Yep sounds like my understanding - the chip puts out complementary PWM at 3.3V amplitude, this can be fed to logic-level FETs to get any output amplitude you want, limited only by the working voltage of those FETs. I figure its worth seeing if higher voltage (and higher step-down ratios of output transformer) lead to better dynamics. My target in such experiments would be of the order of 100V, above that power supply caps start to have poorer specs in ESR.
It only needs to be filtered out for efficiency's sake, otherwise power's potentially wasted in the transducers.
That's the rate for all the internal processing - it has lots of signal processing options available (high/low pass for XO, EQ, limiter etc) prior to the PWM conversion which I think probably uses 8X OS.
Doesn't look to me that the ASRC is bypassed at any time, so upsampling on your computer won't have much effect I think. Notice its an async converter so its a little different from the software running on a PC which is just an SRC.
A USB interface for a DAC is a little diffrerent coz in that application you'll want to isolate the DAC from noise coming from the PC. In this application no such isolation is called for because you're only driving speakers (which are isolated). So I reckon for a FDA the USB interface quality is immaterial, providing it delivers the right bits.
Save
Well, there is also Devialet and Lyngdorf among others. But much more expensive ...
When I add up all the costs of buying and modding semi complete cheap FDA (just one never is enough 
) and I take the high sound quality and useful features of the 151 into account it is the amp with best price/quality ratio. I took the time to compare and sold many FDA I had after that. I looked carefully for new designs but the 151 has quite modern electronics. For the first time I bought an older device as it simply sounds better and it is user friendly. Built in PSU against those power bricks, 4 inputs, small and good looking case, clear display, metal remote control etc. etc. It is sold here for under 500 Euro as the importer went bankrupt.
) and I take the high sound quality and useful features of the 151 into account it is the amp with best price/quality ratio. I took the time to compare and sold many FDA I had after that. I looked carefully for new designs but the 151 has quite modern electronics. For the first time I bought an older device as it simply sounds better and it is user friendly. Built in PSU against those power bricks, 4 inputs, small and good looking case, clear display, metal remote control etc. etc. It is sold here for under 500 Euro as the importer went bankrupt.
Last edited:
Same thoughts. It has solid aluminium heatsinks without fins. I don't bother too much as 2 x 25W is more than enough for me but it is uncommon indeed.
In normal operation it becomes warm not hot. Excellent quality electrolytic caps so apart from the strange RCA SPDIF input connectors it is a well built device. The holes in the back cover are way larger than normal and I found only panel mount Neutrik NBB75SI silver plated BNC connectors that fit. The original ones are mounted from the outside and plugging and unplugging makes them come loose. Slight enlargening the holes from 12 mm to 12.7 mm with a round file was necessary. Full aluminium case so that was a breeze.
Only annoyances are the power switch at the back side and the lack of SPDIF isolation transformers. I can't reply any detailed question as I had an accident.
Anyway this device is worth the reduced price big time. They are dumped here for prices between 400 and 475 Euro.
In normal operation it becomes warm not hot. Excellent quality electrolytic caps so apart from the strange RCA SPDIF input connectors it is a well built device. The holes in the back cover are way larger than normal and I found only panel mount Neutrik NBB75SI silver plated BNC connectors that fit. The original ones are mounted from the outside and plugging and unplugging makes them come loose. Slight enlargening the holes from 12 mm to 12.7 mm with a round file was necessary. Full aluminium case so that was a breeze.
Only annoyances are the power switch at the back side and the lack of SPDIF isolation transformers. I can't reply any detailed question as I had an accident.
Anyway this device is worth the reduced price big time. They are dumped here for prices between 400 and 475 Euro.
Last edited:
So are you saying this chip's PWM switching doesn't generate EMI/EMF on the speaker wires, unlike say a TPA3116? How is the switching between the two different such that one creates EMI and one does not?
Are you sure the actual PWM conversion does 8X OS? That just seems like a strange design to me... if you're ultimately going to oversample to 352.8kHz (or whatever), and the input signal is 352, why down-sample to 48, only to re-up-sample back to 352? Plus I would assume most DSP algorithms produce better results with higher-resolution data...
If the chip wants to work at 48kHz, and you feed it 48kHz, why would it need to manipulate the waveform at all? I would think the ASRC at that point should simply be a "pass through".
At any rate, what I was thinking was: going from 44.1kHz to 48kHz requires re-constructing a lot of data points in the wave form. The Intersil chip has to do this basically in real-time, and is confined to whatever algorithm and processing power is built into the IC. But comparatively speaking, I have effectively "unlimited" processing power if I do the SRC in software, as I can do it as a batch job... I can try all kinds of fancy re-sampling algorithms.
Dunno, just thinking out loud here.
Going strictly off hearsay of others' subjective comments, but... one example, I've seen several people now say the new XMOS USB receivers improve the Soekris dam1021, which already has I2S isolation. I've also seen people say their DAC is improved with non-isolated USB receivers using the newer XMOS chips. Just guessing here, but I wonder if it's simply the re-clocking that makes it better? If the Intersil chip does it's own reclocking---and does a good job of it---then yeah, fancy inputs might not make much of a difference.
At any rate, from the pics, the USB receiver looks modular, or at least has pins for easy experimentation. FWIW, I noticed the Alientek D8 also has a modular USB receiver. I was debating between it and the Popu.
When I add up all the costs of buying and modding semi complete cheap FDA (just one never is enough
Do you still have and use the Wadia?
Ignoring cost, of all the cheap FDA devices you played with, which came closest to the Wadia in terms of sound quality? And how close did it come?
Also interesting is that it uses the TAS5162. I've personally never heard a device that uses that chip, but based on subjective internet hearsay, I feel like I've generally read negative comments about TI's TAS-series FDA chips. Though, perhaps like so many things (everything?) it's more about the implementation than the chip itself.
Based on JP's comments, it sounds like the diyAudio community has been tasked with designing a Wadia 151-inspired FDA board.
Since the best one of my cheap ones is the modified D802 I compared it to the 151. I kept my modified D802 as I like that one and it sounds very good but it is no match for the Wadia. I particularly was fond of Nuforce DDA-100 but it was also no match for the Wadia. DDA-100 has better looks and the way of operating it is the best I experienced till now with one button control.
I made NO real life comparisons with newer gear as I didn't have the time or the 151 would be gone. The idea was to buy an Alientek D8 and I even thought of the QULOOS QA690 to compare those but then the 151 came around. I still use the 151 as my main amp. I think it is a very complete package that just delivers. The reason it is different probably is the Wadia DSP and its up sampling as IMO this TAS chip is not very special. It seems there is more to FDA than just the power stage but we already knew that
* Please note that I only use SPDIF. The 151 has isosynchronous USB (supposedly Wadia style) but I haven't even tried it.
I made NO real life comparisons with newer gear as I didn't have the time or the 151 would be gone. The idea was to buy an Alientek D8 and I even thought of the QULOOS QA690 to compare those but then the 151 came around. I still use the 151 as my main amp. I think it is a very complete package that just delivers. The reason it is different probably is the Wadia DSP and its up sampling as IMO this TAS chip is not very special. It seems there is more to FDA than just the power stage but we already knew that
* Please note that I only use SPDIF. The 151 has isosynchronous USB (supposedly Wadia style) but I haven't even tried it.
Last edited:
Ha, I wasn't thinking about EMI when I wrote that, mainly because I use relatively short speaker wires. The filtering needed for EMI reduction is much milder than the normal output filtering which attacks the carrier frequency (typically 300-400kHz). Speaker wires make poor antennas in that frequency band so EMI filtering can use smaller value (and size) inductors to attack mostly higher harmonics of the carrier.
So answer your question, there's no difference that I'm aware of in EMI except that the TPA3116 is capable of more than modulation mode, some of which are worse for EMI than others.
Makes sense to me - if you processed the 352 rate internally you'd have to switch off so many processing options because the MIPS are fixed at a fixed clock. Which would be a nightmare, so everything's down-sampled to a rate which the DSP can cope with.
That looks to be an assumption too far in this instance. Check out IIR filters and how many bits are required to do bass EQ with sufficiently low noise (Kendall Castor-Perry wrote a nice blog article about this I recall). As the sample rate increases you need more bits in the math to maintain the same noise because one of the operations involved is differencing (subtracting consecutive samples). At higher rates those differences reduce, leading to a requirement for higher precision to capture them.
Its a good question and as always the devil's in the details. The idea that all 48kHzs are identical is just an assumption - in practice they're not identical down to infinite numbers of decimal places coz they're normally crystal derived. Crystals have tolerances and they have drift- digital watches using them drift apart over days. Which matters in digital audio as when sample rates drift against one another glitches get introduced.
So in practice there's never a pass-through when the ASRC is asynchronous.
Yep, makes sense but you'll still be bottlenecked by the ASRC no matter which sample rate you choose. And this assumes that the ASRC is indeed the bottleneck in terms of SQ.
Based on JP's comments, it sounds like the diyAudio community has been tasked with designing a Wadia 151-inspired FDA board.
I'd be interested in helping with this - I reckon what's needed is open source PCM->PWM conversion code. Then the main works can be handled by a DSP-type microcontroller (Infineon XMC4200 comes to mind) and we'd have a very nice platform for DIYers to play with both hardware and software. From a quick glance at the internal pics of the Wadia seemed they're using a custom chip which perhaps only Wadia themselves have access to (it has their logo on top).
Save
I apologize if this counts as thread hijacking, but I am really interested in fully digital amps (just seems like a more logical technology to me) and have a few questions for ya'll.
In general, what do you guys use for input? Optical, coax, USB? It appears that most of the amps mentioned here can run a higher sampling rate on optical/coax rather than USB. In your opinion, does this drastically change the music quality? Or is it above what is the limiting factor?
If opti/coax is significantly better, what is the best way to get that from my computer (main source)? I don't think most affordable sound cards can output digital that high a resolution anyways right? And getting an external DAC with opti out seems like a redundant solution.
In general, what do you guys use for input? Optical, coax, USB? It appears that most of the amps mentioned here can run a higher sampling rate on optical/coax rather than USB. In your opinion, does this drastically change the music quality? Or is it above what is the limiting factor?
If opti/coax is significantly better, what is the best way to get that from my computer (main source)? I don't think most affordable sound cards can output digital that high a resolution anyways right? And getting an external DAC with opti out seems like a redundant solution.
That's not true. Many/most of the amplifiers discussed in this thread support USB sources up to 192kHz - FX-Audio D802, I.AM.D V200, Alientek D8-XMOS.
And the SMSL Q5 PRO will do USB up to 96kHz.
It's only the JC-SZ80 and entry-level Alientek D8 which have a USB limitation of 48kHz.
But as jean-paul has hinted, SQ is not ultimately defined by maximum supported bitrate.
And the devil is in the implementation details
E.g both my v200 and d802 do 192khz on both USb and coax. Yet the USB implementation on the d802 is not good while the coax spdif is rather fine. This matches how they sound to me i.e. on USB I prefer the v200, on coax spdif I prefer the d802. Alientek D8 (Xmos USB version) should be good for both - but my newly arrived one is on temporary loan elsewhere and I haven't listened to it yet.
Last edited:
Well, it hasn't reached home yet or been plugged in for listening but I have taken it apart to have a look and it is interesting
Construction-wise it's very neat and tidy.
The Xmos USB board is easily detachable and connected to the main board on one side by four USB pins, labelled on both boards. On the other side of the Xmos board are nine pins, however the main board socket only connects six of these. We have a potentially easy direct 12s entry point for this amp using that socket!
As the mainboard socket does not have those pins labelled (although the xmos board does) I'll enumerate them here for anyone purchasing the cheaper non-xmos version of the amp. Starting from 'P4' label (closest edge of socket to RCA connectors)
1) GND
2) SPD
3) MCLK
4) SCLK
5) SDIN
6) LRCLK
then missing connections for these on the Xmos board
7) RST (I think, first letter hard to read)
8) GND (first letter hard too read)
9) STA
With the xmos board removed there is Pulse PE65612NL transformer revealed labelled USB_SPDIF on the main board. Hopefully this means that both the USB and coax/spdif inputs are isolated, I couldn't visually trace the lines to establish this but its physical positioning on the board would fit such usage.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.