Author Topic: Audio DAC output Low Pass Filter.  (Read 23193 times)

0 Members and 1 Guest are viewing this topic.

Offline paulcaTopic starter

  • Super Contributor
  • ***
  • Posts: 4003
  • Country: gb
Re: Audio DAC output Low Pass Filter.
« Reply #50 on: March 18, 2018, 11:08:03 am »
Make sure your IC smd footprints are DIY assembly friendly.  They don't look it.
It may be just hard to tell from the photo...

Good point.  I've got fairly good with SMD soldering, but longer pads do make things easier. 

I wish KiCad's auto router could handle dragging components with the traces still attached.  It can become a pain in the neck moving a component after it's fully laid out.
"What could possibly go wrong?"
Current Open Projects:  STM32F411RE+ESP32+TFT for home IoT (NoT) projects.  Child's advent xmas countdown toy.  Digital audio routing board.
 

Offline BrianHG

  • Super Contributor
  • ***
  • Posts: 7661
  • Country: ca
Re: Audio DAC output Low Pass Filter.
« Reply #51 on: March 18, 2018, 10:18:05 pm »
Make sure your IC smd footprints are DIY assembly friendly.  They don't look it.
It may be just hard to tell from the photo...

Good point.  I've got fairly good with SMD soldering, but longer pads do make things easier. 

I wish KiCad's auto router could handle dragging components with the traces still attached.  It can become a pain in the neck moving a component after it's fully laid out.

Don't just make them blindly longer.  For your SSOP-28 pin, download this datasheet : http://www.microchip.com/mymicrochip/filehandler.aspx?ddocname=en552327
Use the footprint recommended layout on page 341.  Switch the rectangular pads to rounded pads and increase the recommended width of each pad by 1x the pad's height.  (Recovers loss due to pad rounding + gives you a tad more.)  Also, I still like making pad 1 a rectangle for additional visual ID in case silkscreen is damaged.

If your other IC is an SO-8 (I'm assuming it's an SO-8), download this data sheet: http://www.ti.com/lit/ds/symlink/msp430g2230.pdf
Use the footprint recommended layout on page 45.  Switch the...............
 
« Last Edit: March 19, 2018, 12:38:31 am by BrianHG »
 

Offline paulcaTopic starter

  • Super Contributor
  • ***
  • Posts: 4003
  • Country: gb
Re: Audio DAC output Low Pass Filter.
« Reply #52 on: March 19, 2018, 07:51:49 am »
Make sure your IC smd footprints are DIY assembly friendly.  They don't look it.
It may be just hard to tell from the photo...

Good point.  I've got fairly good with SMD soldering, but longer pads do make things easier. 

I wish KiCad's auto router could handle dragging components with the traces still attached.  It can become a pain in the neck moving a component after it's fully laid out.

Don't just make them blindly longer. 

I made them 0.5x1.75 and rounded.  They look slightly easier.

Anyway I went looking for enclosures and the one I like that stands a chance of fitting is this one:
http://www.hammondmfg.com/pdf/1551K.pdf

But it's 34mm x 74mm max PCB size which presents a challenge.  The board as screenshotted above is 47x77, but there is unused space.  Without the amps/filter it would be easy!  The little chinese PCM270x I have here is on a board that must be 25mm x 40mm. 

The reason I like these enclosures is they come with PCB screw mounting holes in the base, so no fannying around with hot glue or drilling holes in the back and putting spaces in, which would mean buying hardware.

I was trying Dave's idea of making small blocks off board and dragging them on.  Regulator, 2x Amps, 1x DAC + connectors.  However the trouble with this approach is that when the small blocks are dragged onto the board they don't make efficient use of space.  Also, once dragged onto the board they are a pain to select and move again.

Anyway I can move the DAC block down to the lower left corner more, rearrange the amp blocks to make them more space efficient.

I'm also considering moving all connectors to one side, which will make getting it into the case easier and less critical.

