Topping Mini 300 Amplifier Review

[Dave TR3A]

*** EDIT: This post has changed significantly to incorporate information from @NTK and Topping that explained the 600 kHz signal ***

In this post I will present the last of the findings from the measurements I took of the Topping Mini300 amplifier.

In previous posts (here and here) I discussed:

1. current requirements for the 12V trigger
2. the maximum input signal amplitude that would be noise-free (and if that differed from the 1.3 Vrms input sensitivity)
3. maximum gain settings before the onset of noise/clipping
4. power into 4, 8, and 16 ohm loads
5. channel balance (how much did the L and R channels differ in output when presented with the same input)
6. how "clean" is the Mini300 output

In this post I'll discuss how "clean" the Mini300 output is, offer some thoughts on changes I would like to see, and draw some conclusions.

How "clean" is the Mini300 output?

When one compares these oscilloscope traces of the Mini300 and the Aiyima A07, one can't help but notice that the Mini300 trace is a broader line.

An oscilloscope provides the means to zoom into the trace. Doing so reveals that a ~500 mVpp (0.18 Vrms) signal at roughly 600 kHz has been superimposed on the Mini300's output.

That 600 kHz signal is consistent at all gain settings. It isn't coming from the input; it's present even when no input is attached. It must be coming from either the Mini300 itself, or from its power supply.

*** EDIT: @NTK explained the presence of the 600 kHz signal in this post. I've summarized what he said below. ***

The output of a class D amplifier when it is idle is not a flat line at 0 V (at least for those that use AD modulation, which is the case for amplifiers based on the TI TPA 32XX and other chips). The idle output from the output bridge is a 50% duty cycle square wave, which is composed of the fundamental frequency (the ~600 kHz switching frequency) and its odd harmonics, i.e. H1, H3, H5, etc. They are, or should be, blocked by the output low pass filter (LPF).

The typical class D output filter is a 2nd order LC (inductor-capacitor) LPF with a cutoff frequency at ~50 kHz, about 1 decade below (i.e., 1/10 of) the switching frequency. The output filter therefore attenuates the amplitude of the switching frequency by 40 dB (= 0.99 or 99% attenuation). For an amplifier with a rail (supply) voltage of 38 V (such as the Mini300), a residual of about 1% is expected.

For a 38 V square wave, the amplitude of the fundamental component is (4/π)×38 = 48 V. Therefore the ~500 mVPP residual is right in the ballpark.
An idle ripple of ~500 mVPP is normal and is considered harmless.

I've written to Topping's support team asking if the superimposed signal is present on all Mini300 amps, or if what I'm seeing might be due to a defective amp or SMPS. I'll update this post if I hear back from Topping.

*** EDIT: Topping's service team sent this response to the e-mail I sent them about the 600 kHz signal ***

Thank you for the detailed description and oscilloscope measurements.

1. The two oscilloscope screenshots were taken with different output amplitudes and vertical scales: the Mini300 was measured at 1 V/div with 4.00 Vpp, while the A07 was measured at 5 V/div with 15.6 Vpp. Therefore, the apparent trace thickness cannot be used directly to conclude that the Mini300 has higher noise. This comparison is not technically rigorous.
2. The Mini300 is a Class-D amplifier, whose power stage switches at approximately 600 kHz. Although the output LC filter significantly attenuates the switching carrier, it does not eliminate it completely. As a result, a small amount of residual switching component around 600 kHz may still be measurable at the output. Since this frequency is far above the audible range and the effective response range of typical loudspeakers, it normally does not result in an audible effect.

*** EDIT: The information from both @NTK and Topping explains the presence of the 600 kHz signal. I am content that it can be ignored. ***

Changes I'd like to see

The Topping Mini300 is a fantastic amp. But that doesn't mean it couldn't be better. Here are some changes that I would like to see:

1. Put markings on the volume knob. Having no markings makes it hard to set gain consistently. My preference would be to have it marked in dBFS, but even unlabelled tick marks would be OK.
2. More power. Even though the Mini300 should be fine for my highly sensitive speakers, I'd like to see more power -- perhaps twice as much.
3. Eliminate the 600 kHz signal that is present on the Mini300's output. Even though that signal should be inaudible, I'd still like for it to be gone.
4. Consolidate the unit into fewer boxes. I'm enamored of the PAM15x design (several channels, integrated power supplies, trigger in and out) and would very much like to see a design like that from Topping.
5. Accompany the TRIGGER IN with a TRIGGER OUT

Conclusions

As mentioned above, I think the Topping Mini300 is a fantastic amp.

• It is completely silent when no input is playing; there is no audible hiss whatsoever.
• With a little fussing about, its 12V trigger is quite versatile
• Topping has mainly delivered to the Mini300's published specs. Although I think that it works better with a lower input signal amplitude than the 1.3 Vrms that Topping specifies as its input sensitivity that's likely because my noise threshold is more stringent than the 1% THD+N that they are using.
• Because I believe that a lower input signal amplitude is required, that means the Mini300's power is lower than claimed. Again, that's because I'm using a more stringent noise threshold.
• The 38V/4A power supply is adequate for my requirements.

I would position the Mini300 as a very nice 50W/channel (into 8 ohms) amp. But it may not be a good choice for those with low efficiency or current hungry speakers.

Since my speakers are unusually efficient they don't need a lot of power. This slide (EDIT: Updated to correct typo) explains why I think the Mini300 will be adequate for my system:

My main listening position (MLP) is about 3m from the speakers and I typically listen at 80 to 90 dB SPL.

My next task will be to integrate the Mini300 into my system, replacing the Aiyima A07 that currently drives the tweeters. This will involve setting the gain appropriately on my DSP to keep my system well matched.

I have not yet decided whether or not to use Mini300 amps with my midrange and woofers. Although I'd like the convenience of the TRIGGER IN to turn the amps on and off, the A07s are doing a good job and I don't know if I'd be able to discern any audible difference from the change.

*** EDIT: I have decided to use 2 Topping Mini300 amps (1 for my tweeters and 1 for my midrange) and a 3e Audio A7 amp for my woofers. I’ve already configured the Mini300 and the A7 and am awaiting for delivery of the second Mini300.

I had been concerned that the Mini300 might not have enough power for my 16 ohm midrange drivers, since by my measurements it would max out at about 25.4W. Since the drivers are rated at 40W I thought I should have more power. However, when I examined my configuration, I found that my available power was apportioned:

• 82% to 87% to woofers
• 10% to 15% for midrange
• 2% to 3% for tweeters

Knowing that the midrange will only require about 15% of the system power convinced me that 25W would be enough. To use all 25W at the midrange, I'd need to be driving nearly 150W into the woofers (which are only rated for 100W). Putting it another way, if I were to drive my woofers flat out at 100W, I'd only be pumping 18.75W into the midrange, which is certainly well within the Mini300's capability. ***

 

[Dave TR3A]

NICE!
A question from a noob:
If I'm using a simple multimeter on the RCA cable to check the signal coming out of my source, say a PC running a 40hz signal, should I then make sure I should not go beyond 1.2V on the multimeter?

And what do you mean by 'high gain' in this context? Did the minidsp flex have a high gain setting? Just trying to relate back to my case using PC's integrated audio as source

Yes. However, it will depend on the capabilities of your multimeter. You will want a TrueRMS DMM that can take accurate readings at the frequencies you're interested in. Some inexpensive multimeters will only be accurate at 60Hz (i.e., 120 Vac line frequency). In this case, I was testing at 1 kHz and couldn't use my Brymen 235 DMM because it's only accurate to 800 Hz.

When I referenced high gain on the Mini300 I was particularly referring to gain at or above 3:30 on the dial (if one viewed the volume knob sitting in front of a clock face). That was pretty much anything around -2 dBFS or higher.

The MiniDSP Flex allows me to set gain in dB for each channel. It doesn't have a high gain setting per se.

 

[Dave TR3A]

Amir doesn't measure wide band class D amplifiers output often, but here is the spectrum of the PA5, a predecessor of the Mini 300.

