New HQ DIY DAC

[Deleted]

Jan 13, 2010 15:08:38 GMT bpgoa said:

Jan 13, 2010 2:34:11 GMT @sandyk said:

Alfred
Do we really need the 100nF polyprop capacitors anyway ? Typical SPDIF input capacitor values range from 10nF to 100nF. The 68nF capacitors are already fairly big as can be seen in my photo, and the 5mm lead spacing is a joke. It is what they shaped the leads to in the factory. Jaycar sells packs of 6 ferrite beads for $2.95
Alex

ferrite beads in the uk.. here... www.modecomponents.co.uk/

...even at RS!
uk.rs-online.com/web/search/searchBrowseAction.html?method=retrieveTfg&Ne=4294922322&Ntt=ferrite+beads&Nr=AND%28avl%3auk%2csearchDiscon_uk%3aN%29&Ntk=I18NAll&Ntx=mode%2bmatchallpartial&N=4294956443+4294919435&Ns=stockPolicy_uk%7c1%7c%7cnew_uk%7c1&Nty=1&binCount=222&multiselectParam=4294956443&selectSubRange=Beads - PCB#breadCrumb

[Will]

Hi Alfred, nice to see you here too!

Thanks for chasing up RS over in Australia, it'll be good to know when others start building it, and the supply decoupling pdf is not something I've seen before. Having gone to town on the supply side, do you think the extra is worth?

Also, if you have any pictures of your build, we tend to be suckers for that sort of thing..

[Deleted]

Will, Chris and Adrian
The polyprops have now been posted to Will via Registered Post.
Alex

[bpgoa]

Jan 13, 2010 22:46:53 GMT @sandyk said:

Will, Chris and Adrian
The polyprops have now been posted to Will via Registered Post.
Alex

Thanks Alex, maybe we should return the favour and post you some of our snow.. i hear it's hot there at the mo..

[Deleted]

Jan 13, 2010 23:50:22 GMT bpgoa said:

Jan 13, 2010 22:46:53 GMT @sandyk said:

Will, Chris and Adrian
The polyprops have now been posted to Will via Registered Post.
Alex

Thanks Alex, maybe we should return the favour and post you some of our snow.. i hear it's hot there at the mo..

It's now a comfortable 19.5C and a max today of 24C.
A few more comfortable days ahead hopefully!
So far this summer, at least in Sydney, is not as hot as a few previous summers. I reckon that Pommy guy that came over here a couple of centuries ago in his sailing ship picked the best place in Oz for a settlement !

[Deleted]

Alex,

many thanks.

Would you care for some cooling rain as well ;D

Chris

[Deleted]

Would you care for some cooling rain as well

Chris
We already have a little of that since yesterday afternoon, and perhaps a little more for another couple of days.
Alex

[Deleted]

I am using the referance voltage from the chip I deduced that a zener "may" be too noisy, the 5v one come from the ripple eater with a LM317 set to 5v and the 3.3 is the LM3940T 3.3, I used 10uf tants for the filtering, referance and the output, plus a .1 across each chipAlfred

Alfred
I checked the data sheet for the LM3940 5V to 3.3V regulator,and typical noise voltage is 150uV.
The 3.3V Zener diode reference is likely to be considerably better, especially as it will be bypassed by a capacitor before the CR network.
Alex
P.S.
Looks like I need to get another HyQ 24M576 module to compare versions !

[bpgoa]

Jan 14, 2010 9:08:32 GMT @sandyk said:

I am using the referance voltage from the chip I deduced that a zener "may" be too noisy, the 5v one come from the ripple eater with a LM317 set to 5v and the 3.3 is the LM3940T 3.3, I used 10uf tants for the filtering, referance and the output, plus a .1 across each chipAlfred

Alfred
I checked the data sheet for the LM3940 5V to 3.3V regulator,and typical noise voltage is 150uV.
The 3.3V Zener diode reference is likely to be considerably better, especially as it will be bypassed by a capacitor before the CR network.
Alex
P.S.
Looks like I need to get another HyQ 24M576 module to compare versions !

Voltage regulation..

let's see if I've got my head around this...

the circuit you posted uses an op- amp to effectively feed the output 3.3v by using a zenner diode as the reference voltage but sourcing it's supply from the +15v rail. the output is then fed back into the op-amp in a feed back loop to further minimise any fluctuations...

is that correct? and what current will it be able to source? how do we choose the op-amp? are you recommending two of these, one set to 3.3v and one set to 5.1v?

I've never used op-amps before - is there any recommended reading or is this 6 months of reading?

one more anyone - is there a UK source for the clock units.... ?

(thanks for the mentoring )

[bpgoa]

part of me thinks that we're close to the point of building a new pcb to site all of the supply upgrades ..

if the designer of this dac is "lurking", do you have any opinions?

