What is your goal in using the transformer between DAC and amp? Typical cases would be balanced/unbalanced conversion, DC-blocking, or voltage step-up/down.
Which class D amp?
By 'this one' do you mean the lingDAC? If so its not suitable for driving a transformer, not enough current capability. A classA buffer's needed after it to drive into a lowish impedance load.
Interesting that OEP has 'upgraded' their name. I'd not recommend that one as its only a 600ohm, for line level we'd really need 10k or greater working impedance.
These look to be very good bang for the buck - 10K:10K????????? ???????? ???????-???. I am just about to order some myself to get their measurements.
The going rate on Taobao looks to be about 0.18rmb each. In other words around 35 chips to the USD. At that price I can't see how faking would be profitable.
RPI HAT SE v1.2 Files
Here are the files for v1.2 of the RPI HAT SE (single-ended). This is the board pictured in post #58, and post #59. Both of the boards I built are working.
There are three zip files attached to this post. The BOM zip includes Excel, ODS (open office/libre office), and CSV formats. The Gerber files are the ones I sent to Elecrow for manufacturing. And the KiCad files are for hacking on the design in KiCad. Note: I use a fairly recent nightly build of KiCad; these files cannot be opened by the current release version.
Let me know if you have any questions. The build is very straightforward. The only possible issue is the footprint I used for the tda1387. I used a stock KiCad footprint, which works, but the pads are a little smaller than they could be. Soldering is not terribly hard, but you'll need a steady hand and some precision tweezers. I have since created a custom KiCad footprint for the tda1387, which should be easier to use. However, as it's currently untested, I don't want to release it. But feel free to PM me if you'd like to take on the risk.
Hope somebody enjoys these!
Here are the files for v1.2 of the RPI HAT SE (single-ended). This is the board pictured in post #58, and post #59. Both of the boards I built are working.
There are three zip files attached to this post. The BOM zip includes Excel, ODS (open office/libre office), and CSV formats. The Gerber files are the ones I sent to Elecrow for manufacturing. And the KiCad files are for hacking on the design in KiCad. Note: I use a fairly recent nightly build of KiCad; these files cannot be opened by the current release version.
Let me know if you have any questions. The build is very straightforward. The only possible issue is the footprint I used for the tda1387. I used a stock KiCad footprint, which works, but the pads are a little smaller than they could be. Soldering is not terribly hard, but you'll need a steady hand and some precision tweezers. I have since created a custom KiCad footprint for the tda1387, which should be easier to use. However, as it's currently untested, I don't want to release it. But feel free to PM me if you'd like to take on the risk.
Hope somebody enjoys these!
Single-ended RPI HAT works with Allo Kali
I bought an Allow Kali Raspberry Pi I2S reclocker a long time ago, finally got around to using it. I paired it up with the tda1387 rpi hat pictured above. Now I've got it playing through my main system: TI tpa3251evm amp, Solstice MLTL speakers. The DAC I've been using in this system for months now is the Allo Boss. I can't tell if it's the inherent "sweetness" of the tda1387 DAC chip, or simply my "baby bias", but I really think this DAC sounds great. It's surprisingly clean as well. I'm even using software volume control, which has known limitations. Yet it all sounds good to my ears!
I'm going to go ahead an build up the rest of these single-ended RPI HAT boards. PM me if you're interested in a fully populated board.
I bought an Allow Kali Raspberry Pi I2S reclocker a long time ago, finally got around to using it. I paired it up with the tda1387 rpi hat pictured above. Now I've got it playing through my main system: TI tpa3251evm amp, Solstice MLTL speakers. The DAC I've been using in this system for months now is the Allo Boss. I can't tell if it's the inherent "sweetness" of the tda1387 DAC chip, or simply my "baby bias", but I really think this DAC sounds great. It's surprisingly clean as well. I'm even using software volume control, which has known limitations. Yet it all sounds good to my ears!
I'm going to go ahead an build up the rest of these single-ended RPI HAT boards. PM me if you're interested in a fully populated board.
Attachments
Congrats on your new baby, Matt. 
If SQ stands up to the Allo Boss, you must be proud indeed. Have you listened to other NOS DACs before, like any based on 1543? I have read many subjective listening impressions of 1543 DACs, but haven't found much for 1387 yet.
You are generous to share your Gerber files. I have been itching to switch over from my Mac to a RPi as a source, and I want to give the NOS sound a try. I'm tired of the harsh treble of my oversampling DACs. This looks like a killer setup, and at an unbelievable value. What PSU are you feeding your Kali with?
If SQ stands up to the Allo Boss, you must be proud indeed. Have you listened to other NOS DACs before, like any based on 1543? I have read many subjective listening impressions of 1543 DACs, but haven't found much for 1387 yet.
You are generous to share your Gerber files. I have been itching to switch over from my Mac to a RPi as a source, and I want to give the NOS sound a try. I'm tired of the harsh treble of my oversampling DACs. This looks like a killer setup, and at an unbelievable value. What PSU are you feeding your Kali with?
Thank you! Although I can't take much credit here. It's basically just a more "polished" version of didiet78's Orange Pi DAC. Not to mention, I've had lots of hand-holding from Abraxalito.
I received seven total RPI HAT boards from Elecrow (only ordered five), and built the remaining up this weekend. I made a few mistakes, so need some debugging time. But once that's done I'm thinking of doing a "mini" group buy with the extras, just to make it easy for folks to get a feel for the tda1387 sound. I haven't thought about price yet, but it will basically be the BOM cost give or take a bit (plus shipping).
I have only listened to one other NOS DAC, and it was a tda1543. In particular, it was the Muse 4x1543 USB DAC. I didn't really like it. I tried one or two of the simplest mods, it never did it for me. But this was many years ago. I don't claim to know much now, and I knew even less then. Knowing what I do now, I've always wanted to go back and give the tda1543 another try. (And with that Muse DAC in particular, I think there were some necessary mods that I was too unskilled to attempt at the time.)
Anyway; I don't remember what compelled me to try, but it was the TDA1387 x8 DAC that turned me on to this chip. That x8 DAC sounds ok out of the box, and a few simple mods make it quite enjoyable (IMO). Ever since then, I've wanted some purpose-built tda1387 PCBs specifically for my particular use-case. This thread is the result of me "scratching that itch".
The iFi iPower 5V supply. It's basically your classic wall-wart SMPS, but with some extra circuitry to clean it up. I found someone did some measurements on it, it actually measures pretty well. At some point, I'll probably switch to a Sigma11. I have some "miniaturized" S11 PCBs I designed a year or so ago, and have yet to put them to good use.
The enthusiasm for NOS seems to come in waves, as far as I can tell. Several 1543-based DACs seem to have caused quite a stir around 8 to 10 year ago, even among professional reviewers, before I was even aware of such matters. I found your thread after researching some of these older commercial implementations, such as the DAC AH and Teradak, both of which are still in production and sold on eBay. I also came across the Monica DIY project. More recently, the single chip Starting Point Systems DAC has brought the 1543 back into the limelight.
So far, I don't really see many folks (at least on the English hifi forums) talking about the 1387. Yet when I look at the eBay and Ali offerings, I do see a few Chinese sellers claiming that they sound closer to like the old Philips flagship, 1541, than anything else (and everyone seems to agree that 1541 was the cat's meow). Abraxalito, I assume you have heard several implementations of both chips. How do you compare their sound?
I have my eye on the iPower myself. I have read more than one user say they could not hear any difference between the iPower and a boutique linear power supply when paired with an Allo/Pi stack.
As it happens, I've not played at all with TDA1541. I do have one PCB from Taobao with it on but as I found out it has a design flaw I've not given it more than a very cursory listen and that was many years ago, before I'd started on 1387 development. So I have only other DACs as points of reference to compare against TDA1387, like 1543 and various S-D types.
Hi folks, I'm following this thread with interest and just wanted to thank matt_garman and abraxalito for their efforts and input so far.
I bought Malefoda's 8x DAC (now 4x) a couple of weeks ago and have been enjoying it but also wondering what a TDA1387 implementation might sound like without op-amps.
I've been chatting with stellarelephant over on Head-fi about the SPS single-TDA1543 DAC3. Richard's description of 1543 as 'beguiling' is a good one.
And I would vouch for iPower being a good option for a quiet SMPS. Some users have reported reliability concerns; I have 3 of different voltages which have all worked without problems over the past couple of years.
Anyway, hi.
I bought Malefoda's 8x DAC (now 4x) a couple of weeks ago and have been enjoying it but also wondering what a TDA1387 implementation might sound like without op-amps.
I've been chatting with stellarelephant over on Head-fi about the SPS single-TDA1543 DAC3. Richard's description of 1543 as 'beguiling' is a good one.
And I would vouch for iPower being a good option for a quiet SMPS. Some users have reported reliability concerns; I have 3 of different voltages which have all worked without problems over the past couple of years.
Anyway, hi.
Matt, I see you are using a ferrite bead and a film cap as a low pass filter at the output. You credit Abraxalito for informing some of your design choices...have you tried his lingDAC droop correction filter with your creation?
I can see that this filter would be much more time-consuming to accurately implement, but I am nonetheless curious about if/how I could retrofit it on your board down the road...if I ever get a nicer multimeter that can measure inductance...
I can see that this filter would be much more time-consuming to accurately implement, but I am nonetheless curious about if/how I could retrofit it on your board down the road...if I ever get a nicer multimeter that can measure inductance...
I think Matt's been waiting on availability of a PCB - happy to report that my gf's PCBs have now arrived and I'm into building up the first one. Once I'm sure its fine the manufacturing data will be available on request.
I'm also interested in getting some listening feedback from people in a position to make comparisons to other DACs, particularly multibit ones. In such cases I may send out built up filter boards in exchange for listening reports
