MSGEQ7 at 3.3V

Other Hardware General Electronics
1

For those who do not know what the MSGEQ7 is:
It is a 8pin IC that presents the frequency(7 bands) amplitude as an analog voltage from 0V to 5V/VCC. The 7 bands are multiplexed to one Output Pin.

I need to use a MSGEQ7 at 3.3V and I have heard of problems with the MSGEQ7 at 3.3V and the datasheet says that it is operating better at 5V. I would need to convert the 0-5V range of the output to 0-3.3V.
I need an analog conversion so I thought of an OpAmp or a simple resistor divider.
I would prefer the resistor divider because it is simpler but would it work? And what resistances should I chose?
In the range of 1k,10k or 100k?

2

The datasheet says it has an output impedance of 700 Ohms. You generally want a load 10 times (or more) of the output impedance so a voltage divider with resistances that add-up to 10K-100K should be fine.

But... The strobe & clock inputs might need 5V (when the chip is running from 5V).

3

They make logic level converters

4

The outputs are not logic level INTP, They are analog levels that signify the 'volume' of a frequency band.
Resistor divider seems like a good start.

5

The analogness doesn't carry through the converter? I didn't know that. Just figured it shifted proportionally.

6

An luck? I am facing the same issue. Works fine on an arduino but i need to move to a higher spec micro controller which has 3.3v IO.

This guy wen't into allot of detail and testing on this, but it ended up with an inconclusive work around.

There was a hint that maybe a 2.2k, 4.7k voltage divider could do it, but as pointed out the strobe and clock may still need 5V.

7

I think the author of that thread is fooling himself. Yes those chips are noisy and shifting it to a Teensy does not make them any noisier it is simply the higher resolution A/D that appears to produce more noise.
Also these are just second order filters so thinking you can have any sort of meaningful results by having five chips slightly off tuned is pure folly.

They work quite well at three volts as shown by my project running off a Raspberry Pi.


Also check out how I use them for dancing skeletons

8

Grumpy_Mike:
I think the author of that thread is fooling himself. Yes those chips are noisy and shifting it to a Teensy does not make them any noisier it is simply the higher resolution A/D that appears to produce more noise.
Also these are just second order filters so thinking you can have any sort of meaningful results by having five chips slightly off tuned is pure folly.

They work quite well at three volts as shown by my project running off a Raspberry Pi.

Spectrum Display on Vimeo
Also check out how I use them for dancing skeletons
Danse Macabre on Vimeo

Nice, the dancing skeletons are quite well done. Any one who didn't know would probably assume their movement was preset to the music.

Sounds like 3.3v will work just fine for what i'm doing then. As you'd know the MSQEQ7 data sheet just says 5v provides "the best performance". That's pretty much the reason for people like myself and the op trying to keep it at 5v. By performance are they referring to the amount of potential noise?

9

By performance are they referring to the amount of potential noise?

Possible but who knows it is quite a poor data sheet. What you can say is that with 5V you get a better dynamic range. But at the end of the day is is only a second order switched capacitor band pass filter.