[Deleted]

Jan 14, 2010 14:00:08 GMT bpgoa said:

Jan 14, 2010 9:08:32 GMT @sandyk said:

Alfred
I checked the data sheet for the LM3940 5V to 3.3V regulator,and typical noise voltage is 150uV.
The 3.3V Zener diode reference is likely to be considerably better, especially as it will be bypassed by a capacitor before the CR network.
Alex
P.S.
Looks like I need to get another HyQ 24M576 module to compare versions !

Voltage regulation..

let's see if I've got my head around this...

the circuit you posted uses an op- amp to effectively feed the output 3.3v by using a zenner diode as the reference voltage but sourcing it's supply from the +15v rail. the output is then fed back into the op-amp in a feed back loop to further minimise any fluctuations...

is that correct? and what current will it be able to source? how do we choose the op-amp? are you recommending two of these, one set to 3.3v and one set to 5.1v?

I've never used op-amps before - is there any recommended reading or is this 6 months of reading?

one more anyone - is there a UK source for the clock units.... ?

(thanks for the mentoring )

Adrian
These are only possible scenarios for experienced constructors to explore and report back. It would be nice if we could still get the relatively expensive PFM Flea oscillators, but we can't.
Many opamps should be capable of supplying the 35mA max. needed for the Clock supply. However, the supplies may also be able to supply other individual 3.3V and 5V areas of the DAC.
These modules would be getting a clean +15V supply from a JLH, and then using the reference voltage,output a higher current at a lower impedance. I suggest that you get the basic unit working first and worry about these areas later.
Yes, there is quite a learning curve in this area, and I am far from as knowledgeable as Robert from Rockhampton in such matters.
Alex

[Deleted]

Jan 14, 2010 14:04:24 GMT bpgoa said:

part of me thinks that we're close to the point of building a new pcb to site all of the supply upgrades ..

if the designer of this dac is "lurking", do you have any opinions?

Adrian
Nicholas Vinen from Sydney is the designer of the DAC, in collaboration with Silicon Chip magazine. He doesn't even believe that the 24.576MHZ oscillator needs to be any more than a simple reference. I proposed such an oscillator module to him in the DIYAudio Forum and he thought it was a waste of time. However, far more experienced members in PFM Forum and other Forums such as the Sabre DAC thread in DIY Audio have vastly different views about the importance of very clean supplies to such areas. Ask Leo or Jkeny for example.
Alex

[alchemy]

Alex, I did look up the noise figures on the 3 pin regs and they all seem to be the same 150uv, my understanding of the circuit is that we are using the PSRR of the chip used to drive the more "clean" voltage, the referance may be incidental, but you are right we should explore it.
People may be interested in the following as a background and yes there are pictures,
www.lcaudio.com/index.php?page=4
The following drawing the latest clock module from our friend above illustrates the use of separate supplies for each digital chip, it seems the evolution of his clock module is pointing the direction that we have taken to relise the outcomes possible, he uses an elaborate referance chip, I guess the zener wasn't his choice. (see last DWG)
www.lcaudio.com/index.php?page=62
I working on some pictures, but I am not a fluent as Alex in confuser talk and I have a Mac so I can't ask Alex for much help, so patience for this muddleheaded Wombat.
For another opiion about the flea try this site from some of your bretherin
www.acoustica.org.uk/
Alfred

[Deleted]

Alex, I did look up the noise figures on the 3 pin regs and they all seem to be the same 150uv, my understanding of the circuit is that we are using the PSRR of the chip used to drive the more "clean" voltage, the referance may be incidental, but you are right we should explore it.

Alfred
All the higher PSRR will do, is ensure that the output signal will not be further degraded by PSU supply noise. You will have higher current capability, but any noise present from the reference will still be there at the output of the I.C.
In the case of the PFM Flea, this noise will be amplified by the gain used to step up the reference voltage to that required.
Yes , they have taken other measures to try and reduce some of it again. Incidentally, because VREGs have such a low output impedance,(35 milli ohms for the LM3940) output bypass electrolytics are pretty well ineffective at further reducing noise, until much higher frequencies where their output impedance rises markedly.
A 3.3V Zener has a much higher dynamic impedance,(around 85 ohms) so bypass capacitors are more effective.
Alex

[Will]

one more anyone - is there a UK source for the clock units.... ?

I've found one that meets the specification (voltage/frequency/drive) , but it's £20, half the cost of the kit!
I did find a couple more in keeping with the kit, but they aren't quite the same, and would be harder to implement (thanks for the advice, Alex)

I'll have a dredge through some of those shops you listed

such as the Sabre DAC thread in DIY Audio have vastly different views about the importance of very clean supplies to such areas

rockgrotto.proboards.com/index.cgi?board=talk&action=display&thread=3322&page=25#59195 Leo gives a nice, condensed version as to why the Sabre chip needs extremely good supplies to sound it's best. There is a link in the same thread earlier, where the designer explains why the chip needs good supplies in greater detail.

[Will]

I working on some pictures, but I am not a fluent as Alex in confuser talk and I have a Mac so I can't ask Alex for much help, so patience for this muddleheaded Wombat.

