I don't know of one, no. But its not so difficult to DIY a pair for yourself, and I'm more than happy to provide hand-holding for that endeavour. Making your own definitely meets the criterion of cheap and able to do the job. You can find details on my blog.
I bought one out of curiosity. Let's see what we can do to it
I think you'll like it very much, its like the D802 but without the noise.
I just got this board, add a 16V PS with LT1803 regulator (which I already have running for another amp) and sound comes out of this little board.
I was disappointed at first as the sound is tiny and grainy, then I realize I had push a few buttons including the Mute (this sets the tone control treble and bass). So I power off and restart it and a big difference in sound. Much more smoother and balanced. It does not have the resolution of the tpa3116, but quite musical. So like the FX-Audio D802, DON'T fool with the tone control (the Mute button)
My D802 has no noticeable noise !?! Do you have the D802C ?
I think a manual is missing so what is the sequence of buttons ? Push the muting button and then ? Without a display it is somewhat hard to see what settings are chosen isn't it ?
just D802, it has noise if I put my ears close to the speaker.
There is a very brief instruction on the sales web page:
push Mute button long press 5 sec sets mute, short press alternates treble, bass, total volume. Use Vol+/- button to increase/decrease treble/bass/volume.
I hope not much is lost in translation.
The speaker connectors are magnetic, I am going to replace that with a direct cable. And some major modding.
STM datasheet gives you the 0.2% at 1 watt distortion and the 1% and the 10% points. 18V psu, 8 ohm load, 1kHz, 384kHz switching. You can assume they used a brick made in China they bought on Ebay in 2006, if you want Blue line is just a quick impression of how much distortion you gain by skipping the dac, less is more, in every way
Blue line is just a quick impression of how much distortion you gain by skipping the dac, less is more, in every way
STM has some software you can download (I've installed it but not run it yet) which allows setting up all the internals with a GUI. But it needs an external USB adaptor to work. I plan to hack one sometime with an FT2232D board I got cheap on Taobao. If you have the cash to splash you can go for a STEVAL-CCA035V1 which is the offical USB interface.
http://www.st.com/content/ccc/resou...pdf/jcr:content/translations/en.flapw0713.pdf
http://www.st.com/content/ccc/resou...pdf/jcr:content/translations/en.flapw0713.pdf
The new (hardly available) $3 tpa3245 has lower distortion into 8 ohm load than the graph I attached, so difference is bigger than the graph showed. Just to put lowfi in perspective of a TI lowfi chip. Prices of this chip will drop, introductionpricelevel I believe is lower than where tpa3116 was introduced at.
STA326 FX audio is better than the blue line, just a little worse than tpa3116, but that TI tpa doesn't need a very good powersupply to perform ok. The FX audio will give amazing satisfaction when upgrading to a good powersupply, because only at best it is just a little worse than tpa3116.
STA326 FX audio is better than the blue line, just a little worse than tpa3116, but that TI tpa doesn't need a very good powersupply to perform ok. The FX audio will give amazing satisfaction when upgrading to a good powersupply, because only at best it is just a little worse than tpa3116.
@irribeo: I know, I use a different PSU with the D802 and I am a very satisfied user. The amp has features that make sense like multiple inputs but SPDIF performs best. Speaker protection by relays is unheard of in this class. I have some TPA boards but the D802 is way easier to use and it is easily modded to a better level.
IMHO all audio devices so also TPA3116 benefit from a very good power supply.
IMHO all audio devices so also TPA3116 benefit from a very good power supply.
Last edited:
A further improved spec mod86 would only need some smd components instead of an all through hole design. It still will be limited in speakerchoice going down in impedance.
(I just saw the mod86 spec now already benefit from a new designed smps powersupply. Need to try that with my composite ampboards).
(I just saw the mod86 spec now already benefit from a new designed smps powersupply. Need to try that with my composite ampboards).
