Inside the iHP-120...
Jan 27, 2004 at 8:39 AM Post #18 of 40
Hi All,

this are the semiconductors built in the iHP (Summary)
(type - manufacturer)

K4S561632D-TL75 - SAMSUNG
256Mbit SDRAM

SCF5249VF140 - SAMSUNG
Integrated Audio Controller/Decoder Microprozessor

39VF160 - SST
16 Mbit (x16) Multi-Purpose Flash

LVX245-F - Toshiba
Low-voltage logic IC

LVCH16245A - Texas Instruments
16-bit Bus Tranceiver

TLV0834I - Texas Instruments
3-Volt 8-Bit Analog-to-Digital Converters With Serial Control (Rev. B)

LM052A - Texas Instruments
?

UDA1380TT - Philips
Stereo audio coder/decoder

There are some other chips between the mainboard and the
harddisc-connector circuit-board, but the two boards are soldered
together, so I can't identify the components. :wink:

I've downloded the datasheets for all semiconductors listed above. Also the manufacturers links. I will place all information
within the relaunch of site www.ihplounge.com.


cheers
Juergen
 
Jan 27, 2004 at 11:23 AM Post #19 of 40
Quote:

Originally posted by Buckaroo Banzai
...UDA1380TT - Philips Stereo audio coder/decoder...


Now we are getting somewhere (thx, Juergen):

"The front-end of the UDA1380 is equipped with a stereo line input, which has a PGA control, and a mono microphone input with an LNA and a VGA. The digital decimation filter is equipped with an AGC which can be used in case of voice-recording.

The DAC part is equipped with a stereo line output and a headphone driver output. The headphone driver is capable of driving a 16 Ohm load. The headphone driver is also capable of driving a headphone without the need for external DC decoupling capacitors, since the headphone can be connected to a pin VREF(HP) on the chip.
...
The UDA1380 also supports an application mode in which the coder-decoder itself is not running, but an analog signal, for instance coming from an FM tuner, can be controlled in gain and applied to the output via the headphone driver and line outputs.

The UDA1380 supports the I2S-bus data format with word lengths of up to 24 bits, the MSB-justified data format with word lengths of up to 24 bits and the LSB-justified serial data format with word lengths of 16, 18, 20 or 24 bits (LSB-justified 24 bits is only supported for the output interface).

The UDA1380 has sound processing features in playback mode, de-emphasis, volume, mute, bass boost and treble which can be controlled by the L3-bus or I²C-bus interface."


(Edit: See link below for detailed info)

So, now all we need is someone with an audio-engineering background to put it all together & "draw" a complete pic for EE-impared idiots like me...
tongue.gif


My main interest is in the quality of the DAC, ADC, analog i/o stages, whether the volume is controlled in the digital or analog domain (is it gain, or attenuation - VIP for line-out) and when having the option, is it better to play back / record via digital or analog domain. That's all I'm askin' for... (lol!)
 
Jan 28, 2004 at 11:35 AM Post #20 of 40
Quote:

Originally posted by Traveller
Now we are getting somewhere (thx, Juergen)


Ok, some initial details (& plz keep in mind, I'm EE-impared!).

The first (& most obvious) assesment is that there are some redundant circuits between the core Motorola Coldfire processor & the Philips Audio Coder-Decoder chips (not to mention the separate Ti ADC chip!).

But here's a few assumptions on my part. They are based on the extremely detailed documentation Philips provides (s/b available @ Juergen's iHP Lounge).

I'm guessing that the Coldfire is only responsible for decoding audio files at which point the digital datastream is sent to the Philips Audio chip (via Data-in). The Coldfire also sends commands to the Philips chip (via the L3/ [ed. I2C, not I2S] interface).

The Philips chip is responsible for all audio I/O (as seen by both the block diagram & PCB traces). ADC, DSP processing, etc - could be performed by either chip, but I'm betting again Philips' is doing the work. The Coldfire chip looks like a CPU with a few built-in goodies that on a typical PC would be done by either peripheral HW or by SW.

While the Coldfire has, for example, an ADC component, I think that it's there to spare the [portable] designer the need to add extra chips (lowering costs) but the quality will certainly be lower than dedicated audio chip(s).


The key components of the Philips chip are:

Analog & digital [ed. I2S] i/o, gain ("PGA" - Programmable Gain Amp), A/D + "Decimator", DAC components (Interpolation filter, Noise Shaper (5th order) & "Filter Stream" D/A, amps), and finally, a headphone-driver "stage".


Possible data paths

HDD->Coldfire (decoded, perhaps complex DSP routines,...) -> Philips (Din, additional DSP, Interpolation Filter, Noise Shaper, FSDAC, Vout -or- HP Driver -> VoutHP)

Philips (Vin @12KOhm, PGA, ADC/ Decimation Filter/ DC-Cancellation Filter), DSP-optional, Dout) -> Coldfire (encoding) -> HDD.

Those are two possible paths. There are many others, considering the iHP-120 has Line-in, internal-Mic-In, External-mic-in, S/PDIF-in, Line-out and S/DIF-out.

There are also different analog (i/p signal) gain "stages" depending on i/p-type (eg. Low-Noise Amp w/29dB gain for mic), Vol Ctrl can be performed in the [ADC] Decimator "stage" (+24dB, -63.5dB) and again in the Interpolation stage (0, -78dB max suppression).

