Music Server with Improved Apple Airport Express

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hi,

I'd like some advice on building a quiet power supply that could deliver 1.2A in 3.3v and 0.7A in 5v.
This is a project to improve the Apple Airport express as an Audio Source. On the web you can find that the Apple Airport Express has been used apparently very successfully to work as a High End wifi audio source. For instance Apertura Rosita is a modified Airport express (http://www.larosita.eu/english/). The modification seems related to the Power supply, the clock and some caps or dig output of the pcm 2705.

My objective is to use the AE Dig output to High end dac (Studer D424) and I can envison a project in 3 phases:
- 1) improving the power supply
- 2) convert the optical output to a SPDIF coax or AES/EBU output (the PCM2705 has SPDIF output)
- 3) Improve the clock (Apparently without trying to make any publicity Selectronic has a good clock)

Phase 1: Power supply.
From what I can gather the AE I have 802.11g has a power switching power supply that can deliver 5V 0.7A and 3.3V 1.2A.
This stock power supply is noisy and a quiet one should improve a lot the Analog and dig out.
Link to a dissected AE: http://www.vonwentzel.net/ABS/Dissection-Express/index.html
I am looking for something more advanced than the std LM317, and I found the two schematics in old magazines (LED):

-Schematic A

- Schematic B

However I have some questions:
- Can any of those schematics with a LM317 and a transistor handle up to 1.2A with 3.3V and 0.7A with 5V?
- What reference of transistor should I use?

Any advice/help on this is welcome.
 

Attachments

  • schematic_a.jpeg
    schematic_a.jpeg
    27.7 KB · Views: 3,177
Values for the supply

I think I am going to try schematic B as I think I was able to calculate the right values.
If I understand correctly the schematic B, the transistor is a pre-regulation stage then the IC1 is the regulator.
I plan to use a 30VA Transformer with 2x 9v on the secondary. 1 secondary will be used for the 5v 0.7A and the other for 3.3v 1.21A.
So here are the values that I have calculated:

For 5v 0.7A
PR1 6A 600v
C1 0.1uF63v
4 caps for the rectifier 10nF 63v
C6 4700uF 40v Electrolitic
C7 0.1uF 63v
Z1 Zener 9v1 1.3W
R1 82 Ohm 1/2 Watt
T1 TIP31C
C8 10uF 25v Tantal
C9 1uF 25v Tantal
Rv1 Variable 1KOhm 1/4 Watt (Set at 706 Ohm)
IC1 LM317T
D1 & D2 1N4001
R2 240 Ohm 1/4 Watt
C10 47uF 25v Tantal
C11 1uF 63v

For 3.3v 1.21A
PR1 6A 600v
C1 0.1uF63v
4 caps for the rectifier 10nF 63v
C6 4700uF 40v Electrolitic
C7 0.1uF 63v
Z1 Zener 6v8 1.3W
R1 82 Ohm 1/2 Watt
T1 TIP31C
C8 10uF 25v Tantal
C9 1uF 25v Tantal
Rv1 Variable 1KOhm 1/4 Watt (Set at 386 Ohm)
IC1 LM317T
D1 & D2 1N4001
R2 240 Ohm 1/4 Watt
C10 47uF 25v Tantal
C11 1uF 63v

I have included the schematic in attachment.
I'd appreciate if anybody can confirm that my calculations are correct.

Thanks,

cdfr
 

Attachments

  • power.jpg
    power.jpg
    14.5 KB · Views: 3,220
Administrator
Joined 2004
Paid Member
Well it's interesting to know that the AE uses the good old pcm2705.

It's a good chip, but will not do anything over 48Khz - 16 bit.
For the most part our music is 44.1/16, so no worries. But can you force iTunes and/or the OS to send 44.1 and not resample to 48K? Maybe on a Mac or Vista?

Funny to see that the AE, which has a USB port, uses a USB DAC. Hmm.....

Cool project - have fun! I'll be interested to know what you come up with.
 
Hi, thanks for the comments. I'll try to provide some answers...

glt:
what I read was that the pre-regulator with the zener would actually help reducing the noise. I made sense to me on the "Garbage in, Garbage out" principle. If the regulator input is less noisy than its output should be a bit better. But true it does not change the fact that the LM317 will provide at best 80db ripple rejection.

Schematic A which I think is a Sultzer could be better but I am not sure how to calculate the values for 3.3 & 5v as well as not sure if it can wistand 12.A draw.
From looking at the forum, Jung redulator would be even better but there again the complexity increases and I could not find if actually i works for 3.3v.

WT:
You have to break the AE to open it. It is sealed no other way. I cut it with a dremel at the seam between the 2 half. Be careful though not to cut too much. You can find plenty of explanations on how to do that. I don't have a link handy just google it.

Your idea of the Teddyreg seemed appealing, but I read that it is for small current. 1.2A is not small. I do not think it will work well.

panomaniac:
Apparently itunes does not resample PC or Mac. I have used both. The AE I have which is a 802.11g version has the 2705 so as you said limited to 16/44.1. However the newer ones support 802.11n and use a Marvel chip, not sure of their capabilities but supposedly better...

There is a lot of buzz on this rosita and db system modified AE:
http://www.db-system.fr/bd_system/Aiport_Express_modifiee_-_WiFi.html
http://www.homecinema-fr.com/forum/viewtopic.php?f=1195&t=29882747&start=1335
 
Update on AE Mod

I just wanted to post some updates on this project.