I'm anticipating my STA333BW boards arriving shortly so I'm starting to sort out what I need to try the abaxalito headphone project.
First, I currently have 32ohm Goldring headphones but am expecting to get some 600ohm BeyerDynamics in the near future. Will I be able to use the Goldrings with this project or are they a bit too low impedance? I'm guessing that accommodating 32ohm and 600ohm headphones will be problematic anyway?
So, my plan of action is (just want to check my understanding):
If it works I'll turn to improving the power supply and digital source.
Cheers
Ray
First, I currently have 32ohm Goldring headphones but am expecting to get some 600ohm BeyerDynamics in the near future. Will I be able to use the Goldrings with this project or are they a bit too low impedance? I'm guessing that accommodating 32ohm and 600ohm headphones will be problematic anyway?
So, my plan of action is (just want to check my understanding):
- Remove the onboard filter inductors/caps.
- Wind a 1:7 transformer as per the instructions in the abraxalito blog - are there any good tutorials on winding these as I have never tried it before? - when done, solder the primaries to the PCB where the filters were previously.
- Build the filter (http://www.diyaudio.com/forums/blog...8227410-using-digital-amp-dac-tokofiltera.png) as specified in the blog and hook it up to the secondaries of the step-down transformer.
- Hook up a headphone socket to the output of the filter.
If it works I'll turn to improving the power supply and digital source.
Cheers
Ray
If you read some more details on my blog you'll note that the 600ohm cans need to have a lot more turns of wire than 32ohm ones. So you can either make two pairs of transformers, one for each or you can create multiple secondaries which you put in series for the high impedance ones and in parallel for the low. But this last option (paralleling secondaries) requires you to count turns accurately when you're winding - what you do NOT want to do is parallel windings with different numbers of turns.
One of the beauties of using transformers to drive headphones is that no pair is too low impedance, you simply wind fewer turns of thicker wire to accommodate them. The more tricky ones are higher impedance as they seem to be rather less sensitive in terms of SPL/mW.
As regards winding just wind like you're winding cotton onto a reel, nothing special is required other than keeping the turns relatively neat and doing your utmost not to double back on yourself. This last aim is to minimize the space taken by the windings - doubling back makes the packing of the windings less optimal.
I haven't mentioned it on the blog but trafos do need DC blocking in order not to saturate the core (which makes them sound terrible). I am using a 4.7uF film cap, nothing specially audiophile. You can use electrolytics but preferably bipolar ones unless you know in advance the polarity of the DC offset. Put the cap in series with the primary. I've not soldered the primary directly to the IC's outputs myself, rather I've still used a filter but with higher values of inductance (8.2mH) and lower values of capacitance (5nF). Certainly its possible when using the trafo to do away with the output filter entirely but then you're rather at the mercy of the leakage inductance of your trafo to handle the filtering. Its not easy to design to a specific leakage inductance.
On point 3, the filter isn't required for headphones. They really don't care about being fed ultrasonics, but amplifiers do. Just hook the cans to the outputs of the trafos and enjoy.
One of the beauties of using transformers to drive headphones is that no pair is too low impedance, you simply wind fewer turns of thicker wire to accommodate them. The more tricky ones are higher impedance as they seem to be rather less sensitive in terms of SPL/mW.
As regards winding just wind like you're winding cotton onto a reel, nothing special is required other than keeping the turns relatively neat and doing your utmost not to double back on yourself. This last aim is to minimize the space taken by the windings - doubling back makes the packing of the windings less optimal.
I haven't mentioned it on the blog but trafos do need DC blocking in order not to saturate the core (which makes them sound terrible). I am using a 4.7uF film cap, nothing specially audiophile. You can use electrolytics but preferably bipolar ones unless you know in advance the polarity of the DC offset. Put the cap in series with the primary. I've not soldered the primary directly to the IC's outputs myself, rather I've still used a filter but with higher values of inductance (8.2mH) and lower values of capacitance (5nF). Certainly its possible when using the trafo to do away with the output filter entirely but then you're rather at the mercy of the leakage inductance of your trafo to handle the filtering. Its not easy to design to a specific leakage inductance.
On point 3, the filter isn't required for headphones. They really don't care about being fed ultrasonics, but amplifiers do. Just hook the cans to the outputs of the trafos and enjoy.
Thanks abraxalito; this is a learning experience for me so I really appreciate your patience. Your information prompted me to reread your blog entries and I think things are a bit clearer now.
I think in the first instance, as I'll be using the 32ohm Goldring headphones initially, I will stick with the stock filters and just implement the stepdown transformer (primary = 2000turns of 0.19mm; secondary = 290turns of 0.65mm - correct?)
I'll wind separate stepdowns for 600ohm headphones as per your blog when I need them.
What impedance headphones is the 8.2uH/5nF filter optimised for?
I've also noted the DC blocking requirement and your upgrade to the power supply rail caps.
Thanks again
Ray
I think in the first instance, as I'll be using the 32ohm Goldring headphones initially, I will stick with the stock filters and just implement the stepdown transformer (primary = 2000turns of 0.19mm; secondary = 290turns of 0.65mm - correct?)
I'll wind separate stepdowns for 600ohm headphones as per your blog when I need them.
What impedance headphones is the 8.2uH/5nF filter optimised for?
I've also noted the DC blocking requirement and your upgrade to the power supply rail caps.
Thanks again
Ray
You might find the Goldrings play fairly loud with the stock 7:1 transformer, but as I've not looked up the sensitivity figures for them I'm not completely sure. I went for the 7:1 to drive my Superluxes (HD668B) which are 56ohm. It all depends how loud you like to listen, certainly this amp will go extremely loud without any sense of strain.
How I went about the design choices (core size, wire sizes and ratios) for this trafo was based initially on the rule of thumb that 'phones shouldn't be driven by source impedances higher than an eighth of their impedance. So for 56ohms that's 7ohms. The rest of the design stemmed from that initial foundation in terms of how many turns of copper would be needed - its the copper in the trafo which determines the DCR. If was going initially for 32 ohms I'd have gone most likely for a higher step down ratio as lower impedances normally need higher current and lower voltage.
Since I'm using a transformer to match the filter to the headphones the filter's not designed for any particular headphone impedance, rather the filter's targeting an impedance which can be provided by the headphones or can come from a combination of an RC snubber (like a zobel network) in parallel with the headphones. In this case the target impedance is 1100ohms for the 8.2mH/5nF. This target impedance can suit a wide variety of cans so long as the transformer ratio is chosen appropriately.
You've got the details for the trafo correct, yep.
How I went about the design choices (core size, wire sizes and ratios) for this trafo was based initially on the rule of thumb that 'phones shouldn't be driven by source impedances higher than an eighth of their impedance. So for 56ohms that's 7ohms. The rest of the design stemmed from that initial foundation in terms of how many turns of copper would be needed - its the copper in the trafo which determines the DCR. If was going initially for 32 ohms I'd have gone most likely for a higher step down ratio as lower impedances normally need higher current and lower voltage.
Since I'm using a transformer to match the filter to the headphones the filter's not designed for any particular headphone impedance, rather the filter's targeting an impedance which can be provided by the headphones or can come from a combination of an RC snubber (like a zobel network) in parallel with the headphones. In this case the target impedance is 1100ohms for the 8.2mH/5nF. This target impedance can suit a wide variety of cans so long as the transformer ratio is chosen appropriately.
You've got the details for the trafo correct, yep.
The Goldrings are 110dB sensitivity
https://earmass.com/2013/01/25/goldring-dr150-hundred-years-of-excellence/
So fewer turns of thicker wire for the secondary?
https://earmass.com/2013/01/25/goldring-dr150-hundred-years-of-excellence/
So fewer turns of thicker wire for the secondary?