The 1-bit DAC ("Filter Stream DAC") is designed to provide low-clk jitter and has "on-board" amps to drive the subsequent stages (Vout, HP Driver). This stage also provides a form of filtering that eliminates the need of an external filter on the o/p side.

The HP Driver "stage" can o/p 35mW into 16 Ohms (given a 3.0V supply) and is designed to eliminate the need for external caps (DC-coupling).

Last but not least, the FSDAC stage can accept analog signals (which are "forwarded" to the next stages) and this is (most likely) how the iHP's FM signals are sent to the Vout & HP Driver.

Again, I am basing my assumptions on Philip's documentation: how iRiver uses the chip's functionality is still an unknown variable, but I think my assumptions are reasonable. I would be more than pleased if someone with an EE background could back up or refute the assumptions made so far
smily_headphones1.gif


But if nothing else comes of this amateur analysis, some of the more simple questions asked by myself and other posters here (such as what is the difference between Line-out & HP-out) have been answered...
biggrin.gif
 
Jan 29, 2004 at 10:13 AM Post #21 of 40
Btw, a few links describing the Philips' audio chip while Juergen is finshing up his site:

Philips Audio Codecs catalog

A not-so-informative news brief from Arrowasia.

I don't know how it compares to iPod G3's Wolfson Microelectronics WM8731L, discussed @IPoding (& btw, here's a "insder's" @Ipodding as well, in case your interested).

Anyone care to compare? I need a little break from this EE-business, lol.
 
Jan 29, 2004 at 6:13 PM Post #22 of 40
from one non-EE to another, you seem to be doing great! you're a little more on top of this stuff than i, so i can barely follow.

i was wondering, where does the S/PDIF signal out come from? is it one of the pins on the philips chip?
 
Jan 29, 2004 at 7:12 PM Post #23 of 40
Quote:

Originally posted by faustian bargain
...where does the S/PDIF signal out come from?


Good one! As I2S is an internal digital interface & S/PDIF external, there must be a "convertor" between the S/PDIF TX&RX terminals & the Philips. That convertor looks to be the Coldfire as it has both interfaces. I'm not certain what interface the data I/O stage [audio+clk] of Philips uses, but I'm assuming it's I2S ("I squared S"). The I2C interface seems to be for passing instructions & states between the two.

I'd guess that the Coldfire handles digital I/O comm between the outside world & the Philips. What it does with the data before passing it on is an entirely different question....

(& thx for the compliment - but I'm afraid you'll have to take everything I've written with caution
wink.gif
)
 
Jan 30, 2004 at 7:49 PM Post #24 of 40
Last but not least, a look @ my a la cocktail-napkin, EE-Impared audio-trace. I'd like to think of it as an educated guess, but take it with a grain... no, make that a bag of salt. Family size.

Note: The UDA1380TT can transmit data from either the Decimator or Interpolator stages (hence the two paths to the I2S bus). Data is only received by the DSP group.




It's been real & it's been fun, but...
cool.gif
 
Jan 31, 2004 at 2:18 AM Post #25 of 40
looks good to me, but what do i know. :/

my next level of interest is, i have read about some people's preference for tapping into the I2S signal, rather than the spdif. granted this is in the context of cd players, so i don't know if it makes a difference in this case. but i wonder if it would be possible to hack a connection to it in the iHP.

if i'm not mistaken Elso Kwak from the diyaudio forums has a DAC schematic that uses the I2S.
 
Feb 2, 2004 at 3:40 PM Post #26 of 40
Hi,

after an hard job I've relaunched the iHP@Lounge (www.ihplounge.com).

At that moment you will find some known information about the semiconductors.
Some more new stuff I've found in the web (ColdfireDesign, programming, components) I'm still put into graphs. Even the develompent suit for the processor.

http://www.ghs.com/products/68k_development.html

It could be the key for firmware-hacks?!

Among other things there's the external bat-pack. DIY-Guide will follow up :wink:

Dockingstation is still under construction but it goes on (-:

So can someone make a day more than 24h?!

cheers
Juergen
 
Feb 3, 2004 at 2:05 PM Post #27 of 40
Quote:

Originally posted by Buckaroo Banzai
after an hard job I've relaunched the iHP@Lounge (www.ihplounge.com).


Wow, Juergen, and an excellent job you did, too! *****
biggrin.gif


It's a simple structure, so I would think that even non-German speakers should get around w/o much effort!

I hope that some day the iRiver crowd will be just as big & badass as those ipod-junkies! One thng's for certain, Apple doesn't have a monopoly on portable HDD players any more...
very_evil_smiley.gif
 
Feb 3, 2004 at 9:27 PM Post #28 of 40
yeah nice site!!!! umm i have a few ?'s tho... what the hell would the iHP need 256mb of SDRAM for? Is is this usual for an mp3 player? sorry im kinda new to this sorta stuff but i'm tryin to learn
tongue.gif
... also are these the specs of a iHP-120 or iHP-100? because they have different amounts of buffer so probably have different hardware...
 
Feb 3, 2004 at 10:41 PM Post #29 of 40
Quote:

Originally posted by jpburton5150
yeah nice site!!!! umm i have a few ?'s tho... what the hell would the iHP need 256mb of SDRAM for? Is is this usual for an mp3 player? sorry im kinda new to this sorta stuff but i'm tryin to learn
tongue.gif
...


256 Megabit, not megabyte. Divide by 8.
smily_headphones1.gif


-Jeff
 

Users who are viewing this thread

Back
Top