Need sugestions about virtual ground or symetric supply for OPAMP
- Thread starter wuemura
- Start date
audioguru2
- Apr 6, 2004
- 12,026
- Joined
- Apr 6, 2004
- Messages
- 12,026
My Halfler surround sound was passive and used no electronics. The rear speakers were wired in series and were powered from the HOT terminals of the left and right amplifiers. So the rear speakers played the ambience difference sounds between the channels and the main speakers played the in-phase front and mono sounds.
Many years ago I made a Motorols full-logic SQ surround sound system that has its front-to-back gain changed by the signals. It worked from specially encoded vinyl records.
Many years ago I made a Motorols full-logic SQ surround sound system that has its front-to-back gain changed by the signals. It worked from specially encoded vinyl records.
Last edited by a moderator:
Do you still have SQ vinil records?
I was reading a page about Quadraphonics here, maybe you will like.
I've never heard any type of vinyl records that uses technology like that, SQ, Quadradisc, Q4, Quad-8, Quadraphonic Stereo etc.
http://en.wikipedia.org/wiki/Quadraphonic
But i know from some friends that this type of equipament was very expensive, the price that you paid at the time you could buy a new car, so he said.
So, i did a real life test of the circuit in my car, it worked out fine but had problems with the old Ace of Base CD, i was getting peaking at high frequency on the ambience channels. Looks like the problem was the power amp load, measuring the virtual ground again there was fluctuations.
So i came up with the changes attached.
I've found more information about virtual ground and opamp here:
http://tangentsoft.net/elec/vgrounds.html
With that changes and using the virtual ground as negative reference i do get a stable +6.04V and -6.00, the voltage used in the circuit is 12.04V, at the R1/R2 divider the voltage is 5.99V.
The circuit looks stable now.
I was reading a page about Quadraphonics here, maybe you will like.
I've never heard any type of vinyl records that uses technology like that, SQ, Quadradisc, Q4, Quad-8, Quadraphonic Stereo etc.
http://en.wikipedia.org/wiki/Quadraphonic
But i know from some friends that this type of equipament was very expensive, the price that you paid at the time you could buy a new car, so he said.
So, i did a real life test of the circuit in my car, it worked out fine but had problems with the old Ace of Base CD, i was getting peaking at high frequency on the ambience channels. Looks like the problem was the power amp load, measuring the virtual ground again there was fluctuations.
So i came up with the changes attached.
I've found more information about virtual ground and opamp here:
http://tangentsoft.net/elec/vgrounds.html
With that changes and using the virtual ground as negative reference i do get a stable +6.04V and -6.00, the voltage used in the circuit is 12.04V, at the R1/R2 divider the voltage is 5.99V.
The circuit looks stable now.
audioguru2
- Apr 6, 2004
- 12,026
- Joined
- Apr 6, 2004
- Messages
- 12,026
I made my SQ full logic surround sound decoder with the inexpensive ICs from Motorola. With ordinary stereo signals the surround effect was weird because it frequently caused sounds to go around the room randomly.
Hi audioguru,
Tell me something, do you know what is happening with this sample?
http://www.sendspace.com/file/f2fh8j (1.2Mb)
Looks like the sound is cracking up at the both ambience channels, the voice channel is fine, this is one of the "special situations" because the circuit works great with others CD's, there is no variation at the virtual ground so the problem is not there, adjusted the levels and opamp gain with no success, try to use signal levels at 500mv but no change. I did change the CD player with 3 different models and again with no success. The CD is original and analysing the signal without the circuit it shows that it is no clipping or distortion in it.
Following the signal it get bad before and after R2 and R25, removing R5 and R22 the signal became good before and after R2/R25, but at the output of the opamp's the sound is very bad and loud. So i try to buffer after R2 and R25 and use a 100K resistor at the opamp output but with no success.
Do you have any idea what is causing this?
Thanks
Tell me something, do you know what is happening with this sample?
http://www.sendspace.com/file/f2fh8j (1.2Mb)
Looks like the sound is cracking up at the both ambience channels, the voice channel is fine, this is one of the "special situations" because the circuit works great with others CD's, there is no variation at the virtual ground so the problem is not there, adjusted the levels and opamp gain with no success, try to use signal levels at 500mv but no change. I did change the CD player with 3 different models and again with no success. The CD is original and analysing the signal without the circuit it shows that it is no clipping or distortion in it.
Following the signal it get bad before and after R2 and R25, removing R5 and R22 the signal became good before and after R2/R25, but at the output of the opamp's the sound is very bad and loud. So i try to buffer after R2 and R25 and use a 100K resistor at the opamp output but with no success.
Do you have any idea what is causing this?
Thanks
audioguru2
- Apr 6, 2004
- 12,026
- Joined
- Apr 6, 2004
- Messages
- 12,026
Hi Wellington,
The "music" is rap which is talking to a beat. No music. I didn't hear any cracking up.
The signals before R2 and before R25 are the input signals. They should be clean.
The signal after R2 is attenuated R channel and extremely low level distortion from the L channel.
Your differential amplifiers are adding the two channels out-of-phase. Are the L out and R out the rear channels?
You have resistors between the opamps that don't do anything. The IC3B and IC3C opamps also don't do anything.
I have shown one differential amplifier with a 1V signal on one channel then the 1V signal on the other channel. I have marked it with signal violtages.
The "music" is rap which is talking to a beat. No music. I didn't hear any cracking up.
The signals before R2 and before R25 are the input signals. They should be clean.
The signal after R2 is attenuated R channel and extremely low level distortion from the L channel.
Your differential amplifiers are adding the two channels out-of-phase. Are the L out and R out the rear channels?
You have resistors between the opamps that don't do anything. The IC3B and IC3C opamps also don't do anything.
I have shown one differential amplifier with a 1V signal on one channel then the 1V signal on the other channel. I have marked it with signal violtages.
Hi audioguru,
I dont know if "cracking" was the correct name of that but if your hear only the right channel you will listen something "cracky" or "cranky" distortion, i dont know the exact name, sorry.
The strange is that distortion is coming from R5 and R22, looks like a "Feedback Noise"?
Yes, they are ambience channels, one is L-R (L channel) and the other is R-L (R channel), when this two are connected + the standard L and R channels inside the car it gives the effect of 360 degrees rotation if you build a "stereo rotation effect" in Adobe Audition, if you record a train passing by and apply the stereo rotation effect you will have the impression that the train running in a complete circle inside the car, very nice. Together phase analysis show a -180 degree phase shift.
But what is very nice about it is that when you turn it on, since we are getting a -180 degree out of phase from the original sound, they cancel each other right? If you put a CD with empty tracks and play it inside the car you can feel a really strange silence, is good to help relax when you are trapped on heavy trafic.
The resistors was added just to do signal attenuation, the buffers IC3B/C (ambience channel) was added and so the IC2C/D (voice channel) because inside the car there is a 6 meter long RCA cable from the instalation point to the power amp so they are capable of driving long cables and capacitive loads from the power amp.
What simulation program that you use?
Thanks
I dont know if "cracking" was the correct name of that but if your hear only the right channel you will listen something "cracky" or "cranky" distortion, i dont know the exact name, sorry.
The strange is that distortion is coming from R5 and R22, looks like a "Feedback Noise"?
Yes, they are ambience channels, one is L-R (L channel) and the other is R-L (R channel), when this two are connected + the standard L and R channels inside the car it gives the effect of 360 degrees rotation if you build a "stereo rotation effect" in Adobe Audition, if you record a train passing by and apply the stereo rotation effect you will have the impression that the train running in a complete circle inside the car, very nice. Together phase analysis show a -180 degree phase shift.
But what is very nice about it is that when you turn it on, since we are getting a -180 degree out of phase from the original sound, they cancel each other right? If you put a CD with empty tracks and play it inside the car you can feel a really strange silence, is good to help relax when you are trapped on heavy trafic.
The resistors was added just to do signal attenuation, the buffers IC3B/C (ambience channel) was added and so the IC2C/D (voice channel) because inside the car there is a 6 meter long RCA cable from the instalation point to the power amp so they are capable of driving long cables and capacitive loads from the power amp.
What simulation program that you use?
Thanks
Last edited by a moderator:
audioguru2
- Apr 6, 2004
- 12,026
- Joined
- Apr 6, 2004
- Messages
- 12,026
Hi Wellington,
I use Microsoft Paint program to modify schematics.
Your circuit has R23, R24, IC3B and IC3C that don't do anything.
I use Microsoft Paint program to modify schematics.
Your circuit has R23, R24, IC3B and IC3C that don't do anything.
audioguru2
- Apr 6, 2004
- 12,026
- Joined
- Apr 6, 2004
- Messages
- 12,026
The buffer opamps don't do anything. They are ordinary opamps that have a max output current of only 15mA peak. If the load capacitance is very high then the load is the 100 ohm resistors which overloads ordinary opamps.
Hi audioguru,
Did a few more experiences with the circuit and have found a document that show me part of the solution for the problem:
http://tangentsoft.net/audio/hs-opamp.html
Paying close attention to the distortion it has to do with the bass of the sound, removing R5/R22 and replacing with a 120K trimpot to adjust the fase and gain, as i reduce the effect the sound was becoming more clear and the distortion was the booming of the bass. The problem is that if i remove the distortion, the effect got killed but if i increase it so the distortion.
Analyzing the bass part of the song the problem was a frequency between 52Hz and 71Hz, the solution was to add a simple R/C high pass filter at the input. I set the filter to 72Hz with a 220nF with a 10K resistor grounded, that gives me a 6db per octave, calculations gives a 72.34Hz cut and a error of -0.07.
It worked fine, did not kill all the distortion but is much better than before.
In respect of audioguru golden ears i choose a real music, with lots of melody and high notes to test and hear if there is too much changes at the output and how the notes above that get affected. The music is "The Winner Takes It All" from ABBA.
http://www.sendspace.com/file/iwlms4
Is the same deal, left channel is voice and right channel is ambience, play with the balance to hear what each channel is doing.
I change the configuration of IC3B/C, for now it has the gain of 1, latter i will try a active bass boost at that frequency range and made a few more tests to see if it is necessary bost the bass or not.
Did a few more experiences with the circuit and have found a document that show me part of the solution for the problem:
http://tangentsoft.net/audio/hs-opamp.html
Paying close attention to the distortion it has to do with the bass of the sound, removing R5/R22 and replacing with a 120K trimpot to adjust the fase and gain, as i reduce the effect the sound was becoming more clear and the distortion was the booming of the bass. The problem is that if i remove the distortion, the effect got killed but if i increase it so the distortion.
Analyzing the bass part of the song the problem was a frequency between 52Hz and 71Hz, the solution was to add a simple R/C high pass filter at the input. I set the filter to 72Hz with a 220nF with a 10K resistor grounded, that gives me a 6db per octave, calculations gives a 72.34Hz cut and a error of -0.07.
It worked fine, did not kill all the distortion but is much better than before.
In respect of audioguru golden ears i choose a real music, with lots of melody and high notes to test and hear if there is too much changes at the output and how the notes above that get affected. The music is "The Winner Takes It All" from ABBA.
http://www.sendspace.com/file/iwlms4
Is the same deal, left channel is voice and right channel is ambience, play with the balance to hear what each channel is doing.
I change the configuration of IC3B/C, for now it has the gain of 1, latter i will try a active bass boost at that frequency range and made a few more tests to see if it is necessary bost the bass or not.
Last edited by a moderator:
audioguru2
- Apr 6, 2004
- 12,026
- Joined
- Apr 6, 2004
- Messages
- 12,026
Hi Wellington,
Vinyl records and tapes had terrible high frequency phase balance between channels.
It causes distortion when you use a circuit like you have that extracts the difference (ambience) between channels.
Reducing the value of the input capacitors killed the deep bass. Each 10k resistor is in parallel with two 162k and also in parallel with 62k, resulting in a -3dB (half-power) frequency of 93.5Hz. It is flat down to about 470Hz, then the bass begins to roll off.
Vinyl records and tapes had terrible high frequency phase balance between channels.
It causes distortion when you use a circuit like you have that extracts the difference (ambience) between channels.
Reducing the value of the input capacitors killed the deep bass. Each 10k resistor is in parallel with two 162k and also in parallel with 62k, resulting in a -3dB (half-power) frequency of 93.5Hz. It is flat down to about 470Hz, then the bass begins to roll off.
Hi audioguru,
This circuit was design to be used in my car, so i think i am safe from vinil and tapes ;D
The idea is to kill thouse deep bass because it does no good over the ambience channel, i did generate a 0db sine wave from 50Hz to 100Hz and in the ambience channel it get only distortion, i did not go any further with this tests but around 560Hz i can completly remove any distortion over the ambience channel.
If in you simulation it get's a flat response down to about 470Hz sounds good to me, because in my car i have a pair of 2-way components system from Diamond (D661a) and a pair of 125W JL subwoofers in a sealed box, the head unit has 4 outputs 2 output for front and 2 output for rear channels and a stereo subwoofer output, so inside the car the subwoofers will take care of thouse deep bass.
This circuit is installed to work with the head unit rear channel only, the front channels are connected in a standard way so i can hear the effect behind me, in the voice channel i use a Alpine SBS-05DC.
When all that stuf is set it gives a incredible sound inside the car, if you hear a live recorded CD it gives the effect that you are there, you can feel it.
I'm open to sugestions, what can be done to increase the cut, for example, -12db or -24db and tweek this a bit more.
This circuit was design to be used in my car, so i think i am safe from vinil and tapes ;D
The idea is to kill thouse deep bass because it does no good over the ambience channel, i did generate a 0db sine wave from 50Hz to 100Hz and in the ambience channel it get only distortion, i did not go any further with this tests but around 560Hz i can completly remove any distortion over the ambience channel.
If in you simulation it get's a flat response down to about 470Hz sounds good to me, because in my car i have a pair of 2-way components system from Diamond (D661a) and a pair of 125W JL subwoofers in a sealed box, the head unit has 4 outputs 2 output for front and 2 output for rear channels and a stereo subwoofer output, so inside the car the subwoofers will take care of thouse deep bass.
This circuit is installed to work with the head unit rear channel only, the front channels are connected in a standard way so i can hear the effect behind me, in the voice channel i use a Alpine SBS-05DC.
When all that stuf is set it gives a incredible sound inside the car, if you hear a live recorded CD it gives the effect that you are there, you can feel it.
I'm open to sugestions, what can be done to increase the cut, for example, -12db or -24db and tweek this a bit more.
audioguru2
- Apr 6, 2004
- 12,026
- Joined
- Apr 6, 2004
- Messages
- 12,026
The ambience circuits produce the difference between channels.
Mono is cancelled because it is the same in both channels, no difference.
Bass is usually in mono. So the bass is cancelled which leaves nothing but its harmonics which is distortion.
A Butterworth 2nd-order filter is 12dB/octave, a 3rd-order is 18dB/octave and a 4th-order is 24dB/octave. Sallen and Key filters are the easiest.
Mono is cancelled because it is the same in both channels, no difference.
Bass is usually in mono. So the bass is cancelled which leaves nothing but its harmonics which is distortion.
A Butterworth 2nd-order filter is 12dB/octave, a 3rd-order is 18dB/octave and a 4th-order is 24dB/octave. Sallen and Key filters are the easiest.
Similar threads