The output of a class D amplifier when it is idle is not a flat line at 0 V (at least for those that use AD modulation, which is the case for amplifiers based on the TI TPA 32XX chips). The idle output from the output bridge is a 50% duty cycle square wave, which is composed of the fundamental frequency (the switching frequency) and its odd harmonics, i.e. H1, H3, H5, ... They are, or should be, therefore effectively blocked by the output low pass filter.

The typical class D output filter is a 2nd order LC LPF with a cutoff frequency at ~50 kHz, about 1 decade below the switching frequency. They therefore attenuate the switching frequency by 40 dB (= 0.99 or 99% attenuation). For an amplifier with a rail (supply) voltage of 38 V, a residual of about 1% is expected*. Therefore an idle ripple of ~500 mV is normal, and is considered harmless.

Note: *For a 38 V square wave, the amplitude of the fundamental component is (4/π)×38 = 48 V. Therefore the ~500 mV residual is right at the ballpark.

That's an excellent contribution. Thank you.

I agree that the 600 kHz signal ought not to be audible. But I'd still rather that it wasn't there.

 

[NTK]

That's an excellent contribution. Thank you.

I agree that the 600 kHz signal ought not to be audible. But I'd still rather that it wasn't there.

The Purifi 1ET400A was worse.

 

[fpitas]

Here are some of the results of testing that I have done on the Topping Mini300 amplifier.

Background

I run tri-amped Klipsch LaScala speakers with 2 Bill Fitzmaurice-designed DIY horn-loaded subwoofers, configured as per this diagram:
View attachment 554243
Since I wasn't sure how well tri-amping and DSP would work, I used inexpensive Aiyima A07 class D amplifiers in my initial configuration. Pleased with how my system performs, I have been looking at upgrading the amps. The Aiyima A07s are fine amps, but I was keen to try a 12V trigger (for convenience), PFFB, and better SINAD.

My speakers are very efficient (about 104 dB SPL @1W @1m). As such they do not require a lot of power but they do need clean signals.

I tried 3e Audio A7 amps and a 3e Audio/ALSOZONE PAM158 amplifier but neither worked out. A major problem that I had with the PAM158 was a hiss that was audible from my seating position, roughly 3m from the speakers. There was some hiss from the tweeters with the Aiyima A07 amp but only when I put my ear right up against the horn.

Attracted by the claims of exceptional THD+N, SNR, and a very low (<15UVrms) noise floor, I decided to give the Mini300 a try.

Testing

After spending an evening listening to the Mini300 in my system, where I verified that I could hear no hiss from the tweeter, even with my ear pressed against the horn, I took it to my workbench to run some tests.

This slide shows my testing setup and describes the methodology that I used.
View attachment 554245The 1.3 Vrms input level was used because Topping specs that as the input sensitivity of the Mini300

What I was looking to find out

I wanted to determine:

1. current requirements for the 12V trigger
2. the maximum input signal amplitude that would be noise-free (and if that differed from the 1.3 Vrms input sensitivity)
3. maximum gain settings before the onset of noise/clipping
4. power into 4, 8, and 16 ohm loads
5. channel balance (how much did the L and R channels differ in output when presented with the same input)
6. how "clean" is the Mini300 output

12V trigger current consumption

I was keen to know how much current the Mini300's 12Vdc trigger required.

I had hoped to have three amps all controlled by a single 12V trigger for on/off (one amp for my tweeters, another for my midrange drivers, and the third for my woofers). Since the Mini300 has a trigger IN but not a trigger OUT, I was going to have to split the trigger from my source (a Bluesound Node) and send it to all three amps. That would only work if the cumulative current drawn by the 3 amps did not exceed the Bluesound Node TRIGGER OUT's 50 mA capacity.

I measured the current drawn by the Mini300's trigger at 7 mA. The Bluesound Node would certainly be able to support 3 Mini300 amps.

Maximum noise-free input signal amplitude

I have found that noise or clipping is often evident on oscilloscope traces even when amps are driven by a signal at their rated input sensitivity at high gain settings. I look to determine the highest noise-free input signal amplitude so that I can attenuate the signal going to the amp so as not to exceed that level.

I found that, at high gain and at the rated 1.3 Vrms input signal amplitude, the output of the Mini300 always showed noise. The noise wasn't always large, but it was clearly visible on the oscilloscope.
View attachment 554252

The noise would disappear if the gain was turned down, or if the magnitude of the input signal was decreased.

Since it would be convenient to "set and forget" the gain on the Mini300 (which is controlled by an unmarked volume knob), my preferred approach is to establish a maximum input signal amplitude and then set the gain (typically at MAX) and leave it there.
View attachment 554253I will continue presenting my findings in another post in this thread.

I wonder if you're seeing the dead time for switching the FETS as "noise".

 

[fpitas]

That's an excellent contribution. Thank you.

I agree that the 600 kHz signal ought not to be audible. But I'd still rather that it wasn't there.

You would need a higher carrier or a higher order output filter to reduce it. Perhaps a notch filter would work, but there may be excess phase consequences if you include it in the feedback loop.

 

[fpitas]

That's an excellent contribution. Thank you.

I agree that the 600 kHz signal ought not to be audible. But I'd still rather that it wasn't there.

One sneaky idea would be to turn the output filter into an elliptical with the zero at 600kHz. Once again, the feedback loop might get unhappy, but it's an easy thing to try.

 

[Dave TR3A]

I wonder if you're seeing the dead time for switching the FETS as "noise".

I don't think so.

Here's another example of noise prior to the onset of clipping. It was generated at high gain with a 2.0 Vrms input signal amplitude.

If you were referring to the 600 kHz signal overlaid on the output, I think @NTK gave a cogent explanation for that.

 

[fpitas]

I don't think so.

Here's another example of noise prior to the onset of clipping. It was generated at high gain with a 2.0 Vrms input signal amplitude.
View attachment 554356

If you were referring to the 600 kHz signal overlaid on the output, I think @NTK gave a cogent explanation for that.

Not exactly what I meant, although it would occur at the 600kHz rate of switching as the FETs transition.

 

[Dave TR3A]

I have modified my post discussing the 600 kHz signal overlaid on the Mini300 output to incorporate information from @NTK and @TOPPING-Service.

 

[ziggurcat]

Another follow-up from Topping regarding the tick/pop:

"Thank you for carrying out the requested isolation test and for providing both videos with clear timestamps. Is this distortion occurring in both the left and right channels simultaneously? If so, please provide the specific model and specifications of your speakers for our reference.

If you have other passive speakers, you can also try connecting them. We will also report this issue to our testing department for further testing."

They still haven't brushed it off as normal, so there might be something there.

 

[Dave TR3A]

Another follow-up from Topping regarding the tick/pop:

"Thank you for carrying out the requested isolation test and for providing both videos with clear timestamps. Is this distortion occurring in both the left and right channels simultaneously? If so, please provide the specific model and specifications of your speakers for our reference.

If you have other passive speakers, you can also try connecting them. We will also report this issue to our testing department for further testing."

They still haven't brushed it off as normal, so there might be something there.

I haven't heard a tick or pop sound from the Mini300 I received this month.

What were the answers to the questions:

• is the tick/pop on both channels, or just one channel?
• what model speaker are you using?
• does it happen with other speakers?

I can't see why the tick/pop would happen with one set of speakers and not another, unless the first set is exquisitely sensitive.

 

[ziggurcat]

I haven't heard a tick or pop sound from the Mini300 I received this month.

What were the answers to the questions:

• is the tick/pop on both channels, or just one channel?
• what model speaker are you using?
• does it happen with other speakers?

I can't see why the tick/pop would happen with one set of speakers and not another, unless the first set is exquisitely sensitive.

Most people who own this amp have said that they don't hear a single thing from their Mini300 (which, to me, corresponds with the power on/off graph Amir has shown in this review), but I have seen two people here say that they experience the same thing as I do. It feels like it's out-of-spec to be hearing any pops when powering the amp on/off, which is why I finally got around to asking Topping about it after 8 months of use .

- Yes, it's in both channels. I have tested with only one channel connected at a time, and it is just as audible in either of them. I've also held my ears close to each speaker channel when both are connected, and it's the same level of audibility. I have also made sure that my cables are fully seated on either end, and that the binding posts on the amp as well as the binding posts/jumper plates are tight/secure. I have tried a few different speaker cables as well.

