Continue to Site

Welcome to our site!

Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

  • Welcome to our site! Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

Problem with chained LM3915

Status
Not open for further replies.
Your new peak detector has its output reference voltage at half the supply voltage which is not compatible with the LM3915 that needs it at ground.
It does not have the current-boosting transistor at the output of the opamp so it might not charge the capacitor fast enough for high frequencies.
The output of the opamp will suddenly go as low as it can for each negative half-cycle of the input and the delay for it to come back up to normal ruins it at high frequencies, it is missing a second diode to prevent its output from going too low. The 100k at the output should be 330k for good low frequency detection.

Why does your 'scope show one frequency but "says" two frequencies?
 
Your new peak detector has its output reference voltage at half the supply voltage which is not compatible with the LM3915 that needs it at ground.
It does not have the current-boosting transistor at the output of the opamp so it might not charge the capacitor fast enough for high frequencies.
The output of the opamp will suddenly go as low as it can for each negative half-cycle of the input and the delay for it to come back up to normal ruins it at high frequencies, it is missing a second diode to prevent its output from going too low. The 100k at the output should be 330k for good low frequency detection.
hehe, perhaps it's better idea to stick to your topology. There were some variations that used two diodes. I suppose it would be possible to add current-boosting transistor but dunno.
How would tlo71 be hooked with split supply? split supply is quite new for me, surprise....
that i know, artificial ground can be made with two resistors and opamp as buffer, and some more better topologies...
Why does your 'scope show one frequency but "says" two frequencies?
hmm, good point, haven't noticed that before, mostly i measure only one frequency so i really didn't notice that one before...But, i tested that again, and it has same frequency, around ~92hz for both, maybe single-shot blew it up somehow or then there was trigger issue. I'll try to keep an eye out for that one, thanks!

yeah, farnell and TME have quite difference in price for TLE214x, i'm not buying these, untils as last resort.....
Snow felled here too, first snow actually for this winter.....and 20cm, there were, still are quite many households without electricity.
 
Last edited:
With a single +9V supply its input can go from 4V to 5V.
missed this one here, seems tlo71 is no-go then, well, my supply can go only to 15v but 5-12v would be best supply voltage range
 
Somebody on one of these threads is going to try using an ICL7660 voltage converter to make -5V from +5V for a dual polarity supply for his opamps. But the converter produces 10kHz interference that might be heard.
The negative supply pin on a TL07x connects to the negative power supply voltage and the connection where the positive supply's ground and the negative supply's ground connects to the circuit's ground. Then the input and output of the opamp can swing above and below ground without going near a power supply voltage.
 
hehey, seems now there is output we're after (at least much closer), except if my circuit is wrong but why it would work then hehe....so, here's circuit i have now (your circuit) for tlo71 & shot from skope
blue=input from generator, yellow=output from peak detector.
i'm not entirely sure did i hook this correctly, but you tell me....what i wonder is that virtual ground is tied to normal ground and shorting that other resistor?i used two 10k resistors to create virtual ground, green stuff at schematic is virtual ground
also, without signal applied, peak detectors output stays at 2.52, supply voltage is 12v one sided
small addition, virtual ground is now made with opamp buffer which doesnt connect to power supply ground now
vfb-opa.png

NewFile15.jpg
peak detector_audioguru's_split_supply.jpg
 
Last edited:
so here's now circuit, measured towards power supply ground as reference, +9v
again, blue input, yellow output
peak detector_audioguru's_split_supply-with_buffer.jpg
output, same as earlier, only changed both channels to same volts/div
NewFile16.jpg
and output without signal
NewFile17.jpg

Edit: small error in schematic i posted, i DO have virtual ground tied to power supply ground, if i take those apart it goes haywire
 
Last edited:
With the virtual ground voltage at +6V then the peak detector signal to the LM3915 ICs can be no lower than +6V so 18 of the LEDs will always be lighted. That is why using opamps with inputs and outputs that work all the way down to ground (single supply opamps) or using a plus and minus supply with the bias at 0V must be used. If you tie the virtual ground to the power supply ground like you said then you do not have a virtual ground and you are shorting the output of the virtual ground opamp.
A virtual ground can be used when the input and output of an opamp has a series coupling capacitor. But the output of the peak detector is an important DC voltage so a coupling capacitor to the LM3915 input cannot be used.

In post #85 you show an opamp wrongly because it has no negative feedback and its output will be latched as high as it can go.
 
hmm, okay
If you tie the virtual ground to the power supply ground like you said then you do not have a virtual ground and you are shorting the output of the virtual ground opamp.
ah, well then it seems i understood this following bit wrong it seems:
The negative supply pin on a TL07x connects to the negative power supply voltage and the connection where the positive supply's ground and the negative supply's ground connects to the circuit's ground.
But, when i un-short opamps output that is virtual ground from real ground, output goes haywire. well, luckily opamps don't get destroyed by short circuits...

