Setting constant gain of line level audio amp for max. gain without causing clipping

Thread Starter

PeteHL

Joined Dec 17, 2014
337
In the shown schematic diagram, my question is what can be the maximum closed loop gain of the inverting amplifier including op amp U3 which also will not result in clipping by the inverting amp? The pre-amplified program material can be any type of music, and voltage gain of the preamplifier can be at the high end. What I need is RMS voltage taken at output pin 6 of U3 to be as high as possible, but not so high that under certain circumstances the amplifier is driven into saturation.

Regards,
Pete


AAC-Problem1.png
 

Audioguru again

Joined Oct 21, 2019
1,289
Clipping of U3 depends on the input levels times the gain of U3. If the load on U3 is 10k ohms or more then its output will begin clipping when it reaches 13.5V peak.
 

Thread Starter

PeteHL

Joined Dec 17, 2014
337
That question cannot be answered without knowing the maximum value of the preamp output voltage.
In trying to arrive at an answer, I thought that maybe I could safely assume the maximum output voltage of the preamp equal to 1.0V RMS. From Rod Elliott, the typical sensitivity of an audio power amp for full power output is 1.0V RMS.

Thanks,
Pete
 

crutschow

Joined Mar 14, 2008
24,730
In trying to arrive at an answer, I thought that maybe I could safely assume the maximum output voltage of the preamp equal to 1.0V RMS. From Rod Elliott, the typical sensitivity of an audio power amp for full power output is 1.0V RMS.
But that doesn't mean the preamp output is limited to that value.
 

Thread Starter

PeteHL

Joined Dec 17, 2014
337
Clipping of U3 depends on the input levels times the gain of U3. If the load on U3 is 10k ohms or more then its output will begin clipping when it reaches 13.5V peak.
I'm assuming a high impedance load for the output of U3 such as 10 k Ohm. It boils down to I think figuring out what would be the maximum peak voltage at input to the inverting amp. Then voltage gain of U3 can be such that the corresponding voltage output taken at pin 6 of U3 is not more than 13.5V.
 

Thread Starter

PeteHL

Joined Dec 17, 2014
337
But that doesn't mean the preamp output is limited to that value.
Very true, but it does mean that when the output of the preamp is at input to a typical audio power amplifier being put to use, then usually the output voltage of the preamp won't exceed 1.0V RMS. This is assuming that the user is not trying to destroy his/ her equipment.

Arriving at an answer I think is not easy, so I'm trying to make some assumptions that will simplify arriving at what voltage gain can be.
 

crutschow

Joined Mar 14, 2008
24,730
So if you assume the the preamp signal is no more than 1Vrms, and the output is limited to 13.5Vpk (9.5Vrms) then the total closed-loop gain should be no more than 9.5 or a value of 95kΩ max for R6.
 

Thread Starter

PeteHL

Joined Dec 17, 2014
337
So if you assume the the preamp signal is no more than 1Vrms, and the output is limited to 13.5Vpk (9.5Vrms) then the total closed-loop gain should be no more than 9.5 or a value of 95kΩ max for R6.
That would be true, I think, only if you neglect the high-pass filter in the circuit. To come up with a practical solution, which is what I'm after, also the output of the preamp should be considered to be 1.0V RMS of pink noise.
 

Audioguru again

Joined Oct 21, 2019
1,289
Some bands play loud and some play soft and singers too. Their average level is line level.
The amplifier has a volume control for you to turn down to avoid clipping.

Music and pink noise have a wide range of levels between peak and RMS. To avoid clipping you need at least 20 times the headroom. Then your 1V RMS of pink noise is causing the output of U3 to be clipping.
How will you know that the pink noise is not being clipped?
 

