Hoffman Amplifiers Tube Amplifier Forum

Amp Stuff => Tube Amp Building - Tweaks - Repairs => Topic started by: hesamadman on February 21, 2017, 06:02:11 am

Title: measuring frequency response
Post by: hesamadman on February 21, 2017, 06:02:11 am
I am diving into some tone stack modifications. What kind of program or equipment do I need to measure the frequency response of the different tone stack values and plot them on a graph? I know Duncans TSC gives a plot for values but Im just curious how to go about doing this.
Title: Re: measuring frequency response
Post by: shooter on February 21, 2017, 08:35:08 am
You'll need a scope, 2channel preferably, a tone generator.
my basic steps;

Calibrate the input signal, measure your objective signal.
change freq, repeat.

Sine waves are pretty hard to see nuanced changes, so if your scope has a subtract feature, you can subtract the input from your tested point to see "differences" 
Title: Re: measuring frequency response
Post by: Bluemeany on February 21, 2017, 12:07:30 pm

Various Picoscope models have a frequency response measurement feature.


Or, as shooter suggests above, get yourself some log graph paper, signal generator, scope and RMS voltmeter (the scope may have this feature) and do it the old school way by stepping the signal generator frequency and plotting the results!
Title: Re: measuring frequency response
Post by: jjasilli on February 21, 2017, 01:02:29 pm
A scope is ok if you're willing to test a few targeted frequencies.  But you really need a spectrum analyzer.   These used to be expensive pieces of dedicated equipment.  These days you can do it with free software on a laptop.  Checkout e.get. REW = Room Equalization Wizzard.
Title: Re: measuring frequency response
Post by: PRR on February 21, 2017, 01:40:05 pm
IMHO, a basic audio sine signal generator and a known-good audio voltmeter. And graph-paper.

Sweep generator and pen-plotter is classic but always expensive.

A hiss source and a spectrum analyzer gives an "instant" view, used to be expensive but getting dirt-cheap.

RightMark RMAA (http://audio.rightmark.org/products/rmaa.shtml) is for sound cards, so you need the PC and soundcard plus possible interfacing so your tone control can be inserted in the sound card out/in loop. Lovely quick graphs.
Title: Re: measuring frequency response
Post by: trobbins on February 21, 2017, 03:09:29 pm
TrueRTA also does a rapid frequency response sweep, using a 'chirp' signal, that is excellent for tone control performance meazurements - instant results.
Title: Re: measuring frequency response
Post by: hesamadman on February 22, 2017, 03:30:29 pm
A scope is ok if you're willing to test a few targeted frequencies.  But you really need a spectrum analyzer.   These used to be expensive pieces of dedicated equipment.  These days you can do it with free software on a laptop.  Checkout e.get. REW = Room Equalization Wizzard.


This is what I was hoping to find. I was hoping to try and do it and plot it on my computer.


I know the hardware with true RTA only has a + and - of a couple volts. I have heard that it isnt as useful for that reason but maybe there are other versions??
Title: Re: measuring frequency response
Post by: jjasilli on February 22, 2017, 08:36:03 pm
With software-on-a-laptop spectrum analysis you'd be playing the amp though a speaker.  A mic would pick up the sound > computer.  True this does not give the isolated response of the amp on its own.  It would measure the amp + speaker combination.  This is good info.

Alternatively, to measure the amp alone, you'd need a dummy load and a line-out, a DI box, or some such thing to feed the amp's output to the laptop at an acceptable input voltage.  For hi-fi testing a fixed dummy load is used because the amp tester cannot know what speakers will be used by various owners of the amp.  Unlike a fixed dummy load speakers are reactive devices, so different amp-speaker combinations will yield different frequency responses, etc.  The fixed dummy load test procedure eliminates the variables caused by the "personalities" of different speakers. The fixed dummy load test procedure is repeatable and verifiable by other testers.  But you have the option to test how your amp behaves with your speaker. 
Title: Re: measuring frequency response
Post by: shooter on February 22, 2017, 08:39:35 pm
Quote
plot it on my computer.
I'm fortunate to have a scope that takes a "snap-shot", but also samples all data points that can be plotted with Excel.  but I've also used it manual!

as an aside, if you're gonna do full audio spectrum, 20-20k, might be good to do a quick "sound-check" of your hearing.  I spent to much time troubleshooting hissy, staticy highs, that were my ears, not the system! :think1:
Title: Re: measuring frequency response
Post by: trobbins on February 23, 2017, 01:23:24 am
TrueRTA, like REW etc, is just software that interfaces with a soundcard. 

The soundcard can be a beaut 192kHz sampling ASIO driver low noise external box, just as much as an on-board motherboard device, just as much as a $1 USB plug in that has been tweaked for grounded input and output (which is what I use). 

Normally, a soundcard (MIC/line-in, and line-out) isn't directly connected to you amp's input and output, to measure amplifier response, but rather through a probe or preamp or something with a limiter or buffer, so you don't kill the line/mic input of the soundcard.  There are a few diy designs and techniques around, including just a simple resistive divider.  I use an old magazine preamp kit designed for soundcard interfacing.  You then get the full software power of something like TrueRTA or REW etc at your disposal (up to the bandwidth of the soundcard).

But that is diy, rather than a purchased do-it-all-for you path such as a Rigol with FFT.