LED Spectrum Analyzer

[lasvegasmale31]

I have been working on my own spectrum analyzer for a few months now and have collected several schematics from others just to get ideas on how they solved certain obstacles. I feel my design is evolving pretty well and the circuit I have breadboarded right now runs surprisingly well with little more than 4 degrees F rise in the chips after hours of use.

I enjoy comparing other peoples methods of construction, component selection, and filter design. Not looking to steal a design, Rather I am looking to collaborate and share what I have learned if others are interested and post my results.

I have also purchased a Spectrum Analyzer kit from Gavin Curtis (sells them on Ebay) as I like his design and I hope to use it in my home (very nice looking).

If anyone else would like to collaborate on designs, feel free to PM or Email me.

 

[MrDEB]

In the mean time

perhaps you have run across this schematic?
https://www.electro-tech-online.com/attachments/audio-spectrum-analyzer-en-pdf.15065/

 

[lasvegasmale31]

No, I actually had not run across that one yet. Thank you for that info.

 

[d3dreaper]

perhaps you have run across this schematic?
https://www.electro-tech-online.com/attachments/audio-spectrum-analyzer-en-pdf.15065/

I have run across that one several times actually while i was trying to find a solution to the velleman issue. interesting stuff. I think i might try to build this one as well....

 

[MrDEB]

The multiplexing issue will come up

audioguru and I discussed the multiplexing issue several months ago and I came to the conclusion that multiplexing is out of the question.
I went with 10 lm3916's
have yet to actually build but the boy scout I am working with is presently building said unit.
hopfully by jan 1st the unit will be running??
I have schematices done in expresspcb that I would be willing to share.
Have pcb design as well

 

[Aristarco]

If you multiplex a single LM3916 then the 30mA output becomes an average of only 3mA in each LED which is very dim.
If you use millions of transistors then the LEDs will be brighter and only one LM3916 can be used with a switcher from each bandpass filter.

Use 10 LM3916 ICs.

At the output of LM3916 use a high current driver, ULN2003 or transistors, then multiplex those outputs. Might work.

 

[MrDEB]

If you have EXPRESS PCB software I would be happy to Email you the present design of the entire schematic along with my current PCboard design.
The unit I am working on uses the TILT FILTER
over at DIY AUDIO they have a detailed description of this filter design.

 

[MrDEB]

Here's a link to description of a TILT filter design. about half way down the page.
**broken link removed**

 

[Aristarco]

I'd like to see the design, but haven't Express PCB. If you can export to PDF or image format, I'd like to see it. Greetings!

 

[MrDEB]

expresspcb is a free download
much easier to send the files as the schematic is 7 sheet long.

 

[Aristarco]

I'll download it asap. Can you send the schematics? Greetings.

 

[MrDEB]

YES but go to PM and ssend me your Email address or post it here and take out the @ and put X or ?
the website can't send to many files unless maybe I zip the files.
Will try that first

 

[MrDEB]

spectrum analizer etc

here are the compressed schematic files and the pcboard files.
The pc board file needs more work

 

[Aristarco]

Thank you. I'll check 'em out.

 

[MrDEB]

I redid the LED resistor arrangement. Going with a bussed 330 ohm resistor array
4611X-101-331LF Bourns Resistor Networks & Arrays
about the same prie as a 3 watt resistor. Should work better.
Got wild idea of using RGB LEDs for display but needs work.
CONCEPT - lets say for discussion you have 10 LEDs that are separated by a 1db input signal (10db all the green leds with 20db all the yellows are lit and 30db all the reds are lit. Using only 10 leds per LM3916
wold perhaps require a micro controller.

 

[audioguru]

I have always used LM3915 ICs for showing audio levels because each step is 3dB which is heard as a small change and is exactly half or double the output power of an amplifier.
An LM3914 is a linear voltmeter that should not be used for showing logarithmic audio levels.
An LM3916 is a "toy" VU meter IC that has a different amount of level change for each step.

Professional recording studios use a VU meter with very small step changes of 1dB that can barely or cannot be heard.

At home I have my Sound Level Indicator project that uses a single LM3915 for a range of 30dB plus an automatic-gain-control stage that gives an additional range of 20dB for a total range of 50dB which is plenty. It shows a pin dropped on the floor in the next room (if I hold my breath and the refrigerator is not running) and also shows pretty loud music playing or a loud TV.

Years ago I made a VU meter with two LM3915 ICs in series for a total range of 60dB that was too much because either the low level LEDs were always lighted or the high level LEDs were never lighted.

 

[Aristarco]

A "toy" VU meter? Please elaborate. I find that interesting. In the good sense, I mean.
I want to use the 3916 in an audio project and that statement made me doubt.

Greetings.

 

[MrDEB]

maybe add this

using perhaps a 555 outputting a sawtooth signal

"For increased resolution, it’s possible to obtain a display with
a smooth transition between LEDs. This is accomplished by
superimposing an AC waveform on top of the input level as
shown in Figure 14. The signal can be a triangle, sawtooth or
sine wave from 60 Hz to 1 kHz. The display can be run in
either dot or bar mode."

just a thought.
working on refining the PCboard design. Found some SMD LEDs for the displays.
Superbright RGB LEDs PLCC2K type in red, green and yellow at $.08 per.
Thinking of separate board just for the displays and either ribbon cable or edge card connectors.

 

[MrDEB]

PCBoard design ???

trying to get the SMD LEDs on one side of a pcboard with the other side to mount the LM3916"s and resistor arrays.
KEEPING the LED arrays as close together as possible
the SMD leds are 2.8mm x 3.5mm
in one single row = 10 x 2.8mm = 28mm long x 3.5mm wide.
one 18pin dip and 11 pin SIP resistor array.
This would be a one sided board. LEDs on bottom side.

 

[audioguru]

A "toy" VU meter? Please elaborate. I find that interesting. In the good sense, I mean.
I want to use the 3916 in an audio project and that statement made me doubt.

The LM3916 has a different amount of signal powewr between its steps.
It has only 1dB between its 10th LED and its 9th. 1dB is a very small change in level that might not be noticeable.
It has another 1db between its 9th LED and its 8th, another 1dB from its 8th to its 7th and another 1db from its 7th to its 6th.
It has a little more change in level of only 2dB between its 6th and 5th LEDs.
It has another change in level of only 2dB between its 5th and 4th LEDs.
It has a change in level of only 2dB between its 4th and 3rd LEDs.
It has a noticeable change in level of 3db (half or double the power) between its 3rd and second LEDs.
It has a change in level of 10db (1/10th or 10 times the power) between its 2nd and 1st LEDs.
The difference in level between most of the LEDs is very small except for the lowest couple of steps.

The LM3915 has exactly 3dB between each step which is exactly half or double the power that sounds noticeably a little less or a little more loud. You can see how much output power at a glance. Its total 10 steps range is 30dB which is very noticeable.