Sorry if my search skills have failed. I'm looking for a good sounding not especially high power set of amp modules for some three way active speakers.
I want to keep it all digital up to the drive units.
I've looked at a few of the best known names, but they all seem to be designed for an analog input...
Anyone want to give me some other names to look for?
I'm looking for boards with or without power supply.
I want to keep it all digital up to the drive units.
I've looked at a few of the best known names, but they all seem to be designed for an analog input...
Anyone want to give me some other names to look for?
I'm looking for boards with or without power supply.
Boards based on MA12070P might fit the bill. There are some configured as RPi hats.
There is also this PCB layout : MA12070p amplifier -
EasyEDA
and this eval module : HIFICLASSDMERUS MA12070 2X80W-
There is also this PCB layout : MA12070p amplifier -
EasyEDA
and this eval module : HIFICLASSDMERUS MA12070 2X80W-
I read some reviews that suggested that while nominal output goes as high as 80w it starts to distort over 20w.so for my purposes it's wildly underpowered.
The obvious solution is to look for a higher power chip with the same architecture. But there isn't one.
A more crippling problem is that it has a built in DAC. The idea is to bypass the DAC and go straight from I2S to PWM. However I'm reading that only a tiny handful of devices do this.
This is the reason for choosing class D, so it looks like I'm out of luck and it's back to the class A/B :-(
The obvious solution is to look for a higher power chip with the same architecture. But there isn't one.
A more crippling problem is that it has a built in DAC. The idea is to bypass the DAC and go straight from I2S to PWM. However I'm reading that only a tiny handful of devices do this.
This is the reason for choosing class D, so it looks like I'm out of luck and it's back to the class A/B :-(
Further there's a long piece on here, which I can't link to on this phone, written by Bruno Putzeys, who really ought to know because he designs this stuff, explaining why, actually, skipping the analog stage causes more problems than it solves. Hence why there are almost no designs that do so.
So my clever idea is a bust :-(
So my clever idea is a bust :-(
There is STA323 which is NRND but probably not enough power for you. But avoids the built-in DAC.
https://www.st.com/resource/en/datasheet/sta323w.pdf
https://www.st.com/resource/en/datasheet/sta323w.pdf
So part of your difficulty - the difficulty - is that you need lots of power, along with the I2s to PWM, DSP, etc in one of these amps. Is there any way a lower powered amp would be useful to you, if only to judge the sonic character of the digital to analog process it uses?
I say this because as we know, not all DAC processes sound the same. It's validating when a particular process implementation has the better sonics logic and reason say it should. However, as in many things audio, the final judgement must be by the ear. Because sometimes the ear distinguishes things logic and reason say it shouldnt, well this means to me you have to try it for yourself.
You dont have a set of nicely efficient two way speakers your could try these designs out on, just to get the sonic flavor for what they're doing with this "Power" DAC implementation. If how it sounds to you is a go, then you can start to pursue ways to get the power levels you're looking for.
I see two ways to do it. One is that an H bridge has optimum loading for the lowest distortion it can do. Provide that loading and just...tap off one of the output terminals with an attenuator down to RCA line level and capacitively couple that to the input of the big A/B analog amplifier. You could even transformer couple it to provided a balanced output signal.
The other is to find someone willing to implement a board with a TAS5634 that'll take the output from any existing class d amplifier's H bridge terminals, and boost that out into the 300W range. I'm glad you brought it up; perhaps there'd be a real market for such a thing, 2 channels in a box with power supply. Or just 2 channels on a built PCB you could do your own integration with.
I wouldnt say you're a bust without further investigation.
I say this because as we know, not all DAC processes sound the same. It's validating when a particular process implementation has the better sonics logic and reason say it should. However, as in many things audio, the final judgement must be by the ear. Because sometimes the ear distinguishes things logic and reason say it shouldnt, well this means to me you have to try it for yourself.
You dont have a set of nicely efficient two way speakers your could try these designs out on, just to get the sonic flavor for what they're doing with this "Power" DAC implementation. If how it sounds to you is a go, then you can start to pursue ways to get the power levels you're looking for.
I see two ways to do it. One is that an H bridge has optimum loading for the lowest distortion it can do. Provide that loading and just...tap off one of the output terminals with an attenuator down to RCA line level and capacitively couple that to the input of the big A/B analog amplifier. You could even transformer couple it to provided a balanced output signal.
The other is to find someone willing to implement a board with a TAS5634 that'll take the output from any existing class d amplifier's H bridge terminals, and boost that out into the 300W range. I'm glad you brought it up; perhaps there'd be a real market for such a thing, 2 channels in a box with power supply. Or just 2 channels on a built PCB you could do your own integration with.
I wouldnt say you're a bust without further investigation.
Last edited:
I was thinking about the issue at one point and it became very obvious why it has to be converted to analog somewhere along the way - negative feedback. Ultimately it needs some sort of mechanism to compare the output voltage to the input voltage. Unless it is open loop.