So, the current circuit i have now is in post #86, i didn't quite understood what is current situation....
In post #85 you show an opamp wrongly because it has no negative feedback and its output will be latched as high as it can go.
oh, which of those schematic you mean?
 
You cannot use a virtual ground for the peak detector circuit because its output needs to go down to 0V so that the LM3915 ICs work properly.
 

Attachments

  • virtual ground.png
    virtual ground.png
    23.6 KB · Views: 153
whaa? well that makes sense now that you mention it, i thought opamps that need split supply need also virtual ground....sorry!
But, since i have only single-side supply, where do i get positive and negative ground? don't get gray hair with me...
 
The negative supply pin on a TL07x connects to the negative power supply voltage and the connection wThat is why using opamps with inputs and outputs that work all the way down to ground (single supply opamps) or using a plus and minus supply with the bias at 0V must be used.here the positive supply's ground and the negative supply's ground connects to the circuit's ground.

^these two confuse me at the moment, if that makes sense
 
A single supply opamp has inputs and an output that work all the way down to 0V (ground) that you want for the peak detector.
But an ordinary dual-supply opamp will have inputs and an output that work all the way down to 0V (ground) only if it has an additional negative supply so that its inputs and output stay in their working voltage range which is around 0V.
0V is where the negative of the positive supply joins the positive of the negative supply.
A virtual ground is used when an opamp has a single positive supply but needs its output to swing up and down around its biased point like with audio.
 
0V is where the negative of the positive supply joins the positive of the negative supply.
like putting two power supplies in series and middle point is ground between?
A virtual ground is used when an opamp has a single positive supply but needs its output to swing up and down around its biased point like with audio.
so i actually do need virtual ground now, as we're dealing with audio yes?
A virtual ground is used when an opamp has a single positive supply but needs its output to swing up and down around its biased point like with audio
^nevermind
So, the goal is to get output from peak detector at zero volts when signal is not applied?
few years ago i did play bit with dual-supply opamps placing two power supplies in series
but, now i deal with only one power supply
 
Last edited:
like putting two power supplies in series and middle point is ground between?
Exactly. then you will have a dual polarity supply with a plus voltage output and a minus voltage output.

so i actually do need virtual ground now, as we're dealing with audio yes?
So, the goal is to get output from peak detector at zero volts when signal is applied?
You need the output of the peak detector to be 0V with no input signal, a small positive voltage with a small audio input signal and a large positive voltage with a large audio input signal.
If your positive supply is +12V and the virtual ground voltage for the peak detector is +6V then you need a voltage shifter to change all voltages down 6V. The voltage shifter can be a single supply opamp. BUT then you should use the single supply opamp as the peak detector in the first place so that you do not need to shift the voltages.
 
BUT then you should use the single supply opamp as the peak detector in the first place so that you do not need to shift the voltages.
hehe, i would use single supply opamp, but seems this lm358 and similar i have won't cut it, or then it's some other reason like those capacitors you pointed eariler and use of breadboard.
Unless these are fake lm358's, which i hope they are not, but they do have worked well in other stuff i have used them, well, somewhat slowly
 
If you use only one LM3915 with a reference voltage of 3V, then its 1st LED will light with an input peak of 134mV which is well above the 7mV max input offset voltage of an LM358. If you do not care about detecting frequencies above maybe 8kHz then you can use one LM358 but not two chained ones unless you are lucky to find an LM358 with very low input offset voltage.

Most music is compressed so that the dynamic range is not much more than the range on one LM315 anyway.
 
okies!
I'll let this be for now, and try to stock better opamps like those TLE's (probadly TLE2142) they seem to be much better alternative for this circuit
thanks for help and your patience, it might not be easy task to try and teach always! althought this issue is not done yet, we'll see what happens when i get better opamps....
 
this i also found while looking, should be adequate for this application, only power supply is 7v absolute maxinum.....but at least it's single-supply, low noise & offset and fast enought
https://www.farnell.com/datasheets/630323.pdf
 
You found a very nice modern opamp. They recommend 5.5V maximum supply voltage.
 
and it's cheaper than TLE one, i think i'll buy both, surely both'll find use. Meanwhile when i wait for those to arrive, i can start building 8x8x8 cube, made 4x4x4 earlier. As i don't have job, gotta find some other way to pass time, and electronics reward in all way possible
 
Status
Not open for further replies.

Latest threads

New Articles From Microcontroller Tips

Back
Top