Differentiator help..

[huricaine]

So, I noticed, with the volume down it dosn't distort. What is funny is that with the dual supply I was able to get the volume up high where the preamp starts clipping. Now I cannot. When I turn it high the positive half of the waveform is clipped at 0v and negative has some waveform which looks bad.

Should I use a larger power supply resistor?

Also, I am going to try using 3M and 4M resistors on the output to see if it works.

 

[audioguru]

I forgot the EXTREMELY important input coupling capacitor and an output coupling capacitor.

 

[huricaine]

That's right I forgot about that aswell.

I have decided to construct a twin T filter. I can use a potentiometer to change the q of the filter. I constructed one already today but I used a dual supply and I got bad results very similar to the last circuit as when I have it too high the waveform distorts. I will redo the power supply for a single rail and report back in the morning what I get.

 

[huricaine]

I figured out that my car I was testing in has a DEEEAD battery. I was doing tests with it not running because they were pretty short tests but I guess that wasn't a good idea. Now I have to wait for it to charge.

I have my twin t notch filter setup already in it waiting...So I will let you know what happens after I test. I was wondering why when I had a dual supply the waveform wasn't distorted and was with the single supply, and then distorted still again after going back to the dual supply. It was my electrical power supply!(my batt)

 

[huricaine]

Okay, today I witnessed it working. I used a dual supply, and a 9v home power supply. It was the twin T filter.

I got 51mV at 50hz
134mV at 40hz

but a few odd things:
the upper side(above 50hz) was not the same as the rear side behind 50. The front side had less steep of a slope. For instance a 60hz I got like 70mV and also the hill kept climbing and climbing intill it got to about 1.1V near 1khz! Basically, no symetry.

I'm wondering why it came out this way.

I used 47n for C1,C2
and 2 47n caps in paralell for 2*C
68k for R1, R2
34k for the 2/R combining 30k, 3k, and 1k in series for the value.

I'm also trying to figure out what I will do for making the filter compatible with other peoples audio setup since their amplifier's power will varry.

 

[huricaine]

Is 80mV for the notch and about 800mV at the 2nd harmonic(150hz) a good enough difference? My problem is that I need the input of the device to be compatible with other amps for other people. For instance I am using an 800w sub amp and another guy may be using a 2000w subwoofer amp. I have heardof the term auto ranging but I do not understand how to design an autoranging circuit. How exactly does it work? I'd like to keep it analog.

 

[audioguru]

If the amplitude of the 2nd harmonic frequency measures only 10 times higher than the fundamental frequency then the Q of your notch filter is too low or it makes a lousy notch.

The levels of the 3rd, 5th and 7th harmonics are important to see if a symmetrical amplifier is clipping, not the 2nd harmonic. when the 2nd harmonic is 150Hz then the fundamental frequency is 75Hz, not 50Hz. 150Hz is the 3rd harmonic of 50Hz.

800W into 4 ohms is an amplitude of 56.6V RMS. 2000W into 4 ohms is an amplitude of 89.4V RMS which is not much more. Then maybe you do not need auto-ranging.

 

[huricaine]

It's bugging me how the front curve is not as steep as the rear curve. The rear curve is nice and steep. I need to figure out what I did to make it turn out this way.

Do you have any ideas on how I am going to feed the output to an LED after figuring it out? I was thinking something like a peak detector that goes to a Zener, or an opamp with alot of gain that acts like a comparator which will also regulate my voltage from going too high...theres alot of different ideas.

 

[huricaine]

So I just figured out that the symetry wasn't off on the filter. My center notch is actually at 54hz instead of 50 which should be simple to fix. On the circuit I am using I have a potentiometer which adjusts the q of the circuit(twin T filter). But as that page says, when adjusting the Q, the depth of the notch will change too so it is like diminishing returns. If I tighten the bandwidth by increasing the Q, the notch depth reduces(notch frequency is not as deep).

I have heard of getting notch attenuation as high as about 100db or so....I also heard the exact matching of components makes the notch deeper etc...I'm just wondering what I can do.

I am using a breadboard with wires for quite a couple connections, do you think the wires have anything to do with it?

After reading more on it, i figured I need to match my components much better. My "1%" resistors varied more then I thought after measuring them, i heard a difference of 10ohms can even cause a decent difference in notch depth and that matching all components are important for a deep notch. Below 1% is what I am trying to acheive with potentiometers and I need to find some lower tolorance caps around the house or something.