Can't finish the layout until the parts arrive and I get to check the footprints are correct.  So expect a few redos before I'm happy.  In the meantime I will double check the schematic against both the PCM2904 datasheet and the 2906 demo board.  The later is a little more confusing as it's explicitly shows GND and power connections everywhere.
"What could possibly go wrong?"
Current Open Projects:  STM32F411RE+ESP32+TFT for home IoT (NoT) projects.  Child's advent xmas countdown toy.  Digital audio routing board.
 

Offline BrianHG

  • Super Contributor
  • ***
  • Posts: 7661
  • Country: ca
Re: Audio DAC output Low Pass Filter.
« Reply #53 on: March 19, 2018, 08:27:13 am »
Make sure your IC smd footprints are DIY assembly friendly.  They don't look it.
It may be just hard to tell from the photo...

Good point.  I've got fairly good with SMD soldering, but longer pads do make things easier. 

I wish KiCad's auto router could handle dragging components with the traces still attached.  It can become a pain in the neck moving a component after it's fully laid out.

Don't just make them blindly longer. 

I made them 0.5x1.75 and rounded.  They look slightly easier.

No, you pads as so fat that the solder will bridge between them if there is any vertical alignment error on the IC.  Go for the 0.45 x 1.75mm, if not, 0.45 x 2.0mm.
The extra 0.125mm extra length helps give you a bit of horizontal play when placing down the IC + a little pad surface to allow extra solder to wick out, as well as you will be able to use solder wick in an emergency touchup job if necessary.

See my example SO14.  Basically, your SO8 should appear similar to the top 8 pins.  Zoom into my hand soldering job and you will see what I mean about extra meat for your weller, or paste, to solder onto.
 

Offline BrianHG

  • Super Contributor
  • ***
  • Posts: 7661
  • Country: ca
Re: Audio DAC output Low Pass Filter.
« Reply #54 on: March 19, 2018, 08:35:50 am »
Take a closeup look at where my IC pins end:
With the stubs only under the IC in your footptints, you will not be able to solder the pins easily with a simple weller like the footprints I have with all that extra solder-able surface coming out after the end of the IC pin.
 

Offline paulcaTopic starter

  • Super Contributor
  • ***
  • Posts: 4003
  • Country: gb
Re: Audio DAC output Low Pass Filter.
« Reply #55 on: March 19, 2018, 09:04:31 am »
Hmm.  SOICs are not exactly difficult.  SSOP-28's are much finer pitch though.  I did my binary clock SSOP28 on Thursday night and one side went absolutely perfectly, the other side bridged badly (too much solder which was about 2mm of 0.6mm solder wire!), but I was still able to clean it up with flux and it works, which was good as I only had one chip!

If you click the direct link to this and zoom in (it's 12MP or similar) you will see how the pads look.  Inspected with a uScope and tested with a multi-meter it all checks out.


My soldering set up has progressed quite a bit from the days of my 25W Weller. 
* TS-100 Iron w/ 6.5A PSU
* 898 Air reflow
* MBO 0.6mm 60/40 solder
* Circuit works flux pen
* Mechanic 60/40 Solder paste
* Andostar M201 microscope
* Rolson x36 jewelers visor

Actually paste and drag might actually be easier, I have found on 1.6mm boards with ground planes the hot air method struggles sometimes.  I cooked an ATTiny85 trying to get the paste to reflow.  After binning the chip with heat mangled PLL clock the second one was done with a single drag of the iron.  I also basically melted the plastic base of a 100uF SMD cap trying to get it to air reflow and again, next board I used the iron and it took barely a touch to reflow it.  I think for air-reflow on ground planed thick boards I will put them into the oven for a while first to heat soak the board first, before reflowing.  Need to research correct soak temps as I'll need to do it with the paste and components in place.  But... the iron is often easier for ICs and large caps.

Interestingly, that ATTiny85 was wider than the foot print I used, didn't check the datasheet (doh!) it's pins ended beyond the end of the pads and I still got it soldered and tested though it was a b0ll0x.

"What could possibly go wrong?"
Current Open Projects:  STM32F411RE+ESP32+TFT for home IoT (NoT) projects.  Child's advent xmas countdown toy.  Digital audio routing board.
 

Offline paulcaTopic starter

  • Super Contributor
  • ***
  • Posts: 4003
  • Country: gb
Re: Audio DAC output Low Pass Filter.
« Reply #56 on: March 19, 2018, 09:41:16 am »
So, I have modified the TI schematic for the DAC to remove the HID buttons and force enable the chip and regulator.  (I don't need suspend mode or buttons).

It's slightly confusing though, but I believe I have worked it out.

http://www.ti.com/lit/ds/symlink/pcm2904.pdf
Figure 36 Page 26

VDDI is an internal power supply "output" which is connected to D+ via a 1.5K as a pull up.  It also pulls SEL0 and SEL1 up.  If I connect it to SSPND it should force the chip on but ... will VDDI be active at power on, would it be better to directly link it to the V3.x rail?  The datasheet is unspecific, although the block diagram shows VBus going directly to the internal regulator where VDDI comes from.

The regulator I just connect IN and EN to force it on.

The only other optionals in that scematic worth note are that TEST0 is tied low and TEST1 is left unconnected.

I did my best to pick the correct Schottky diode with a matching Vf but the worst case would be ordering different resistors for the regulator adjust divider if it's Vf is too high.

Anyone see anything glaringly wrong or blonde?



"What could possibly go wrong?"
Current Open Projects:  STM32F411RE+ESP32+TFT for home IoT (NoT) projects.  Child's advent xmas countdown toy.  Digital audio routing board.
 

Offline paulcaTopic starter

  • Super Contributor
  • ***
  • Posts: 4003
  • Country: gb
Re: Audio DAC output Low Pass Filter.
« Reply #57 on: March 19, 2018, 02:29:56 pm »
Well I got it to fit on a board designed for the linked case.

But I had to go darkside.





Still going to be a pig to solder.

EDIT: Grrrr.  Left and Right are swapped on the jacks :(

EDIT2:  Grrrr.  Wrong foot print for the DAC!  This will take some rework.
« Last Edit: March 19, 2018, 04:02:36 pm by paulca »
"What could possibly go wrong?"
Current Open Projects:  STM32F411RE+ESP32+TFT for home IoT (NoT) projects.  Child's advent xmas countdown toy.  Digital audio routing board.
 

Offline paulcaTopic starter

  • Super Contributor
  • ***
  • Posts: 4003
  • Country: gb
Re: Audio DAC output Low Pass Filter.
« Reply #58 on: March 20, 2018, 08:25:27 am »
So those issues addressed, board looks good.

However I have become a little concerned that I paid little attention to the grounds.

The PCM290x has not one but 5 different grounds.  3 analogue, 2 digital + the Vcom for the opamps.  Vcom I have kept separate as a starred to the amps, but the others are just dumped onto the flood fill ground.

* Digital USB Ground
* Digital Ground
* Analogue side DAC ground
* PLL ground
* Oscillator Ground

Separating even the analogue and digital grounds into two planes with stars back to the USB plug, but will take a lot of rerouting.  I might be able to split off the Oscillator ground and (I assume) use it just for the crystal.

The question is... on such a small low powered board which is getting it's ground from the single noisy USB source anyway (excluding of course the audio grounds from the jacks)... how critical is the ground paths?

I expect it will work, but I might suffer some minor noise issues.

Any suggestions?
"What could possibly go wrong?"
Current Open Projects:  STM32F411RE+ESP32+TFT for home IoT (NoT) projects.  Child's advent xmas countdown toy.  Digital audio routing board.
 

Offline BrianHG

  • Super Contributor
  • ***
  • Posts: 7661
  • Country: ca
Re: Audio DAC output Low Pass Filter.
« Reply #59 on: March 20, 2018, 08:28:39 am »
Since you PCB isn't 4 layer, and now you have parts on both sides, sorry, no recommendations.  Otherwise, I would say reserve the entire bottom as 1 single ground fill except for the odd trace here and there you need to pass.
 

Offline paulcaTopic starter

  • Super Contributor
  • ***
  • Posts: 4003
  • Country: gb
Re: Audio DAC output Low Pass Filter.
« Reply #60 on: March 20, 2018, 09:08:17 am »
Since you PCB isn't 4 layer, and now you have parts on both sides, sorry, no recommendations.  Otherwise, I would say reserve the entire bottom as 1 single ground fill except for the odd trace here and there you need to pass.

Thanks.  I went back and reviewed the TI schematic and to the best I could the gerber layouts and they do not appear to be handling the individual grounds any different than I am.

So all I need to do now is wait for a few critical components to arrive to check the foot prints and it's good to go :)

Sprinkled a little bit of freeware art on it.





Of course a power LED would help.  As it's going into a translucent blue base I'm wondering if a white LED bent over to light the board would look good.
« Last Edit: March 20, 2018, 09:09:49 am by paulca »
"What could possibly go wrong?"
Current Open Projects:  STM32F411RE+ESP32+TFT for home IoT (NoT) projects.  Child's advent xmas countdown toy.  Digital audio routing board.
 
The following users thanked this post: blueskull

Offline paulcaTopic starter

  • Super Contributor
  • ***
  • Posts: 4003
  • Country: gb
Re: Audio DAC output Low Pass Filter.
« Reply #61 on: March 20, 2018, 09:52:01 am »
That looks very well, though a bit over engineered with some extra wet capacitors.

True, there are a lot of them.  But when you list out the essential ones the "over engineered ones" are only a couple.

Essential:
4 x Coupling/DC block caps - 2 x Stereo channels (one in, one out) for the jacks
1 x 3.xV power decoupling/bypass cap

Less essential on such a small board:
2 x OpAmp power rail decoupling caps - 1 or even 0 would probably have been fine.
1 x Virtual ground decoupling/bypass cap

Probably overkill, but would require double checking the analogue audio reference ground is indeed VCOM:
4 x Coupling/DC block caps 2 x Stereo channels (one in, one out) from the DAC analogue side.

At least they are all 10uF so cheap to buy in a pack of 50.
"What could possibly go wrong?"
Current Open Projects:  STM32F411RE+ESP32+TFT for home IoT (NoT) projects.  Child's advent xmas countdown toy.  Digital audio routing board.
 

Offline BrianHG

  • Super Contributor
  • ***
  • Posts: 7661
  • Country: ca
Re: Audio DAC output Low Pass Filter.
« Reply #62 on: March 20, 2018, 10:16:32 am »
You do realize that 10uf 1206 or 10uf 0805 in ceramic has better performance and costs less than half as much as electrolytic and takes up 1/8th the PCB area.

Also, with that output capacitance and resistance, do not expect to drive headphones with those parts.

Also, you might consider powering your op-amps with 5v, not 3.3v, you will get better performance and power draw if you decide to use headphones will not modulate the the 3.3v bus powering your dac.  If you do so, don't forget to optionally power the VCOM of the output opamps with 2.5v if you want the voltage drive head room.  You are powering your opamps from the power supply for the analog internal of the PCM2904.  This is wrong, read the PCM2904 data sheet.  They are powering the opamps not from that source where there is a schottky diode in series.  Varying current load on the opamp load will modulate that voltage due to the v-drop across the diode.  This will mess up the PCM2904's analog out.

Also, since you went with polarized caps, your line input cap's polarity are backwards.
Also, your line input impedance is 12k, not 50k or 100k.
Also, your audio inputs will be inverted.
Also, your Audio inputs will clip your op-amp with standard consumer grade 1vrms audio line level.
Also, your Audio input will over drive the input of the PCM2904 by 2x, even with a -10db source audio.
Also, your audio input will be inverted.  (Initial bass drum strike will be sucked in instead of pushed out.)
Also, your audio output will be inverted.  (Initial bass drum strike will be sucked in instead of pushed out.)
Also, your audio output will be around -10db of consumer level 1vrms.  (-10db is weaker than my cell phone headphone out.)
Also, a number of DC filter caps are not necessary if you are using 'VCOM' from the PCM2904.
Also, non of the power filtering caps are on the same side and right at the PCM2904's power input pins like the recommendations in the data sheet.

That's enough for now.  You should have checked & reviewed your schematic first.
« Last Edit: March 20, 2018, 11:03:03 am by BrianHG »
 

Offline paulcaTopic starter

  • Super Contributor
  • ***
  • Posts: 4003
  • Country: gb
Re: Audio DAC output Low Pass Filter.
« Reply #63 on: March 20, 2018, 12:14:22 pm »
Cap polarity and powering the opamps after the diode were indeed mistakes, which are fixable.  Thanks.

I think somethings come down to design decisions.  The question then becomes one of requirements.  Remember the primary purpose of this project is not to make a USB sound card.  Rather it is to make a prototype module that will be part of a larger project.  Yes, the one that has the pre-amps, mixer and headphone amp. 

I will likely make several stereo pre-amps PCBs next and inter-connect things on the bench.  Then I'll make a multi-input mixer and headphone amp board.  At some point a BT module.  From each I expect to learn things, as I am now.

Thinking about it, I could save myself some money by replacing the 3.5mm Jacks with headers and interconnect with 3 conductor shielded sound card leads.  For initial tests I can always solder a jacks onto those.  Depends on what's cheaper.

The output levels might be wrong though.  I wanted to aim for 0.615Vrms  internal "nominal" level for the project.  With the configuration shown in the TI schematic it's 0.6Vpp.  For 0.615Vrms I need 1.74Vpp. 

Alternatively I could just leave the output levels as they are and rely on the pre-amp to lift it to 0.615Vrms, which is kind of their purpose, to normalise multiple input levels before mixing.

The input levels are a bit of a weirdness.  Feeding high line level into an amp with a gain of 1 and then a ADC with a input level of 0.6Vpp will clip it.

The confusing thing is... if you look at the user guide:
http://www.ti.com/lit/ug/sbau195/sbau195.pdf

It states the input AND output levels are 1.86Vpp full scale LINE level.  However if I run a spice simulation on those amps they appear to have a gain off 1.  The 2906 datasheet says it's analogue output is only 0.6Vpp.  So something is a miss somewhere.

The capacitors are now bugging me, for some reason I didn't think 0805/1206 multi-layer ceramics went up to 10uF.  I must have been looking for higher voltages or something, but even then 0805 10uF 16V MLC caps are only 4p each.

That will make it much easier to place the caps where they will do better good and save space.

So most of it needs redone.  All good practice though.
"What could possibly go wrong?"
Current Open Projects:  STM32F411RE+ESP32+TFT for home IoT (NoT) projects.  Child's advent xmas countdown toy.  Digital audio routing board.
 

Offline BrianHG

  • Super Contributor
  • ***
  • Posts: 7661
  • Country: ca
Re: Audio DAC output Low Pass Filter.
« Reply #64 on: March 20, 2018, 12:58:58 pm »
The confusing thing is... if you look at the user guide:
http://www.ti.com/lit/ug/sbau195/sbau195.pdf
Warning, that is not a user guide to the PCM2904 IC which was made by BurrBrown.  That's a user guide for a cheap eval board made by TI.  And TI, if you didn't already know, makes a shit load of mistakes in their data sheet for their own IC and their eval boards.  TI has one of the worst documentations I have ever seen, where examples a lot of the times only work with 1 specific setup, or software configuration and don't exhibit proper functionality throughout the devices proper intended range of functionality.

That being said, the eval design, besides flipping/inverting the audio up-side-down, isn't absolutely non-functional.  You just need to be aware of the signal levels being totally messed up and do not expect to get any audio but a soft weak low volume if you connect headphones directly to this device as is.

10uf, 6.3v 0805, 100pcs for 2.25 as cheap as digikey gets:
https://www.digikey.co.uk/product-detail/en/samsung-electro-mechanics/CL21A106KQCLRNC/1276-2405-1-ND/3890491
10uf, 16v 1206, (X7R) 100pcs for 4.72 as cheap as digikey gets:
https://www.digikey.co.uk/product-detail/en/samsung-electro-mechanics/CL31B106MOHNNNE/1276-6641-1-ND/5961500
10uf, 6.3v 0603, 100pcs for 4.58 as cheap as digikey gets:
https://www.digikey.co.uk/product-detail/en/tdk-corporation/C1608X5R0J106M080AB/445-4112-1-ND/1975462
10uf, 25v 0805, 100pcs for 4.79 as cheap as digikey gets:
https://www.digikey.co.uk/product-detail/en/taiyo-yuden/TMK212BBJ106KG-T/587-2985-1-ND/2714178
10uf, 10v 0603, 100pcs for 7.04 as cheap as digikey gets:
https://www.digikey.com/product-detail/en/taiyo-yuden/LMK107BBJ106MALT/587-3258-1-ND/3662218
 

Offline paulcaTopic starter

  • Super Contributor
  • ***
  • Posts: 4003
  • Country: gb
Re: Audio DAC output Low Pass Filter.
« Reply #65 on: March 20, 2018, 02:53:23 pm »
The output level being 0.6Vpp I can probably live with, but should the input amps not attenuate a little?  Are we assuming this board will expect a person to be able to adjust the input level to where it doesn't clip?

Then again, the gain on the amps is easily adjust able by swapping in some replacement resistors when I choose.

On caps...

I think I'll go for these:
https://uk.rs-online.com/web/p/ceramic-multilayer-capacitors/7984621/
£1.45

Or probably wise to buy them capable of being used in other projects so 16V ones:
https://uk.rs-online.com/web/p/ceramic-multilayer-capacitors/8467313/
£4.10

RS Components are not the cheapest, but they don't charge for shipping and are nearly next day delivery.

Changing the capacitors to 0805s allows the board to be moved around a lot so I might as well see if I can address splitting the grounds in two.  KiCad can handle two grounds, but there are pitfalls when they interconnect which might allow the ground plane to flood over both.  Need to have a play.  I'm thinking if I take the ground point as the voltage regulator, everything "before" it is digital, everything "after" is analogue.  Obviously "before" and "after" are abstract.

Anyway, my FY6600 arrived this morning, so I can actually reliably test my op amps on breadboard etc. for the wider project.  Just need to check a few things first... mainly do I need to replace the PSU immediately or can I use it as is until I feel like modding it.  96V phantom leakage voltage w/ 68uA current one the BNC ground.
"What could possibly go wrong?"
Current Open Projects:  STM32F411RE+ESP32+TFT for home IoT (NoT) projects.  Child's advent xmas countdown toy.  Digital audio routing board.
 

Offline paulcaTopic starter

  • Super Contributor
  • ***
  • Posts: 4003
  • Country: gb
Re: Audio DAC output Low Pass Filter.
« Reply #66 on: March 20, 2018, 02:59:36 pm »
If I'm straying from the schematic... should I not flip the amp inputs to non-inverting?
"What could possibly go wrong?"
Current Open Projects:  STM32F411RE+ESP32+TFT for home IoT (NoT) projects.  Child's advent xmas countdown toy.  Digital audio routing board.
 

Offline Bassman59

  • Super Contributor
  • ***
  • Posts: 2501
  • Country: us
  • Yes, I do this for a living
Re: Audio DAC output Low Pass Filter.
« Reply #67 on: March 20, 2018, 06:05:48 pm »
The confusing thing is... if you look at the user guide:
http://www.ti.com/lit/ug/sbau195/sbau195.pdf
Warning, that is not a user guide to the PCM2904 IC which was made by BurrBrown.  That's a user guide for a cheap eval board made by TI.

You do realize that TI bought Burr-Brown prior to the introduction of these devices? The old BB buildings in Tucson (down by the airport) were emptied out soon after as they moved the operation to Dallas and elsewhere. TI keeps the Burr-Brown name around for marketing reasons.
 

Offline Bassman59

  • Super Contributor
  • ***
  • Posts: 2501
  • Country: us
  • Yes, I do this for a living
Re: Audio DAC output Low Pass Filter.
« Reply #68 on: March 20, 2018, 06:15:27 pm »
The PCM290x has not one but 5 different grounds.  3 analogue, 2 digital + the Vcom for the opamps.  Vcom I have kept separate as a starred to the amps, but the others are just dumped onto the flood fill ground.