- Monitor Audio Silver 50 (6G)

- I haven't checked that yet, but it's on my to-do list. The Silver 50s aren't particularly sensitive (87dB, so fairly average/below average). I have a pair of speakers that are 89dB sensitive, and 85dB sensitive, so I can connect both of them to see if that changes anything. That won't be for a while, though since it's going to be a little busy in the next 7 days (I'll also be out of town for 3-4 days).

I have posted a recording of this here:

IMG_4952.MOV

drive.google.com

In this video, I am holding my phone pretty close to the speaker to be able to pick up the pops (you may have to turn the volume up, though). You'll also hear some remote button clicking because Topping just added a new feature on the DX5 II to be able set the 12v trigger to sync with the line output rather than power, and I was testing the feature out. For reference, the initial tick when the amp is powered on occurs at the 9 second mark (this happens about 3 seconds after the light on the amp has turned on), and the more audible pop when the amp is powered off occurs at the 12-13 second mark in the linked video. The tick when powered on is not really audible when standing 6+ feet away, but the powered off pop is still quite audible even further away.

To be perfectly clear: it's not the DX5 II causing the pops, it happens with only speakers connected to the amp, and I have sent video documentation to Topping with the amp in view as well as up close to the speakers, all showing that nothing else is connected to the Mini300.

mini300_Pop.MOV

drive.google.com

mini300_Pop2.MOV

drive.google.com

In the video showing the amp, I had to carefully turn the amp off because the clicking of the toggle switch would drown out the actual pop, but in the other video, I'm close enough to the speaker that the toggle switch doesn't drown it out (you sort of hear both the toggle switch and the pop at the same time in that video).

@Dave TR3A in case you miss my edit...

 

[Janox]

I don't have very sensitive speakers, but I can hear noise from the tweeter with the Mini 300 amplifier about 15 cm away.

 

[fpitas]

I don't have very sensitive speakers, but I can hear noise from the tweeter with the Mini 300 amplifier about 15 cm away.

Sounds defective. Maybe it's from the input source. Have you tried putting the input switch to the other position, or removing the input cables? Do both channels hiss?

 

[Janox]

Sounds defective. Maybe it's from the input source. Have you tried putting the input switch to the other position, or removing the input cables? Do both channels hiss?

I have two Mini 300s. Both make the same noise. In my opinion, this is not a defect, but a characteristic of this amplifier. The Topping Mini 300 is not a noise-free amplifier.

 

[fpitas]

I have two Mini 300s. Both make the same noise. In my opinion, this is not a defect, but a characteristic of this amplifier. The Topping Mini 300 is not a noise-free amplifier.

It makes no audible noise connected directly to my horn drivers, which are 111dB/w.

I'll add, Amir would have seen such high noise during testing if his unit did it.

 

[NTK]

I have two Mini 300s. Both make the same noise. In my opinion, this is not a defect, but a characteristic of this amplifier. The Topping Mini 300 is not a noise-free amplifier.

Amir has measurement of the noise. The idle noise is about 80 μV, which is 20 log10( 80e-6/2.83 ) = -91 dB down from the standard voltage loudspeaker sensitivity is measured. Therefore, if your speaker sensitivity is 89 dB @ 1 m / 2.83 V, you'll hear -2 dBSPL in free field at 1 m distance, compared to the typical room noise level of 30-40 dBSPL.

 

[fpitas]

I have two Mini 300s. Both make the same noise. In my opinion, this is not a defect, but a characteristic of this amplifier. The Topping Mini 300 is not a noise-free amplifier.

Have you tried what I suggested in post #815?

If your units are that badly out of spec, Topping will probably swap you for good ones. But Topping will ask you to perform the tests I suggested.

 

[staticV3]

The idle noise is about 80 μV

With extended measurement BW to catch the pop I'm assuming.

With 20kHz BW Amir measured 107dB SNR at 5W/4Ω, so sqrt(5*4) * 10^(-107/20) = 20μV of Z-wt noise.



Topping claim 15μV A-wt, so 2.5dB less which checks out for A-wt white noise.