Thread Starter

PeteHL

Joined Dec 17, 2014
337
What is the gain of each of the filter op amps with their outputs connected to their minus inputs?
The filter output gain can be no higher than that.
The op amps of the filters are configured as voltage followers, so that means that the gain can be no more than 1. But the point is that if the input is 1 Vrms of pink noise, 20 Hz- 20kHz, if you decrease that range to 400 Hz - 20 kHz, then this reduces the RMS voltage of the signal (taken at the output of the high- pass filter). The question is, by how much is the RMS voltage reduced, and also the relationship of peak voltage to RMS voltage of pink noise. I suggest a pink noise source as pink noise is a regular signal source similar to recorded music.

-Pete
 

Thread Starter

PeteHL

Joined Dec 17, 2014
337
Some bands play loud and some play soft and singers too. Their average level is line level.
The amplifier has a volume control for you to turn down to avoid clipping.

Music and pink noise have a wide range of levels between peak and RMS. To avoid clipping you need at least 20 times the headroom. Then your 1V RMS of pink noise is causing the output of U3 to be clipping.
How will you know that the pink noise is not being clipped?
13.5V divided by 20 = 675 mV RMS. Not being very experienced with audio electronics, I would still say that a ratio of 20 to 1 is not correct. Certainly I would think the ratio of RMS to peak voltage of music is more than that of a sine wave or 1.414, but not that much.

-Pete
 

Thread Starter

PeteHL

Joined Dec 17, 2014
337
Here is a bench test that I did to try to get an answer to my question- see the below circuit. Amazingly, I was able to get unwavering RMS voltage readings of the pink noise with a HP3478A multimeter. Try to do that with your typical better quality DMM.

If the conditions of this test are indicative, then the gain of the inverting amplifier of post #1 could be slightly less than 10 and clipping should not occur. Let me know if you think that this testing holds water or not

Thanks,
Pete

PK-E-LL-Audio.png
 
Last edited:

Audioguru again

Joined Oct 21, 2019
1,289
13.5V divided by 20 = 675 mV RMS. Not being very experienced with audio electronics, I would still say that a ratio of 20 to 1 is not correct. Certainly I would think the ratio of RMS to peak voltage of music is more than that of a sine wave or 1.414, but not that much.

-Pete
No. 13.5V peak/20= 675mV peak which is 477mV RMS.
Music and pink noise have loud parts, soft parts and quiet parts. You are talking about a continuous tone having a ratio of 1.414 times. The level of short duration drum beats in music is far more than the average RMS level. The peak levels in pink noise are similar.
 
Last edited:

eetech00

Joined Jun 8, 2013
1,869
Here is a bench test that I did to try to get an answer to my question- see the below circuit. Amazingly, I was able to get unwavering RMS voltage readings of the pink noise with a HP3478A multimeter. Try to do that with your typical better quality DMM.

If the conditions of this test are indicative, then the gain of the inverting amplifier of post #1 could be slightly less than 10 and clipping should not occur. Let me know if you think that this testing holds water or notView attachment 200817.

Thanks,
Pete
If you really want to know, why not do a an FRA of your filter stages and eliminate the guess work.

eT
 

Thread Starter

PeteHL

Joined Dec 17, 2014
337
Earlier this evening I changed the voltage readings for the circuit attached to my post # 16.The new readings are what I got using a different CD player. The first CD player that I originally used I believe was not reproducing the pink noise track correctly.

Is a 100 mV output from the CD player about what I should be getting playing recorded pink noise?

Huber and Runstein (in their book Modern Recording Techniques, p. 245), discussing the VU meter, say that short term peaks of a music recording are ignored by the VU meter and can be from "8 to 14 dB higher than the RMS value indicated". So given that the measured voltage output from the high-pass filter in the circuit of my post #16 equals 655 mV, then that would mean that a 14 dB peak would equal 5 times 655 mV or 3.28V. Is that correct?

Then gain of the inverting amplifier of my first post should be not more than 13.5V/3.28V or about 4.
 

Audioguru again

Joined Oct 21, 2019
1,289
A modern music recording is compressed, a live performance and pink noise are not compressed. 14dB is only 5 times the level.
Most multimeters measure 50Hz and 60Hz but not wideband audio levels.
Your peak rectifier circuit cannot charge the capacitor fast enough because the opamp's output current is too low.
 
Top