I completed Phase 1 (Power supply)
- The results were very good on both analog output and surprisingly on the digital output. It gave a sound with more detail, more subjective dynamic and wider soundstage. All in all much better than the original.

I have not started phase 2.

I also completed phase 3
- This turns out to be a big disappointment so far. I am not sure why but removing the built in crystal and replacing it with a low jitter clock (10ps 1ppm; suposedly as good as tent's clock) has not provided any positive results. I cannot compare with phase 1 anymore but it seems worse, still better than the original AE but....
May be the input wire is too long? I used a separate transformer, rectifier and a 1000uf cap just for this clock which is then regulated with a low noise regulation circuit (150db ripple). Any hints, advice welcome here.

A picture is attached.

So far the CD drive (Studer D730) is still a far better source.
 

Attachments

  • aemod_cdfr4.jpg
    aemod_cdfr4.jpg
    68.2 KB · Views: 2,879
Hi, very nice. Regarding your previous post, theory says that the zener will help reduce the noise of the 317 if you use it in the adjust pin (see TNTaudio). But you already have it running.

I also plan to upgrade the PS of my AE. How easy was to remove the casing? Did you have to destroy it?

Since my dac can take I2S (TwistedPear Opus). I plan on tapping the USB inteface an feed that to a PCM2707, thus bypassing the internal USB to SPDIF (at least that was how the old AE was configured)

If possible, can you show the DAC chip in your AE? Thanks...
 
Hi glt,

>I also plan to upgrade the PS of my AE. How easy was to remove the casing?
cdfr: Not very hard with a dremel, but you have to be careful not to go too deep.

> Did you have to destroy it?
cdfr: Yes you basically destroy the casing since it is welded.

>I plan on tapping the USB inteface an feed that to a PCM2707, thus bypassing the internal USB to SPDIF (at least that was how the old AE was configured)
cdfr: Ok so if I understood you well, instead of the std pcm2705 of the AE 802.11g you want to put a pcm2707 which can deliver I2S. Keep me posted on your progress.

For the pictures I refer you this link:
http://www.vonwentzel.net/ABS/Dissection-Express/index.html
http://www.vonwentzel.net/ABS/Dissection-Express/DSCN8883-pp.jpg
http://www.vonwentzel.net/ABS/Dissection-Express/DSCN8884-pp.jpg

This is the same model as the AE I am using (i.e. the old one 802.11g). If you find pictures of the new AE 802.11n, let me know I am interested to find out more.

My immediate worry for now is how to get the expected results from the clock, and adding the phase 2 (AES 110 Ohm output).
 
Thanks glt. I took time to shoot a new picture.
I reduced the clock path an filtered more the clock AC-DC supply. Seems to be improving the sound. Difficult without a scope to tell for sure. Ears are subjective :)
 

Attachments

  • aemod_cdfr_1_20090321small.jpg
    aemod_cdfr_1_20090321small.jpg
    65.2 KB · Views: 1,858
Yet an another update of this project.

I completed the last planned modification (Phase 2). Creating an AES/EBU compatible Digital Output.

Basically this consisted in by passing the digital optical output of the AE and soldering directly on the pin 5 of the PCM2705 a cap and pulse transformer 75:110 Ohm.
This resulted in another improvement in sound quality when connected to my DAC.
After spending sometime listening to this last mod, I believe that the customised AE is now on par with my Studer D730 (used as a drive on the same DAC).
Not better (unfortunately) but equivalent, which is already a very good result since the Studer was a great CD player/drive and that the AE is cheaper, and gives the convenience of streaming audio from hard drive.

In summary:
A stock AE (802.11g) with:
- 1) improved power supply
- 2) Electrical digital output (pseudo AES/EBU in this case)
- 3) Improved clock

is equivalent to what used to be a very good drive with a CDM mechanic.
 
I just tried a new Airport Express change. I kept the two stage regulated power supply, but instead of feeding with rectifier/transformer, I used a car battery 12V 50Ah.
The change seems quite dramatic, the treble seems clearer and the bass tighter. At one point I actually wondered if I had less bass, but I think it is just less confused.

Has any body experience with DAc or preamp driven by batteries. What improvement did you notice?
 
After modifying the Apple Airport Express 802.11g, I modified the Apple Airport Express 802.11n

Reports were saying the the AE N was much better. Indeed the stock version was already very good and superior to the G version.
Instead of a TI USB DAC, this version actually uses a Broadcom processor with (I think) an integrated 24/96 DAC.
All the voltages are 3.3v instead of 5v & 3.3v.
This N version is what is used in the db system ZardoZ, La rosita and the Micromega Air Stream.
I described (in franch) the modification I made on the homecinema-fr forum.

Basically I made 3x 3.3v supplies with LM317 & a pre-regulation stage, with a 12v battery as the source to remove all noise.
This combination with my Studer DAC beats my previous combination with a CD player as a source.

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


Has any one tried to perform changes to the AE?
What results have you gotten?
What changes did you do?
 
Let me know, how it goes. I am very pleased with the results, but as usual I am looking for more :)
With the pre-regulator & LM 317 I think I am getting more than 80db ripple rejection on the supply.
There might be changes worth while to do on the AE itself, but would need more knowlege of the schematic.
 
Thanks Promitheus. I am very happy with the results.

Technically the new AE does support 24/96 (I read the broadcom spec), however Apple has not enabled it with the SW flashed or Airfoil, so it is currently limited to 16/44.1 .
Itunes does support 24/96.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.