LED VU meter not sensitive enough at low volumes

[Super-Dave]

Never mind, I just reread the entire thread. Using a 1N4148 introduces a forward voltage drop I don't want, killing my low volume sensitivity.

 

[audioguru]

I had a thought that woke me up early this morning from a dead sleep: the phase inversion happens when the input sine wave dips too deep below zero volts. If I were to make a full wave bridge rectifier using 1n4148 diodes upstream of the op-amp, it would never swing that low. Wouldn't that work? And if not, why not? A single diode could also do the job but then I lose half the energy of the audio input signal. As long as I don't wire the op-amp's up to invert the signal, it should be fine, right?

But then I wonder if that would effectively double my hertz, requiring that I increase the bandpass filter frequencies to reflect that for the spectrum analyzer? Or would that even matter?

The Jfet inputs on a TL07x produce phase inversion if the input voltage becomes within a few volts more positive than its negative supply. That is why it needs a negative supply so that the inputs never go too much negative.

If you use a bridge rectifier then it will have no output until the audio peaks exceed 1.2V or 1.3V.
You need an ACTIVE rectifier using an opamp with the diode or diodes inside its negative feedback loop.
The datasheet for the LM3915 shows half-wave and full-wave active rectifier circuits. Those opamp circuits need a positive and negative supply.
On my VU meter circuit I use opamps that need only a positive supply and they have inputs that work down to 0V. I use an active half-wave rectifier/peak detector that feeds the LM3915 with a very wide range of levels.

 

[audioguru]

I use an opamp preamp that feeds my peak detector.
Here is my active half-wave peak detector circuit. It inverts but it can be simply changed to be non-inverting.

 

[crutschow]

Here is my active half-wave peak detector circuit. It inverts

Looks non-inverting to me.

 

[audioguru]

Looks non-inverting to me.

When the input to my peak detector goes negative then the output goes positive.
When the input goes positive then the output of the opamp goes to ground, and the diode does not conduct.
I made it inverting so that my non-inverting preamp will not have 360 degrees of phase shift for oscillation.

 

[rjenkinsgb]

Have a read through this - it covers various active rectifiers, including full-wave:

https://nicadd.niu.edu/~fortner/course/phys475/lect/p575_01b.pdf

 

[Super-Dave]

If I do go with the peak detector, which should I chose? The one with the diode or the transistor? What are the advantages/tradeoffs of one over the other? Would this replace the input bias network, or be placed before or after it?

And in the interest of building a spectrum analyzer, would the bandpass come before it or after?

 

[audioguru]

1) Use the simple diode., it is fast enough. My peak detector circuit is designed to drive the high input resistance of an LM3915, not your transistor.
2) My peak detector is biased at 0V and must use an opamp that has inputs and outputs that go all the way down to its negative supply voltage that can be 0V. then an opamp that works without a negative supply must be used. Your TL07x opamps must have a negative supply or use a bias voltage.
3) The output of a peak detector is not audio so the bandpass filter must feed its input.

 

[Super-Dave]

I thought the reason you posted the rectifying peak detector was as a solution to phase inversion from using the TL074 op amp.

 

[audioguru]

Without a filtered rectifier the LEDs will look like a dim blur. The rectifier diode blocks all low levels because it does nothing until the signal level reaches 0.7V. An active rectifier with an opamp and a diode in its negative feedback loop cancels the 0.7V from the rectifier diode but if it does not have a peak detector then the LEDs will also be a dim blur.

You want the audio level to show LED light from a signal that is 0V (LEDs are off) to a few volts of signal level (all LEDs on or just the highest LED is on). A TL07x needs an additional negative supply for its inputs to go down to 0V.

An LM3915 with a peak detector feeding it will do exactly what you want and you should use an single-supply opamp that does not need a negative supply for the peak detector.

 

[Super-Dave]

I'm not using a 3915 driver, I have a biased transistor driving a cascading series of 1N4148s, resistors & LEDs. The fact the signal must pass thru resistors rectifies the signal already, and a 47uF electrolytic bypass cap between the transistor & the lights serves as my peak detector by holding the voltage long enough so they don't blur. Higher cap values hold it to a longer decay rate but reduce the maximum number of LEDs to light up, lower values tend to cause flickering. Experimenting has led me to choose 47uF, I get a quick decay with minimal flicker.

If I were using the LM3915, yes I would most likely need a rectifying peak detector. But I haven't had any luck testing or prototyping them on a breadboard. For some reason led #4 always stays lit up regardless of input. I don't know why, bad breadboard, resistor wires touching, maybe a bad batch of Chinese manufactured chips from Amazon. I don't know. But the biased transistor driven diode ladder works for me. So I'm going to run with it.

I really do appreciate your input Audioguru, in my research your name comes up an awful lot in different forums offering wisdom & advice. I grant you respect in regards to your experience and knowledge. But the wife is already pitching a fit over money already spent in acquiring parts and components for this project. I gotta make do with what I have on hand already. I have perf boards, a couple hundred LEDs in assorted colors, 2 boxes of assorted 1% resistors, 3 boxes of assorted caps (ceramic, film & electrolytic), transistors, op-amps (324s, 074s & 3900s), a couple dual rail power supplies (+15 & -15), and some pots. I also have an under-dash DIN box and a lens cover to house them all in once completed. I also have several dozen LM3914 & 3915 drivers that refuse to cooperate with me, some of which are toast.

Thanks to Rjenkinsgb's suggestion of incorporating a bias, I now have a functional light set that works suitably for my needs. It may be crude, uses a lot of components, but it's simple and it works with what I have on hand. 'Nuff said.

 

[audioguru]

The simple biasing of the transistor is affected by temperature.
Our hearing's sensitivity is logarithmic (which is why the steps on an LM3915 are logarithmic) so the steps with your diodes will look wrong.
Amazon sells shoes, clothes and some cheap fake electronic parts from "over there". A local real electronic parts distributor like Digikey has high quality reliable parts.

 

[Super-Dave]

I'm less concerned with accuracy than I am with remaining within my budget. But I can live with it. Those who get to witness it won't notice or care.

Not many products are locally made anymore, & the ones that are are priced outside my reach. So what if my components are Chinese? I designed & assembled it myself right here in the USA, and I doubt many folks can brag about that.

Believe me, I scrounged the interwebs for an audio spectrum analyzer for my car, no one makes them anymore, especially not in a stereo display (separate bands for left & right channels), not for 12 volts, not even as a kit. The few kits that are available aren't 12 volts nor in stereo. So if I want one, I must design & build it myself and do it within my means. If that means Chinese components sourced from Amazon, so be it. Too many manufacturers have closed their doors or moved their factories overseas & the few that remain are priced outside the reach of budget minded enthusiasts. Digikey is too expensive for me.

Kudos for beating the "Buy USA" drum. Unfortunately that's not a beat I can afford to dance to.

 

[audioguru]

I see poor people smoking extremely expensive cigarettes and drinking very expensive beer and whiskey.
They eat out frequently at restaurants and fast food places, but they are poor. I cannot understand that.

I had good jobs during my career so I afforded high quality electronic parts then I never needed to buy cheap Chinese fakes or defective junk. In another forum I was helping a member with a cheap Chinese kit that did not work.
Its defective ICs had no part number writing and were fakes. The same kit is sold on a few Chinese online sites.

 

[Super-Dave]

Maybe they're poor because they spend a majority of their funds on expensive stuff so there's nothing left to pay for decent things.

A couple years back my wife lived 4 states away taking care of her cancer stricken mother. When momma died she stayed to be a caretaker to her dad. Meanwhile I was working 7 am to 9:30 pm, 7 days a week, 2 full time jobs & 2 part time gigs (since there wasn't anyone at home waiting for me). I was pulling in a solid grand a week, yet I was eating Ramen noodles or dollar menu TV dinners every night. Every spare penny I had went into Project Turd to fix it, make it roadworthy and badass. Because I KNEW the wife would ask me to pack up and leave for Texas, and I wasn't going to leave my hotrod behind.

Sure enough, I'm in Texas today. The car isn't finished yet, but nearly all the important stuff is. New motor, trans, brakes, suspension, exhaust, wheels & tires, lights, most of the electrical is done. Interior needs work, carpet, heater, AC, dash. But it runs, drives and turns heads everywhere I go. I struggle to bring in a 3rd of what I did in Florida but I'm doing okay. My hotrod is legal, legit, dependable & safe. And it barely sees a 100 miles of road a month. Maybe 20 miles during the winter months, gotta keep the battery charged and the fuel fresh or it'll rot. Just last week I got antique Texas plates for it. $9k invested so far & counting. Soon as my audio display is finished, I'm pulling the dash & installing new performance gauges into it.

The point I'm making is this: we all have something we cherish, and are willing to make sacrifices for. You pick & choose what you love & what you don't. Today I'm back with the wife and she gets most of my time & money. My electronics "hobby" has taken me a year to build just 5 rows of lights, gather a drawer full of components and a case to house it all. I'll finish it this year, hopefully.

 

[audioguru]

You spent a lot on that car.
In 60 years I had seven 4 cylinder cars and one had a fun turbo, also I had one with a straight 6 and one with a V6.
I learned to drive on my dad's '58 Chevy.
36 years ago I bought my home for 108 thousand dollars and today it is worth 1.1 million or more.
I have three kids and three grand-kids.
I am triple vaccinated for Covid and the 4th shot should be soon.

Did you see the ship burning with 4000 luxury new cars on it? I thought it would sink but now they are towing it. I wonder if any of the cars survived, its crew did.

 

[Super-Dave]

You spent a lot on that car.

New cars cost a lot more. Several magnitudes even. But not many could look, sound or earn respect like mine does for even twice the price.
"Built, not Bought". Had I payed someone to do all this, I probably could have bought a new car. It's a lot more satisfying to claim "I did this!" Which is why I'm building a custom spectrum analyzer. One of a kind, not available in stores.