brndvnrdn

lm386 with bigger input
« on: July 25, 2019, 08:40:38 AM »
Hi all,

I'm working on a project which incorporates the lm386. The circuit works, but I'm having trouble with the input of the lm386. Its clipping at an input of about 300mV, and I would prefer it if I could just use the line level coming out of my interface.

Does anyone have a suggestion for another ic that accepts greater input voltages?


Harpo

Re: lm386 with bigger input
« Reply #1 on: July 25, 2019, 09:16:31 AM »
abs.max.input voltage according to LM386 datasheet is +/-0.4V.
A simple voltage divider/pot/trimmer as shown in the parts datasheet should fix this for your line level voltage feed.
A common mistake that people make when trying to design something completely foolproof is to underestimate the ingenuity of complete fools.  -Douglas Adams

PRR

Re: lm386 with bigger input
« Reply #2 on: July 25, 2019, 11:51:36 AM »
Reducing signal level is the simplest problem in audio. You should know how to do it.

JMFahey

Re: lm386 with bigger input
« Reply #3 on: July 27, 2019, 04:15:54 AM »
Hi all,

I'm working on a project which incorporates the lm386. The circuit works, but I'm having trouble with the input of the lm386. Its clipping at an input of about 300mV, and I would prefer it if I could just use the line level coming out of my interface.

Does anyone have a suggestion for another ic that accepts greater input voltages?
Which is? .......  ;)

Please state available signal/desired sensitivity.
Design - Make - Service Audio Equipment since 1969.

Re: lm386 with bigger input
« Reply #4 on: July 27, 2019, 04:56:13 AM »
Thats something Ive wondered about too on smaller speaker driving IC's , is there anyway to change the gain of the device itself ?
on ic's  like TDA2030/2050 etc

JohnRoberts

Re: lm386 with bigger input
« Reply #5 on: July 27, 2019, 09:08:13 AM »
Thats something Ive wondered about too on smaller speaker driving IC's , is there anyway to change the gain of the device itself ?
on ic's  like TDA2030/2050 etc
Look at the data sheet to see... some have feedback networks built in so difficult to tweak. They often default to high gain for better stability. Negative feedback amplifiers are more likely to oscillate at low gain than high, because at high gain the NF is being divided down. 

There are ways to trick NF amps to be stable at lower gains.  An RC from - input to ground, or sometimes between - input and + input can trick the amplifier into thinking it is running higher gain than it is (it is high gain only at very high frequency). Using an external pad is generally safer and less intrusive approach.

JR
Visit https://circularscience.com to hear what properly "cleared" drums sound like.

Re: lm386 with bigger input
« Reply #6 on: July 27, 2019, 01:49:20 PM »
Cheers John,

Ive been wanting to make a very minimalist halfway decent amp to run off 12 volts for ages
I found an old Nokia 'wired' handsfree kit , TDA 72xx or 79xx ,forgot which ,will put out 30+watts from a single chip off a single rail 12 volt supply , typically driving a 2 inch 4ohm speaker  to very realistic volumes for voice reproduction without clipping or blowing things,
I have  tried the modern class 'D' 'Sure' amp , if its input level is strictly controlled ,performance is acceptable , but any peaks or transients can be  ear splitting  , old style class B to my ears handles overload better .

Oh Id been meaning to say, I now know what you meant by 'clearing' heads on the drum tuner , the audio demonstrates better than words , and yes thats exactly what you want , less frequency modulation ping pong  , and just a good sustained 'boooooom' with a smooth tailing  off.
Is there a way of measuring the fundamental of the shell alone with your gadget ?
I know a few drummers who insist that for max volume you need to tune the heads to excite the shells fundamentals, Ive seen it done by a few pro's and the level difference is huge ,maybe 10 db . Of course its not all about volume ,but having that extra dynamic punch on tap is awesome in the hands of a skilled player .

JohnRoberts

Re: lm386 with bigger input
« Reply #7 on: July 27, 2019, 05:49:38 PM »
Cheers John,

Ive been wanting to make a very minimalist halfway decent amp to run off 12 volts for ages
I found an old Nokia 'wired' handsfree kit , TDA 72xx or 79xx ,forgot which ,will put out 30+watts from a single chip off a single rail 12 volt supply , typically driving a 2 inch 4ohm speaker  to very realistic volumes for voice reproduction without clipping or blowing things,
I have  tried the modern class 'D' 'Sure' amp , if its input level is strictly controlled ,performance is acceptable , but any peaks or transients can be  ear splitting  , old style class B to my ears handles overload better .
I am not familiar with a 'sure' amp...  I really like the hypex class D in my home theater set up, and for battery operation in my tuner I use a cute SMD class D amp about the size of a wood tick who just ate, that just sips current.   There is not reason in theory why class B would overload graciously at all... No doubt there are good and bad implementations of any technology.
Quote
Oh Id been meaning to say, I now know what you meant by 'clearing' heads on the drum tuner , the audio demonstrates better than words , and yes thats exactly what you want , less frequency modulation ping pong  , and just a good sustained 'boooooom' with a smooth tailing  off.
clear is good  8)
Quote
Is there a way of measuring the fundamental of the shell alone with your gadget ?
To be honest I never tried and you are not the first to ask... DW who makes fine drums BTW have a very successful marketing promotion called “Timbre-Match     (R)”.
Quote
I know a few drummers who insist that for max volume you need to tune the heads to excite the shells fundamentals, Ive seen it done by a few pro's and the level difference is huge ,maybe 10 db . Of course its not all about volume ,but having that extra dynamic punch on tap is awesome in the hands of a skilled player .
Funny I recall when we were marketing the Peavey drums (made with Steve Volpp's rigid rim patent) we routinely got complaints that the drums were too loud. If they're too loud you might not be a drummer.  ::)

OK to answer your question no, I never even tried to measure that directly... but a standard feature in RESOTUNE, that works the first few passes in Tune Lug also indicates the amplitude of the signal coming off the drum.  So with a little patience you could tune your drum higher or lower pitch and note the level change.. Keep tuning in the direction that gets louder, if that is what you desire.

I am not sure that is a good use of anybody's time, but I just make the tuning equipment, you guys get to figure out how to use it.   

here is some more general info about drums https://circularscience.com/about-drums

JR

PS: The DW drums sound good even if tuned to other notes than the (cough) shell resonance.  That said if the drum manufacturer gives you tuning advice, it is worth giving that a listen first.
Visit https://circularscience.com to hear what properly "cleared" drums sound like.

Re: lm386 with bigger input
« Reply #8 on: July 28, 2019, 04:36:38 AM »
Hey John,

A bit off topic but I read through the drum tuning. My 'day job' background is in turbomachinery,  and we do dynamic anlaysis similar to what you're describing for discs and impellers. One thing that I noted, though, in your overview was that a certain type of disc nodal response is missing. If you consider a disc in polar notation (r, theta) you can have nodes in either dimension - constant theta, which you show, but also constant R. So much as you can cut the pi radians by two, four, etc... you can also cut the R dimension in the same way...making nodes at constant radius.

What gets more complex is that you have interaction of the modes... so like a matrix, you have to consider the "0" (umbrella) mode with 1, 2, 3 etc linear (constant theta) nodes, as well as the first radial node (1/2 of the radius) with the 1,2,3 etc linear nodes, then the second, and onward... making *very* complex mode shapes, albeit at increasingly higher and higher harmonics.

I was trying to think about this. Do you think these modes are bound to be multiples of the fundamental frequencies vs the lug?  The interaction of these modes seems to suggest that the lug modes and fundamentals should be in some way interrelated, as you populate the matrix of modes. Maybe this is over-doing the analysis though!  :o

JohnRoberts

Re: lm386 with bigger input
« Reply #9 on: July 28, 2019, 10:44:16 AM »
Hey John,

A bit off topic but I read through the drum tuning. My 'day job' background is in turbomachinery,  and we do dynamic anlaysis similar to what you're describing for discs and impellers. One thing that I noted, though, in your overview was that a certain type of disc nodal response is missing. If you consider a disc in polar notation (r, theta) you can have nodes in either dimension - constant theta, which you show, but also constant R. So much as you can cut the pi radians by two, four, etc... you can also cut the R dimension in the same way...making nodes at constant radius.

What gets more complex is that you have interaction of the modes... so like a matrix, you have to consider the "0" (umbrella) mode with 1, 2, 3 etc linear (constant theta) nodes, as well as the first radial node (1/2 of the radius) with the 1,2,3 etc linear nodes, then the second, and onward... making *very* complex mode shapes, albeit at increasingly higher and higher harmonics.
Sorry about feeding the veer.

I am not sure I completely understand everything you are saying but I get enough...   Yes in the wild drumhead vibration is more complex than just a single vibration mode, but for my measurements I am exciting the drum head with a single pitch, so the drum head will be reasonably well behaved in its vibration behavior (for tuning measurements).

During my "find both" mode fast scan I excite the drum head with octave spaced notes and can usually read them all accurately enough to isolate the resonances across a few octaves on one sweep.

Drum design is part art, part science. I mentioned the Volpp drum that used a massive rigid rims system that in my judgement was louder than typical because it kept more of the drum hit energy in the drum head vibration, and less lost in rim flexing and vibrating the shell. The Volpp drum also supported thin shells, but that's another story.
Quote
I was trying to think about this. Do you think these modes are bound to be multiples of the fundamental frequencies vs the lug?  The interaction of these modes seems to suggest that the lug modes and fundamentals should be in some way interrelated, as you populate the matrix of modes. Maybe this is over-doing the analysis though!  :o
I think I talked about this in the link I provided.  For a typical two head drum the fundamental note  (mode 0,1) and 1st lug overtone  (mode 1,1) are roughly 1.6x apart. They do not follow a linear multiple like string or wind instruments because of the round drum head and non-integer path lengths. This 1.6x ratio is not cast in stone and typical for equal weight/equal tension batter and resonant heads. You can pull this ratio somewhat by detuning the resonant head sharp or flat, but I wasn't able to comfortably get them spaced one full octave apart, in my attempts. There are drumheads (dot? heads) made with an added mass right at the center of the head...  This is ignored by the vibration modes with nodes or no movement at the exact center, while pulling the pitch of the fundamental lower due to the extra mass. This changes not just the fundamental but all resonance series with anti-nodes at exact center.

Alternately with one head concert toms, the first overtone mode (1,1) is suppressed and lug overtone mode (2,1)  becomes the first apparent overtone at 2.1x.  With no resonant head to shift the resonant series around, the concert tom has a pretty identifiable and consistent sound character. 

I try to remain agnostic about subjective tuning/voicing preferences. I just make an objective measurement device.

JR

Visit https://circularscience.com to hear what properly "cleared" drums sound like.


PRR

Re: lm386 with bigger input
« Reply #10 on: July 28, 2019, 10:55:53 PM »
> not sure I completely understand everything you are saying

I thought he was categorizing circular and diameter modes.

But I hit this dang page which covers the subject without making it simple:
https://www.acs.psu.edu/drussell/Demos/MembraneCircle/Circle.html
You may need sea-sick pills......

JohnRoberts

Re: lm386 with bigger input
« Reply #11 on: July 29, 2019, 01:28:58 AM »
> not sure I completely understand everything you are saying

I thought he was categorizing circular and diameter modes.

But I hit this dang page which covers the subject without making it simple:
https://www.acs.psu.edu/drussell/Demos/MembraneCircle/Circle.html
You may need sea-sick pills......
I used to have a link to a similar Dan Russel page on my website years ago but the link broke a year or two back (maybe longer if/when he left  Kettering)... now the old link asks for a donation.  http://www.kettering.edu/~drussell/demos.html  I guess I can just tell people to google Dan Russel's demos... or ask you for the current link.  ;D



I have a lot of info about drums on my website... I am not familiar with the jargon Dogears used... but use the same (d,c) numbering nomenclature for describing vibration modes in my writing as Russell. 

I believe Dogears was talking about more than just single note pitch vibration  modes.

Again I apologize for the veer...

JR
Visit https://circularscience.com to hear what properly "cleared" drums sound like.

Re: lm386 with bigger input
« Reply #12 on: July 29, 2019, 06:24:18 AM »
Well, it’s all the same we just don’t talk about notes but instead frequencies, and their corresponding mode shapes. And we don’t excite these mode shapes by hitting them with things (or at least we try very hard not to).  For my brain this whole thing is kind of like working backward. Since  every structure has infinite mode shapes, each corresponding to precisely one certain resonant frequency (or note) based on its 3 dimensional stiffness and mass, we have to be concerned about exciting mode shapes by frequencies (like passing with an upstream structure or self-excitation at the rotating frequency). With a drum you excite every mode at once by hitting it (though some mores more than others depending on where you hit). Completely different way to think.

But yes, I was talking about diameter modes vs circular modes.  A drum skin is basically 2d, you have nodes that are lines instead of points. And those lines can be straight (folding the thing in half) or circular (Like the corner in the bottom of a Dixie cup or the bottom of a coffee filter).

It seems like all multiples of the circular modes would be the same amount of sympathetic to the effects of the lugs.

It also seems if you mapped a frequency response plot to an impulse, like a Fourier transform, you could “see” which modes were over-represented.. which should correspond to any cyclic asymmetry (ie uneven lugs). Really fascinatingly complex to think about. Actually because of the lugs it’s quit a bit more complex than what we do with a rotor since cyclic symmetry is assumed to be ideal.

Completely useless discussion from a practical sense. Has no effect on what one does when tuning a drum or otherwise. But neat.

Re: lm386 with bigger input
« Reply #13 on: July 29, 2019, 06:36:55 AM »
Sorry to clarify in my original post I was talking about the nodal diameters and nodal circles in polar notation. So (theta,R) corresponds to (d,c). It is considered in your original post, John, I just missed it since your analysis focuses on the (0,1)  (1,1) and (2,1) - as they should, since  lower harmonics should mask increasingly higher ones.

Sorry for the derail.

JohnRoberts

Re: lm386 with bigger input
« Reply #14 on: July 29, 2019, 09:51:00 AM »
Well, it’s all the same we just don’t talk about notes but instead frequencies, and their corresponding mode shapes. And we don’t excite these mode shapes by hitting them with things (or at least we try very hard not to).  For my brain this whole thing is kind of like working backward. Since  every structure has infinite mode shapes, each corresponding to precisely one certain resonant frequency (or note) based on its 3 dimensional stiffness and mass, we have to be concerned about exciting mode shapes by frequencies (like passing with an upstream structure or self-excitation at the rotating frequency). With a drum you excite every mode at once by hitting it (though some mores more than others depending on where you hit). Completely different way to think.

But yes, I was talking about diameter modes vs circular modes.  A drum skin is basically 2d, you have nodes that are lines instead of points. And those lines can be straight (folding the thing in half) or circular (Like the corner in the bottom of a Dixie cup or the bottom of a coffee filter).

It seems like all multiples of the circular modes would be the same amount of sympathetic to the effects of the lugs.
I first started looking at the problem of how to measure drum tune quality electronically almost 20 years ago when I was at loose ends after leaving my last day job, and drummer friend shared that he was told "it couldn't be done". That was all I needed to hear.  8)
Quote
It also seems if you mapped a frequency response plot to an impulse, like a Fourier transform, you could “see” which modes were over-represented.. which should correspond to any cyclic asymmetry (ie uneven lugs). Really fascinatingly complex to think about. Actually because of the lugs it’s quit a bit more complex than what we do with a rotor since cyclic symmetry is assumed to be ideal.
My early attempts used computer FFT reading of struck drumheads and couldn't make sense of the output (perhaps a personal problem).
Quote
Completely useless discussion from a practical sense. Has no effect on what one does when tuning a drum or otherwise. But neat.
Drums are alternately simple to play (you hit them with a stick), and complicated to tune (well). Of course playing well is more than just hitting them.

You may find the mechanism that I use to measure "clear" quality interesting. I have two loudspeakers firing directly at the drumhead to vibrate it. After a scan/sweep I determine the first overtone, mode (1,1) for two head drums, or mode (2,1) for one head drums. Playing a steady tone at this overtone resonance frequency I set up a standing wave in the drum head.  With a microphone centered over the lug I want to clear, I measure the phase shift between the sine wave I am sending to the drumhead and the sine wave coming back. As I tighten or loosen the lug being measured the phase of returning sine wave advances or lags. Making the phase shift at each lug the same as each other, has the additional benefit fine tuning the lugs to each other clearing up higher overtone dissonance.

There's lots more but this is already TMI for OP and people interested in LM386, again sorry about the veer.

JR
Visit https://circularscience.com to hear what properly "cleared" drums sound like.

PRR

Re: lm386 with bigger input
« Reply #15 on: July 29, 2019, 02:57:35 PM »
Russell's current page has many interesting discussions and animations.
https://www.acs.psu.edu/drussell/demos.html

Of interest to folks here:
Clink of a beer-bottle (high treble content)
Vibration of an electric guitar
Crankshaft Motion showing the effect of a short conn-rod (including the limiting case where it hits)

JohnRoberts

Re: lm386 with bigger input
« Reply #16 on: July 29, 2019, 04:45:07 PM »
Russell's current page has many interesting discussions and animations.
https://www.acs.psu.edu/drussell/demos.html

Of interest to folks here:
Clink of a beer-bottle (high treble content)
Vibration of an electric guitar
Crankshaft Motion showing the effect of a short conn-rod (including the limiting case where it hits)
I vaguely recall the sound of a broken connecting rod cutting my oil pan in half and punching holes in both side of my engine block... 

I may have been somewhat above the engine red line at that moment.  ::)

JR 
Visit https://circularscience.com to hear what properly "cleared" drums sound like.


 

Related Topics

  Subject / Started by Replies Last post
3 Replies
3180 Views
Last post January 27, 2007, 12:08:54 AM
by skipwave
6 Replies
1903 Views
Last post November 01, 2007, 07:09:03 AM
by jensenmann
4 Replies
1359 Views
Last post May 09, 2008, 07:29:32 AM
by rotation
2 Replies
1007 Views
Last post January 23, 2010, 12:29:45 PM
by diyfanatic