There is a module by one of our diyer that does that including DSP. Might fit your bill. Look in Vendors bazaars ZOaudio I think.
Oon
I think there are mainly 2 factors limiting "pure digital" power amps when the digital pwm signal enters the analogue world at the very end: The analogue supply voltage - including rdson of the switches - must be "perfect". And the pwm timing must be equally "perfect" as well. Here the negative feedback comes in, compensating all these analogue imperfections.
When I look at the distortion curves vs power out for these amps, they all go "straight up" at the end. I assume this is all the analogue factors coming into play at a point where there's no longer any room to compensate. So my tricks for more power out operate the I2S input amp ideally, while something else takes the burn.
Does "skipping the analog stage" still cause problems when such an amp is sweet spotted in operation? I've read where some dont like them, but I highly suspect that is when operated right out at their maximum power capability; see? It sucks! Well, yeah...
Of course everyone wants a monolithic circuit that's perfect in every way. Being this is DIY, I'd think there'd be some room for "cement half A to half B" - particularly if an I2S input with DSP and high power out is needed.
Of course, there's also "what fringe of the audio marketplace would understand an I2S input amplifier - versus those who get what a pair of RCA inputs do?". That just may have something to do with why so few are willing to stick their neck out on it. Even my Zoudio amp was provisioned with analog in, but that's not its forte' - digital in is.
Does "skipping the analog stage" still cause problems when such an amp is sweet spotted in operation? I've read where some dont like them, but I highly suspect that is when operated right out at their maximum power capability; see? It sucks! Well, yeah...
Of course everyone wants a monolithic circuit that's perfect in every way. Being this is DIY, I'd think there'd be some room for "cement half A to half B" - particularly if an I2S input with DSP and high power out is needed.
Of course, there's also "what fringe of the audio marketplace would understand an I2S input amplifier - versus those who get what a pair of RCA inputs do?". That just may have something to do with why so few are willing to stick their neck out on it. Even my Zoudio amp was provisioned with analog in, but that's not its forte' - digital in is.
Yup. My interest in Class D was in eliminating the DAC. But it seems eliminating the DAC is not a smart thing to do. So we'll write that idea off.
Thanks all!
Thanks all!
The best amp I have heard is the tact Millennium mk3. It’s digital open loop high end build. Digital input with closed loop have analogue input from the feedback. The true digital still need conversion, but it’s in the digital domain at input but the PWM stage is still analogue. Open loop also requires a very good stabilized PSU since all the bad stuff from the PSU goes directly to the PWM stage.
Try search for the article
PSRR: The Real Story about Closed- and Open-Loop Class-D Amplifiers
Try search for the article
PSRR: The Real Story about Closed- and Open-Loop Class-D Amplifiers
Last edited:
One cant get from D to A without a DAC. Class-D is not D, it's A. But I have (had?) also the hope that some gains would be possible by going D->PWM, but alas, there seems not to.
The Millenium i as I understand it a DAC with a lot of current capability - a "power" DAC?
There should be a need for a better, modern STAxxx in say 2-300W territory.
//
The Millenium i as I understand it a DAC with a lot of current capability - a "power" DAC?
There should be a need for a better, modern STAxxx in say 2-300W territory.
//
It’s equibit as you will also find in some Texas instruments chips. The chips were used in some Panasonic AV receivers
There is also the ddx Technology from ST
Hifime UDP320 USB and SPDIF TRUE DIGITAL POWER AMPLIFIER 2X150W - Hifime Audio
There is also the ddx Technology from ST
Hifime UDP320 USB and SPDIF TRUE DIGITAL POWER AMPLIFIER 2X150W - Hifime Audio
Last edited:
I don't think that is entirely true either. I think the thing to gain, off all thing.... is volume control. Having a digital volume control certainly has its benefit vs the pot. Unfortunately I don't have data to prove it. The problem with the pot is the wiper to resistor contact is always a suspect. And anybody who has done low level electronics will tell you, if it ain't gold, there is a lot of **** that is going to happen....
Oon
In an integrated amp, sure. But here the chain is digital streaming source, digital DSP crossover (with over all volume control and software channel volume control), three two channel power amps, three way speakers.
Or, as it turns out between DSP and amp there is a DAC, and if there is a DAC then the choice is between one I can't choose, built into the chip, or one I can spec myself. I prefer the latter.
So then you are left only with the traditional advantage of class D, which is cheaper ancillaries for a given power output. As against which it doesn't have a wonderful audiophile reputation, and while that may be conservatism there are enough technologies new to me in this project and I'm not keen to introduce another variable.
I'm very open minded though. I've got ages to think about this before I have to power up an amp. I could do class A tweeters, trad A/AB mid, class D bass. We'll see.
PowerDACs or FDA can sound quite good. I tried a few in the past like Nuforce DDA100, Wadia 151 etc. and liked them so I continued that path (although nothing is perfect of course). Personally I think this is the future so I have one for everyday use. Developments go fast so eventually it will become dominant (as it is cheapest technology too).