* Digital USB Ground
* Digital Ground
* Analogue side DAC ground
* PLL ground
* Oscillator Ground

Separating even the analogue and digital grounds into two planes with stars back to the USB plug, but will take a lot of rerouting.  I might be able to split off the Oscillator ground and (I assume) use it just for the crystal.

The question is... on such a small low powered board which is getting it's ground from the single noisy USB source anyway (excluding of course the audio grounds from the jacks)... how critical is the ground paths?

I expect it will work, but I might suffer some minor noise issues.

Any suggestions?

The various ground pins are called out by their function, but they have to all be tied together on the board for the circuit to actually work. Ancient recommendations were to “keep analog and digital grounds separate and tie them together at only one point under the converter,” but that advice fails when you have more than one converter in the design.

The better advice is to pour a single ground plane and route the board with the idea that your digital and analog sections are in their own areas. That means, don’t route digital signals in the analog area (to the extent possible), and the converse. This implies that your parts placement is important.

Henry Ott discusses this in his books.

Chips might have particular recommendations for PLLs and the like.
 

Offline paulcaTopic starter

  • Super Contributor
  • ***
  • Posts: 4003
  • Country: gb
Re: Audio DAC output Low Pass Filter.
« Reply #69 on: March 20, 2018, 07:34:00 pm »
I should be able to keep things on different ends of the board, well sort of.

I tried flipping the opamps over into non-inverting, but I get a different cut off profile.  It cuts off way too early.  I tried a dozen different configs of the same values, moving the 330p capacitor around, but it looks like it would need recalculated completely.  The non-inverting variant seems a lot more sensitive and the cut off never as sharp.  I can get the same cut off with different cap values, but not the sharpness.

I suppose this comes from the inverting amps virtual ground behaviour, but again getting out of my depth.

I'm not really sure how much I care about true phase anyway.
"What could possibly go wrong?"
Current Open Projects:  STM32F411RE+ESP32+TFT for home IoT (NoT) projects.  Child's advent xmas countdown toy.  Digital audio routing board.
 

Offline BrianHG

  • Super Contributor
  • ***
  • Posts: 7661
  • Country: ca
Re: Audio DAC output Low Pass Filter.
« Reply #70 on: March 20, 2018, 11:21:02 pm »
Or probably wise to buy them capable of being used in other projects so 16V ones:
https://uk.rs-online.com/web/p/ceramic-multilayer-capacitors/8467313/
£4.10
Those are X6S dialect.  :(

Go for these, they are top of the line X7R dialect:
https://uk.rs-online.com/web/p/ceramic-multilayer-capacitors/7661104/     (35v, Free shipping next day)
https://uk.rs-online.com/web/p/ceramic-multilayer-capacitors/1721635/

For 0805 16v, X5R second best dialect, this is you best bet.
https://uk.rs-online.com/web/p/ceramic-multilayer-capacitors/8851698/

I know the 25v/35v ones are 1206, a fraction of a mm wider, and a little longer, but, they are X7R dialect and 25v.  Good for decoupling on your pre-amp as well with a +/- 18v supply.
« Last Edit: March 20, 2018, 11:28:54 pm by BrianHG »
 

Offline paulcaTopic starter

  • Super Contributor
  • ***
  • Posts: 4003
  • Country: gb
Re: Audio DAC output Low Pass Filter.
« Reply #71 on: March 21, 2018, 07:34:51 am »
Or probably wise to buy them capable of being used in other projects so 16V ones:
https://uk.rs-online.com/web/p/ceramic-multilayer-capacitors/8467313/
£4.10
Those are X6S dialect.  :(

Go for these, they are top of the line X7R dialect:
https://uk.rs-online.com/web/p/ceramic-multilayer-capacitors/7661104/     (35v, Free shipping next day)
https://uk.rs-online.com/web/p/ceramic-multilayer-capacitors/1721635/

For 0805 16v, X5R second best dialect, this is you best bet.
https://uk.rs-online.com/web/p/ceramic-multilayer-capacitors/8851698/

I know the 25v/35v ones are 1206, a fraction of a mm wider, and a little longer, but, they are X7R dialect and 25v.  Good for decoupling on your pre-amp as well with a +/- 18v supply.

