LM39XX LED meter. Freq dependent issue

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

mkiijam

Well-known member
Joined
Aug 25, 2017
Messages
171
I'm working with the old LM3915 and I'm finding that it is too frequency dependent. It is in an existing design so I'd rather not have to change too much, but is there a easy way to get it to respond more evenly to AC frequency? Right now 100hz or so barely shows up, then 1khz is almost maximum.
 
The standard LM3915 circuit should be quite accurate. It's more likely that the problem is something external like the rectifier circuit or just a misunderstanding of how it works. Note that if you supply a pure tone, the standard rectifier circuit will output a higher value with a wide-band input as opposed to a pure tone. So test it with pure tones first. Try 1kHz and adjust the level so that it's in the middle of one of the upper LEDs. Then try 500Hz and see if it changes. Then 200, 100, etc. It should produce the same result. If it does, there's nothing actually wrong.
 
There is no rectifier circuit in this circuit. The AC signal is being fed directly into PIN 5, which is the input.
 
It is just the circuit from the datasheet
 

Attachments

  • lm3915 - Copy.JPG
    lm3915 - Copy.JPG
    32.1 KB · Views: 26
mkiijam said:
There is no rectifier circuit in this circuit. The AC signal is being fed directly into PIN 5, which is the input.
Ok. Then it's a peak meter. So it will only tell you the level of the frequency with the largest amplitude. So it's probably working just fine. Try test tones like I described to confirm. But usually there is a rectifier circuit (which could be as simple as a resistor, diode and capacitor).
 
ruffrecords said:
OK, but that is only part of the story. What is the circuit of the part feeding audio to the chip?

Cheers

Ian

??? It's just an audio input. You can send it whatever you like. I'm sending it an AC signal from my signal generator. It seems to be more "sensitive" to lower frequencies.
 
With audio material you're most likely interested in peak values in the low frequency range as well as peak hold in some cases (it leaves the highest measured value LED lit, a feature found in some high end mixers/consoles)

Other content edited by Abbey because mostly wrong and irrelevant.
 
efinque said:
With audio material you're most likely interested in peak values in the low frequency range as well as peak hold in some cases (it leaves the highest measured value LED lit, a feature found in some high end mixers/consoles)
Which the proposed circuit does not provide, because the illumination of the LED's is not maintained long enough to impress the retina.
 
mkiijam said:
??? It's just an audio input. You can send it whatever you like. I'm sending it an AC signal from my signal generator. It seems to be more "sensitive" to lower frequencies.
Yes, but you said it was part of 'an existing design' from which I deduce there is already something there doing the driving?

Cheers

Ian
 
mkiijam said:
It is just the circuit from the datasheet
https://www.mouser.com/datasheet/2/405/lm3915-443929.pdf

Fig 17 works a LOT better than it deserves, a lot better than feeding raw audio at the "signal" pin (which is intended to get DC).  Figs 18 to 20 are reasonable improvements, none real expensive.
 
I think some very early IC example circuits were (from a functional viewpoint) copied directly from the datasheets by hobbyists/engineers into their designs but due to component prices/availability the ICs were protected with all sorts of components which means you can't directly calculate certain parameters  (say, THD in opamps etc, in this case the measured range for example)
 

Latest posts

Back
Top