Raspberry-Pi-Play-It-All-Luxury-Bamboo-3-Way-Musicbox Project

Status
Not open for further replies.
Hi,

I started sourcing parts for my latest project and I would like to hear some
feedback from you guys about whether I am on the right track or not so..
To justify some of my picks, first I wanted to use some parts I have anyway, like
the 5" display, the Pi3 (I have an old model B pi1 but apparently It doesn't work
with i2s DACs), the OLED display,.. and second, I am a polymer materials
scientist guy and therefore I am totally into Aramide speakers and fancy Polyimide AMTs ... 😀.
I also wanted quite some high quality sound without paying more than 200€ for
the whole thing.

I have a CNC router where I am going to carve the Bamboo wood plate and metal front with.

Ok, so here's the mockup I made since I wanted to see how It's going to look upfront without doing too much CAD work first:

An externally hosted image should be here but it was not working when we last tested it.


Ok, first let's start with the speakers:

I have chosen this woofers which I have a mixed feeling with, first they are quite
cheap, have a good size for the project and they look damn sexy , on the other hand they might not be as Hi-Fi as I wanted it to be .. let's see.

Skytronic 5 1/4" Woofer, 100W RMS, 8Ohms, Range: 35 Hz – 8 kHz

As for the midrange/full (let´s see If I use filters) I am going to use this (I don´t know if its original H/K or chinesium fake):

Harman/Kardon 2 3/4", 12W RMS, 6Ohms, Full-Range

And for the highs, I fell in love with this little beasts:

No-Name Polyimide Ribbon ("Air Motion Transformer") 2" , 15W, 8Ohms, 3,3 KHz-40 kHz

I have little knowledge about speakers, but to me It looks like this combination
will be useable. The "full spectrum"/midrange could be driven with the full
spectrum or just the mids.. I have no clue what´s better in this case.
Also, I will use two 2.1 class-D-amps which have a crossover built in (can be
adjusted too) and so the bass is already separate. I could then put the full sound
range straight out of the amp to the full range speakers and then use just one
highpass for the AMT-tweeters. Would that be feasible?

Let´s see what I will be using as an amplifier:

TPA3118 2.1 30W/30W/60W Digital amplifier
I have chosen this since the new TPA3116/3118 have excellent reviews and seem
to be superior to the already good TDA7498 class-D-Amps. The advantage here
is that it has an active crossover already built in, so I will be using two of this
which will give me 4.2 channels (with stereo audio off course). So I have 2x 60W
for the woofers and 4x30W for the rest, which is more than I need. I am not sure
if I will use two channels for the tweeters and two for the midrange or just use
two channels and omit the unused two.
See picture:
planlmss6.png

I think I need some advise at this point since I am not quite sure if this is a good
Idea or not.
Maybe an other amp? Different wiring? DSP? Crossover?

Next up is the DAC. I already have this one:

HifiBerry DAC 2.0 clone with TI PCM5122
It produces great sound and I am satisfied. Had this part already in use and so I
will likely keep using it. I have an eye on this one:

AK4495SEQ I2S DAC

So, I am not sure if this will actually make a difference with the complete setup.
The DAC seems superior, I could swap opamps If I wanted and such. But I am not
sure If I would actually hear a difference for twice the price. Advice?

For the other stuff, I have a yellow 128x64 OLED screen which would display
track info or an EQ or so.. Had this in my parts bin and never found a use for it.
on the mockup it looks great.

The raspberry pi is either going to be a pi1 model B rev 2.0 or a pi3. I have both
and initially wanted to use the pi3 since I thought the pi1 model B didn't have an
I2S port. It turns out that the pi1 rev 2 DOES have an I2S port, although there is
no header. I would have to solder it.

Does anybody have experience running volumio on a pi? I think that using the pi3
instead of the pi1 would be overkill for just a music player and the pi1 is unused
anyway at the moment, so It would be a shame not to use it. So it´s likely going to be
the pi1 which should be fine for volumio. Btw, I want to use Amazon Alexa with it
too. Amazon recently opened the Alexa API and it can be used on DIY projects.
COOL!