Alfred, Mike has made a good 'howto' for posting pictures, if I catch you right.

rockgrotto.proboards.com/index.cgi?board=general&action=display&thread=4876

[bpgoa]

Jan 14, 2010 20:27:43 GMT @sandyk said:

Jan 14, 2010 14:04:24 GMT bpgoa said:

part of me thinks that we're close to the point of building a new pcb to site all of the supply upgrades ..

if the designer of this dac is "lurking", do you have any opinions?

Adrian
Nicholas Vinen from Sydney is the designer of the DAC, in collaboration with Silicon Chip magazine. He doesn't even believe that the 24.576MHZ oscillator needs to be any more than a simple reference. I proposed such an oscillator module to him in the DIYAudio Forum and he thought it was a waste of time. However, far more experienced members in PFM Forum and other Forums such as the Sabre DAC thread in DIY Audio have vastly different views about the importance of very clean supplies to such areas. Ask Leo or Jkeny for example.
Alex

i must admit that from my simple digital background, upgrading the clock to the receiver chip would make far more sense if the said clock is then fed back to the transport as a world source( if the transport clock was a division). if you could do that then it would make a massive difference to jitter especially with regard to sync and cycle recovery.

my reading of this design is that the clock on the dac is only used for "start up" then the receiver chip syncs onto the clock overlaid on the spdif bus (with all of it's jitter errors)

after saying that... if it sounds better then it's worth doing.

it'd be interesting to put a modded DAC back on the original test bench to see how it compares with the original.

[Deleted]

Adrian
Not everything is immediately obvious from looking at a block diagram on a Data Sheet.
It is possible that no further gains will be made after just isolating the power supplies as with using the LM3940 VREG version of the outboard clock, but we are trying to find a common PCB design that can be used in other areas as well as the clock.

To quote from Ray's Audio page about the PFM flea.
members.quicknet.nl/ra.vdsteen/flea_en.html

On the left it is shown how the oscillator is often constructed in most digital audio-equipment. With only a few discrete parts and an inverter, the clock signal is generated. The inverter is configured as an amplifier by means of Rf, and it is usually integrated in one of the other IC's that are present in the equipment, like the DAC, the decoder or the digital filter. But this way, it obtains it's power from the same supply-pins, or through the same voltage regulator as the rest of the circuit. This causes high-frequency noise signals from other parts of the IC to have easy access to the oscillator through the common supply-lines, where they can cause jitter. The other way around, the harmonics generated by the oscillator can easily reach other parts of the IC, or other parts of the circuit. This situation is far from ideal. By fitting the oscillator externally and feeding it with a low-noise voltage regulator and a separate power supply-module, the cross-talk is greatly reduced and jitter is reduced.

[Will]

A parcel arrived from Australia this morning.



I've never had a jaycar kit before (my scha was made on the epemag pcb, when Jaycar ran out of kits), and I'm pretty impressed as to what you get, which is everything! Even a bit of solder. Just to reconfirm as well, the SMD bits do come soldered onto the PCB, and not as the Jaycar site states, that you have to solder them yourself.

Just need to make it now ;D

[Deleted]

Will
Have the other little bits arrived yet ?
Alex


P.S.
See if you can line up those 5mm switch panel LEDs a little better, as discussed.

[Will]

Not yet Alex, but I'm expecting them to be delayed a little, after what happened last times capacitors were sent to me from Australia, as they were delayed in customs (our end) for 15 days for 'security reasons'

Also meant to say the kit comes with good documentation and a little errata sheet. They have also fixed the regulator silk mistake with dalo pen!

[Deleted]

Will
Yes, the documentation is pretty good. I didn't need to refer to the scanned copy on my P.C.
It could be worthwhile pretinning around the holes for the Blue LEDs in the switches to aid solder flow when they are soldered while in situ.
Alex
P.S. Check that the little bit of track isn't missing on the DAC board's output I.C. Either pin 2 or 3 on one of them from the 180R. I can't remember which.I imagine they have corrected that by now, though.

[Spirit]

Jan 19, 2010 9:53:06 GMT Will said:

Not yet Alex, but I'm expecting them to be delayed a little, after what happened last times capacitors were sent to me from Australia, as they were delayed in customs (our end) for 15 days for 'security reasons'

It's because I hollowed one out and put a bomb in it...

Wait what? ;D



A 1k celebration is just about in order eh Will ;D

[Will]

It's because I hollowed one out and put a bomb in it...Wait what?

Careful, Phil, they are 'watching' you now!

Alex, the errata they sent describes what to check on the dac output and how to fix it. From the quick flick through of the instructions, and what can be provided here, anyone who made the Jaycar SCHA could make this.

[Deleted]

From the quick flick through of the instructions, and what can be provided here, anyone who made the Jaycar SCHA could make this.

Will
The only tricky bits are the mounting of the front panel switches and LEDs, and getting the Blue LEDs to work. The IDP cables can cause a few headaches too. You need to make sure that the cable goes in straight against the pin 1 side, as there is a small amount of play when the wires are pushed through.
Alex