After a 2 year on and off battle with these power meters ive fought with from badly failed microcontrollers to descretes to comparators, ive decided to go back to the LM3915 for my amplifier build that i have all but forgotten about.
Heres what i have drawn up so far.
Ive really just been copying right from the datasheet.
Im going to cascade 2 of them but only use 15 levels. im having trouble with these parts, i want to meters to run on BAR MODE. heat dissipation isnt an issue i can slap some heatsinks onto the chips. 2nd is the rectifier AND "sensitivity" adjust to connect it to speaker level, about 200W into 8Ohm or 40V RMS (higher than the IC can take)
For the rectifier i was looking at "Figure 18 Precision half-wave rectifier" from the datasheet then i get stuck on what op amps to use because the LF351 isnt around anymore and it has to run on a single 12V rail.
Another issue that might be with that rectifier, maybe not but is with the rubbish ebay bit a have is low frequencies like 40Hz the entire bar can be easily seen flickering with the AC signal. i was able to fix it on a breadboard with a 22uF cap and a 1.5K resistor on the signal input to act as a filter.
Another part but isnt a must have is the brightness control seen in Figure 25. I Believe this allows me to adjust current for ALL LEDS at once for a brightness adjust but im not certain
The plan in the end is to get PCB's made and hopefully finally have a design that works like ive been planning for months and months
Heres what i have drawn up so far.
An externally hosted image should be here but it was not working when we last tested it.
Ive really just been copying right from the datasheet.
Im going to cascade 2 of them but only use 15 levels. im having trouble with these parts, i want to meters to run on BAR MODE. heat dissipation isnt an issue i can slap some heatsinks onto the chips. 2nd is the rectifier AND "sensitivity" adjust to connect it to speaker level, about 200W into 8Ohm or 40V RMS (higher than the IC can take)
For the rectifier i was looking at "Figure 18 Precision half-wave rectifier" from the datasheet then i get stuck on what op amps to use because the LF351 isnt around anymore and it has to run on a single 12V rail.
Another issue that might be with that rectifier, maybe not but is with the rubbish ebay bit a have is low frequencies like 40Hz the entire bar can be easily seen flickering with the AC signal. i was able to fix it on a breadboard with a 22uF cap and a 1.5K resistor on the signal input to act as a filter.
Another part but isnt a must have is the brightness control seen in Figure 25. I Believe this allows me to adjust current for ALL LEDS at once for a brightness adjust but im not certain
The plan in the end is to get PCB's made and hopefully finally have a design that works like ive been planning for months and months
Last edited:
To mitigate LED flicker, you need a peak detector circuit. Just a follower with a diode inside the feedback, and a cap to hold the charge. Play with C and the load resistance till you get “hold” times that you like. 1 uf and a meg ohm or two works well.
The 3915 does not work in dot mode with audio without the peak detector. Without it, every cycle illuminates all the LEDs below the top because the signal goes through all those values periodically. Dot mode needs it to be held at the top and only look at the signal envelope.
Do not sweat over which op amp to use. Crazy high slew rates are not required, especially when using the peak detector. LF351s were suggested because of the low input bias currents. If making precision DC measurements you may want that, but for audio who cares? If you want to run off a single supply, an op amp which will common-mode all the way to its negative rail (ground) is preferred since it’s output will go to zero on a single supply, and more importantly can sense input voltages down to ground. Any bipolar op amp with a PNP input stage will do this. I use LM358’s or their quad versions for things like this because they are good enough and quite cheap.
The 3915 does not work in dot mode with audio without the peak detector. Without it, every cycle illuminates all the LEDs below the top because the signal goes through all those values periodically. Dot mode needs it to be held at the top and only look at the signal envelope.
Do not sweat over which op amp to use. Crazy high slew rates are not required, especially when using the peak detector. LF351s were suggested because of the low input bias currents. If making precision DC measurements you may want that, but for audio who cares? If you want to run off a single supply, an op amp which will common-mode all the way to its negative rail (ground) is preferred since it’s output will go to zero on a single supply, and more importantly can sense input voltages down to ground. Any bipolar op amp with a PNP input stage will do this. I use LM358’s or their quad versions for things like this because they are good enough and quite cheap.
How do you think this circuit will work? im probably going to use a jelly bean TL071/72 for my opamps. im going to make BOTH the peak and RMS versions of the rectifier to pick which i like better.
Imgur: The magic of the Internet
Does it look good so far?
Imgur: The magic of the Internet
Does it look good so far?
When I pull up the image, I don’t see anything to the left of the two 3915’s (where all your input circuitry would be). Looks like the pdf is cut off.
And another item - both 3915s should use a common reference divider, set for the high end of the range (requiring a large input voltage). Then amplify the signal going into the lower range unit. it says to do that in the old long form version of the data sheet. With only 2 stacked units and only using part of one of them, you may get away with it like it is.
And another item - both 3915s should use a common reference divider, set for the high end of the range (requiring a large input voltage). Then amplify the signal going into the lower range unit. it says to do that in the old long form version of the data sheet. With only 2 stacked units and only using part of one of them, you may get away with it like it is.
That's because the way EasyEDA works i have to have both rectifier circuits as a separate file or it will put them all of the same PCB. the inputs is going straight to the output of the datasheet copy of one of the rectifiers. i think all thats left to do is add a trimmer on the input to adjust the sensitivity. How do i change it for the common reference divider? i copied it all from the datasheets for the adjustable brightness, mainly because i dont want soft white LEDS lighting up my entire room. If your talking about pin 8 RED ADJ they are ties together before the trimmer pots. i might try to clean the schematic up it is a little messy
Last edited: