help regarding audio spectrum analyzer using active filters and lm3915

[japanjot]

Sir can you tell me any alternate single supply op- amp for this audio circuit? The mc33171 ic is not available to the shop near me

And sir as you said

Mine has a fast attack and a fairly slow decay. Increase the value of the 0.1uf filter capacitor to make both slower.

I want to slow only decay rate not attack rate, so how to do that?

And Sir you are very experienced and helpful to me so thanx,
Wrs ur age must be above 50-55

 

[audioguru]

I am 67 years old. I retired 12 years ago. Most of my career was with designing, repairing and installing audio systems.
My first job was fixing then designing car radios with Philips for Chrysler cars. I saw their first LED and I worked with their new audio cassette recorders and players (Philips invented the cassette tape system). Then Philips office computers (DTL logic ICs and millions of ferrite cores were its RAM memory) and punched cards were its operating program. Then Philips huge intercom systems. Then I designed, fixed and installed large very high quality intercom systems and telephone systems for a smaller company for 15 years.
Then I joined an upcoming American Tele-conferencing company in Canada. Then I joined a Japanese Pro audio manufacturer company. Then I retired but continued making changes with audio and intercom systems for a few years.

Now I since I am retired, I do whatever I want whenever I want and it is WONDERFUL.

 

[dougy83]

hello frnd, what are single switched-capacitor filters ???? u mean passive filters??? or they are different?

Switched capacitor filters are discrete-time filters (like DSP) using capacitors to set the filter coefficients. Unlike DSPs, they use analogue (i.e. unquantised) signals.

In Circuit Cellar magazine a while back, there was a design using a DSP. In software they made a single filter and tuned it from one frequency to another. A single filter that tunes for 20 to 20khz is common. The problem is that if the filter is looking at 50hz it is not looking at 1khz and may miss information. With 7 to 15 separate filters, running all the time, no information will be lost.

That adjustable frequency filter could be something as simple as a Goertzel filter; they are often suggested for implementation on simple 8bit microcontrollers. Apparently the filter can be implemented in 45 cycles on an AVR.

Another option is using a cascade of low pass filters and 2:1 decimators. Each filter would use the exact same coefficients but be driven at half the previous filter's sample rate as shown in the attached image... this process is continued for each extra frequency range with each requiring half the CPU cycles of the previous.

 

[ronsimpson]

I love the thought of doing this in a DSP.

After seeing the "8 pin IC" in post #11 I changed direction. It must be using switched capacitor filters or DSP because the filters are related to a master clock.

 

[japanjot]

I am 67 years old. I retired 12 years ago. Most of my career was with designing, repairing and installing audio systems.
My first job was fixing then designing car radios with Philips for Chrysler cars. I saw their first LED and I worked with their new audio cassette recorders and players (Philips invented the cassette tape system). Then Philips office computers (DTL logic ICs and millions of ferrite cores were its RAM memory) and punched cards were its operating program. Then Philips huge intercom systems. Then I designed, fixed and installed large very high quality intercom systems and telephone systems for a smaller company for 15 years.
Then I joined an upcoming American Tele-conferencing company in Canada. Then I joined a Japanese Pro audio manufacturer company. Then I retired but continued making changes with audio and intercom systems for a few years.

Now I since I am retired, I do whatever I want whenever I want and it is WONDERFUL.

Wow cool sir, may god bless u nd u live 100 more years,wts your real name? U have any website ?

Sir can you tell me any alternate single supply op- amp for this audio circuit? The mc33171 ic is not available to the shop near me and also tell me dat can i use 074 op amp with dual supply for peak detector as u said its a good op amp??

And sir as you said

Mine has a fast attack and a fairly slow decay. Increase the value of the 0.1uf filter capacitor to make both slower.

I want to slow only decay rate not attack rate, so how to do that?

 

[dougy83]

I want to slow only decay rate not attack rate, so how to do that?

Just increase the cap size. The attack rate will still be almost instant.

If you're interested, the attack rate is governed by the 1k resistor and C2, while the decay rate is affected by the 1M resistors and C2.

 

[japanjot]

Just increase the cap size. The attack rate will still be almost instant.

If you're interested, the attack rate is governed by the 1k resistor and C2, while the decay rate is affected by the 1M resistors and C2.

ok ,thnx i got it

 

[ccurtis]

Originally Posted by ronsimpson
The problem is that if the filter is looking at 50hz it is not looking at 1khz and may miss information. With 7 to 15 separate filters, running all the time, no information will be lost.

With a multiplexed display, and several filters, only one filter output is being seen at any given time anyway while the other filters are just sitting there. Of course, if sweeping and stablizing a single filter to look at a new frequency takes a significant amout of time, more than sweeping the display, then more information would be lost. The question of using a single filter comes down to how fast it can be changed and stablized to the next frequency. Thus, my inquiry about switch-capacitor filters.

 

[audioguru]

Wow cool sir

Please stop calling me a SIR. A SIR was Knighted by the Queen of England and I have never met her so I AM NOT A SIR!
Instead I am just an ordinary cool dude.

A TL074 quad (or TL071 single opamp) needs a dual-polarity supply if it is used in my circuit. Instead you could use half of an LM358 dual opamp but its response above 2kHz will be poor. I think there was a single opamp like an LM358 dual but I have never seen one because its performance is awful.

My circuit has been working well for 6 years. Why do you want to change it?

 

[ronsimpson]

With a multiplexed display, and several filters, only one filter output is being seen at any given time anyway while the other filters are just sitting there. Of course, if sweeping and stablizing a single filter to look at a new frequency takes a significant amout of time, more than sweeping the display, then more information would be lost. The question of using a single filter comes down to how fast it can be changed and stablized to the next frequency. Thus, my inquiry about switch-capacitor filters.

Most of the circuits I have seen, and built some of, there is a peak detector behind each filter. If the display is looking at the low frequency filter (and peak detector) and a high frequency note comes by, hits the high frequency filter and detector, then quits before the display moves to that detector the information is held in the detector for some time. This way even very short tone burst are captured and displayed. While in a single filter they will not be displayed.

In the 8 pin IC I used, the detector holds the information (peak) until the display reaches that detector. After the display moves to the next detector the old peak is deduced to 0.9 of what it was. So the attack time is fast and the decay time is 0.9x per display time.

This IC has 7 switched capacitor filters.

The switched cap filters I used did not work well over a wide range of frequencies because they need anti- aliasing filters ahead of the filter. So you need to build a filter followed by a filter each needing tuning.

 

[japanjot]

Can we change the frequency of that 8 pin ic (msgeq7 right ? Or else)??
I want to do this so that i can connect two of those ic to make 14 band equalizer

 

[ronsimpson]

All the frequencies are related to the 'clock'.
The 14 band example I found uses two MSGEQ7s. One shifted up and one shifted down in frequency. The math should be easy. We know the 7 bands and the clock frequency. If we want to shift up 1/2 or 1/4 notch then the clock frequency should be shifted the same percent. I also thought about using two clock capacitors. The micro computer will do one sweep then switch timing caps and do the next sweep. The MSGEQ7 is so simple that using two is best.

https://www.youtube.com/watch?v=EpLLqgySCD4 here is a picture!

 

[japanjot]

Please stop calling me a SIR. A SIR was Knighted by the Queen of England and I have never met her so I AM NOT A SIR!
Instead I am just an ordinary cool dude.

A TL074 quad (or TL071 single opamp) needs a dual-polarity supply if it is used in my circuit. Instead you could use half of an LM358 dual opamp but its response above 2kHz will be poor. I think there was a single opamp like an LM358 dual but I have never seen one because its performance is awful.

My circuit has been working well for 6 years. Why do you want to change it?

well i call u sir cz ur elder from me in terms of age and knowledge, since i was a child i was taught to respect every one...........so i guess now i should call u "COOL DUDE SIR"

I don't wanna change ur circuit , the thing is MC33171 is not available to me nearby so was just asking alternate...

one question~~> shall i multiplex ten outputs from ten filters to one LM3915 with peak detecktor OR
shall i multiplex ten outputs from ten filters with ten peak detectors to one LM3915

after this can you tell me frequency list which is present in 32 band equalizer, i am thinking to do something more big

 

[audioguru]

Each filter needs to have its own peak detector.

 

[japanjot]

hey buddy can u help in cascading two of 3915 ic, i tried the way shown in its datasheet but it seemed me bit confusing...thnx

 

[audioguru]

hey buddy can u help in cascading two of 3915 ic, i tried the way shown in its datasheet but it seemed me bit confusing...thnx

Figure 6 in the datasheet shows how an opamp is used to amplify the input to the low level LM3915. Then it shows low levels.
The second LM3915 is fed directly from the input signal to show high levels.

Page 9 shows how to connect two LM3915 ICs so that the 10th LED turns off when a higher LED turns on in the DOT mode.

 

[japanjot]

Figure 6 in the datasheet shows how an opamp is used to amplify the input to the low level LM3915. Then it shows low levels.
The second LM3915 is fed directly from the input signal to show high levels.

Page 9 shows how to connect two LM3915 ICs so that the 10th LED turns off when a higher LED turns on in the DOT mode.

ya i tried this circuit but the tenth led does not light brightly and does not completely gets of when the 11th led is turned on.....
and it didnt worked like with the opamp showed in page 9, i used single polarity opamp is that the problem?

then i tried without opamp, it worked but the 10th led problem comes as i stated earlier

 

[japanjot]

hey can this flip flop circuit be used to divide audio frequency ? or its used only for digital ones
View attachment 66963

 

[ronsimpson]

Digital only.
Not analog or audio.

 

[japanjot]

Digital only.
Not analog or audio.

ok thanx, can you help in cascading two 3915