 

[huricaine]

Okay yesterday I figuredout it was the tolerance because I replaced all the resistors with pots(rheostat syle) and I was able to get 59mV at the notch, and 1.1v at about 82hz. 1.1V is the basically the voltage comming from the pre-amps source to the input of the circuit so it goes to show that it's working much better. I fine tuned the rheostats. Well I couldnt find any 1% capacitors at home or even online...

 

[huricaine]

Do I need to do better then 59mV - 1.1V from 52hz - 82hz? What exactly does nulling by a given db mean and how do I calculate it? I know the more, the better. Do I multiply the notch voltage x2 and add 3db each time I multiply to figure out how much db attenuation I have? If so, I think I need to do better...

 

[audioguru]

This is for voltage amplitudes:
A -20dB notch is 1/10th the amplitude of the highest level. Each additional 6dB reduces the level to half so -26dB is 1/20th.
-40dB is 1/100th.
-60dB is 1/1000th.
A half-decent notch filter attenuates to -60dB or better.

Look at the LMF100 by National Semi/Texas Instruments. It is a switched-capacitor lowpass/notch filter IC that has VERY accurate parts inside and makes a notch that is -80dB to -100dB. But it is stocked only in a surface-mount package today.

Power measurements are -3dB for half the power, -6dB is 1/4 the power and -10dB is 1/10th the power.

 

[huricaine]

Thanks audioguru that helped alot. I had a feeling there was a difference between voltage and power db calculation. I am going to see what I can do with this circuit, if I cannot get it up there in attenuation then I will def look into that IC.

 

[huricaine]

Okay, I realized a pot leg looked in on my breadboard but was not conducting so I pushed it the correct way, now I get 1mV at the notch(48.23hz)and full 1.1V at the 3rd harmonic...So basically -60db! Much better...I am still going to further fine tune it with smaller pots this time, smaller pots will make it easier to settle on a specific resistance down to ohm vs large one right? By using a smaller pot I mean using a fixed resistor and rheostat style pots.

 

[huricaine]

oh...What about the noise floor of the circuit? Isn't the noise floor value the lowest I can detect? How do I calculate the lowest % of distortion I can detect given the noise floor and the notch circuit's db attenuation?

 

[audioguru]

There are many low noise audio opamps available. The TL08x is a fairly good opamp but the TL07x is one that is selected for low noise.

TL071 single, TL072 dual and TL074 quad opamps. Since they are used in millions of audio circuits they do not cost more.

Some people make a low distortion generator with a square-wave oscillator and filter the heck out of the harmonics with a switched-capacitor lowpass filter IC.
I did better because I used a Cmos oscillator/counter to make a 10 steps "sinewave" and used a switched-capacitor lowpass filter IC to reduce its harmonics to almost nothing. Using TL072 ICs in the circuit it can measure distortion and noise down to 0.02%.

 

[huricaine]

Sounds interesting...about the LMF100 you mentioned, I looked it up earlier and I don't understand it, it says that it can filter upto like 4th order, isn't that 24db or so?

 

[audioguru]

The datasheet for the LMF100 switched capacitor filter IC shows how it can be an allpass, highpass, notch, bandpass or lowpass filter.
I used it as a notch filter in my distortion analyser. The notch can have a very high Q.

 

[huricaine]

After reading into it, I realized that the THD figure of amps is in power right? And the voltage THD is more then the power THD. So for an average amp is rated say .001% THD, we were talking about 3rd order content only so this would be even less then the THD rating of an amp since there are no other harmonics beeing looked at right?

Since we were talking about only 3rd order harmonic content and in voltage, does this mean that I do not need to be able to detect as low as .001% 3rd order harmonic voltage to be able to detect if an amp is at it's limits? I am just trying to make sure the amp is not running far past it's rated THD.....(well, I want to make sure it's detected as early as .001%, or maybe .01% THD or so)

 

[audioguru]

After reading into it, I realized that the THD figure of amps is in power right? And the voltage THD is more then the power THD. So for an average amp is rated say .001% THD, we were talking about 3rd order content only so this would be even less then the THD rating of an amp since there are no other harmonics beeing looked at right?

You make no sense.
0.001% distortion is very low. 10% distortion is very high.
Usually an amplifier has a symmetrical output (but some do not) so the odd harmonics (3rd, 5th, 7th etc.) are produced when the amplifier is clipping.

Some amplifiers have extremely low distortion when they are not clipping but other amplifiers have fairly high distortion all the time.