I have just shared a new project: “Spectral Animations with a DIY MSGEQ7 I2C Device”
When I was starting out on this hobby I would buy something that looked interesting and play around with it to see what I’d learn. One victim of my ‘pretty thing consumption spells’ was this MsgEq7 which I bought, received, put away, and promptly ignored for twelve months.
Read more
This is great! Word counts are more of a suggestion anyway. I’m excited to have a proper read through of this later.
Thanks Pix!
Hi Everyone.
Some quick updates gathered between putting the pen down and the article being published.
What a great write up and project!
Also overshooting the 1k word count in the first 1/5 of the guide is gold.
Hi Pix
Congratulations on a good project and in particular the documentation, excellent. I have not gone through properly yet but I feel there are bits there of use in other applications so will look closer when I have a chance.
But there is one glaring error which may or may not be important. Depends on yow well up a person is with “preferred values” and what actual values are important and which are non critical.
Your Audio jack input arrangement in the schematic calls up two 4k6 and one 26k resistors. Now I don’t know where these values came rom but they for a start are not preferred values even in the E48 or E96 1% series. I note that you have used 5% resistors and they are in fact 4k7 and 27k which ARE preferred values.
The Violet or purple ring means 7, not 6. We will excuse the error as it is pretty obvious.
I bring this up because it is all very well for me and many others to see this schematic and say “well that is NOT a preferred value and the actual value is not particularly critical so I’ll use 4k7 and 27k”. All OK.
Then there are some who don’t know all this and will rush out and try to purchase 4k6 and 26k resistors. Good luck with that. This discrepancy should be pointed out at the point of sale but some confusion has been established and I could imagine someone going away wondering who is right and who is wrong. Pix or the counter sales person.
I am not trying to be critical here but while this is pretty unimportant to me and many others it could be quite confusing to the inexperienced person and I stress the importance of accuracy here because of this. I also note that no one seems to have picked up on this yet.
Cheers Bob
Thanks Bob. I put a lot of work into it.
Fair point.
If I took the audio coming through that jack and those resistors back into a speaker I imagine the quality of the sound would be subpar.
I chose those values through experimentation and they produce the most dynamic range out of the MsgEq7s Peak detectors.
I probably should have at least mentioned the preferred values and discussed why I didn’t adhere to them. Good feedback. Thanks 
Cheers. I still use this occasionally with my home system for some sweet vibes.
Hi Pix
You did adhere to them. In the pics they look like 4k7 and 27k to me. That is a violet band (the second one) is it not??. They are 5% types anyway and to my knowledge 4k6 and 26k don’t exist. My point was you appear to have used 4k7 and 27k but the schematic uses otherwise. The difference in practice would be just about nil.
The audio out of a speaker would be as good as the amplifier and speaker would allow. It would just be a Mono signal instead of stereo that is all.
Cheers Bob
PS. Violet = 7, Blue = 6 in the resistor colour code.
I think the lighting in the photos are are deceptive.
Looking at them in person, I think it’s sliightly more blue than violet and there supposed to be 4k6. It’s actually hard to tell even looking right at them. 
Measurements show it to be riiight in the middle. 
Roses are red
Resistors are blue
Violets are also resistors
I see the confusion.
When I said I was “not adhering” to the recommendations I was talking about this schematic from Sparkfun’s MsgEq7 Datasheet which asks for 22k resistors.
I found the datasheets 22k suggestion didn’t work for my circuit. The peak detectors of the multiplexer would either do nothing or came on full blast. That’s why I went with the resistors I chose.
Hopefully that clears that up for future readers trying to replicate my results.
Hi Pix
A question. When you tried the 22k (R1, R2) did you have the 27k (R3) in circuit.
I can see what happened. I checked the data sheet and found the input resistance (ZIN) of that IC to be 1MΩ which is pretty low. If you had R3 in circuit that would be an effective 49K as the upper resistance of a voltage divider with the ZIN of 1MΩ the lower resistor. This 49k would be significant. Pencil in a 1MΩ resistor from pin 5 to ground and you will see what I mean.
When you reduced R1 and R2 to 4k7 you get a bit closer to the suggested 22k (31k7). All the above assumes that the 27k R3 was present.
As per the suggested circuit R3 is not required but is a convenient way to adjust the aforementioned voltage divider and thus the actual level of signal reaching the input. Which you have effectively done.
R1 and R2 isolate L and R channels from each other and combine the 2 for a mono signal. I would not go below the 4k7 as you might lessen the stereo effect with cross talk.
As for colour of the bands. Sometimes the background resistor colour can disguise the actual colour a bit. I personally find the blue background on the 5 band 1% metal film resistors a nightmare to identify. Sometimes just easier to measure. But I can assure you that yours look like 5% types and 4k6 and 26k do not exist normally. That is not even a preferred value in the E48 and E96 1% type range.
That 4k65 you measure is within the 5% of 4k7 and if you factor the measurement meter tolerance maybe going the wrong way it would be well within spec. The difference is really academic and I only bought it up due to the possible confusion with a non experienced hobbyist due to the difference between pictures and schematic, however slight and of no real consequence.
BUT all in all a great project which I will be looking closer at as I think there is some useful programming (which I am pretty much a novice at) material I could use elsewhere.
Cheers Bob
Oh Maybe!!
The early prototyping stage was mid last year so I don’t remember the details but I think your theory is pretty likely because it explains the behavior really neatly.
I didn’t know that. I assumed they just made every value resistor. Guess it makes sense for the industry to settle on a few favorites to cash in on that sweet sweet economies of scale.
Glad you liked it Bob.
Hi Pix
Not quite but sometimes close.
The whole “Preferred value” system is arranged so that tolerances overlap. For instance your 4k7 lower value at 5% is 4k465Ω and the next lower value is 4k3 with an upper 5% value of 4k515Ω which overlaps and so on.
As the tolerances become smaller (1% and 0.1%) so the number of preferred values becomes larger as the values get closer together.
There is a good web site with lots of information.
Go to “technical data”/Resistor preferred values.
And there is heaps more. Probably enough to answer most future questions.
Cheers Bob
That’s a killer website Bob.
Good place for me to go find all the simple formulas I’m needing more and more.
Thanks for the link. I’m sure I’ll re-use this a lot.
Hi Pix
Yes i think so. I have just found it and have not looked properly yet but seems to have lots of info.
Formula simple and complex. Still requires a basic understanding of what is happening and where to use these formulae. I think a lot of that could be covered here also.
Cheers Bob
[/quote]
Wow man, you really had a proper crack at this! 
Looking forward to having the chance to get back to it, larger project (parenthood) permitting…
Anyone had a go at any of this? Have you done any more tinkering with it, Pix?
Hey @Kimmo
Lent the final product to a mate who’s a bit of a synth nerd and he loves it.
Seems to work perfectly.
I personally haven’t tinkered much since, but you must understand by the time I had finished writing my thesis on i2c msgeq7 devices the tinkering was really rather accomplished. 
And you can get our latest projects and tips straight away by following us on:
