Comments on my cs8416 pcm1730 dac project

cviller · 2006-09-10 4:54 pm

Finally working

I just finished soldering and got the dac working. I have made a D1 I/V stage based on promitheus schematics.
I think there is something wrong with my psu, but in between the scratch noise I can hear superb vocals and a very powerfull bass.

Problem:
High transients sounds like the scratch noise from a turntable. If the cd is recorded in a high volumen, the noise is more constant. Keith Jarrett really knows how to discharge my caps! I really hope someone on the forum knows how I can solve this problem.

I have based my psu on this:
Lars Nielsen psu

I have also tried a atx psu with a quite good result, but only after I added a 100uF cap to the three voltages: 3.3V spdif, 3.3V digital and 5.0V output driver. The sound was however not as good - bass was less powerful and after 5-10 minutes of listening, some static noise appeared.

I don't think the problem is related to my I/V stage which runs on +/- 30V and a powerful supply.

With the original psu, the voltages was around 3.2V and 4.7V - I suspect the 5V supply to be the problem, but I just tested another supply design using lm317 and some diodes which gave me 5.2V - and less scratch noise, but it didn't disappear. The voltage was not very constant - ranging from 5.1 to 5.3 while playing.

I have attached my schematics for the digital circuit. I think I have some ground loops, but I'm not completely sure on how to get rid of them. I have four boards two psu boards, one with the digital circuit and one I/V stage. I'm running unbalanced.

ezkcdude · 2006-09-10 5:05 pm

Maybe your DAC is just very revealing? I'd love to see some pics, when you get the chance.

cviller · 2006-09-10 5:33 pm

Here is a picture of the digital board.

cviller · 2006-09-10 5:35 pm

And here is the setup. (Before I added caps to my board)

singa · 2006-09-10 7:20 pm

cviller wrote:
Problem:
High transients sounds like the scratch noise from a turntable. If the cd is recorded in a high volumen, the noise is more constant. Keith Jarrett really knows how to discharge my caps! I really hope someone on the forum knows how I can solve this problem.

cviller wrote:
I have attached my schematics for the digital circuit. I think I have some ground loops, but I'm not completely sure on how to get rid of them.

I think there is a error in the schematic of PCM1730 datasheet and you copied it wholesale.Look at vcom3 /c101@ 10uF and compare vcom1,both are internal bias decoupling pins.(vcom2 is I/V ref bias).In this case vcom3 is connected to Vcc 5V and polarity of c101 is wrong when it should be like vcom1.Technically noise is being pumped back in dac.Vcom1 seem more correct.

ezkcdude · 2006-09-10 7:46 pm

One thing I noticed on a related note is that you have VCOM3 bypassed before the connection to VCC1 (at least, in your schematic). In the data sheet, VCOM3 is connected through the cap (C101 in your schem), and then bypassed. I'm not sure if this makes an audible difference, but I've been given the same advice before on this forum.

cviller · 2006-09-10 7:54 pm

singa said:
I think there is a error in the schematic of PCM1730 datasheet and you copied it wholesale.Look at vcom3 /c101@ 10uF and compare vcom1,both are internal bias decoupling pins.(vcom2 is I/V ref bias).In this case vcom3 is connected to Vcc 5V and polarity of c101 is wrong when it should be like vcom1.Technically noise is being pumped back in dac.Vcom1 seem more correct.

Thank you for your comments. So you say that vcom3 should be connected to ground through the capacitor?

ezkcdude said:
One thing I noticed on a related note is that you have VCOM3 bypassed before the connection to VCC1 (at least, in your schematic). In the data sheet, VCOM3 is connected through the cap (C101 in your schem), and then bypassed. I'm not sure if this makes an audible difference, but I've been given the same advice before on this forum.

What do you mean about being bypassed before connected to vcc1? Are you talking layout or schematics?

ezkcdude · 2006-09-10 8:04 pm

What do you mean about being bypassed before connected to vcc1? Are you talking layout or schematics? [/B]

In your schematics. Compare to page 13 of the data sheet. I don't know what your PCB layout looks like. Can you post a pic?

singa · 2006-09-10 8:09 pm

cviller wrote:
Thank you for your comments. So you say that vcom3 should be connected to ground through the capacitor?

Follow same like Vcom1 with cap polarity and see if noise goes away.I think it will.Technically c103/105 or c101/105 become non/bi polar in function. :bigeyes:

so noise can easily go back into dac.

edit-- connection of Vcc 5V came as datasheet.(your schematic)

cviller · 2006-09-10 8:12 pm

I tried to change c101 according to singa's tip, but nothing really changed - the scratching is still there.

singa · 2006-09-10 8:20 pm

cviller wrote:

I tried to change c101 according to singa's tip, but nothing really changed - the scratching is still there.

