is this DTS/AC3 dac worth modding?

[linuxworks]

here's a 'movie dac' (dts, dd5.1, spdif linear) that I got on ebay a few months ago.

some internal shots:

and what it looks like from the outside:

it was about $150. it seems ok but I'm curious what folks think.

the internal build seems like a disaster to me, to be honest. that trafo hums a bit (physically) and does get hot. the reg chip gets VERY hot and the case has holes on the top for that (not sure I even trust it).

they stacked that small green board on top of the orange board and I'm not a fan of stacking esp. when the lower board has regulators on it generating heat! ;(

the output chips (3 for 6 channels, hence 5.1) are ne5532. I assume those should be yanked out and upgraded right away? anything in mind? 8620 maybe? technically I only use this for 2 channel listening (it can downmap 5.1 to 2.1) and so I'd only have to unsolder that first main chip, I think (out of the 3 5532's).

there's a whole bunch of electrolytics on that green board. I wonder what those are doing and if any are in the signal path...

what I'm thinking of doing is removing this entirely from its box, putting in a decent torroid (or 2 since its dual +/-12v and +5 v) and laying it out horizontally in a 1U rack mount so that nothing sits on top of another board.

one more question: that fluorescent display is ANNOYING - it scrolls once every 30 seconds or so and I want it to stop. I know there's no firmware option to stop it so I'm thinking of some hack I can do to remove power from it after a certain amount of delay. ie, listen in on the IR receive line and if there is no activity seen for X seconds, turn off power to the display. when someone presses a button, IR will show activity and this should turn the power on to the fluor. display for that timeout again. ideas? or am I trying to re-work too much on this poor dac?

 

[error401]

Those caps look to be in the signal path to me. Possibly some of those ceramics as well, which is even worse! You can probably replace the opamps with pretty much whatever you fancy, though I doubt this design is worth AD8620's cost, the DAC's specs are not very good. There also looks to be an analog MUX after the opamps that's probably not great for sound quality either. This DAC has voltage output, you can take it directly from the DAC pins and cut the traces to add your own analog stage. Needs a coupling cap.

Interesting that it's got two microcontrollers, the one on the phenolic board is probably just for driving the VFD. If it's just the scrolling that bothers you, what might work instead of cutting power is just shorting the enable lines to 'disable' after your delay (or breaking the data lines), as long as this doesn't blow up the microcontroller.

Where does the DAC clock come from?! I guess generated by the CS8415?

 

[linuxworks]

I got the dac from the 'wsz' guy on ebay, who also sells the zero-dac units. not lawrence but another seller.

looking closer, it does look like there are 6 electros in the audio path. doh! why would anyone do that on a so-called 'high end' (I know its not, but...) dac?

the chips at the end do seem to be hcf4053 which is an analog mux.

what could they be using this for? for blocking the rear l/r and center if you choose '2.1' mode instead of 5.1 ? that's all I can think of, just a cheap mute of extra chan outs?

you say the dac chip itself is no great shakes? so its not even worth an analog upgrade or hardware change-out on the other things?

considering that I HAVE to figure some way to turn off that annoying display - maybe its not worth all this hassle. shame.

for movie watching, its quite ok. but I hate having to have a dac for movies and a dac for all my other 'good source' things.

this dac also has some annoyances. it does not remember where you left some of the control settings on power off. on power-on it does not resume and powers on with a line of dashes in the display, so a switched-ac hurts you in this way ;( it also switches on at full brightness and has no mem function to save the dimmer setting.

finally, it also gets confused on auto-mode selection. it supports that annoying old dolby analog format and sometimes tries to use that instead of a plain old analog input. I wish you could disable prologic (that's it - prologic, couldn't remember the name until just now)

so when you switch to analog-ins it can't quite tell whether you have prologic (you won't!) or regular 2 channel analog.

there was very little doc about this unit on the ebay ad and I could find only ONE online review (in french) about it. so I took a chance and ordered one. maybe I should just resell it since it has more drawbacks than bene's, for me.

