some practice in REW I used a PE class D, which likely has some performance concerns listed in the thread. Thanks folks!
I use an older version of this Emotiva amp. Light small easy to move around.
Rob 🙂
https://emotiva.com/products/basx-a2m-stereo-flex-amplifier
Rob 🙂
https://emotiva.com/products/basx-a2m-stereo-flex-amplifier
I send the signal straight from my laptop to the driver for driver measurement and thru my system for microphone measurement.
I use an IcePower 125ASX2. Link to Test Amp. This amp has a stable high frequency response and low distortion. I made my own case for it, but if I had known about Ghent Audio cases, I would have just bought one of those.
Voice Coil’s test bench of Purifi PTT8.0X04 playing for @94dB (measurer at 10cm)
https://audioxpress.com/article/test-bench-the-purifi-audio-ptt8-0x04-nab-02-8-aluminum-cone-woofer
Note @1KHz
H2 ~ -52dB
H3 ~ -63dB
Reference:
https://sengpielaudio.com/calculator-thd.htm
Note @1KHz
H2 ~ -53dB
H3 ~ -60dB
Reference:
https://www.erinsaudiocorner.com/driveunits/purifi_ptt80x04-nab-02/
In 2024; it took some effort to see what
what state of the art midrange performance looked like:
Purifi PTT8.0XNAA playing at 96dB/1m:
Here is the harmonics in -dBr. There are minor differences due to the mouse cursor position placed at ~1KHz, to show the harmonics in the legend.
Here at 1KHz
H2 ~ -72dB
H3 ~ -87dB
This is an order of magnitude (20dB) lower than what others have demonstrated.
In 2014 Jeff Bagby did a review of the
ATC SM75-150S Midrange Driver, first announced here
on page 7 he shoes the distortion of this 3" dome "as mounted on the actual baffle used for the monitor I designed. This data was taken from about one foot away"
I traced the charts-
at 96dB@1ft -> 86dB@1m...
Our dear Jeff (R.I.P.) did acknowledge that "These levels are very close to the noise floor in my room and I would consider this to be in the state-of-the-art range for midrange distortion.”
Unfortunately he was not aware that it was noise floor limit of his microphone Omnimic and/or amp.
This ATC driver is 95dB/W driver, It makes it hard to compare this data against other data…
https://audioxpress.com/article/test-bench-the-purifi-audio-ptt8-0x04-nab-02-8-aluminum-cone-woofer
Note @1KHz
H2 ~ -52dB
H3 ~ -63dB
Reference:
https://sengpielaudio.com/calculator-thd.htm
Note @1KHz
H2 ~ -53dB
H3 ~ -60dB
Reference:
https://www.erinsaudiocorner.com/driveunits/purifi_ptt80x04-nab-02/
In 2024; it took some effort to see what
what state of the art midrange performance looked like:
Purifi PTT8.0XNAA playing at 96dB/1m:
Here is the harmonics in -dBr. There are minor differences due to the mouse cursor position placed at ~1KHz, to show the harmonics in the legend.
Here at 1KHz
H2 ~ -72dB
H3 ~ -87dB
This is an order of magnitude (20dB) lower than what others have demonstrated.
In 2014 Jeff Bagby did a review of the
ATC SM75-150S Midrange Driver, first announced here
on page 7 he shoes the distortion of this 3" dome "as mounted on the actual baffle used for the monitor I designed. This data was taken from about one foot away"
I traced the charts-
at 96dB@1ft -> 86dB@1m...
Our dear Jeff (R.I.P.) did acknowledge that "These levels are very close to the noise floor in my room and I would consider this to be in the state-of-the-art range for midrange distortion.”
Unfortunately he was not aware that it was noise floor limit of his microphone Omnimic and/or amp.
This ATC driver is 95dB/W driver, It makes it hard to compare this data against other data…
Last edited: