I've built an amp based on the TAS5518 and a PIC, but I'm having a hard time getting the TAS5518 to do anything.
When I look at the I2C communications, pretty much all the time I don't get back an ack bit, it just stays high. But occasionally it appears to receive an ack bit, but there is no change in the results.
Also, the valid and PWM outputs are all at ~2-3mV, and don't change.
I already changed out the chip, in case I had damaged it previously, and there is no change in response.
Anyone have any ideas?
Thanks,
William
When I look at the I2C communications, pretty much all the time I don't get back an ack bit, it just stays high. But occasionally it appears to receive an ack bit, but there is no change in the results.
Also, the valid and PWM outputs are all at ~2-3mV, and don't change.
I already changed out the chip, in case I had damaged it previously, and there is no change in response.
Anyone have any ideas?
Thanks,
William
I did not use a digital input since I wanted to be able to use it with a variety of sources. I also wasn't convinced I could output an decoded/uncompressed stream from my computer and I didn't want to deal with decoding...
I'm not sure that I have the time to develop/sell modules, maybe if it's small or no time crunch. Which part were you interested in, in particular? The output stage?
I can try and upload pictures this weekend.
I'm not sure that I have the time to develop/sell modules, maybe if it's small or no time crunch. Which part were you interested in, in particular? The output stage?
I can try and upload pictures this weekend.
I guess you could easily use a miniDSP miniStreamer for USB, or miniDSP miniDigi for SPDIF
MiniDSP - miniDIGI
MiniDSP - miniSTREAMER
I'd personally be using SPDIF..
Regarding the parts of the amplifier I'm interested in, I'd say everything but the AD converter, but I guess it wouldn't be much of a problem to have it onboard if it makes it easier for you.
I'm wondering about the ability of this "power D/A" to get rid of incoming jitter from the SPDIF, Is jitter rejection a task for the miniDigi board or not? what would be the ideal clocking strategy on such a design?.. forgive my newbie questions.
BTW, a TAS5631 design would be very exciting as well
MiniDSP - miniDIGI
MiniDSP - miniSTREAMER
I'd personally be using SPDIF..
Regarding the parts of the amplifier I'm interested in, I'd say everything but the AD converter, but I guess it wouldn't be much of a problem to have it onboard if it makes it easier for you.
I'm wondering about the ability of this "power D/A" to get rid of incoming jitter from the SPDIF, Is jitter rejection a task for the miniDigi board or not? what would be the ideal clocking strategy on such a design?.. forgive my newbie questions.
BTW, a TAS5631 design would be very exciting as well
Those are interesting boards. I hadn't heard of their products before. What is the advantage of a TAS5518+TAS5631 amp connected to one of the miniDIGIs over their amp+dsp combo? Better dynamic range? Lower noise?
I imagine the SPDIF input stage would be the part that should remove jitter, but I don't know.
I imagine the SPDIF input stage would be the part that should remove jitter, but I don't know.
I think you wouldn't need their DSP board in order to make a power D/A, their miniAMP uses TI purepath chipset as well, I'm not sure if it si "purepath HD" or just plain purepath chipset.. but anyway the difference would be POWER.. the miniAMP is 2x20W and as far as I can understand from TI site TAS5631 could output 2x300W, the last PurepathHD are suposed to have more power and more fidelity..
I'd love to get a board with I2S input and TAS5518 + TAS5631, I guess all I'd have to do is to get hold of a miniDIGI board and a power supply.
My only concerns are the volume control and the jitter rejection..
The only power D/A's I've found at a moderated price are sumoh's, but still underpowered for my needs
| SUMOH | True digital amplifiers
wadia, TacT, NAD and Steinway-Lyngdorff are out of my budget, and nobody is selling DIY modules yet, besides the very low power miniAMP from miniDSP.
I'd buy on blink of an eye a TAS5518 + TAS5631 I2s input module, if anybody was making it.. I guess purepath HD are still pretty new chips, the funny thing is that I see very nice TAS5630 analogue amps like these ones:
Class D Audio TI-600 Amplifier - Class D Audio Amplifiers - PRODUCTS
I guess digital implementation with digital PWM modulators is more complicated than analog PWM modulators, that's why I got interested on this thread, cos you were using TAS5518..
but to be honest I'm a complete newbie regarding all this stuff, I just find it a very interesting technology I want to try, since my only source is a computer, so a power D/A makes sense to me.
I'd love to get a board with I2S input and TAS5518 + TAS5631, I guess all I'd have to do is to get hold of a miniDIGI board and a power supply.
My only concerns are the volume control and the jitter rejection..
The only power D/A's I've found at a moderated price are sumoh's, but still underpowered for my needs
| SUMOH | True digital amplifiers
wadia, TacT, NAD and Steinway-Lyngdorff are out of my budget, and nobody is selling DIY modules yet, besides the very low power miniAMP from miniDSP.
I'd buy on blink of an eye a TAS5518 + TAS5631 I2s input module, if anybody was making it.. I guess purepath HD are still pretty new chips, the funny thing is that I see very nice TAS5630 analogue amps like these ones:
Class D Audio TI-600 Amplifier - Class D Audio Amplifiers - PRODUCTS
I guess digital implementation with digital PWM modulators is more complicated than analog PWM modulators, that's why I got interested on this thread, cos you were using TAS5518..
but to be honest I'm a complete newbie regarding all this stuff, I just find it a very interesting technology I want to try, since my only source is a computer, so a power D/A makes sense to me.
Last edited:
I guess my amp is like the sumoh, same output stage at least.
What kind of time frame were you looking to experiment with it?
I have also been curious about cleaning up my design and adding a digital input... but I don't think I have the time right now to work on it so I probably wouldn't get serious with it until February and even then it could take a few weeks. I guess it gives some time to really figure out the design and requirements before starting the board modifications.
I'm curious, what are you planning on using all that power for? I think I've measured my power output at around 1-2W for normal listening.
What kind of time frame were you looking to experiment with it?
I have also been curious about cleaning up my design and adding a digital input... but I don't think I have the time right now to work on it so I probably wouldn't get serious with it until February and even then it could take a few weeks. I guess it gives some time to really figure out the design and requirements before starting the board modifications.
I'm curious, what are you planning on using all that power for? I think I've measured my power output at around 1-2W for normal listening.
Regarding my power needs I have a humble recording studio at home.. I have Yamaha NS10M studio and Alesis Monitor one MK2, both passive nearfield studio monitors
NS10M is 8ohm and rated at maximum of 60W program
Monitor 1 MK2 is 4ohm and rated at a maximum of 120W program
In my experience the minimum power to drive them comfortably is a bit over the maximum rated, otherwise amps usually tend to work "forced" and they act as a dynamic compressor when you ask for a bit of loudness, right now I'm driving them (not at the same time) with an alesis RA300 power amp (150W@4ohm/90W@8ohm) and I wouldn't really want to drive them with less power.
I think the difference between nearfield studio monitors and home speakers is that at mixing time a mixing engineer usually goes up and down in volume in order to check out things and for the decision making, we don't listen at the same volumen all the time, we listen at moderated levels, but we have to check from time to time up there, very loud, and down there very soft in order to make sure it will translate good for any other listening conditions.. It is not desirable to have a compressor on my power amp at high levels, I want to keep dynamics, even at high levels.
Regarding my time frame.. don't worry dude, Im not in hurry at all, I have and amp already, and I can work.. Al this power DAC stuff just got my attention, since I think it would be very interesting in order to get rid of my DAC, now a days I record and mix everything in the computer and I output my master output to SPDIF > DAC > passive volume control > poweramp.. I was just wondering if I could shorten my listening path and gain some more "detail" and at the same time making it more affordable.. I've been even thinking about linear phase xovers within the computer, bi-amping and bypassing the passive xovers on the speakers.. that would get rid of one more analog section and would make a more accurate xover..