 

[error401]

CS4360 is the DAC, the specs are nothing special but it's not terrible either. Passable, but definitely a budget DAC IC for cheap DVD players and the like. I think if you took the output straight from the DAC with quality coupling caps (and cut the traces to the onboard stage) it would be considerably better, but I wouldn't bother with a fancy output stage for it. The DAC has full line-level outputs.

You could also consider tapping the I2S lines - they seem to be available on vias - and adding an S/PDIF transmitter so you can just send the signal to your main DAC, if you only care about stereo.

 

[linuxworks]

Quote:

Originally Posted by error401 /img/forum/go_quote.gif CS4360 is the DAC, the specs are nothing special but it's not terrible either. Passable, but definitely a budget DAC IC for cheap DVD players and the like. I think if you took the output straight from the DAC with quality coupling caps (and cut the traces to the onboard stage) it would be considerably better, but I wouldn't bother with a fancy output stage for it. The DAC has full line-level outputs.

that sounds useful. if the output is line level (like, 2v level?) right off the chip I'd love to just tap into the L/R and be done with it. I can run them into a pair of my own rca's or steal 2 from the block of them in the back and that's all I really want. I never wanted the real surround outputs and had always planned to reduce n.m to 2.0

I'll take a look at the datasheet, I'm assuming it will be pretty obvious which pins I should tack some greenwires to?

Quote:

You could also consider tapping the I2S lines - they seem to be available on vias - and adding an S/PDIF transmitter so you can just send the signal to your main DAC, if you only care about stereo.

if I understand correctly, I could get just the 'stereo' L/R pair in i2s and send that to a dac that can deal with i2s inputs? essentially making this just a digital break-out box for n.m to 2.0?

is THAT what you are saying? if so, I LIKE!! I hope that's what you are saying

a lot of people would like to be able to hardware-decode DTS into spdif or something like that..

 

[error401]

Quote:

Originally Posted by linuxworks /img/forum/go_quote.gif I'll take a look at the datasheet, I'm assuming it will be pretty obvious which pins I should tack some greenwires to?

There's no 'L' and 'R' in the datasheet, just 3 stereo DACs with numbered outputs. It's pretty likely that A1 is left and B1 is right from the layout, but I'd verify that by following the traces. But yea, should be simple to figure out. It's specified as 0.6*VA peak-to-peak (VA is probably 5V), so slightly more than 2Vrms. It will be biased around VA/2 (2.5V) so you really need those coupling caps. Cut the traces going to the existing output stage.

Quote:

if I understand correctly, I could get just the 'stereo' L/R pair in i2s and send that to a dac that can deal with i2s inputs? essentially making this just a digital break-out box for n.m to 2.0? is THAT what you are saying? if so, I LIKE!! I hope that's what you are saying a lot of people would like to be able to hardware-decode DTS into spdif or something like that..

Yup, the digital data is certainly available (they've even put convenient vias on all the lines to make it even easier), though it might not be 'standard' I2S. The DIF0 and DIF1 lines go to vias so I can't guess what format is set, but if you send it to a DAC you build (like a TPA Opus or something) you should be able to match the correct format. It's unlikely you'll be able to change the format the decoder chip is sending though, so you might need some control over the DAC or S/PDIF transmitter to make it match.

Indeed, a useful widget, you could generate 3x SPDIF with it or just put it in a box with a better DAC.

 

[linuxworks]

it sounds like you are saying the i2s isn't fully 'plug and play' as spdif is?

ideally, I'd be 'ok' with just 3 spdif streams. I've been asking the popcorn hour guys (the device) to consider having 3 toslinks on the back so that we have have 3 dacs for 5.1 and really have some fun

they don't listen to me though...

if a box like this could break-out into 3 spdifs that any old dac (3 of them) can just read, that would be UBER cool!

is there any other device that natively does this? its a shame to hack to much if there was a native 5.1 decoder that didn't cost an arm and a leg and broke out the streams to things we all could more easily deal with in a DIY sense.

personally, I like being able to 'discard' the rear and center channels and listen to 2.0 of the 5.1. sometimes you need to map the center to a phantom and sometimes not. I'd like to have total control over all that and it seems DIY would be one way (maybe the only way) to have that control.

