| janneman |
| quote: | Originally posted by FLECOM
according to the datasheet the 96 dosent need any dc blocking on the output at all... |
Where in the data sheet does it say that? Can you show that part? I didn't find it. AFAICS, the outputs sit on 2.5VDC. You can't make that disappear.... Unless you do a balanced to SE conversion as noted before.
Jan Didden |
|
|
| zanash |
| stupid question, how do I open eso.rar ? |
|
|
| ds23man |
It does not work!
GRG |
|
|
| Ken L |
| quote: | Originally posted by zanash
stupid question, how do I open eso.rar ? |
use winrar -
google to find |
|
|
| Ryssen |
| quote: | The right Power connector pinout:
1 Red = +15V
2 = -15V
3 = 9V
4 = analog ground
5 = digital ground
6 = 5V
7 = 3V
|
| quote: | | I believe the 9V goes to the regulator LM317 that powers the dac and other chips, so this supply is very important. But somebody please double check. |
Is this true?I am planing on making a better supply for the DAC,is it the 9v I shoul supply NOT the 5V?How much is the current draw?
Anyone done this? |
|
|
| Ryssen |
| quote: | | I believe the 9V goes to the regulator LM317 that powers the dac and other chips, so this supply is very important. But somebody please double check. |
Sommeone figured this out yet??I really would like to know wich of the regulators goes to the DAC??:cannotbe: |
|
|
| CLS |
Yes, I know this is an old thread...
Anyway, inspired by this & this , I recall the wonderful DIY DAC I used several years ago, which came with transformer line stage.
That DAC was an evaluation board of CS4397 & I skipped the whole analog stage on the board with a pair of Lundahl 7902. That was inspired by this.
Tranformer-coupled DAC is awsome, period.
By comparison, the OP analog stages sound mechanical & grainy. On the other hand, transformer-couple sound lifelike & fluid. In good recordings, the sparkle & airy decay of the cymbals are simply amazing. In those records without noticeable trebles, the "atmosphere" are also remarkable....
Sadly the evaluation board died after only about 1year of use, and I could not fix it. (tried but failed)
Later I bought DEQ2496. Using it with digital input & analog output, it serves both EQ & DAC functions. In the EQ portion, it's doing great. In the DAC portion, it's not.
Looking into the boards inside the DEQ2496, those tiny SMD scared me away everytime. (you guys who can deal with them just amaze me! ) So, despite the imperfection of the analog stages, eventually I get used to its sound.
Until now.
I think, I made a tranformer-coupled analog stage before, why not now?
I traced the analog output pins of AK4393, I found they connet to R13, 14, 16, 17 (2k in value) on the back of output board. R13/R14 are L-/L+; R16/R17 are R+/R-.
I just attach thin wires on them & connet to pin2/3 of the output connectors. (Of course those connector are disconnected from the board) I need the AUX output for other usage so I leave everything else untouched. Now the DA chip is driving both downstream coupling transformer & the analog stage of AUX output. Maybe or maybe not, I'm not sure. With extra load of 10k, I'm not worried.
Anyway, all those tiny SMD and copper foil traces are so complex to me and I'm not willing to destroy any of them. So I'd better leave it that way...
Now the output of the XLR connetors is differential with 2.5V DC offset. Luckily those wonderful Lundahl 7902 are now sitting at the input of my preamp with floating primary winding. So the DC offset dose not matter.
The fluid, airy, open sound came back immediately. Oh, how I miss that. So, with only 4 short & thin wire, the sound of this box transforms a great deal.
The drawback is, the output level drops a lot. I must turn up a lot more on the knob. And the punch, dynamic, energy of the bass seem weaken a bit. This dosen't bother me very much, since the textures & overtones are now much much better which overshine. |
|
|
| ampsarus |
| quote: | Originally posted by seoman
I seccond that!!!!!!
Remove the glue dot off the connector en wiglle it.
That wil solve ít.
I had the DEQ for 3 days and it started switching on-off.
When removing the power supply, in search for bad kontakts, the problem was fixed.
I never have had a power failure again.
But I also agree on the bad PCB layout.
I only took a closer look on the analog part and i think that can be done much better.
Regards Simon |
I am having the same trouble with my DEQ after 1 day of use. Was it the DC or the AC connection(s) that was the culprit? Thanks! |
|
|
| ampsarus |
| Fixed it with tape on the bottom of the chassis, same as in one of my DCX's. Tape can work wonders in shorty Behringer gear. |
|
|
| stef1777 |
I got a new DEQ2496 v2.4. it seems that the DSP board have been heavily modified. A first look show that crystals have been moved to another locations, LM317 to produce the digital 2.5v removed (and replaced by a dedicated regulator), new uController, some capacitors better placed close to the chips, etc...
No yet checked analog board.
. |
|
|
| ampsarus |
| I'd like to "modify" my DEQ so that it boots and works when I turn it on instead of seeing the flashing display and hearing relays doing an endless clicking cricket imitation. *sigh* |
|
|
| stef1777 |
| quote: | Originally posted by ampsarus
I'd like to "modify" my DEQ so that it boots and works when I turn it on instead of seeing the flashing display and hearing relays doing an endless clicking cricket imitation. *sigh* |
First check each voltages. Relays are powered with 5v. |
|
|
| ampsarus |
| quote: | Originally posted by stef1777
First check each voltages. Relays are powered with 5v. |
TY Stef! ;) |
|
|
| stef1777 |
I added your link on my web site.
Below is a summarize of the changes found in the new DEQ v2.
On the DSP board:
- new specialized uController: an Analog Device ADSP-BF531 remplace the Triscend TAS04
- new SDRAM memory chips
- LM317 removed and replaced by a dedicated circuit close to the BF531. Seems to be SOT223 reg but not checked yet in detail. Ref printed on reg google nothing.
- crystals moved to a new location, better placement (shorter traces)
- complete new traces and ground plane (a lot better than old version)
- large smd coils on the 9v input
On the input/output analog board: nothing except RoHS. I hate new solders. Need to be clean-up a lot to solder something else.
On the switched mode psu: shielded box around the transformer
No more thermal adhesive on the connectors (good !!!)
AC input connector inverted.
More stickers and signs on the TOP (a real marketing tool for Behringer !!!).
Sound same apparently. |
|
|
| Ryssen |
| quote: | | LM317 removed and replaced by a dedicated circuit close to the BF531. Seems to be SOT223 reg but not checked yet in detail. Ref printed on reg google nothing. |
Would this work with a LT1763 2,5v?Different pinout ofcourse. |
|
|
| stef1777 |
| quote: | Originally posted by Ryssen
Would this work with a LT1763 2,5v?Different pinout ofcourse. |
I don't know. The DEQ v2 was not mine. Just one to upgrade and it left.
In fact, I don't realy know what is this component. I took a picture.
It can be read: F D65AB and 11120 05/06.
Seems to be from Fairchild.
Concerning the 3.3v, I also discovered that inside the DEQ v1, Behringer mixed analog and digital 3.3v. They produce analog 3.3v with the same 3.3v and just insert a coil just before the chip pin... As in the DCX in fact.
Have a nice digital low noise psu is important in fact.
. |
|
|
| Ryssen |
| Ok,but in DEQv1 (the one I have)maybe the LM317 could be substituted with a LT1763 2,5v on one of your miniregPCB`s? |
|
|
| stef1777 |
| quote: | Originally posted by Ryssen
Ok,but in DEQv1 (the one I have)maybe the LM317 could be substituted with a LT1763 2,5v on one of your miniregPCB`s? |
We need to look what is powered first with 2.5v. In theory, you can remove the 2 resistors and fit a fixed reg in place. But good luke with the legs. ;)
Let me finish do draw the dsp board psu. I'll post it tomorow.
Stef... |
|
|
| janneman |
Stephane,
Good job! Isn't this the same supply that is used in the DCX2496?
Jan Didden |
|
|
| stef1777 |
| quote: | Originally posted by janneman
Stephane,
Good job! Isn't this the same supply that is used in the DCX2496?
Jan Didden |
No exactly the same. If I understand the DCX schematics, they produce the +3.3vA with the +5vD using a transistor. On the DEQv1, they directly use the 3.3vD with a coil to ovoid high frequency go back.
The big difference between the DCXv1 and the DEQv1 is ground plane. Only one mixed for the DEQ (AGND and DGND pins from AKM chips directly connected to the same ground). One analog and one digital for the DCX. DEQv2 DSP seems completely redesigned on this.
This may be explain why with the best possible psu, we still have a low level ground noise.
. |
|
|
| zanash |
stef1777
nice web site ! |
|
|
| AR2 |
| quote: | Originally posted by stef1777
No exactly the same. If I understand the DCX schematics, they produce the +3.3vA with the +5vD using a transistor. On the DEQv1, they directly use the 3.3vD with a coil to ovoid high frequency go back.
The big difference between the DCXv1 and the DEQv1 is ground plane. Only one mixed for the DEQ (AGND and DGND pins from AKM chips directly connected to the same ground). One analog and one digital for the DCX. DEQv2 DSP seems completely redesigned on this.
This may be explain why with the best possible psu, we still have a low level ground noise.
. |
Stef, with that knowledge wouldn't be the best to redesign your board and include all those voltages, including 9V regulated, or make separate board that will solve these issues. After that we just have to look where to cut the traces and completely abandon Behringer PS mess. It seems like that there is no other option. That way we could have one universal power supply for DCX and DEQ, that could incorporate your nice regulators. I have schematics for DCX that shows DSP power supply regulators (not switching supply) and will be happy to send them to you if you need. |
|
|
| zanash |
| did you consider discrete psu for each voltage a la audio synthesis ? |
|
|
| badduh |
That looks very interesting, i was looking for information to upgrade the opamps and caps, but i like to try this mod first. unfortunately I'm a newbie to dacs so i can use a little help :)
I can't see on the photo how the caps are wired up, could someone please explain that to me?
Photo |
|
|
| stef1777 |
| quote: | Originally posted by badduh
That looks very interesting, i was looking for information to upgrade the opamps and caps, but i like to try this mod first. unfortunately I'm a newbie to dacs so i can use a little help :)
I can't see on the photo how the caps are wired up, could someone please explain that to me?
Photo |
You'll find the pinout on this page: http://www.awdiy.com/index.php?page=references
If you wait a little, I'm working on a board to output the ADC and DAC signals without the need to cut something on the flat cable.
. |
|
|
| badduh |
Thank you very much.
i read on your site that the caps are .47uf each. so am i correct that you put 5 caps parallel with each other so you get a total amount of 5x0.47=2.35uf, and you put this stack between the new wires from aoutl+ and aoutr+? to pin 2 of the xlr chassis or in my case a new rca chassis.
Wat is the advantage of using a stack of caps instead of lets say one with a value of 2.2 uf?
Unfortunately here in Holland i can't get the Dayton caps, what i can get is the following:
-AUDYN-CAP plus foil capacitor 2.20uF 800V 25 x 43mm, costing $12 each, a would like toe use one a side in stead of 5x0.47 as that costs 5x$7,5= $37,5 a side, this should be a very good sounding cap.
-AUDYN-CAP KP-QS foil capacitor 0.47uF 400V 11 x 19mm, costing $1.90 each so it would cost me 5x$1.9=$9,50 a side, what seems reasonable toe me, of course a could also use one of 2,2 uf costing $3,00 each, this is a normal sounding cap, not bad but nothing special either.
-Black gate 50 volt Nx-Hi-Q 0,47 uf, costing 3 euro each so 5x$3=$15,00 a side, what is the maximum a would like to spend. the 2,2 uf is not available.
Which cap would you recommend? |
|
|
| badduh |
An other question comes on my mind.
Do i really have to cut the output wires from the dac to the output stage so that the original output stage doesnt have any influence on the direct output stage, or can i just leave everything as it is and put the wires to the coupling caps on the back of the x1 connector? |
|
|
| Ryssen |
| How is the AUX (teleplug) output connected internaly? |
|
|
| Ryssen |
| The AUX out sounds more open (better)than XLR out,at least on my rig,anyone tried the AUX out? |
|
|
| ChuckT |
I wish there is a more clear thread detailing each mod.
I just did a simple mod and find it very effective in improving bass (no, I don't mean pushing up eq of the lower fz)
If you study the ak4393 dac datasheet, look at the analog 5V decoupling for the Vref and Avdd pins, it requires separate 0.1uf and 10uf for each of those pins. Guess what, the deq only has only one 0.1uf for both pins, those pins are starving for quality power.
Obviously the best solution is to give it a separate clean PS, but I lack time and too lazy. I did the next next best thing, I put a Oscon SEP 470uf directly on the Vref & Avdd pins.
Nice improvement on the bass, but I am using direct transformer out, so YMMV.
Anyways, the deq is poorly decoupled and adding good caps in other place will likely improve things too.
Check the pic. |
|
|
| ChuckT |
| quote: | Originally posted by badduh
Thank you very much.
i read on your site that the caps are .47uf each. so am i correct that you put 5 caps parallel with each other so you get a total amount of 5x0.47=2.35uf, and you put this stack between the new wires from aoutl+ and aoutr+? to pin 2 of the xlr chassis or in my case a new rca chassis.
Wat is the advantage of using a stack of caps instead of lets say one with a value of 2.2 uf?
Unfortunately here in Holland i can't get the Dayton caps, what i can get is the following:
-AUDYN-CAP plus foil capacitor 2.20uF 800V 25 x 43mm, costing $12 each, a would like toe use one a side in stead of 5x0.47 as that costs 5x$7,5= $37,5 a side, this should be a very good sounding cap.
-AUDYN-CAP KP-QS foil capacitor 0.47uF 400V 11 x 19mm, costing $1.90 each so it would cost me 5x$1.9=$9,50 a side, what seems reasonable toe me, of course a could also use one of 2,2 uf costing $3,00 each, this is a normal sounding cap, not bad but nothing special either.
-Black gate 50 volt Nx-Hi-Q 0,47 uf, costing 3 euro each so 5x$3=$15,00 a side, what is the maximum a would like to spend. the 2,2 uf is not available.
Which cap would you recommend? |
why not just use one single 4.7uf cap?
I recommend the Vishay/ERO MKT1822, they are one of the best caps I have tried and cheap too. You can get it from Farnell. |
|
|
| coloradosound |
| Anybody have impression of the ESS sabre reference DAC? |
|
|
| Ryssen |
I changed opamps on the AUX out from JCR4580 to LM4562 yesterday.
I think it was a downgrade:bigeyes: smaler sound stage life less etc.
So now I am locking for 3 JCR4580 to change back to.:rolleyes:
please,they seems to be hard to find.
Thinking on the shipping etc. |
|
|
| Ryssen |
| quote: | | they seems to be hard to find. |
Texas instruments have them as samples.:) |
|
|
|