Regarding the TI purepath chips, I've been doing a bit of reading on the TI site, and as far as I can tell, the only chips that would have enough power for my needs and can drive 4ohms are:
Purepath HD > TAS5614 (150W)/TAS5631(300W)
Purepath > TAS5162 (210W)/TAS5261 (315W)
both TAS5162 and TAS5261 have more rated dynamic range, but both are open loop
both TAS5614 and TAS5631 have less dynamic range but are closed loop
as far as I can tell TI suggests that the Purepath HD series are more "audiophile", less distorsion, more PSRR.. blah, blah, blah... but to be honest, I'm not a proper amplifier designer at all... so I can't really tell
There is just one importan point which I'd want clarify and that is the volumen control. How is it supposed to work on a power DAC based on these chips? Should it be DSP controlled like NAD M2? Has it anything to do with power rails voltage? I could digitally control it from the computer, but as far as I know I would loose bit resolution and that is not good.. NAD made it with a smart DSP solution
Thanks legoman for paying attention to my thoughts, I was starting to think that nobody was finding interesting all this powerDAC stuff, i was feeling kind of lonely on this. I'll be following all your progress.
NS10M is 8ohm and rated at maximum of 60W program
Monitor 1 MK2 is 4ohm and rated at a maximum of 120W program
In my experience the minimum power to drive them comfortably is a bit over the maximum rated, otherwise amps usually tend to work "forced" and they act as a dynamic compressor when you ask for a bit of loudness, right now I'm driving them (not at the same time) with an alesis RA300 power amp (150W@4ohm/90W@8ohm) and I wouldn't really want to drive them with less power.
I think the difference between nearfield studio monitors and home speakers is that at mixing time a mixing engineer usually goes up and down in volume in order to check out things and for the decision making, we don't listen at the same volumen all the time, we listen at moderated levels, but we have to check from time to time up there, very loud, and down there very soft in order to make sure it will translate good for any other listening conditions.. It is not desirable to have a compressor on my power amp at high levels, I want to keep dynamics, even at high levels.
Regarding my time frame.. don't worry dude, Im not in hurry at all, I have and amp already, and I can work.. Al this power DAC stuff just got my attention, since I think it would be very interesting in order to get rid of my DAC, now a days I record and mix everything in the computer and I output my master output to SPDIF > DAC > passive volume control > poweramp.. I was just wondering if I could shorten my listening path and gain some more "detail" and at the same time making it more affordable.. I've been even thinking about linear phase xovers within the computer, bi-amping and bypassing the passive xovers on the speakers.. that would get rid of one more analog section and would make a more accurate xover..
Regarding the TI purepath chips, I've been doing a bit of reading on the TI site, and as far as I can tell, the only chips that would have enough power for my needs and can drive 4ohms are:
Purepath HD > TAS5614 (150W)/TAS5631(300W)
Purepath > TAS5162 (210W)/TAS5261 (315W)
both TAS5162 and TAS5261 have more rated dynamic range, but both are open loop
both TAS5614 and TAS5631 have less dynamic range but are closed loop
as far as I can tell TI suggests that the Purepath HD series are more "audiophile", less distorsion, more PSRR.. blah, blah, blah... but to be honest, I'm not a proper amplifier designer at all... so I can't really tell
There is just one importan point which I'd want clarify and that is the volumen control. How is it supposed to work on a power DAC based on these chips? Should it be DSP controlled like NAD M2? Has it anything to do with power rails voltage? I could digitally control it from the computer, but as far as I know I would loose bit resolution and that is not good.. NAD made it with a smart DSP solution
Thanks legoman for paying attention to my thoughts, I was starting to think that nobody was finding interesting all this powerDAC stuff, i was feeling kind of lonely on this. I'll be following all your progress.
Last edited:
Ah, that explanation of power requirements is interesting, thanks.
Yes, a digital crossover would definitely have a better phase response and have a very sharp cutoff at exactly the frequency you want... Could tweak it for each speaker too, that would be neat. I've recently started to get into some digital filtering and it is pretty cool stuff, very powerful...
Theoretically the PurePathHD chips will have lower distortion and better PSRR because they're closed loop. The only question would be whether or not it's enough to notice, but based on their test results on their evaluation boards, I'd say that it definitely improves the performance (quite possibly a noticeable amount). I would have used one for mine, but when I built it they only had one where it took an analog input and directly output to speakers but it was only 2 channels. I would go with the HD chips despite their slightly reduced dynamic range just for the reduced distortion and flatter frequency response since you're dealing with studio monitors and want a flat response.
