Loxjie A30 Amplifier Review

[Froggy23]

Thanks a lot for your answers, I just bought another one!

 

[Neddy]

What level do you keep the amp at?

Amp is at 35, TV (analog/HP output) is at 50 usually, but this is for totally non-critical cheap TV listening thru 1970 JBL L25s (but sounds pretty amazing, considering!).

 

[Laserjock]

Amp is at 35, TV (analog/HP output) is at 50 usually, but this is for totally non-critical cheap TV listening thru 1970 JBL L25s (but sounds pretty amazing, considering!).

Yeah, I have same situation.
Had some KEF X300A powered speakers I used that way but they died on me and was looking to do the same with the amp and some passives.

Something easy the wife can do without using the AVR etc

 

[Neddy]

I was going to use Kali IN8s there, but found the JBL Prima's for a song, in pristine, re-foamed woofer, condition on flea-bay at that same time, so went with the much nicer looking option. (I was going to build custom cabinets to house the IN8s in.)

So I ended up needing a small amp - and had lots to choose from. Finally kind of tossed a coin (well, actually, for the bucks, the A30 had the most features, least cost, and decent enough measurements...and was tiny.)
And, it made the whole 'watch TV thing' dope simple - just use the TV remote per normal - no mussing or fussing with fancy gear.

I ran a quick and dirty comparison between the vintage Prima's, IN8s, and JBL 305s, and was astounded that - for TV use - I actually preferred the Prima's!!
They have remarkably deep bass, and have a far far wider 'soundfield' (not suprising, considering the other two were are field monitors), so are much better for off axis listening (eg, while in the kitchen).

A reasonable - and a bit cheaper - option would have been to just use JBL 305s, and just hide them somewhere, but....the vintage JBLs are pretty stunning, I think.

Still haven't figured out what to do with the basically new unused IN8s, tho.

 

[Juampa1989]

Hola querida, quiero preguntarte si puedo amplificar un par de parlantes klipsch rp-160m con este amplificador sin riesgo de dañar los parlantes. ya qué volumen conviene no excederlos y dañarlos por falta de energía.
Gracias

 

[Juampa1989]

descargue el ecualizador de paz, agregue el evitar el recorte y espero que, ya que se autorregula, no dañe mis parlantes.
igualarlo como recomendaron igualar el rp-600m ya que son similares.
Saludos

 

[aktiondan]

This looks like a great little DAC/BT/Amp combo that should work great for a small two-channel setup. Question though, thinking about a potential upgrade path to a better amplifier, has anyone tried using the headphone out as a line out to an external amplifier? Seems like this thing would be great if it had a dedicated line out but the headphone jack might fit the bill? I suppose the option there is to just go with the D30 but how bad would the A30 be as a dedicated DAC/BT/pre-amp and using the headphone out to an external amp later on down the road?

 

[aktiondan]

So after reading through a bit more of this thread and revisiting the original measurements on the headphone side of things, I'll go ahead and answer my own question with a "probably not a good idea". But that got me thinking, why is the headphone output so terrible? What does the architecture of this thing actually look like? It seems like it would be easy enough to split out one of those low-level analog signals after the DAC to a decent op-amp to drive some headphones. Or maybe the headphone output was just an afterthought? So I wanted to dive deeper into the possible signal paths of this thing and see if anything interesting turned out.

Taking info from a couple of posts by Toku and pics and video from Malis, I put together this graphic which illustrates possible signal paths for each of the inputs to the outputs based on datasheets/block diagrams of all of the know parts in the A30. A couple of interesting things stand out, both the Realtek USB chip and the Qualcomm BT chip have internal DACs but they also both support I2S. So do they feed the ESS DAC via I2S along with the LC89091JA chip or do they skip the DAC and go straight to the NJW volume/tone controller? Suspiciously the NJW chip has 4 analog inputs, presumably one for the DAC and one for the RCA analog inputs, leaving the other two inputs either unused (hopefully) or they are fed by the analog signals from the BT chip and the USB chip. Lots of options though. Some worse than others.

Also noticed from Malis' video that the BT chip he showed was a QCC3034 which looks like a newer version of the QCC3008 chip (shown in earlier images by enno) which according to Qualcomm does support aptX HD. The product page for the A30 doesn't really specify exactly which model BT chip is used so I guess they can update it as needed.

