Ah yes, I've noticed that the SSM3J327R has been out of stock at both Digikey and Mouser. I've some here but I realized only a couple of days ago they're probably fakes (the package does not conform to the DS) even though they work fine. I'll check out that suggested replacement.
What does 'share BOMs on Digikey' mean here? Sorry for my complete ignorance, I've not used Digikey to place orders, just to research parts
SSM3J328R will certainly function in the circuit. Its primary difference is in its parasitic capacitances - they're about 3X higher than for the J327. This will mean higher HF distortion at higher signal levels, not sure whether it'll be audible though.
What does 'share BOMs on Digikey' mean here? Sorry for my complete ignorance, I've not used Digikey to place orders, just to research parts
SSM3J328R will certainly function in the circuit. Its primary difference is in its parasitic capacitances - they're about 3X higher than for the J327. This will mean higher HF distortion at higher signal levels, not sure whether it'll be audible though.
Last edited:
I uploaded the xls sheet your provided into the BOM manager on digikey, which creates a list of all the parts and quantities that you can then order from.
I thought there was a way to share this, but I can't seem to find any option to do this at the moment.
geoff.
Edit.
Found it, here are the links:
Filter Bom: Login - My Digi-Key
DAC Bom: Login - My Digi-Key
SEBUFF Bom: Login - My Digi-Key
I should add that I've not verified these yet, just added the list and swapped out the parts that were OOS.
Last edited:
Thanks to @bartb for doing sterling work on getting parts, quantities, codes etc. correct. Any mistakes though are down to me
Oh it was a pleasure... learned alot also !
New DAC design, cheaper and easier to build than lingDAC
Since lingDAC requires a certain amount of component selection (inductors) and measurement on test I had the idea to do a 'lower maintenance' design. This is one of the early prototypes, the BOM cost should be significantly lower than lingDAC - its going on a single PCB for a start. Note though that the low BOM cost depends on sourcing components not from the usual distributors as that'll push the price up to lingDAC territory very easily.
To get this to be a 'low maintenance' design I'm using opamps rather than discretes. The opamp I've gone for is AD8017, for a couple of reasons. Firstly its ultra-cheap on Taobao and second its wide-bandwidth with a very beefy output stage. It can therefore not only be used as an amplifying stage but also as a voltage regulator.
Although the pic I've uploaded shows five DAC chips on the old lingDAC DAC board, I'm only using a single TDA1387 in this. Filtering is done with 0805 size inductors and then there's a traditional I/V stage, albeit one providing the correct loading for the CLC filter. Beyond that, there's a 3rd order MFB low pass which can be configured to taste - in NOS mode it provides for NOS droop correction but for those who'd like to use this DAC with external oversampling can be chosen to be flat.
Power supply is a single rail from 9-12V and current draw is under 70mA. Output is the normal 2VRMS. Schematic to follow in a while....
Since lingDAC requires a certain amount of component selection (inductors) and measurement on test I had the idea to do a 'lower maintenance' design. This is one of the early prototypes, the BOM cost should be significantly lower than lingDAC - its going on a single PCB for a start. Note though that the low BOM cost depends on sourcing components not from the usual distributors as that'll push the price up to lingDAC territory very easily.
To get this to be a 'low maintenance' design I'm using opamps rather than discretes. The opamp I've gone for is AD8017, for a couple of reasons. Firstly its ultra-cheap on Taobao and second its wide-bandwidth with a very beefy output stage. It can therefore not only be used as an amplifying stage but also as a voltage regulator.
Although the pic I've uploaded shows five DAC chips on the old lingDAC DAC board, I'm only using a single TDA1387 in this. Filtering is done with 0805 size inductors and then there's a traditional I/V stage, albeit one providing the correct loading for the CLC filter. Beyond that, there's a 3rd order MFB low pass which can be configured to taste - in NOS mode it provides for NOS droop correction but for those who'd like to use this DAC with external oversampling can be chosen to be flat.
Power supply is a single rail from 9-12V and current draw is under 70mA. Output is the normal 2VRMS. Schematic to follow in a while....
Attachments
Hello Richard,
double question on this one
Is there a trafo you would recommend ( or show an example of just to get an idea) ?