As far is volume control... I use the TAS5518 to do volume control on mine (although, I only have 3 settings - full, low volume -15 or -18 dB, and mute). Within the TAS5518 you can specify an attenuation or amplification value: -127dB to +18dB. I rarely use it since I just change the volume on the computer. I planned on putting a volume knob on the amp (I put a hole in the case for it and the MCU is programmed), but I haven't felt the need. Maybe I'll give it a try and see how it behaves.
The TAS5518 can output a PWM signal for manipulating the supply voltage to achieve higher dynamic range (24dB), but I haven't investigated how well it works or how easy it is to use.
That NAD M2 has some pretty impressive specs... and an impressive price to match...
Yes, a digital crossover would definitely have a better phase response and have a very sharp cutoff at exactly the frequency you want... Could tweak it for each speaker too, that would be neat. I've recently started to get into some digital filtering and it is pretty cool stuff, very powerful...
Theoretically the PurePathHD chips will have lower distortion and better PSRR because they're closed loop. The only question would be whether or not it's enough to notice, but based on their test results on their evaluation boards, I'd say that it definitely improves the performance (quite possibly a noticeable amount). I would have used one for mine, but when I built it they only had one where it took an analog input and directly output to speakers but it was only 2 channels. I would go with the HD chips despite their slightly reduced dynamic range just for the reduced distortion and flatter frequency response since you're dealing with studio monitors and want a flat response.
As far is volume control... I use the TAS5518 to do volume control on mine (although, I only have 3 settings - full, low volume -15 or -18 dB, and mute). Within the TAS5518 you can specify an attenuation or amplification value: -127dB to +18dB. I rarely use it since I just change the volume on the computer. I planned on putting a volume knob on the amp (I put a hole in the case for it and the MCU is programmed), but I haven't felt the need. Maybe I'll give it a try and see how it behaves.
The TAS5518 can output a PWM signal for manipulating the supply voltage to achieve higher dynamic range (24dB), but I haven't investigated how well it works or how easy it is to use.
That NAD M2 has some pretty impressive specs... and an impressive price to match...
Thanks again for all the info legoman! How is your design going?
Someone at Sumoh has told me they have plans for some more powerful amps based on Purepath HD chips, and he has pointed me out to this amp:
HFX - RipAMP 2.1
From what sumoh says it seems they design and sell OEM amps for other brands and this must be one of them.. They say they will sell the same amp + Spdif + volume control + remote on their own line of products. It is based on purepath HD chips, it looks like TAS5614, but not sure.. I don't know about the PCM to PWM modulator.
Regarding volume control on the TAS5518, I had a look at the datasheet and it seems like the -127dB/+18dB is DSP based (DAP as TI names it), I'm not sure, but I guess it is plain LSB truncation (I should ask TI), so AFAIK that should be OK for 16bit audio, but not good for 24bit audio, I haven't seen anything regarding those 24dB of extra dynamic range that could be used for volume control, I'll print the datasheet and I'll give it a deeper read.
Regarding clocking and incoming jitter rejection I've seen in the datasheet that the TAS5518 needs its own crystal clock and it has a master clock input with a PLL, I guess it also is able to use the same PLL for the I2S incoming clock, so it looks to me that the TAS5518 syncs to any incoming clock thru PLL. I guess that getting to know how good is that PLL at rejecting incoming jitter is a question for TI.
Someone at Sumoh has told me they have plans for some more powerful amps based on Purepath HD chips, and he has pointed me out to this amp:
HFX - RipAMP 2.1
From what sumoh says it seems they design and sell OEM amps for other brands and this must be one of them.. They say they will sell the same amp + Spdif + volume control + remote on their own line of products. It is based on purepath HD chips, it looks like TAS5614, but not sure.. I don't know about the PCM to PWM modulator.
Regarding volume control on the TAS5518, I had a look at the datasheet and it seems like the -127dB/+18dB is DSP based (DAP as TI names it), I'm not sure, but I guess it is plain LSB truncation (I should ask TI), so AFAIK that should be OK for 16bit audio, but not good for 24bit audio, I haven't seen anything regarding those 24dB of extra dynamic range that could be used for volume control, I'll print the datasheet and I'll give it a deeper read.
Regarding clocking and incoming jitter rejection I've seen in the datasheet that the TAS5518 needs its own crystal clock and it has a master clock input with a PLL, I guess it also is able to use the same PLL for the I2S incoming clock, so it looks to me that the TAS5518 syncs to any incoming clock thru PLL. I guess that getting to know how good is that PLL at rejecting incoming jitter is a question for TI.
Last edited:
Sorry for not showing up after I've revived this thread. I wasn't aware of the many replies in it.
I am interested in this amp for the same reason as Blinddot, active crossovers. I feel 100w per channel is overkill but it can't hurt
.
However I share the same concerns regarding how this chip is managing the i2s stream. I wasn't aware is using a PLL circuit, this is not entirely a good news, since I was hoping it is not reclocking. I'll take a closer look at the datasheet.
By the way, are you aware of any other interesting digital amplifier chips?
For example i find interesting this one here:
http://koonlab.com/TAS5706/TAS5706.html
You may find on this site also another implementation of TAS5518
http://koonlab.com/TAS4i.html
I am interested in this amp for the same reason as Blinddot, active crossovers. I feel 100w per channel is overkill but it can't hurt
. However I share the same concerns regarding how this chip is managing the i2s stream. I wasn't aware is using a PLL circuit, this is not entirely a good news, since I was hoping it is not reclocking. I'll take a closer look at the datasheet.
By the way, are you aware of any other interesting digital amplifier chips?
For example i find interesting this one here:
http://koonlab.com/TAS5706/TAS5706.html
You may find on this site also another implementation of TAS5518
http://koonlab.com/TAS4i.html
Hi SunRa
when you said active xover, I guess you meant digital xover (which I guess should be considered active anyway).
Regarding clocking strategy I think I'd be very happy if the TAS5518 reclocks the incoming I2S clock to its own crystal.
AFAIK the best clock for any converter is its own crystal converter, any external clock would show worst jitter figures despite how accurate it originally was at the source, specially when the clock has to be embedded within the SPDIF protocol and has to go thru coupling transformers and recovery circuitry in order to convert again to I2S at destination.
that is common knowledge in recording studios, sometimes in recording studios due to the need of having multiple AD or DA conversions simultaneously we use several converter boxes, so the need of a master clock unit sending wordclock to all the boxes can't be avoided, and we need to rely on how good is the jitter rejection circuitry on the converter boxes in order to reject all the jitter induced on its way to the converter. But ideally converters should use its internal crystal clock, or at least that is what I understand from Dan Lavry's comments.
when using a DA converter for listening same principles apply, we can assume (even more if it is spdif clock instead of wordclok) that the incoming clock it is going to be fairly jittery, so we have to rely on the jitter rejection system on the DA, for example Dan Lavry on his hi-end gold series DA converter for mastering and audiophile listening uses a very good reclocking system that implies some kind of data buffering, it causes a slight extra latency of a few milliseconds when is turned on, in return the jitter is virtually eliminated at the DA converter.
My concerns regarding clocking strategy on the TAS5518 is more related to the performance of this PLL circuitry regarding jitter rejection (I think I should ask TI about that), In fact knowing that it uses a local crystal clock is kind of reassuring for me. But to be honest I don't deeply know about all these sync circuits at a design level, so please feel free to correct me if you think I'm falling on any kind of misconception, my source of knowledge are designers comments, not a designer myself, I could easily got them wrong
Regarding volume control, I read from the TAS5518 datasheet:
"32-Bit Processing PWM Architecture With 40 Bits of Precision"
So I'd like to think that this means that its +18dB/-127dB digital gain capability doesn't loose precision for 24bit signal from +18dB/-78dB
my maths are as follows: 40bit -24bit = 16bit
since 16bit is equivalent to 96dB, with 40bit fixed point maths you could digitally control 24bit signal gain from +18dB down to -78dB without loosing anything. If I'm right it'd be great news
In fact, I believe the NAD M2 does it in a similar DSP fashion, I think I should ask TI anyway.
BTW. I got more information about the sumoh amp, they told me is gonna be a TAS5518 + TAS5631 design with the chip's digital volume control, it will have a 36V PSU, don't know yet if it is going to be a SMPS o a linear one, I'm guessing a SMPS, based on their previous designs.
I've been having a look at the TAS5631 evaluation board, it uses TAS5518 as a PWM modulator, Ii think all I need to complete the amp would be a PSU, an SPDIF to I2S board and a USB to I2C board for controlling the TAS5518 with it's own software.. am I right? is that all I would need? would I need to keep the board conected to a computer in order work or control the volume?
thanks
when you said active xover, I guess you meant digital xover (which I guess should be considered active anyway).
Regarding clocking strategy I think I'd be very happy if the TAS5518 reclocks the incoming I2S clock to its own crystal.
AFAIK the best clock for any converter is its own crystal converter, any external clock would show worst jitter figures despite how accurate it originally was at the source, specially when the clock has to be embedded within the SPDIF protocol and has to go thru coupling transformers and recovery circuitry in order to convert again to I2S at destination.
that is common knowledge in recording studios, sometimes in recording studios due to the need of having multiple AD or DA conversions simultaneously we use several converter boxes, so the need of a master clock unit sending wordclock to all the boxes can't be avoided, and we need to rely on how good is the jitter rejection circuitry on the converter boxes in order to reject all the jitter induced on its way to the converter. But ideally converters should use its internal crystal clock, or at least that is what I understand from Dan Lavry's comments.
when using a DA converter for listening same principles apply, we can assume (even more if it is spdif clock instead of wordclok) that the incoming clock it is going to be fairly jittery, so we have to rely on the jitter rejection system on the DA, for example Dan Lavry on his hi-end gold series DA converter for mastering and audiophile listening uses a very good reclocking system that implies some kind of data buffering, it causes a slight extra latency of a few milliseconds when is turned on, in return the jitter is virtually eliminated at the DA converter.
My concerns regarding clocking strategy on the TAS5518 is more related to the performance of this PLL circuitry regarding jitter rejection (I think I should ask TI about that), In fact knowing that it uses a local crystal clock is kind of reassuring for me. But to be honest I don't deeply know about all these sync circuits at a design level, so please feel free to correct me if you think I'm falling on any kind of misconception, my source of knowledge are designers comments, not a designer myself, I could easily got them wrong
Regarding volume control, I read from the TAS5518 datasheet:
"32-Bit Processing PWM Architecture With 40 Bits of Precision"
So I'd like to think that this means that its +18dB/-127dB digital gain capability doesn't loose precision for 24bit signal from +18dB/-78dB
my maths are as follows: 40bit -24bit = 16bit
since 16bit is equivalent to 96dB, with 40bit fixed point maths you could digitally control 24bit signal gain from +18dB down to -78dB without loosing anything. If I'm right it'd be great news
In fact, I believe the NAD M2 does it in a similar DSP fashion, I think I should ask TI anyway.
BTW. I got more information about the sumoh amp, they told me is gonna be a TAS5518 + TAS5631 design with the chip's digital volume control, it will have a 36V PSU, don't know yet if it is going to be a SMPS o a linear one, I'm guessing a SMPS, based on their previous designs.
I've been having a look at the TAS5631 evaluation board, it uses TAS5518 as a PWM modulator, Ii think all I need to complete the amp would be a PSU, an SPDIF to I2S board and a USB to I2C board for controlling the TAS5518 with it's own software.. am I right? is that all I would need? would I need to keep the board conected to a computer in order work or control the volume?
thanks
Last edited:
The TAS5631 evaluation kit comes with a SPDIF input board, and a USB interface for computer control. All you need is a power supply.
Once the TAS5518 has been configured using I2C no further communication needs to happen unless you want to change somethingng. On mine, once the system turns on, a microcontroller sends out the I2C configuration commands. After that nothing gets sent unless the volume or mixing is changed. A USB to I2C board may be overkill... I'm not 100% certain how the TI input board is set up, but in theory, once the configuration is sent (and assuming no volume changes) the I2C interface could be disconnected.
I think I would look at the volume range a little differently. If you assumed your source was capable of producing the maximum output range, then you can't increase the volume without clipping. So then you would get the full 16 bits to attenuate before you lose any data....
Internally, the 5518 uses "48-Bit Processing Architecture With 76 bits of Precision for Most Audio Processing Features". Which is why they can achieve the high volume range without any data loss. You can do nearly anything you want internally without worry about loss of information.
So you wouldn't lose any data until the PWM output, but the output is already insane... With a 36V supply, 24bits is enough to specify 2.1 microvolts... or 0.06ppm!... The output offset will be in the 10 millivolt range, so a few microvolts... a precision 20bit DAC is 1ppm and runs $35 each. $3k ultra low distortion (<.001% THD) function generators test equipment only use 20bits...
What I'm getting at, is once you get analog output, there are far more things that influence the resolution and quality of the signal than the digital processor... Even though 24 bits in the digital world isn't extreme, in the analog world it is quite profound...
For the PLL and local clock. The evaluation boards are using the TAS5518 and their reference PLL filter and 30ppm crystal. They achieve the 0.03% THD and >105dB dynamic range. To me, that says you don't need to change their design in order to achieve similar results. Are those results not good enough? or are you looking to improve their design and get better results? Do you have evidence that their PLL circuit does not sufficiently reject jitter?
Sorry if that came across as harsh... I guess I don't understand the need to have such extremely high output resolution and go into the details of reclocking when the result is already very nice...
Once the TAS5518 has been configured using I2C no further communication needs to happen unless you want to change somethingng. On mine, once the system turns on, a microcontroller sends out the I2C configuration commands. After that nothing gets sent unless the volume or mixing is changed. A USB to I2C board may be overkill... I'm not 100% certain how the TI input board is set up, but in theory, once the configuration is sent (and assuming no volume changes) the I2C interface could be disconnected.
I think I would look at the volume range a little differently. If you assumed your source was capable of producing the maximum output range, then you can't increase the volume without clipping. So then you would get the full 16 bits to attenuate before you lose any data....
Internally, the 5518 uses "48-Bit Processing Architecture With 76 bits of Precision for Most Audio Processing Features". Which is why they can achieve the high volume range without any data loss. You can do nearly anything you want internally without worry about loss of information.
So you wouldn't lose any data until the PWM output, but the output is already insane... With a 36V supply, 24bits is enough to specify 2.1 microvolts... or 0.06ppm!... The output offset will be in the 10 millivolt range, so a few microvolts... a precision 20bit DAC is 1ppm and runs $35 each. $3k ultra low distortion (<.001% THD) function generators test equipment only use 20bits...
What I'm getting at, is once you get analog output, there are far more things that influence the resolution and quality of the signal than the digital processor... Even though 24 bits in the digital world isn't extreme, in the analog world it is quite profound...
For the PLL and local clock. The evaluation boards are using the TAS5518 and their reference PLL filter and 30ppm crystal. They achieve the 0.03% THD and >105dB dynamic range. To me, that says you don't need to change their design in order to achieve similar results. Are those results not good enough? or are you looking to improve their design and get better results? Do you have evidence that their PLL circuit does not sufficiently reject jitter?
Sorry if that came across as harsh... I guess I don't understand the need to have such extremely high output resolution and go into the details of reclocking when the result is already very nice...
No man, it doesn't come across as harsh, don't worry, but I do think you are just misunderstanding me here. I guess it is my fault, I'm not being clear enough.
Regarding clocking I just wanted to know about clocking strategy. Now I know it syncs its internal clock to the incoming clock with a PLL, OK, now I just wonder about the jitter rejection performance of it, but I'm not saying TAS5518 does it good or bad, I do not know, I'm just wondering, you know, spec figures and stuff like that, isn't wondering ok?. I'm not even saying it is a super important issue, in my opinion (not a very reliable one) power supply, layout, analogue (power) stage or reconstruction filters are more important factors for the sound quality than a bit more or less jitter rejection, Im sure the figures are more than decent.
Regarding volume control I also was wondering about performance of digital volume control cos I know I'll be using using it with quite a bit of attenuation, as I already explain to you I want to overpower my speakers a little (or at least to full-power them as much as possible), therefore using a fairly amount of attenuation (about -70dB/-30dB aprox)
I've been reading the datasheet and it seems that for good low level resolution TI is recommending to use PSVC or both digital and PSVC combined. So it looks like the convenience of digital volume control comes at a cost. I'll have to listen to it in order to say (and I'll do), but it seems like PSVC will be better choice for me.
From TAS5518 datasheet:
Thanks for your help.
Regarding clocking I just wanted to know about clocking strategy. Now I know it syncs its internal clock to the incoming clock with a PLL, OK, now I just wonder about the jitter rejection performance of it, but I'm not saying TAS5518 does it good or bad, I do not know, I'm just wondering, you know, spec figures and stuff like that, isn't wondering ok?. I'm not even saying it is a super important issue, in my opinion (not a very reliable one) power supply, layout, analogue (power) stage or reconstruction filters are more important factors for the sound quality than a bit more or less jitter rejection, Im sure the figures are more than decent.
Regarding volume control I also was wondering about performance of digital volume control cos I know I'll be using using it with quite a bit of attenuation, as I already explain to you I want to overpower my speakers a little (or at least to full-power them as much as possible), therefore using a fairly amount of attenuation (about -70dB/-30dB aprox)
I've been reading the datasheet and it seems that for good low level resolution TI is recommending to use PSVC or both digital and PSVC combined. So it looks like the convenience of digital volume control comes at a cost. I'll have to listen to it in order to say (and I'll do), but it seems like PSVC will be better choice for me.
From TAS5518 datasheet:
Thanks for your help.
Last edited:
Ok, I understand now. Thanks for explaining.
I do agree that the power supply, layout, and filters have the biggest impact.
Yes, I would agree that using PSVC will be a major benefit, otherwise you would effectively be losing resolution (because the PWM resolution is fixed, and so the smallest output is then given by the supply voltage) leading to the quiet sounds getting lost. You'll have to use both digital and PSVC to have the resolution you'll want, but I think it'd be worth it.
When I built my amp I found that the power supply was the largest/most expensive part of the system, so I've settled on a pretty weak solution, and I'd like to find a better one...
Are you going to get a TI eval board and try it out?
I do agree that the power supply, layout, and filters have the biggest impact.
Yes, I would agree that using PSVC will be a major benefit, otherwise you would effectively be losing resolution (because the PWM resolution is fixed, and so the smallest output is then given by the supply voltage) leading to the quiet sounds getting lost. You'll have to use both digital and PSVC to have the resolution you'll want, but I think it'd be worth it.
When I built my amp I found that the power supply was the largest/most expensive part of the system, so I've settled on a pretty weak solution, and I'd like to find a better one...
Are you going to get a TI eval board and try it out?
No, I wont get the eval board just yet, I want to wait and see if the purepath HD market has a few more options, I'll wait for sumoh TAS5518 + TAS5631 amp.. I'd also love someone making DIY modules. I've written to Class D Audio Home page in order to find out if they have plans for digital controlled purepath HD amps, I think I'll send a few more emails , maybe miniDSP, it'd be very atractive a DSP xover + two way power DAC from miniDSP with the purepath HD chips
I just think it is a bit too early yet, this technology is just getting ready, I want to see what comes up to the market.. I know I definitely wont be making my own board, so I'll wait til I can buy a DIY board or an already made amp.
I think I'll buy a full range system (sumoh, TI eval board, etc..) so I can evaluate the technology and use it for the living room afterwards, then I'll be very interested in a DSP xover + power DACs with PSVC for a 2way speaker in the same chassis for the studio if I like, since I'll be upgrading speakers to dynaudios or PMCs in the near future
I'm also wondering about intersil D2audio chips, Zetex and apogee DDX, but there is no much info about those chips being used on any fairly affordable amps (NAD M2 is not an option for me, and is not even a 2way system for the price!!)
regards.
I just think it is a bit too early yet, this technology is just getting ready, I want to see what comes up to the market.. I know I definitely wont be making my own board, so I'll wait til I can buy a DIY board or an already made amp.
I think I'll buy a full range system (sumoh, TI eval board, etc..) so I can evaluate the technology and use it for the living room afterwards, then I'll be very interested in a DSP xover + power DACs with PSVC for a 2way speaker in the same chassis for the studio if I like, since I'll be upgrading speakers to dynaudios or PMCs in the near future
I'm also wondering about intersil D2audio chips, Zetex and apogee DDX, but there is no much info about those chips being used on any fairly affordable amps (NAD M2 is not an option for me, and is not even a 2way system for the price!!)
regards.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.