As display I use is this:

5 Inch Touch Screen HDMI 800x480 TFT

I am using this for several projects and I am happy so far. Colors are a bit on the low side since it´s no IPS panel but for the price It´s fine. Size is about perfect for this project.

Knobs and "glue logic" will be made as I go. I will also be updating this thread and make a bit of a log here.

So, before I just start and fail, I first wait and hope for some feedback and possible better ideas regarding amps, speakers, wiring and such.
What do you guys think? 😕

Best regards, Phil
 
Your project, while well intended, is misguided on several levels.

Drivers: The drivers you are buying look "fancy" but could be any junk. How do you know (at all) what their performance is in terms of frequency response, sensitivity, resonance, etc, etc??? Although these parameters are most important for the woofer, all of the drivers need to be matched up and integrated together in the system, tied together by the crossover. Even a good driver may be the wrong driver considering the other parts of the system.

Crossover: it seems you have given very little thought to the crossover. The amp board has built in "crossover" you say? What about it? What can it do? I did not even see a mention of any "crossover" in the link you posted. This is a serious and deep subject. Even people who have spent hundreds and hundreds of (in your case) pounds on drivers can still end up with a speaker that sounds bad if they do not design the crossover appropriately.

Cabinet Design: Your mockup of the boombox looks nice but I noticed that it seems to be lacking some proper scale. The woofer you chose is much smaller than pictured. It's only about twice the diameter of the midrange. You didn't mention how you will size the box for the woofer, and if you will make a separate internal chamber for the midrange (you have to, you know).

Getting it right: I have seen and heard some very nice sounding DIY "boom box" systems designed by a local guy. My own DIY efforts are focused on active loudspeakers. I strongly doubt that you will be able to put together a good sounding system using the specific components and approach that you mention. You certainly can get those parts to make some noise, no doubt, and if that is your lowly goal I think you can throw it all together, tweak some knobs, and with some luck it will sound OK. But there is only a slim chance that it will sound good, and you might as well just go down to high street and get yourself a boomin' system there and save yourself the trouble! No joke. DIY is fun and a great pastime when you REALLY know what you are doing. By your own admission, you don't know anything at all how to design a system like this, and your main goal seems to make something flashy with lots of aramid showing!

MY ADVICE FOR YOU: Start with a super simple design and a proven driver. By using a small full range driver you don't need a "crossover" at all, only some EQ to even out the response. For instance, the Peerless TC9FD-18 driver is very good and inexpensive (not sure where to source in UK). Put that in a properly designed bass reflex box and with the right amp it will really sing for you (meaning it will sound good!). Also, since you are using the Pi, you might consider doing the EQ in the Pi itself before outputting the audio to the DAC. I, and others on this forum, have been doing that in our projects with success and can help you. With only stereo output, you can focus on buying a better quality stereo amp that will work for your mobile project.

Give it some thought. If you want help in choosing speaker drivers and/or with other aspects of your project, post here and people will try to help guide you.
 
Dear Charlie.

Thanks for taking the time and pointing out some important things.
By reading your answer I started realizing that I was not so clear one some important aspects, let me try to pick them up as I reply.

First off, I am not in the UK.. I just happened to find most products on Amazon UK because I wanted you guys to be able to read a thing or two... The product descriptions I had were in german. But thats not important 🙂

