Review and Measurements of Topping DX3Pro DAC and Headphone Amp

[amirm]

Have you seen any 400Hz FFT here?

The answer should be plenty of time in the multi-tone tests:

 

[finneybear]

The answer should be plenty of time in the multi-tone tests:

Multitone is a good glance. But it's the FFT for each frequency which matters the most. We need to see the harmonics for each frequency.

 

[simonchretien]

So you are listening to music with measurement instruments?
Since you believe measurements, can you please tell me why stock DX3 is measured so well yet the sound is not very good hence some owners have started using DX3 as a paperweight?

Read every comment in this thread, I dont remember anyone saying the dx3 sounds bad... Are you talking about the 2 line out comments?
Way too much DIY crap here.

 

[finneybear]

Read every comment in this thread, I dont remember anyone saying the dx3 sounds bad... Are you talking about the 2 line out comments?
Way too much DIY crap here.

Go to Page 150, post #2,996 for instance. Early post also complained about unbearable HF shrills, lack of mid bass, etc.

Well, you call this DIY crap when it solves the DX3 stability problem?

 

[finneybear]

Since most people do not have the tools and experiences working with surface mount parts, here I make out one simple fix list for your reference. Most will be through hole parts. Only basic soldering skill is needed. These will not be considered as "mod", instead, we just follow what 4493's datasheet has called for.

1. Remove the main power supply cap, the one right after the power supply plug socket. Use a 330uf/25V or higher good solid polymer cap. I believe this will resolve 90% of the stability problem. Just remember to connect the power supply to DX3 first, before you plug the supply to power outlet. If you plug the external power supply to power outlet first then connect the supply to DX3, you will see a fire spark show.

2. Remove the 4 100uf yellow caps around the two 4493s. Each 4493 has two. Replace them with 470uf/6.3v solid polymer or even hybrid caps. 4493 datasheet states at least 470uf here. The 100uf used in DX3 is too small.

3. For the LPF OP, I found Muses 8920 was a good match. Sure, you can keep the original OPA1612 if you like that kind of top happy sound. It's the 8 coupling caps which need attention. Change them to 47uf/6.3v nobium oxide cap. 4493 datasheet suggests 100uf here actually. Original DX3 uses 33uf. Anyway, the new caps will bring some bass back to you.

4. If you use the headphone amp a lot, it will be a good idea to replace them 8 coupling caps to 47uf/25v Silmic or Muse KZ. The stock capacitors are very bad, poor measurement result.

 

[finneybear]

Ok, did an AB test between the fixed DX3 and a twin ES9038 Pro based DAC – LKS MH-DA004

It’s not a fair comparison because the LKS has balanced outputs. An ESS USB interface is feeding it with I2S HMDI2.1 cable. Level matching between the two machines was tricky. I tried to get them as close as I could.

LKS was running at DSD512. DX3 was mostly DSD512 and some PCM705.

At DSD512, LKS sounds a bit better than T+A DAC 8 at DSD512.

So here is what I found – both have the stage about as wide. LKS is deeper, you can hear the clear outline of each instrument, the shape is so sharp, sharp to the point probably there’s too much sharpening applied. DX3 sound a bit congested, the shape of each instrument is fuzzier. When the shape is fuzzier, DX3 replaces the overly sharpened sound with smoothness. Oh, it just sounded so smooth. Even a touch of sweetness is associating with the smoothness. In general, DX3 just sounds more realistic, more alive. It is like to compare digital photography against film photos. ES9038 is like digital photo… it’s sharp, probably too much un-sharp mask applied. The images are very clean, colors are bold, strong, but in overall, it just look so digital. DX3 is like film photo – not as sharp, milky color transition, fuzzier details, yet everything looks more real and more alive.

DX3 has very full and rich bass, a lot of it, actually. LKS is not as rich as deep yet is very tight, has better control on drum beats. Obviously LKS has a much better power supply.

Vocals with ES9038 have too much throating. The sound is great, it’s just a touch too aggressive. On the other hand, DX3’s vocal is more full body, has more volume, and again, so smooth and easy. It reminds me of TDA1541 yet not that TDA1541 sexy.

DX3 now sounds the best at PCM 705K, the image definition improves a lot at 705K. It’s more 3D. One thing worth of mention is it sounds very good at CD 44.1K already. The sound is smooth, rich and warm. I can keep listening to it for hours without having fatigue. Have I repeated the word “smooth” again and again?

I would say LKS ES9038 has better dynamics, great image, more micro details. DX3 has the coherence and smoothness in sound.

I am thinking of improving the jitter performance of DX3. This may improve the image a bit. A linear power supply will benefit the dynamic control and bring out more micro details. Yet to push it up to another level, we will have to add local voltage regulator to each 4493, etc, which is too much for DX3. Hey, it’s just a small $200 box! I would say after problems fixed, the DX3 can comfortable compete against DACs around $1000-$2000. It has a very smooth and rich sound. If you like the so called multibit sound, you will have to give it a try. Other than the multibit smoothness, you get better dynamics with DX3.

I am sold on the AKM sound. Now I am on the market looking for a decent twin AK4497 based DAC.

 

[pacman]

The problem in this thread is very clear and simple, radicalism on both sides of spectrum. I was referring to this bizarre form of blindness when throwed some rude words (that wasn't nice btw and I do apologize).

The scientists-wannabes that are excited at numbers and charts showed in a scientific approach by Amir and seems to believe that everything on psychoacoustics can be measured and explained by numbers - for those, every statement will be confronted by "double blind AB test it or go home, you dirty peasant"; on the other hand, the masters of subjectivism, golden-eared-folks that apparently haves superior cognitive skills and talks about strange kinds of alchemystic rituals for improve stuffs that aren't supposed to improve, for delight of naive people. Both are awful.

Radicalism causes blindness. Music is pure mathematic, but in some shady way it affects our state of mind and causes these weird things named emotions, which are not measurable. Live with this, people.

Edit: oh my lord, it's happening again

 

[amirm]

Edit: oh my lord, it's happening again

Your generalizations about this forum and people who believe in audio science are incorrect. I suggest not going there. General debates like this do not belong in review threads.

 

[splkn]

It seems that I have confused you. Those caps are not original. I replaced them with new parts from mouser.com.
Solid polymer caps can have yellow color ones.

View attachment 22899

I can not find a 100uf 16v capacitor with a size of 3.2 mm.

 

[Killingbeans]

I often like to compare this to bridge designs. Do you design a bridge, do a calculation first to see how much weight it can take/withstand test, or design a bridge first then test to see whether it will crash with cars passing? Surely a good engineering is first to ensure your design will have a good test result.

If it was economically feasible for bridge engineers to do multiple full scale prototyping before final implementation, I'm pretty sure it would be considered a felony if they didn't.

I never said that the theoretical design phase is unimportant. I'm just saying that testing and measurements are equally important. A prototype is riddled with uncertainties, no matter how many calculations have been done on its design. Measurements are a great tool for lowering those uncertainties. "Am I really hearing a difference or am I most likely imagining things?" You can't make a bridge stronger via confirmation bias, but you can easily get that illusion going in an audio product.

The only thing that yanks my chain is your complete lack of humility.

You take what should be presented as a qualified guess and turn it into an assumption.

Maybe you are actually that good, and I'm just a lousy amateur who fails to see an expert at work. But until measurements that support your claims are made public, I'll reserve the right to be highly sceptical

 

[ChuckT]

Here is for your reference:

View attachment 22991

Thanks for the detail description.
There are smd to dip socket. You can replace the smd opamp with the socket and play with swaping dip opamps.
You can also get DC-DC isolated converter similar to below to have the input 15V converted into +/-15V instead of the existing +/-10V using the tps54331. The higher voltage should improve the drive for the opamp output.

https://www.mornsun-power.com/uploads/pdf/A_S-1WR2.pdf

 

[Killingbeans]

There are smd to dip socket. You can replace the smd opamp with the socket and play with swaping dip opamps.

Just keep in mind that it will make the tarces to the decoupling capacitors substantially longer. It could cause high bandwith chips to become unstable. At least in theory

 

[Sucram]

Thought this unit supports up to 32 bit/768khz? The highest setting I can select on windows is 384khz.

 

[Krunok]

Thought this unit supports up to 32 bit/768khz? The highest setting I can select on windows is 384khz.

Given that you personally can hear up to 384kHz and that ther's so many music material originally recorded at 768kHz that is a real shame.

 

[Sucram]

Given that you personally can hear up to 384kHz and that ther's so many music material originally recorded at 768kHz that is a real shame.

If the product description says it’s capable of doing that then it should.

 

[Krunok]

If the product description says it’s capable of doing that then it should.

And it can. It's you who can't.

 

[Veri]

If the product description says it’s capable of doing that then it should.

You can use 768Khz if you output that to the driver directly, via asio or wasapi. It is supported if you installed the driver.

If windows doesn't show it in settings, that's just windows being fussy...

 

[Sucram]

You can use 768Khz if you output that to the driver directly, via asio or wasapi. It is supposed if you installed the driver.

If windows doesn't show it in settings, that's just windows being fussy...

Thanks.

 

[Killingbeans]

If the product description says it’s capable of doing that then it should.

I actually agree with that. But I also agree with those who say it's a completely pointless bitrate.

 

[Veri]

Thanks.

You're welcome you're definitely right that wasn't particularly obvious
Newest xmos drivers support up to 1.536Khz btw, but only few manufacturers have licensed the last version yet.