Review and Measurements of WesionTEK Khadas Tone Board DAC

[carlmart]

What host is that Khadas sent to Amir on his Tone test? What does it do?

 

[carlmart]

I have the Tone1. It sounds pretty good. Uses the Apodizing filter in the ESS DAC (not changeable). Here's the frequency and step response that I've measured (1 dB scale, so variations are not audible):

View attachment 140868

View attachment 140869

Some pre-ringing, but not horrible for a $99 (now $79) DAC. The Tone2 Pro has selectable filters, some of which have significantly better time-domain response. (I'll look around to see if I can find my test results. If so, I'll edit this post.)

As others have suggested, the Tone1 is far from the best these days for DACs under $200.
View attachment 140870

Note: The Topping D30 Pro really should not be on this list as it's $399. I guess it's just missing price information in this database.

The Tone2 Pro is in fact, by a fair margin, the best Amir has measured under $200. This is due, in part, to the availability of the novel balanced RCA outputs (balanced outputs are uncommon in DACs under $200). However, using them requires an additional expense of $49 - $79 for balanced RCA to XLR cables (I have both lengths…nicely built and worth the cost, IMHO).

I find the internal headphone amp to be adequate but not great (better, if you use the I2S input for power). What is great is the sound of the Tone2 Pro via balanced connection into a suitable external amp or preamp. I find that the HYBRID FAST filter has the best sound. Although you can buy the Tone1 + case + suitable interconnects for under $125, if you have a resolving system downstream, I think you'll find that the Tone2 Pro + balanced RCA cables ($250) is at least twice as nice.

View attachment 140873

Sorry, but how is the Tone 2 wired to have balanced output using an RCA socket?

 

[somebodyelse]

Sorry, but how is the Tone 2 wired to have balanced output using an RCA socket?

It's a custom connector with an extra contact, backward compatible with existing RCA cables for single end use, but doing balanced with its own cable connectors. Whether it'll end up getting more widely adopted is another matter.
https://www.khadas.com/balanced-rca

 

[Veri]

Sorry, but how is the Tone 2 wired to have balanced output using an RCA socket?

They have made their own connector, with backwards compatibility for unbalanced, and both hot/cold pins to have a real balanced output on this novel connector.

 

[dsnyder0cnn]

Sorry, but how is the Tone 2 wired to have balanced output using an RCA socket?

As others have mentioned, Khadas created a backwards compatible RCA connection by adding a third conductor. The RCA plug has a second insulated collar around the central pin that provides the third connection. They have patented this design but allow other manufacturers to use it without a license fee as long as they comply with their branding requirements.

More info here:

• https://www.khadas.com/post/introducing-the-balanced-rca-jack
• https://www.khadas.com/balanced-rca

Khadas offers decent quality Balanced RCA to XLR cables in 0.2m and 1m lengths for a reasonable price (I have both lengths). They also sell jacks and plugs for DIY enthusiasts and integrators, but, again, anyone can manufacturer their own as long as they agree to brand their products appropriately.

I think the problem they were trying to solve was adding balanced connections to a product that's too small to accommodate traditional 3-pin XLR jacks. They could have gone with the iFi Audio route with a 4.4mm Pentaconn jack or Switchcraft's mini XLR, but neither would have been backwards compatible with commonly available cables. I actually think this design is pretty cool.

 

[dsnyder0cnn]

I also have a tone 1 DAC. I would have preferred a linear phase sharp cutoff filter but I doubt I can hear the difference. The DAC is powered via the USB cable so I get a large pop in my speakers when the streamer is switched on and again when it is switched off. I'm looking into solutions for that issue but apart from that issue I'm happy.

I found my measurement files. I already shared the frequency response of the Apodizing filter on the Tone1. Here's what Linear Fast looks like on the Tone2 Pro (would probably be the same on Tone1 if enabling it were possible):

Here's what Impulse and Step response looks like for this filter:

Not terrible, but of course roughly equal pre and post ringing. There's some question about how audible this is, as you say. In contrast, here's the same plot for Hybrid Fast, which I seem to prefer:

Here, virtually no pre-ringing, but a little more early (up to ~700 μs) post-ringing. It works for me.

 

[kchap]

I found my measurement files. I already shared the frequency response of the Apodizing filter on the Tone1. Here's what Linear Fast looks like on the Tone2 Pro (would probably be the same on Tone1 if enabling it were possible):

View attachment 141614

Here's what Impulse and Step response looks like for this filter:

View attachment 141615

Not terrible, but of course roughly equal pre and post ringing. There's some question about how audible this is, as you say. In contrast, here's the same plot for Hybrid Fast, which I seem to prefer:

View attachment 141616

Here, virtually no pre-ringing, but a little more early (up to ~700 μs) post-ringing. It works for me.

Thanks for the traces. I'm not that concerned by the pre-echo. To me it is just the normal behaviour of a well designed brick wall filter. We often forget that a filter has already been applied during the A to D process; many issues that these novel filter designs are trying to prevent already exist as a result of the A to D phase.

 

[jtatknox]

I'm playing around with some Tone 1 projects and have run into quite a lot of misinformation on the forums. I have read it suggested that the Tone 1 draws about 150 mA current. I found that the output was heavily distorted unless I used a regulator that could supply 600 mA current.

I also don't really understand the I2S options on the Tone 1. Everywhere I read online indicates that I2S is only available on the 8 Channel Ext FPC Connector. I don't understand based on the documentation what pin is the I2S data line.

Has anyone actually gotten I2S signals out of this thing?

 

[Veri]

Has anyone actually gotten I2S signals out of this thing?

You mean into? In any case it seems like a difficult thing to achieve, yes. Not exactly easy/user friendly. Then again, unless you are a DIYer/modder I don't really see the point for I2S.

 

[somebodyelse]

I also don't really understand the I2S options on the Tone 1. Everywhere I read online indicates that I2S is only available on the 8 Channel Ext FPC Connector. I don't understand based on the documentation what pin is the I2S data line.

If you're interested in the DIY side you should really have a look at the schematic, and probably some of the other docs in their download section. There are two distinct I2S connections. The 'I2S Ext' connector leads to the 40 pin GPIO header only, giving access to the I2S on their VIM compute boards. If you don't have a VIM about the only thing you could use it for is an FPC breakout. The '8 Channel Ext' connects to the XMOS for I2S, I2C and some GPIO. IIRC this was entirely unused in the original firmware, and a later update added the option to use some of the GPIO for hardware switches to control volume. If you want to use it for anything else I think you'll have to write some custom XMOS firmware. The hardware volume thread has some info on that, and if you ask nicely occip might send you a copy of the custom source to get you started.

 

[rebbiputzmaker]

You mean into? In any case it seems like a difficult thing to achieve, yes. Not exactly easy/user friendly. Then again, unless you are a DIYer/modder I don't really see the point for I2S.

Just curious does your opinion apply to I2s in general, or just in regards to this DAC? In my system using the the Pi2 AES driving the X16 I am using I2s without issue and with excellent results.

 

[Veri]

In my system using the the Pi2 AES driving the X16 I am using I2s without issue and with excellent results.

In general, really.. I'm sure the set-up you mention works well, but I don't really see the point for I2S interfacing unless you have some very specific use cases..

 

[jtatknox]

If you're interested in the DIY side you should really have a look at the schematic, and probably some of the other docs in their download section. There are two distinct I2S connections. The 'I2S Ext' connector leads to the 40 pin GPIO header only, giving access to the I2S on their VIM compute boards. If you don't have a VIM about the only thing you could use it for is an FPC breakout. The '8 Channel Ext' connects to the XMOS for I2S, I2C and some GPIO. IIRC this was entirely unused in the original firmware, and a later update added the option to use some of the GPIO for hardware switches to control volume. If you want to use it for anything else I think you'll have to write some custom XMOS firmware. The hardware volume thread has some info on that, and if you ask nicely occip might send you a copy of the custom source to get you started.

Very useful link. I think I am not seeing any I2S data line on the 30 pin FPC because the 6 channels it provides, not 8, only come from USB and are not replicated from the SPDIF input. I will test this evening.

 

[jtatknox]

In general, really.. I'm sure the set-up you mention works well, but I don't really see the point for I2S interfacing unless you have some very specific use cases..

It's not surprising to me that there are many enthusiasts with many use cases.

 

[rebbiputzmaker]

In general, really.. I'm sure the set-up you mention works well, but I don't really see the point for I2S interfacing unless you have some very specific use cases..

In my case, although I know this won’t be necessarily well received here, I find it superior to USB. IMO the Pi2AES is an excellent transport, sadly it never was able to be tested.

 

[Veri]

In my case, although I know this won’t be necessarily well received here, I find it superior to USB. IMO the Pi2AES is an excellent transport, sadly it never was able to be tested.

GoldenSound measured it here https://goldensound.audio/2021/07/22/pi2aes-streamer-measurements-and-5v-psu-mod-instructions/

It surely works OK, it's a neat little Pi transport. I just find it a bit of a stretch that these sub -120dB, pico second transport measurements are at all, in any way audible. Your internal DAC clock would have to be bonkers terrible for there to be an "obvious" improvement. Just my .50c

 

[rebbiputzmaker]

GoldenSound measured it here https://goldensound.audio/2021/07/22/pi2aes-streamer-measurements-and-5v-psu-mod-instructions/

It surely works OK, it's a neat little Pi transport. I just find it a bit of a stretch that these sub -120dB, pico second transport measurements are at all, in any way audible. Your internal DAC clock would have to be bonkers terrible for there to be an "obvious" improvement. Just my .50c

You know I really do agree with you… but in my listening, in which I try pretty hard to be reasonable and not crazy thinking that every little thing must makes a difference. I do feel that there is something better sounding using the I2s over USB. With the P2aes I also like the quality of the AES better than USB from either a PC or a raspberry pi. Ymmv.

 

[somebodyelse]

You know I really do agree with you… but in my listening, in which I try pretty hard to be reasonable and not crazy thinking that every little thing must makes a difference. I do feel that there is something better sounding using the I2s over USB. With the P2aes I also like the quality of the AES better than USB from either a PC or a raspberry pi. Ymmv.

I get the impression you're talking about sighted listening, in which case YM will indeed V - that's just how the brain works, even when we're trying to be objective. The various controls in listening tests are needed to separate the genuinely audible differences from the times when the subconscious is misleading us. For well implemented digital interfaces on the same DAC, and assuming no ground related issues (which can certainly be an audible problem), the measurements are about as similar as between different runs with the same interface. Hearing a difference between them in controlled listening tests would be a discovery in itself, and worthy of investigation to find out what measurements were missing.

 

[rebbiputzmaker]

I get the impression you're talking about sighted listening, in which case YM will indeed V - that's just how the brain works, even when we're trying to be objective. The various controls in listening tests are needed to separate the genuinely audible differences from the times when the subconscious is misleading us. For well implemented digital interfaces on the same DAC, and assuming no ground related issues (which can certainly be an audible problem), the measurements are about as similar as between different runs with the same interface. Hearing a difference between them in controlled listening tests would be a discovery in itself, and worthy of investigation to find out what measurements were missing.

You are entitled to your opinion thank you for it.

 

[exwhyzed]

Anyone still using the Khadas Tone Board? I found it through this site a few years ago, and have been using it since. There seems to be a lot of new DACs that have come out since then, even a new Khadas Tone Board (Tone2). I'll probably hang onto it for a little while longer until something else really jumps out at me.