Those 1206 ones are out of stock and the first ones are 50p each!

As the higher voltages in X7R seem expensive, I would probably buy them for specific use.  If I'm using them for decoupling a PSU, such as in the amp/mixer project I only need to buy a few.  I have 12 10us in this project.

What about these? https://uk.rs-online.com/web/p/ceramic-multilayer-capacitors/1526132/
"What could possibly go wrong?"
Current Open Projects:  STM32F411RE+ESP32+TFT for home IoT (NoT) projects.  Child's advent xmas countdown toy.  Digital audio routing board.
 

Offline BrianHG

  • Super Contributor
  • ***
  • Posts: 7661
  • Country: ca
Re: Audio DAC output Low Pass Filter.
« Reply #72 on: March 21, 2018, 09:22:50 am »
Here are X5Rs 10v 0805 (2.0mm x 1.2mm) (+85 degrees instead of X7R +105 degrees)
https://uk.rs-online.com/web/p/ceramic-multilayer-capacitors/9155225/

Here are X5R  10v, 1206 (3.2mm x 1.6mm) Only 0.4mm wider, 1.2mm longer  (ABSOLUTE CHEAPEST by any degree of all searches to date...)
https://uk.rs-online.com/web/p/ceramic-multilayer-capacitors/6698489/

Here are X5R  16v, 1206
https://uk.rs-online.com/web/p/ceramic-multilayer-capacitors/7407565/

« Last Edit: March 21, 2018, 09:32:18 am by BrianHG »
 

Offline BrianHG

  • Super Contributor
  • ***
  • Posts: 7661
  • Country: ca
Re: Audio DAC output Low Pass Filter.
« Reply #73 on: March 21, 2018, 09:41:20 am »
You do realize that 10uf 1206 or 10uf 0805 in ceramic has better performance and costs less than half as much as electrolytic and takes up 1/8th the PCB area.

Type II/III MLCCs in series of a high current (a few mA for headphone drive or later OPAMP stage) audio signal path. Recipe for disaster.

MLCCs should only be used when voltage across them remain the same and change very little to none due to the nonlinearity, in audio, this practically means only for power decoupling or very high Z signal AC coupling (you still have microphonic issues).

For DC blocking and AC coupling, electrolytic is better. If you don't care about cost and just want the smaller size, go with tantalum, and bias them properly.
He is not using this project to drive headphones, just to drive a line level at >=10k load on a 1 v signal.

Quote
And no, don't go MLCC only, even for power decoupling. Use at least one low ESR electrolytic per rail to dampen microphonic noise from MLCCs.
Yes.  I personally might choose tantalum.  Something like this:
https://uk.rs-online.com/web/p/tantalum-capacitors/7190680/

But, I don't think much will be effected in this project.

He can go crazy and go for Polymer Aluminum Capacitors, pay 0.60 per cap, and have almost 100x lower ESR, that's like 2 orders of magnitude better than any of the above, 4x better than Tantalum.
This guy is rated at 45mOhms: https://uk.rs-online.com/web/p/aluminium-capacitors/8387437/
The tantalum is rated at 200mOhms.
Electrolytics are at 2.2Ohm.
The 1206 ceramic I listed is around 0.3 Ohm at 100khz.
Note, these values should be double checked since I don't trust digikey's tables and the damn ESR values are at different frequencies.




« Last Edit: March 21, 2018, 10:24:38 am by BrianHG »
 

Offline BrianHG

  • Super Contributor
  • ***
  • Posts: 7661
  • Country: ca
Re: Audio DAC output Low Pass Filter.
« Reply #74 on: March 21, 2018, 10:33:47 am »
But, I don't think much will be effected in this project.

Sorry for the pedantry from someone who reads diyaudio often and designs high spec DACs on his own ;).

Ok, yes, I would personally use Polymer Aluminum Capacitors exclusively with a top end 24bit, not 32 bit, DAC from Analog devices in mono double balanced mode output.

And this is my go-to decoupling and DC filtering (2 in parallel/channel for headphones) cap: https://uk.rs-online.com/web/p/aluminium-capacitors/6284153/
0.005ohm.  No compromise.

 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf