• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Topping PA3s Review (Desktop Amplifier)

deadwood83

Member
Joined
Sep 1, 2022
Messages
91
Likes
160
Hey ASR! Long time listener, first time caller.

My main HT/ couch speakers are DIYSG Fusion 10 Pures at 98dB/W but I recently realized that when I built them back in 2012-2013 I royally goofed the crossover; and the woofer inductors are feeding each other with EMR so I am seeking to embrace the age of DSP rather than own up to my mistake. The woofers have a sensitivity of 98.6dB/W per Eminence (Delta 10A) and the CD is 107dB/W per B&C but it;s sitting in a SEOS 12. Both specced at 8-Ohms. The woofer has an impedance dip at ~200Hz-1500Hz where it goes from 7.1->5.2->7.1-Ohms. Tweeter is mostly flat from 1300Hz to 20kHz.

Has anybody ever used these to Chi-amp 2-way passives before? Is the componentry from unit to unit consistent enough to have two units form a sort of matched pair?

I am hoping to just bypass the volume pot and let these hang out on the backs of the cabinets (one per speaker) in a chain of: line out -> DBX PA2 -> PA3s (R-channel= tweeter + protection cap ; L-channel= 10" woofer). I have the original XO schematics, vituixCAD, 2-channel o-scope w/ SE probes and function gen, and a multimeter. Calculated spl should be 100+dB up to 5m listening distance using Amir's measurements of 27W real power into 8-Ohms. Just checking if these can handle, or if I am setting myself up to require a week's worth of REW and DBX tweaking. Thanks!
 

ChUmI

Member
Joined
Feb 5, 2021
Messages
9
Likes
16
Location
Stuttgart
Hi @amirm , I tried to measure my Topping PA3s and have a little bit different results. I tried to use similar settings on my analyzer although I can't set them exactly like you because I have an APx515. The SINAD came out at the same level for me but the FFT looks quite different for me and I wonder why ???? I also do not have the original AUX-0025 filters, but my own designs, which are, however, similar to the AUX-0025 - they have exactly the same frequency characteristics:

1664994566920.png


If I look at your FFT it looks much different... noise goes down to -140dB abowe 1kHz...

index.php


Do you know maybe and would you tell me what can cause this difference? I've specially uploaded a screen shot with all the analyzer settings so if something is wrong I would appreciate any comments Thank you :)
 

restorer-john

Grand Contributor
Joined
Mar 1, 2018
Messages
12,579
Likes
38,280
Location
Gold Coast, Queensland, Australia
Hi @amirm , I tried to measure my Topping PA3s and have a little bit different results. I tried to use similar settings on my analyzer although I can't set them exactly like you because I have an APx515. The SINAD came out at the same level for me but the FFT looks quite different for me and I wonder why ???? I also do not have the original AUX-0025 filters, but my own designs, which are, however, similar to the AUX-0025 - they have exactly the same frequency characteristics:

View attachment 235422

If I look at your FFT it looks much different... noise goes down to -140dB abowe 1kHz...

index.php


Do you know maybe and would you tell me what can cause this difference? I've specially uploaded a screen shot with all the analyzer settings so if something is wrong I would appreciate any comments Thank you :)

Amir's reference for the sine is Volts (6.37V at the top of the peak). You've set your Y scale to be Volts RMS, not Watts (over 4R) like Amir. Not comparing apples to apples.

Also, your reference level is about 12-13dB higher. Amir's has normalized to 0dBrA.
 
Last edited:

pma

Major Contributor
Joined
Feb 23, 2019
Messages
4,591
Likes
10,727
Location
Prague
Yes but 5Vrms makes 6.25W/4ohm, That’s not a big difference. And one can also read relative to peak distortion value. The main difference is that the new measurement has much higher high harmonics, the forest at higher frequencies. This is not good. The reason may be that the different production piece has been measured. I have a new explanation what means class D, D stands for dirty.
 

deadwood83

Member
Joined
Sep 1, 2022
Messages
91
Likes
160
I haven't found much objectionable with the PA3s, but the 2 amps I got had rather gross (+/- 10%) channel imbalances on the potentiometers. Both are marked 'CTR' and both are equally bad. Luckily also an easy fix.

MA12070 has a 1s DC protection circuit which seems way too long, but the whole output 'filter' on the PA3s is just a teeeeeny tiny diode to gnd on the + output terminals so it might be okay.
1665846094939.png


It is a very interesting chip which seems to embrace all the SOC advancements of the past decade and try to integrate some new ideas into amp silicon. I have found it to be very very very neutral. Still seems a little odd to have a balanced input when Infineon says that EMI in filterless mode is only good for 60cm or so of speaker cable, and input voltage swing is limited to 1.8Vpp. The PBTL efficiency also seems somewhat poor (but this really is made for mobile devices so... it's not really the strong suit).

But I am surprised that Topping did not seem to take full advantage of the I2C programmable settings on the chip. They used an STM8 on the board but the dedicated I2C lines are used as GPIO (only one pin, other I could not find a termination, maybe floating?) and the actual I2C init on the MA12070 is handled by GPIO pins which I assume are bit-banging the commands on start. Same engineer who designed the LA90 but the implementation of MCU is just weird. Green boxes are STM pin mappings, red text is Topping connections. Both chips are connected via I2C in parallel which is odd since both chips have the exact same I2C Master/Slave configuration (IIRC both are configured as master).

1662650093781-png.1088938


It does look for power good voltage levels before init (both the 26V and the 5V rail are monitored), so that's helping avoid any pop on start. I tried emailing Topping to see if they would provide a balanced pre-select always on firmware but no response. And the binary is read protected. Since the bit-bang is unlikely to allow enough program memory to do many read/write operations it's probably not even in power mode profile 1 or 2 which could be even more beneficial for the design. Instead most likely in profile 0 (default after restart) which is made for power efficiency in mobile applications. Seems to be set up to look for voltage good, assume the MA12070 is on, maybe a short delay, then fire and forget some I2C? My scope has some protocol decoding, maybe I will investigate. It seems like they are leaving another 5-10% performance on the table if the MA12070's aren't getting configured. All the ingredients are there. But maybe they are leaving it in efficiency mode because the thermal management design just straight-up doesn't work well?


One other observation where topping dropped the ball. Heatsink is mounted on the plastic case via what feels like a low W/mK thermal pad. The MA12070 is not designed to dissipate heat this way. It is meant to dissipate via epad on the PCB. On both my amps, the epad was not even connected. The heatsink does basically nothing and in operation it is maybe detectably not cold; but the PCB gets quite warm by comparison. The chips are likely running quite hot and only contact (w/o solder) thermal dissipation is happening on the PCB side. Since silicon dielectric breakdown resistance decreases with heat (hence why NAND lives longer / has more write cycles when run a little warm) this is sort of really foolish error.

1665847519658.png


None of the output blocks were soldered straight either, which caused the PCB to be wedged against the case rails on the inside with both amps. No ground contact points were exposed on the PCB edges (by design), but the case scraped away the solder mask when the amp was assembled, creating ad-hoc ground paths. Perhaps unintentional ground paths (ground plane extends to the scraped areas) may contribute to some unit by unit variance.

1665847762087.png


Granted I have a sample size of 2 which both came from Apos, but they both had identical defects in build quality. The bottom side of the board is also completely unwashed but the top side is washed very clean. The heatsink, from factory, was not quite finger tight on both units.

I still think it is a fine amp for what it is. But for little to no additional cost (since everything is already there) the amp could be better. Ultimately it is held back by assembly quality (or lack thereof) in my opinion. I am still baffled by the STM8 layout. It literally has dedicated I2C pins which have hardware functions baked in. The particular STM8 on display is literally their lowest end model. It seems like more effort, less efficiency, and less overall functionality to bit-bang the I2C. The traces already run on the bottom of the board and pass through vias under the chip which makes bit banging all the more confusing. It's like forcing UDP for a TCP communication when you literally have hardware TCP acting as GPIO.
 

ModDIY

Addicted to Fun and Learning
Joined
Mar 26, 2020
Messages
662
Likes
424
Location
Canada
Hello everyone,

could someone tell me if the balanced entry goes through the pot volume?

If so, is there an unbalanced conversion?
 

deadwood83

Member
Joined
Sep 1, 2022
Messages
91
Likes
160
Hello everyone,

could someone tell me if the balanced entry goes through the pot volume?

If so, is there an unbalanced conversion?
both BAl and SE inputs go through the 4-gang pot. The front button signals an STM8 IC which controls two multi-pole Omron relays to switch between inputs. SE is converted to BAL before the volume pot by an OPA1656 because the MA12070 only works with balanced inputs.

The entire BAL signal chain before the amp ICs:
TRS -> Relays -> Volume Pot -> Input caps -> chips.

The entire SE chain is:
RCA -> OPA1656 -> DC filtering caps -> Relays -> Volume Pot -> Input Caps -> chips.
 

ModDIY

Addicted to Fun and Learning
Joined
Mar 26, 2020
Messages
662
Likes
424
Location
Canada
Thanks for the information. He's a friend who has the pa3s (I have the Sabaj A20a 2019), and we were asking the question.
 

JakeK

Active Member
Forum Donor
Joined
Nov 9, 2022
Messages
137
Likes
122
Location
UK
I joined these forums as I've been reading the reviews and measurements here with interest. These might just be the most impartial reviews out there with actually comprehensive scientific measurements. The only measurement that seems lacking is the power consumption at the wall which to be fair I haven't seen in many other places either.

