More TDA1545A porn...
Someone really went to town with diyparadise TDA1545A kit ...
Again, the point being ... yeah ... duh .... OF COURSE it'll sound good with ALL THAT time/effort/$$ "invested" (although some of would call it wasted).
The TDA1545A /1387 is putatively avail. in SOIC (SO8) and can run on very low power. So, folks, something like ...
Put it up on kickstarter and, then, Tindie ... make a YouTube "unboxing" video ... create that dedicated Facebook or Blogger page ... and Twitter the hell outta it ... oh and, uh, those ubiquitous FANBOY "appreciation" threads on on pre-pubescent headphone-gear communities help, too 😉
A dedicated "crowd-source" or "development" thread on DIY or audio-enthusiast sites means you don't really have to do a lotta brainwork ... let the "crowd" do it ...the loot from kickstarter is all yours no matter what ... Chinese factory will mass-produce your "crowd" project for peanuts.
No, ain't all that funner than plain-jane DIY?
Someone really went to town with diyparadise TDA1545A kit ...
An externally hosted image should be here but it was not working when we last tested it.
Again, the point being ... yeah ... duh .... OF COURSE it'll sound good with ALL THAT time/effort/$$ "invested" (although some of would call it wasted).
The TDA1545A /1387 is putatively avail. in SOIC (SO8) and can run on very low power. So, folks, something like ...
An externally hosted image should be here but it was not working when we last tested it.
Put it up on kickstarter and, then, Tindie ... make a YouTube "unboxing" video ... create that dedicated Facebook or Blogger page ... and Twitter the hell outta it ... oh and, uh, those ubiquitous FANBOY "appreciation" threads on on pre-pubescent headphone-gear communities help, too 😉
A dedicated "crowd-source" or "development" thread on DIY or audio-enthusiast sites means you don't really have to do a lotta brainwork ... let the "crowd" do it ...the loot from kickstarter is all yours no matter what ... Chinese factory will mass-produce your "crowd" project for peanuts.
No, ain't all that funner than plain-jane DIY?
Last edited:
I'd not say 'easily' unless the device feeding the digital stream to the DAC chip has the ability to output I2S (and not just EIAJ). If it can then yeah it should be easy as they're the same pinout - is the TDA1545A in SMT?
Yep - it's item 390922499784 on fleabay. I think the seller used to have a TDA1387 module but it's disappeared.
If someone could come up with a board for the TDA1387, I'd be happy to co-ordinate a group buy.
If someone could come up with a board for the TDA1387, I'd be happy to co-ordinate a group buy.
That board's using the WM8804 S/PDIF receiver and that IC can certainly output I2S. However it might be a bit fiddly to reconfigure it as the WM8804's package is very small (0.65mm pin spacings). Page 19 of the DS shows which pins control the interface format.
Transformer
Hi,
In case of
Balanced -> passive I/V with resistor -> Transformer -> Unbalanced
scheme, what kind of resistor and transformer should I use? I mean a hot spot between a particular I/V resistor and a transformer impedance/turn ratio etc.
Hi,
In case of
Balanced -> passive I/V with resistor -> Transformer -> Unbalanced
scheme, what kind of resistor and transformer should I use? I mean a hot spot between a particular I/V resistor and a transformer impedance/turn ratio etc.
I've never tried using a transformer without a filter first so I can't even have a stab at answering your question. I know some designs have used this kind of scheme, I gather they intend to use the trafo not only for isolation but also filtering. The problem with that approach is that the filtering is a side-effect, meaning its very sensitive to small details about the trafo's construction. So the SQ will be really hit-and-miss.
I recommend filtering first - then the trafo's job is much easier and you can wind it yourself, saving a whole heap of cash 🙂
<edit> If you want to buy a commercial trafo to try this ISTM that the critical parameter is going to be the primary inductance. Without enough inductance the trafo will have an LF roll-off determined by the source impedance driving it (which is the I/V resistor). Trafo's generally don't perform well when driven from a non-zero source impedance, so this is the aspect to watch. In my design I buffer first before the trafo so the trafo's inductance doesn't matter so much, also I believe this gives lower distortion.
A simple example - if your I/V resistor is 1kohm and the trafo's primary inductance is 10H this forms an LR high pass filter which is -3dB at 16Hz. Meaning significant bass roll-off. Ideally for the bass response you'd want an I/V resistor (in ohms) less than 30X the primary inductance (in Henries). This is a reasonable rule of thumb for keeping the bass droop under control. Probably though you'll get considerable LF distortion with this much driving impedance.
I recommend filtering first - then the trafo's job is much easier and you can wind it yourself, saving a whole heap of cash 🙂
<edit> If you want to buy a commercial trafo to try this ISTM that the critical parameter is going to be the primary inductance. Without enough inductance the trafo will have an LF roll-off determined by the source impedance driving it (which is the I/V resistor). Trafo's generally don't perform well when driven from a non-zero source impedance, so this is the aspect to watch. In my design I buffer first before the trafo so the trafo's inductance doesn't matter so much, also I believe this gives lower distortion.