 

[error401]

I2S is a standard published by Philips, but it's not the only way to transfer audio data serially in an inter-ic context. Most other formats are essentially the same, but there are some timing differences (like LRCLK is inverted or data is read at opposite edges of BCLK and such) that make them incompatible. Since you can't control the decoder, you need to conform to what it's doing.

It should be possible to connect a set of S/PDIF transmitters to the serial data lines and send the PCM data off to an external DAC.

I have a PCH too, I set out to design a software decoder board using an AVR32 board, but the processor was too slow to decode DTS or AC3 so I kinda gave up. The reason you won't find outputs like this is because the license the hardware vendor pays to be able to decode Dolby Digital or DTS doesn't allow it. You can't buy the decoder chips (with firmware) either without a license, which is really lame. The PCH can't even decode DTS at all, hence the motivation for my failed project.

 

[linuxworks]

I have not followed the PCH box devel a lot recently but isnt' the a110 supposed to *finally* have license to do DTS in software, at least mixdown to spdif 2ch?

that's all I'd care about. or even a translation from DTS to DD5.1. most things are in dual 'native' dts or dd (that I watch) and so I usually pick dd since dts is 'a PITA' (heh). they can remap all digital formats down to 2.0 on spdif EXCEPT dts (grrr!). but I thought the a110 new model had dts in software for mixdown (at least).

if we can 'steal' the 3 2-ch streams, that would be awesome. do you think this is really do-able? I don't mind franken-modding mine but only if you think it really can work out with some kind of i2s peer.

 

[error401]

Quote:

Originally Posted by linuxworks /img/forum/go_quote.gif if we can 'steal' the 3 2-ch streams, that would be awesome. do you think this is really do-able? I don't mind franken-modding mine but only if you think it really can work out with some kind of i2s peer.

It will work. The CS4360 is a bog-standard DAC with standard serial data input. Chances are that it's probably configured for either 24-bit left justified or 24-bit I2S. Grab a DIT4192/CS8406/WM8804/whatever, hook it up per the datasheet and tap the signals going to the DAC, it doesn't get much simpler. You might want to add a 'proper' transformer coupled 75R output (see this thread). Put some jumpers on the digital format pins and fiddle with them until it works, or get out your meter and measure to find out what mode it's currently configured for.

No guarantees with any of these kind of mods, but this should be pretty fool proof.

 

[linuxworks]

I might just try to probe around and feed signals as you describe. see if I get lucky

...hey wait, the white smoke: is that an op-amp going into class A or a new pope being voted in?

I'll try some stuff on this and let you know what happens.

thanks for the info!

 

[error401]

Oh, and I'm pretty certain that though the PCH A110 has DTS decoding, it will be to the analog outs (or pass through) only. Exposing the decoded digital data is explicitly against the license, as far as I understand. Hence why this is so damn difficult.

 

[linuxworks]

I didn't realize the DTS org was that 'fussy' about its members' bits!

if that's the case then maybe its time I boycott DTS. I don't really need it and I'm perfectly fine with DD5.1 for my multichannel format. I never really saw the huge advantage of DTS and both sound 'fine' to me for movie use.

at least the PCH converts 5.1 DD to 2.0 spdif that normal dacs can handle. I'm able to enjoy nice dacs AND watch movies and I really don't miss all that much.

what I would like is better mix control over the 5.1 streams so that I can 'patch panel' input to output channels. sometimes there is just a center channel with audio and NO others - so I have to map that to L/R mono. most of the time its straight L to L and R to R (front) and drop the rears but sometimes I also want to mix the rears.

is there anything that accepts ac3 and lets me mix the result to a 2.0 final? all with spdif in and out?

 

[error401]

Quote:

Originally Posted by linuxworks /img/forum/go_quote.gif is there anything that accepts ac3 and lets me mix the result to a 2.0 final? all with spdif in and out?

A PC

 

[linuxworks]

actually, I might check out the new atom boards. it might be low power enough to be fanless and fast enough to do the job. that might work.. a cdrom or flash based unix box that is sort of purpose-built.