I love to listen to good music on a good system. As my ears are concerned I can´t stand most compressed music and up to 256kbit/s of current generation encoded mp3 I really hear a difference. I like listening to FLAC files on my fiio x1 with some cheap but kick-*** xiaomi in-ears with dual dynamic/balanced armature drivers. On the other side I THINK I don´t need a 384kHz 32bit player on a 5000$ McIntosh amp on 5000$ Martin Logan speakers 🙂 Just to put my music listening habits in frame.
Also, I don´t want to start a long-time hobby here. I just wanted a fun project for the CNC router I built as my last project 🙂
I also didn´t want to build the next high-end thing here.. as long as it sounds somewhere in between OK and good I´m fine and then It´s mission accomplished for me. But on the other hand you are so damn right, I don´t want to build the next big frustrator either.
Yes, I like something that will be in my living room to look great and as I said.. I have a heart for engineering plastics. I know paper cone speakers can sound absolutely gorgeous and can be expensive too! But you know... It´s hard for me to give up on this idea 🙂 I bet the drivers you have shown are good! BTW, I just got told that the aramide full range speakers are indeed real Harman/Kardon ones and not just a knock-off.
I will really think about this some more time and then decide. But the woofers might really be crap. (aside from lack of component harmonization)

Ok, let´s talk about the cabinet design:
I didn't yet do proper 3D design yet. The scale is not off it's the relations that are off but that´s because the midrange speakers are a different model. I forgot to update this.
I am waiting for T/S parameters for the woofers .. thats the main reason why I didn´t yet do it in CAD. But the fact that you noticed it shows that you really want to help me!

Ok, now comes the crossover part.
For me as an engineer from a different field crossovers seemed to be just higher order bandpass filters that need to be matched to the specifics of the drivers. I didn´t know that there is so much behind it. I would have looked the drivers recommendations where the frequency range is and calculated and built a butterworth filter accordingly. Thats it 🙂

By the way, the amp really has no crossover. Thanks for pointing this out. It just has one piece of dual-channel amp silicon for L+R and a second chip where both channels are laid together. This changes a whole lot for me. If it were just one set of high pass filters for the tweeters that would have to be done I would probably have built them myself. But now it´s also the lowpass and I am still not sure if the "midrange" speakers were just to get the middle range of the spectrum or the full (since they are basically full range drivers).
This really changes some things...
What about I´ll stick to my three-driver crazyness and do the following instead:
You know this piece?:
allo Piaono 2.1 DAC
What about I use this one in parallel to my current HiFiBerry DAC (I2S is one-way, right?)
and then have three DACs ;-) two with DSP and one without.
This way I would omit the filter problem and have all the fun in the world creating the filters in software and iterating my way up towards good sound. Three smaller amps instead of two bigger ones wouldn´t make that much of a difference money-wise.

If thats another good way to the great frustrator, I really should go with the fullrange option 🙂
 
Hi Phil,

If you want a project that you can quickly build I would avoid the crossover and the multiway and go with two full range speakers. There are actually some relatively good ones available these days. There are quality full-range offerings from Mark Audio, Tang-Band, Dayton Audio makes a good 6.5" and 8". This would make the system much more positioned for "success" in terms of sound quality and the amplification and DAC would be simpler in scope.

On the other hand, if you really want to go after the set of drivers that you initially described, then you might want to consider in a little more detail how to implement the crossover. In a simplistic sense the crossover consists of bandpass filtering, but other types can be used as well, and you will need to match the sensitivity of all of the drivers somehow. This can be very complicated, and involve multiple measurements, modeling, etc. or you can get some generic filters and try to do it by ear. Some people are able to design the entire system by ear, but that is far from simple. When you put the drivers in a box you get some different responses than what the manufacturer provides you, and the relative position of each driver can influence whether or not the drivers will sum or interfere at certain frequencies, and with the crossover filters in place. How deep you want to go into this topic is up to you. It really depends on whether you care more about the look of the system than the sound, and are more into the assembly of a working system or getting the sound to be really good. Again, totally up to you and there really isn't a wrong answer - it's your project after all.

As far as the DACs go, I am not sure that you can piggyback two I2S DACs or not. Frankly I am not sure about the implementation from the hardware side. I use USB DACs myself, for ease of use an their low cost but good performance.
 
Status
Not open for further replies.