A simple example - if your I/V resistor is 1kohm and the trafo's primary inductance is 10H this forms an LR high pass filter which is -3dB at 16Hz. Meaning significant bass roll-off. Ideally for the bass response you'd want an I/V resistor (in ohms) less than 30X the primary inductance (in Henries). This is a reasonable rule of thumb for keeping the bass droop under control. Probably though you'll get considerable LF distortion with this much driving impedance.
Last edited:
Before modifying for the TDA1387 I thought I'd have a good listen in its stock form (WM8804->TDA1545) so I put together a simple DC PSU and it does sound great. Very clean treble as a few have already mentioned. One thing I find odd is that the iv resistors are 4k7, if my reading of the datasheet is correct, this is too high??
Yes they do sound too high so probably you're getting clipping at the highest digital levels. Depending on the current setting (pin7 bias voltage) you should have 2k as a maximum for 1mA and 1k for 2mA. The output compliance voltage is 2V - 4V for a 5V supply.
Indeed. Clearly it's been maximised for a larger vrms I think. For pin 7 IREF the voltage is 0.83V, which looks correct as per datasheet. Interestingly they've installed a trimpot to adjust this if necessary.
What about reducing the iv resistor right down and using a trafo for gain?
I have some of these laying around.
Would they do the trick for some additional gain?
What about reducing the iv resistor right down and using a trafo for gain?
I have some of these laying around.
Would they do the trick for some additional gain?
Looks interesting - but its only giving you X5.5 for voltage gain. So to gain any output voltage advantage you'd have to use bigger than 360R for the I/V resistor. Depending on the primary inductance (didn't see that quoted) you might sacrifice bass.
Good point.
I bought them for a PCM1794 project that never got off the ground.
The other concern is obviously the output impedance. If we used an iv resistor of, say, 470R, how would we calculate the output impedance of the trafo stage?
I bought them for a PCM1794 project that never got off the ground.
The other concern is obviously the output impedance. If we used an iv resistor of, say, 470R, how would we calculate the output impedance of the trafo stage?
Last edited:
The output impedance scales as the voltage ratio squared. So for 5.5X that's 30X. So with a 470R you'll be looking at 14k. A bit high.
The solution I'd use is to stack DAC chips to increase the output current. With the TDA1545A you can tweak up the pin7 voltage a bit, but I found problems going over 1.2V.
The solution I'd use is to stack DAC chips to increase the output current. With the TDA1545A you can tweak up the pin7 voltage a bit, but I found problems going over 1.2V.
Ok so there's a few possible mods. I can even just insert an additonal gain stage (non trafo) in between the dac and the power amp (I have a very linear premium that doesn't colour the sound much).
Really I just want to optimise the TDA1545 so it's operating the best possible conditions prior to stacking a few on top, then trying out the TDA1387 etc.
In short - Abrax, if you had this board, (a) what would you change the iv resistor to, and (b) would you tweak up the pin7 voltage or leave it as 0.83?
Really I just want to optimise the TDA1545 so it's operating the best possible conditions prior to stacking a few on top, then trying out the TDA1387 etc.
In short - Abrax, if you had this board, (a) what would you change the iv resistor to, and (b) would you tweak up the pin7 voltage or leave it as 0.83?
If I were playing with this I'd choose an I/V resistor to suit one of my filters. I think I have a filter with 2k5 input impedance. so it presents 1.25k to the DAC. Seems worth tweaking up the Vref a little to get the maximum output from the filter, paying attention to the excursions on the outputs not exceeding the compliance voltage spec.
Beyond that, the listening satisfaction is improved immensely with a very low impedance power supply at LF (paralleled supercaps work a treat), plus a whole heap of caps (at least 20,000uF) decoupling Vref.
Beyond that, the listening satisfaction is improved immensely with a very low impedance power supply at LF (paralleled supercaps work a treat), plus a whole heap of caps (at least 20,000uF) decoupling Vref.
Hi abraxalito,
Have you any conceptual idea abut usage of inverted i2s data on separate dacs? I mean, if we invert i2s signal without separating L-R data, one chip will carry L- and other one will carry L+. Are doing same without any pitfall?
Have you any conceptual idea abut usage of inverted i2s data on separate dacs? I mean, if we invert i2s signal without separating L-R data, one chip will carry L- and other one will carry L+. Are doing same without any pitfall?
I've not had any problems when I've done that and I've done it plenty of times, with an inverter on the data line.
I've been in touch with TekDevice recently about their TDA1387 module and they have a new one in development, due out in a few weeks. It will be a module with four TDA1387 on board. I'll post more when I've got my board....
- Status
- Not open for further replies.