Am I seeing things correctly that if we go balanced and we are then inverting the polarity of one of the threads prior to amplifier, given that the DC component would become negative (and cancel out with the positive component), ...that we could do without the 2200uF polarized caps (C5 and C7) in the buffer stage (and get cleaner sound) ?
Last edited:
Hi Bart - here is a transformer you can try, I have a couple from Taobao and they work very well indeed :
2PCS / 10K:10K permalloy audio isolation transformer, balanced and unbalanced conversion audio isolator-in Transformers from Home Improvement on Aliexpress.com | Alibaba Group
Yes, balanced involves inverting one signal chain (digitally, before the DAC). But as regards offset cancellation it assumes that the DC component is in the signal whereas in reality its not, its an artifact introduced by the electronics. So I wouldn't recommend dispensing with AC coupling caps even when running balanced. Trafos are sensitive to DC, some more so than others, I prefer to keep them entirely free of it.
2PCS / 10K:10K permalloy audio isolation transformer, balanced and unbalanced conversion audio isolator-in Transformers from Home Improvement on Aliexpress.com | Alibaba Group
Yes, balanced involves inverting one signal chain (digitally, before the DAC). But as regards offset cancellation it assumes that the DC component is in the signal whereas in reality its not, its an artifact introduced by the electronics. So I wouldn't recommend dispensing with AC coupling caps even when running balanced. Trafos are sensitive to DC, some more so than others, I prefer to keep them entirely free of it.
New DAC schematic
Here is the schematic for the lower cost, easier-to-build opamp-based DAC which I've decided to call 'phiDAC'. In math, 'phi' is the 'golden ratio' and the PCB's being laid out to be a golden ratio rectangle (approx 5 * 8cm).
I'm hoping the BOM cost for this is going to be under 35RMB here in China (sub $5). AD8017 is quite cheap on Taobao - some of the most expensive parts after those opamps are the NP0 capacitors. You could substitute X7R to save the odd cent but then you'll lose the excellent stability (over time & temp) class1 dielectrics bring.
I'm getting around the inductor measuring issue by using closer tolerance Ls (10% vs 20% on lingDAC) in the DAC-opamp filter and using many in series so that hopefully there will be some averaging out towards the mean (assuming normal distribution, maybe an assumption too far, we shall see). Also the filter is lower Q and hence less sensitive to component variations. I shall build a few and see how consistent the FR turns out to be.
Here is the schematic for the lower cost, easier-to-build opamp-based DAC which I've decided to call 'phiDAC'. In math, 'phi' is the 'golden ratio' and the PCB's being laid out to be a golden ratio rectangle (approx 5 * 8cm).
I'm hoping the BOM cost for this is going to be under 35RMB here in China (sub $5). AD8017 is quite cheap on Taobao - some of the most expensive parts after those opamps are the NP0 capacitors. You could substitute X7R to save the odd cent but then you'll lose the excellent stability (over time & temp) class1 dielectrics bring.
I'm getting around the inductor measuring issue by using closer tolerance Ls (10% vs 20% on lingDAC) in the DAC-opamp filter and using many in series so that hopefully there will be some averaging out towards the mean (assuming normal distribution, maybe an assumption too far, we shall see). Also the filter is lower Q and hence less sensitive to component variations. I shall build a few and see how consistent the FR turns out to be.
Last edited:
...but ...
Huh????? So you've got a $5 DAC that needs a PSU with $42 trafos ... ya' lost me, dude.
Seems you've not been following along very closely. @bartb wants a balanced version, the $5 DAC isn't balanced. "PSU with $42 trafos" - where does that come from? The trafos he's asking about are signal trafos, not power.
Your mention of PSU though is on the money - I realize the $5 price isn't a complete solution so it would be great to have an equally low priced supply for this. I am thinking a stock USB wallwart followed by a voltage doubler based on LM2662....
Your mention of PSU though is on the money - I realize the $5 price isn't a complete solution so it would be great to have an equally low priced supply for this. I am thinking a stock USB wallwart followed by a voltage doubler based on LM2662....
Last edited:
Vcc
Vcc for HC86 is 2.5V? Why so low? (Although some logic chips can operate that low, per datasheet, usually higher Vcc is better).
Dear Toots:
Well, resistor Johnson noise ain't a factor then? Can you get that 74HC from Potato? Get rid of all those in-line I2S R's.