lingDAC - cost effective RBCD multibit DAC design

To celebrate my return to the ferrite fold, here's an intermediate filter idea using TDK off-the-shelf inductors.

Because I already have SLFs binned in various inductances I was able to simplify the prototype not to need the 22uHs. Next step is modifying the I/V resistor of the DAC and then.... listening :)
 

Attachments

  • 2XOSTDKa.jpg
    2XOSTDKa.jpg
    158 KB · Views: 466
I get no satisfaction using a ready made product (call me a masochist)

When ferrite cores arrive, I will try the intentional air gap :)

George

PS. The aspirin trick works, thanks (an outdoors activity)

Thanks for a feedback. I picked up some resources, I think it will also help @abraxalito, if not now, then in the later stage.

This one is for EI core, but interesting, as I thought that it would be sufficient to remove material layer from the centre: Building Your Own Audio Frequency Choke Coils, October 1932 Radio-Craft - RF Cafe

This one is focusing on the optimum air gap (from a Brooks perspective): Air gapped magnetic cores

Forum discussion on air gaps. I also picked up a screenshot from the John Birckhead post: design - Why do we want gap in the core material while designing inductor? - Electrical Engineering Stack Exchange

A Google link to the most recent publication, there is a lot on air gaps:
https://www.ieee.li/pdf/viewgraphs/fundamentals_magnetics_design.pdf

The most of resources focus on tubes (of course), like this part on filters, but it is very comprehensive guide when looking form the index page: audiofilterchokes

A link to the TDK catalogue you posted before is also nice. It seems that distributed airgap is the best from a leakage perspective, but I didn't check prices. :)
 

Attachments

  • linearity.jpg
    linearity.jpg
    184.7 KB · Views: 430
Last edited:
Whether the filter can be used will depend on the output impedance of the 'noDAC'. Its designed to be used with current source output DACs (high output impedance). It can work without much degradation down to 1kohm (the output impedance of PCM1704). Below that I've not explored. Aren't most 'noDAC's based on logic chips? Their output impedance is likely going to be low - below 100ohm.
 
Member
Joined 2009
Paid Member
@abraxalito
What about these inductors: Finest Audio Inductors
Could be custom: Custom Audio Inductors

The RM8 core seems the more interesting, custom:
Max-Pins: 8 (incl. Pin0)
Max-Values: 7 custom windings/values are possible. Pin0 is always 0mH.
Max-mH: max 2000mH
Inductor Tolerance: < 3%.
Tolerance deviation for multitaps with low self resistance: Measured 100mH=2% and 20mH=5% depending on requirements.
Within an order of more than 10pcs we do a measuring in between and optimize the tolerance to < 2%.

Also possible add copper shielding-set for +4,90€

Custom-Inductor-RM8-Core_1.jpg
 
Last edited:
They look good for those who need a European source - over here I can get all that capability at a significantly lower price - though probably I'd need to order larger volumes than at Don Audio. My filter inductors are considerably lower values than their standard ones which are for EQ sections in mixing consoles.
 
Phidac pdf

Revised boards arrived a couple of days ago for PhiDAC SE and the first one I built worked first time - at least it played music, I've not checked the flatness of the filter yet.

The extra caps on this design are fitted underneath the board via 1mm bare copper wires. The DAC still works without the off-board caps though as it has a few ceramics doing the local decoupling.

PhiDAC SE is very much at home driving headphones directly, even down to 30ohm impedance due to the far better dissipation capability of AD815s compared to AD8017s. Plus the advantage of additional caps (up to 26,000uF per channel) to provide greater output current without significant 'sag' on the rails.

Hello, Is it possible to get a pdf copy of this dac to make own pcb.

Regards
 
Banned - Sockpuppet
Joined 2006
XO freq for the Ling

As far as the "ideal" XO freq for the Ling, what might that be? I recall some info at diyparadise noting that for the TDA1545-based Monica DAC, the asynch. reclocker was using an 80.000 Mhz. The owner of that website claimed higher XO freqs. were better.
Dunno.... I see many 4x or 8x TDA1387 Chinese DACs using either 12.000 or 50.000 Mhz. Not sure why THOSE two freqs are used.


An externally hosted image should be here but it was not working when we last tested it.
 
As far as the "ideal" XO freq for the Ling, what might that be? I recall some info at diyparadise noting that for the TDA1545-based Monica DAC, the asynch. reclocker was using an 80.000 Mhz. The owner of that website claimed higher XO freqs. were better.
Dunno.... I see many 4x or 8x TDA1387 Chinese DACs using either 12.000 or 50.000 Mhz. Not sure why THOSE two freqs are used.

I'm not a fan of asynchronous reclocking so I don't recommend anyone use it with any of my DACs. I really can't see the point - it introduces the chance of meta-stability in the reclocking FFs, something that's best avoided.

Here's one of my early prototypes of the 5th order filter - its not on general sale yet as I'm not sure I can make them in sufficient volume. But by all means ask if I have any available :)
 

Attachments

  • 5thorderTDKa.jpg
    5thorderTDKa.jpg
    127.9 KB · Views: 183
Last edited:
Banned - Sockpuppet
Joined 2006
Confusion

Here's one of my early prototypes of the 5th order filter - its not on general sale yet as I'm not sure I can make them in sufficient volume. But by all means ask if I have any available :)
This is long, dense thread with many "sub-projects".

Frankly, I don't know a PhiDAC from a LingDAC. As far as I can tell, the "PhiDAC" is not a DAC at all; it's a reconstruction filter for a TDA1387-based DAC. Correct?

How many of the projects in this thread are offered in "Buy" or "Kit" threads?

Amending POST #1 with a Quick Guide or FAQ would be helpful.

Otherwise, the el cheapo TDA1387 Chinese offerings on ebay or Ali seem much more attractive. Not unless abraxalito projects are on Ali/eBay ... are they?