Did you also cut the Vcc (separate Vcom3 like Vcom1?)

cviller · 2006-09-10 10:10 pm

singa said:
Did you also cut the Vcc (separate Vcom3 like Vcom1?)

I removed the cap 101, which was connected from 5V to vcom3 and soldered it between gnd and vcom3, so that vcom3 is only connected through a capacitor to gnd. (+ to vcom3 an - to gnd)

Does clipping sound like a scratchy noise? The noise gets better after a while, but sometimes it just reappears in the middle of a nice listening session...

singa · 2006-09-11 5:24 am

quote:
Originally posted by singa

Did you also cut the Vcc (separate Vcom3 like Vcom1?)

I removed the cap 101, which was connected from 5V to vcom3 and soldered it between gnd and vcom3, so that vcom3 is only connected through a capacitor to gnd. (+ to vcom3 an - to gnd)

Does clipping sound like a scratchy noise? The noise gets better after a while, but sometimes it just reappears in the middle of a nice listening session...

Hi cviller,
I checked Lars Nielsen's PCM1730/CS8414 dac schematic and pcb layout, he made one "minor" change to 1730 datasheet of Vcom3 that is the capacitor 10uf is not grounded ? Since his is a working circuit may be you should copy/follow?Observe cap polarity to Vcc 5V.Please let us know of the result.

edit-- Lars Nielsen's doc. http://www.diyhifi.dk/DA A1.pdf

cviller · 2006-09-11 7:21 am

singa said:
I checked Lars Nielsen's PCM1730/CS8414 dac schematic and pcb layout, he made one "minor" change to 1730 datasheet of Vcom3 that is the capacitor 10uf is not grounded ? Since his is a working circuit may be you should copy/follow?Observe cap polarity to Vcc 5V.Please let us know of the result.

I can't see the difference between his schematic and the datasheet.
I think I have to check the soldering again - so many smds! It is just strange that it almost works.

singa · 2006-09-11 8:49 am

cviller wrote:
I can't see the difference between his schematic and the datasheet.
I think I have to check the soldering again - so many smds! It is just strange that it almost works.

I apologize if I have confused you and myself.Both of your schematics are identical to the datasheet except Lars Nielsen is of fig.6 page 12 (mcu controlled) and your's fig.7 page 13.Please reinstate back to your schematic (pg.13) but remove c105=10uf to look like Nielsen's circuit of pg.12 since his' is known to work.Hope this time is clear and let us know if your dac work without the noise.

Nielsen's doc-- http://www.diyhifi.dk/DA_A1.pdf

cviller · 2006-09-11 9:17 am

singa said:
I apologize if I have confused you and myself.Both of your schematics are identical to the datasheet except Lars Nielsen is of fig.6 page 12 (mcu controlled) and your's fig.7 page 13.Please reinstate back to your schematic (pg.13) but remove c105=10uf to look like Nielsen's circuit of pg.12 since his' is known to work.Hope this time is clear and let us know if your dac work without the noise.

I don't think removing c105 will change anything - it is just a decoupling cap for vcc.

singa · 2006-09-11 9:36 am

cviller wrote:
I don't think removing c105 will change anything - it is just a decoupling cap for vcc.

In that case,how about the signal cables (input/output).It could also be your cd player has a tracking problem and if not the it's osciloscope time.

edit-- Try other cds to hear if it's the same with other discs.

cviller · 2006-09-11 1:31 pm

singa said:
In that case,how about the signal cables (input/output).It could also be your cd player has a tracking problem and if not the it's osciloscope time.

edit-- Try other cds to hear if it's the same with other discs.

There is a difference between the cds, cos those that are not recorded very loud sounds better.
Do you think it can be something wrong on the digital side of dac chip or is it just an analog problem?
I don't have a scope at home, but I guess I have to use one to see what is going on with the supply voltages and if it is just clipping.

singa · 2006-09-11 2:39 pm

originally posted by cviller
Do you think it can be something wrong on the digital side of dac chip or is it just an analog problem?

If your schematic and datasheet are correct,I don't think it's the dac. Analog side maybe? That is why a scope is needed to confirm.

Comments on my cs8416 pcm1730 dac project

Member

Attachments

Member

Member

Attachments

Member

Attachments

Member

Member

Member

Member

Member

Member

Member

Member

Member

Member

Member

Member

Member

Member

Member