Anyway, that brought me back to headphone output and I can't tell where the headphone signal comes from, likely spilt off after the volume/tone controller but what is the actual amp stage? This may explain why it won't make a good line-level output to a different amp. Curious if Amir would be willing to test the headphone out as if it were line-level and not driving a 300 or 33 ohm headphones. Might be possible to get a better sense of what the DAC (or DACs) are doing in this case without the MA12070 influencing the measurements.

 

[TohuBohu]

Wondering about this device, since the DX3 Pro+ came out.

In his review Amir mentions being disappointed with the current offering and gives a good review for the Loxjie. Wonder if that changed now with the new Topping.

I'm debating which one to get to drive my DT 1990 Pro with my Macbook

 

[VintageFlanker]

Wondering about this device, since the DX3 Pro+ came out.

In his review Amir mentions being disappointed with the current offering and gives a good review for the Loxjie. Wonder if that changed now with the new Topping.

I'm debating which one to get to drive my DT 1990 Pro with my Macbook

I'm pretty sure you're on the wrong thread.

Anyway, I will respond if I guess correctly: DX3 Pro+ > D30

 

[Arkadiusz]

Hello, I bought the Loxjie A30 and I have a problem with it.

Before buying, I made 1.5mm2 (16awg) pure copper wires with a length of 1m (40in) to connect the columns (8ohm) with new amp. After connecting, the loudspeakers didn’t work, they made a clicking sound about every 1s. Heaving in mind that wires are hand-made, I connected to amp the first wires I could find – short wires of standart quality (img below).

The sound was audible but the amp made a hissing sound, a similar soft hissing noise was coming out from the loudspeakers when the music was stopped. I connected short wires x2 (next img), the sound from the speakers was perfect, the amplifier made a very quiet hiss when you put your ear to it.

I bought a pair of pure copper wires with a different cross-section for testing (0.5mm2 - 20awg, 0.75mm2 - 18awg, 1mm2 - 17awg), I cut them to a length of 1.5m (60in). Regardless of the configuration and connection of wires up to 4.5 m (180in), the speakers still did not work (clicking). Only when I connected 4 x 1.5m (60in), meaning: 6m (240in), a nice sound came from the speakers, the amplifier hissed slightly when I put my ear to it, hissing could not be heard in the speakers when the music was not playing (img below).

The loudspeakers are 8ohm DIY project "Overnight Sensation' by Paul Carmody (Specifications: 25 watts RMS/50 watts max, 45-20,000 Hz (-3 dB), 8 ohms, 83 dB 1W/1m). The loudspeakers worked well with the Sony micro system and the previously borrowed Pioneer amplifier.

The volume of the source or the amp does not influence the clicking or the hissing volume. Could applying a filter help? It is not a problem for me to use 6m of wires, but I wonder if the amplifier is broken.

I even measured the resistance of the wires out of curiosity:
- the short, standart-quality wire: 1.0;
- short x2: 1.4;
- copper 1.5m: 0.2;
- copper 3m: 0.2;
- copper 6m 0.2.

The next step will be to disassemble the loudspeaker and check it, then I will send the amplifier back to China. I don't know if this amplifier is so sensitive or the loudspeakers are. Thanks for any tips.

 

[Zek]

Try connecting both speakers at the same time.
What is the source for the A30?

 

[Arkadiusz]

Thank You for Your answer, I connected both speakers. The photos were taken quickly with the amplifier turned off and with example connections that I tested earlier. I tested on a usb connection whit a computer and mainly via bluethooth.

 

[aktiondan]

So this little guy showed up under the Christmas tree this morning. It's an impressive piece of gear and so much smaller than I thought it would be. I had some fun playing around with it (sounds fantastic at first listen) taking some simple frequency response measurements with REW. Thought I'd share some of what I measured because I thought it was pretty interesting and may be helpful to prospective/current owners. Amir did a great job of measuring the amplifier basics, power distortion, etc. but I was very interested in response plots of the headphone output, including what the different EQ settings did as well as what the subwoofer output looked like.

Overall, channel imbalance is excellent, less than 0.02 dB. Frequency response is down 1 dB at 20 Hz but the upper end extends all the way to the limit of my soundcard and is effectively flat. The tone controls provide what I would consider actually usable options. Before taking these measurements one of the first songs I auditioned with headphones was Tears For Fears The Tipping Point and like a sucker I just loved the S-BASS setting. As the graph shows this provides the most bass boost as well as a slight treble kick. Rock provides a little boost in the bass and treble giving you that classic smiley curve. Soft does the same but not as pronounced. Clear pulls out some bass and treble accentuating the mids. This would be great for podcasts or other voice-heavy content. And then of course the tone controls offer ±10 dB in both treble and bass for custom options.

Center frequency for the bass is nicely placed at about 55 Hz. None of that 100 Hz mid-bump stuff you typically see in some EQ's. Q appears to be about 0.22 and affects the response all the way up to 1 kHz. The treble is centered around 12 kHz and likewise affects the response down to about 1 kHz. Direct/bypass and Tones set to 0/0 were measurably identical. I'm not sure if the direct mode is really a bypass of the EQ controls or just an automatic setting to 0/0. Based on the measurements I'm guessing the latter (edit: actually according to the datasheet for the NJW1194, it does contain a bypass circuit for the tone controls). So overall I thought this looks really good, onto the subwoofer output.

This is where things got a little wonky. Subwoofer output f3 is about 220 Hz and rolls at a subtle 12 dB/octave. That's a bit high for subwoofer duty and the roll-off is quite shallow, so you'll probably want to have an external crossover on this (most active subs provide this level of control). But the response is not smooth at all. It's got this ringing that starts at 10 Hz and just gets worse and worse and you go up in frequency. It also gets worse at lower volumes. Not sure what to make of this. It's probably not that big of a deal, especially if you intend to cross it lower, but there still is some wobble even between 10 and 100 Hz. To double-check my setup I have an old pro-style EV XEQ-2 crossover unit that is set it ~85 Hz and has an 18 dB/octave rolloff. And it looked perfect. So they definitely mucked something up on the active low-pass filter on the subwoofer output. I might open this thing up and see if it can be tweaked to improve it and maybe push the crossover point down a bit while I'm at it.

Anyway, thought I'd share and maybe get some feedback on these measurements. I did some distortion measurements as well, but at this point don't trust my setup enough to believe anything beyond a simple frequency response plot (which is calibrated via loopback). But distortion looked good on the headphone output too. Really want to open this up and see what the headphone driver is, it's likely nothing to write home about, otherwise Loxjie surely would have touted it in the product brochure.

 

[aktiondan]

So I ran a few more tests on the A30, this time looking at distortion. Feel free to critique (or mock) my setup, but I've got a basic loopback measurement of the soundcard to get a sense for how good/low I can measure. Anything better than this I can't measure, but anything worse than this is fair game. REW measures and plots distortion across the entire frequency sweep but only provides a calculation of the distortion for a single point on the graph that you select. I chose 3 points and annotated just one graph for each one: 40 Hz, 100 Hz and 1 kHz. I measured the subwoofer output and the headphone output of the A30 (along with some other gear for comparison sake). I used the analog inputs for all measurements and the volume was set to produce a signal level either the same as the loopback or lower (to ensure distortion wasn't introduced just by overdriving the ADC on the soundcard). I also used the Toslink input for a couple of measurements but found no difference in distortion levels compared to the analog. I also ran some of these same measurements as spectrum analysis, which is more consistent with how Amir does his testing, but I like the visual you get from a distortion vs frequency plot as reported by REW. I can share the spectral plots if anyone is interested.

First up is the soundcard loopback check. Since I'm not measuring state-of-the-art gear here, this should do fine. Though it's nothing to get excited about.

Next up is the subwoofer output. Things get decidedly worse. Still trying to decide how significant this is in practice. Most subwoofers will probably have worse distortion that this so maybe not too much to worry about. Still seems odd that even at line level they were unable to produce a decent low-distortion output. Add the unusual ringing on top of that and we've definitely got something interesting going on.

Next up is the A30's headphone output. This most likely accounts for Amir's poor headphone measurements, distortion is dominated by frequencies below 100 Hz. Even still, levels don't exceed 0.142% at 40 Hz with the volume at 45. Which with my headphones is pretty loud. I'm not much of a headphone guy myself, so even though the measurement doesn't look great, they sound fine to my ears. Once again similar to the subwoofer situation, the actual distortion of the headphones it likely to be higher, so you're probably not going to be missing anything.

By comparison here's the headphone output of my UCA222 soundcard. Much better performance in the low-frequency department. But this provides little indication as to how well it can drive a pair of headphones. But distortion is lower than the A30 below 100 Hz.

Also by comparison an Emotiva Little Ego headphone amplifier. While the distortion in the bass regions is better, there's a pretty nasty 1 kHz tone with all its harmonics making its way into the mix. This guys runs off USB so maybe something going on there?

And last we have my trusty ole EV XEQ-2 crossover. This does a much better job at bass duty with much lower distortion than the A30 (even though it uses the same TL072 op-amps) but the crossover response appears to poop out after about 700 Hz.

Also for fun, I traced out the circuit for the subwoofer filter and created a SPICE simulation (there's a great free online version) and can model any filter cutoff frequency desired with just 4 resistor changes. It's also possible to up the slope from 12 to 24 dB/octave but not easily. Here's a modification for example to drop the cutoff to 100 Hz.

 

[David James]

Edit:

I've deleted the review i just posted since it turned out the problem that motivated me to write it in the first place turned out to be related to a problem with my windows machine.

I'm mostly happy with this now it's plugged into my mac.

 

[aktiondan]

Glad you got it all sorted out. USB can be finicky, that's been my general experience. One time just switching to a different USB port solved the problem. But with the A30 the only input port I have not had a chance to really test out yet is the USB. Only because my main two computers are still running Windows 8. Ha! I have a laptop with Windows 10 but haven't tried it out yet. I'll be sure to check for any weirdness though when I do. The coax, Toslink, BT and analog inputs all sound great to me. I've been pretty happy so far with it. Objectively I've measured everything I am capable of and have no real complaints. Took some in-room measurements through a pair of speakers, running them at ~96 dB, distortion is reasonable (above tuning anyway) and is primarily dominated by the speakers (guessing). Going to pair this up with a small DIY 10" sub. Should make a great 2.1 setup.

 

[aktiondan]

I took another stab at deciphering the audio chain in the A30, just to better understand in more detail just how it all works, look for potential pitfalls as well as areas that could maybe be improved. I think I've got it down best I can without actually taking my unit apart (though I'm dying to do it!) I snagged all the high-res photos that @enno posted and watched the video from @Malis a couple of times and this is what I've got. Feel free to comment or correct as needed. I'm posting this just in the general A30 review thread because I feel it's pertinent to the overall discussion about its performance, even though this is more of a detailed teardown post.

Some initial observations, this thing is quite nice! They've done an excellent job of implementing all the various chip/microcircuits per their respective datasheets (though I didn't review them all), lots of ceramic decoupling caps on DC power, some nice Nichicon caps on the MA12070 supply side, signal traces and board layout is logical. All parts are SMT with only a few exceptions. Not sure I understand using one RCA for the coax and analog input thus requiring a complicated 2:1 mux to be implemented and adding additional circuitry to only the right channel. They should have just made the thing a 1/2 inch wider and added a third RCA connector! Also, the main signal path to the MA12070 amp runs through a set of 2.2 uF electrolytic caps where Infineon expressly recommends these be X5R ceramics. The analog signal takes another hit by running through another set of 22 uF electrolytics. Two places where upgraded caps could make a minor difference.

The ES9023 DAC takes on digital-to-audio conversion for the Coax, Toslink and Bluetooth signals. But not the USB. Bummer. The internal DAC and "Headphone Amplifier" of the Realtek ALC4050 drive the NJW1194 volume controller directly. Even though the device supports I2S and could have been fed to the ES9023, it looks like we are left with whatever proprietary DAC exists inside the Realtek part, which probably isn't all that great. It's supposed to support 32-bit/384kHz but yet the DAC poops out at a 100 dBA SNR. Not a big deal, if you want better USB support, you can always get a USB-Toslink adapter and run that into the A30 to take advantage of the ESS DAC. But that's the only way if you're keen on USB audio.

The output filter on the MA12070 forgoes the traditional LC circuit in favor of a a much smaller ferrite filter and pair of 1nF caps to GND. They recommend a cutoff frequency as high as 200 kHz! Explains why Amir measured out to 100 kHz with this thing. Very unusual for class D but nice to see the extra bandwidth. There's a 24 page app note on the output filter too which shows improved THD with a specific type/brand ferrite. No clue what Loxjie chose here though. This type of filter appears is less sensitive to load impedances as well, which means you should get a flat response to 20 kHz no matter the type of speakers.

So that little 2:1 MUX chip that handles the S/PDIF signals and the analog signal for the right channel is just weird. I couldn't quite make out how they were actually performing this logic from the pics and the datasheets. So I just left it vague in the wiring scheme, just know that they are doing some switching of those signals to get things to work. Still seems like an odd architecture choice to begin with. But size may have been the main constraint (and in keeping with the same form factor as the D30).

The subwoofer output takes the outputs from the NJW1194 and sums them with a simple pair of 15kohm resistors. Sleazy. At least it didn't affect the channel isolation, which it easily could have dragged way down by doing this. A proper summing op-amp would have been the thing to do. Also with the dual TL072C op-amp they could have implemented a steeper slope and easily could have changed the cutoff frequency. I don't know how well 200 Hz integrates with satellites, seems a bit high to me. I showed in a previous post that with a few resistor changes, that can be changed if you're up to it.

The last thing I'll comment on the is headphone stage. They picked off the output from the NJW1194 volume control and then fed it to what looks like just some kind of analog switch. Seems like a reasonable way to disable the headphones when in speaker mode. But the chip that appears to be the amplifier I cannot find any information on. It's a complete no-name, no-nothing on the web about it. Which seems like such a wasted opportunity. Everything up until this point looks really good from a line-level perspective. They haven't done anything to completely screw it up until you get this this part - the 4B088. It's got a nice big 470uF cap on what is most likely the DC rail, so they attempted to at least keep it quiet, but based on Amir's testing this part is trash, barely able to push 4mW into 300 ohms. Would it have been so hard to drop any OPA16xx series headphone amp in there and call it day? It's limited to a single supply rail, so maybe that was the driving reason? This might be another area where if you really wanted to up the performance you could mod the headphone amplifier stage. Haven't really looked into this. I understand that it doesn't measure great, but sounded fine to me with an old pair of Audio Technica ATH-M30s. So might not be worth the trouble unless the form factor and features of the A30 are just right and having another separate headphone amp doesn't work for you. You can't exactly use this with a separate headphone amp either since it doesn't have a line out. So you need an almost totally different setup if phones are your jam.

Anyway, enough chatter, here's the pics with as many the signal annotations as I could come up (focusing on audio paths, not the digital logic of the microprocessor or the DC paths so much). Plus I sketched out a new simplified audio signal path block diagram. Enjoy!

 

[Steve.]

The ES9023 DAC takes on digital-to-audio conversion for the Coax, Toslink and Bluetooth signals. But not the USB. Bummer. The internal DAC and "Headphone Amplifier" of the Realtek ALC4050 drive the NJW1194 volume controller directly. Even though the device supports I2S and could have been fed to the ES9023, it looks like we are left with whatever proprietary DAC exists inside the Realtek part, which probably isn't all that great. It's supposed to support 32-bit/384kHz but yet the DAC poops out at a 100 dBA SNR. Not a big deal, if you want better USB support, you can always get a USB-Toslink adapter and run that into the A30 to take advantage of the ESS DAC. But that's the only way if you're keen on USB audio.

After stalking this thread for months both pre and post purchase of the Loxjie A30 to pair with Klipsch RP-8000Fs, I'm a bit disappointed at the USB routing. I tend to use the USB input for my iPad when running Tidal / Qobuz and this is not great news.

Has anyone been able to dig up more info on the Realtek Dac chip? would it be worth using a USB to Toslink converter (douk audio has one with an XMOS chip that looks descent). Or at this point should I just grab a D30 and hook it up.

Inputs on my Loxjie look like this:

iPad > USB
Chromecast Audio > Toslink
Turntable > Preamp > RCA

I was thinking maybe a

iPad > USB to Toslink converter > Toslink
+ <Toslink Switcher>
Chrome Cast Audio > Toslinke

 

[tvrgeek]

I do not see your problem. Do you not like USB? Converting USB to TOSLINK does not improve quality anywhere. It removes the host's ability to get feedback from the DAC. The iPad would not know what it was talking to is you ran to a switch.