I want to upgrade from my Amptastic Mini-T TA2020 based amp to something with a little more power and the PA3s seems like a good candidate. I'm using it to power some Gale G20 (nominal 6 to 8 ohm) bookshelf speakers on my desk with my PC. The TA2020 sounds good but up to about the 10 o'clock position on the volume which is fine for quiet desktop listening to most sources. It's not enough for listening to a classical symphony as it this power level the quiet parts are too quiet. So I want something small and unobtrusive with low power consumption but at least 2x the power of the TA2020 with low distortion. I do plan to upgrade the speakers at a later date as well but the amp seems more important for now.
 

deadwood83

Member
Joined
Sep 1, 2022
Messages
91
Likes
160
I joined these forums as I've been reading the reviews and measurements here with interest. These might just be the most impartial reviews out there with actually comprehensive scientific measurements. The only measurement that seems lacking is the power consumption at the wall which to be fair I haven't seen in many other places either.

I want to upgrade from my Amptastic Mini-T TA2020 based amp to something with a little more power and the PA3s seems like a good candidate. I'm using it to power some Gale G20 (nominal 6 to 8 ohm) bookshelf speakers on my desk with my PC. The TA2020 sounds good but up to about the 10 o'clock position on the volume which is fine for quiet desktop listening to most sources. It's not enough for listening to a classical symphony as it this power level the quiet parts are too quiet. So I want something small and unobtrusive with low power consumption but at least 2x the power of the TA2020 with low distortion. I do plan to upgrade the speakers at a later date as well but the amp seems more important for now.
IMo the speakers are always the most important part. I would do those first. There are also a lot of frankly incredible sounding monitor speakers which require no external amplification. The JBL 308P Mk2 frequently go on sale about this time of year for about the price of a single PA3s per speaker (and sometimes much less, as I got my pair in 2019 for $100/speaker). I cannot think of a better sounding near/midfield setup for less money, but I also do not listen over 80-85dB as I value my long-term hearing.

I put some 308P's on my DX7 Pro (2019 version) and for me.... aesthetics aside it's basically endgame desktop listening. Can't think of what else would really be needed for a consumer who just wants to listen to music; and accurately. PC allows you to run Equalizer APO + Peace or some other backend EQ/DSP if you need to make room corrections too.
 

JakeK

Active Member
Forum Donor
Joined
Nov 9, 2022
Messages
137
Likes
122
Location
UK
IMo the speakers are always the most important part. I would do those first. There are also a lot of frankly incredible sounding monitor speakers which require no external amplification. The JBL 308P Mk2 frequently go on sale about this time of year for about the price of a single PA3s per speaker (and sometimes much less, as I got my pair in 2019 for $100/speaker). I cannot think of a better sounding near/midfield setup for less money, but I also do not listen over 80-85dB as I value my long-term hearing.

I put some 308P's on my DX7 Pro (2019 version) and for me.... aesthetics aside it's basically endgame desktop listening. Can't think of what else would really be needed for a consumer who just wants to listen to music; and accurately. PC allows you to run Equalizer APO + Peace or some other backend EQ/DSP if you need to make room corrections too.
I've given powered speakers plenty of thought but so many of them have audible hiss which seems like a deal breaker for me really. Amirm said this in his review of the one you mention: 'And oh, yes, there is hiss from the tweeter. It is very audible with your ear at the tweeter level' and that exactly where my ear will be if the speaker is on my desk on a little stand. Plus I don't really want to have a DAC.
 
Last edited:

deadwood83

Member
Joined
Sep 1, 2022
Messages
91
Likes
160
Ah, if your speakers are closer than 2 feet (where AMirm said the hiss disappeared) then the 308Ps might be too large anyway. Mine are wall mounted and at about 120-150cm from my ears to the tweeters so there is no hiss I can hear above my breathing/heartbeat. I was saying powered units so you could declutter a desk (dac -> speakers, since they have amplification inside) instead of a more common passive setup of dac -> amp -> speakers. Well, the commonality of passives is decreasing. Active speakers tend to be more consistent in most scenarios so you see sonos, soundbars, etc all moving that way as opposed to the older HT in a box kits that used to be favored by many consumers.

The PA3s is pretty decent. I do believe that the units Amrrim receives are not hand-picked, but I also think his test setup is a best case scenario just to show the technical capabilities of the units.It will also hiss if your source does not have some loading though. E.g. if you have a power outage and the amps flip back to on/SE when you feed them BAL or if they flip to on/BAL when you feed them SE then they will hiss. It is part of the MA12070 chip, maybe the active feedback? I use one for each of my relatively efficient 2-way speakers on an active crossover. If I walk into the room and hear a faint hiss, I need to check each amp to make sure it is on the correct input. They seem to work well for me, build quality problems and inconsistent volume pots aside.
 

JakeK

Active Member
Forum Donor
Joined
Nov 9, 2022
Messages
137
Likes
122
Location
UK
So I got my PA3s and I can add some measurements made with a basic power meter plugged into the wall. With the unit plugged in but not turned on it draws less than 1 watt as the unit shows zero. Turned on but not playing any sound it's 3w, rising to 4w with the music loud enough to get my wife to come and complain she can't hear the TV downstairs. Compared to my TA2020 based Amptastic Mini-T which draws less than 1w even when turned on it's less efficient at low volume more but still very efficient. By comparison my old Z5500 computer speakers draw 11 watts when turned off and 40 watts when playing fairly quietly.
 

Fleuch

Member
Joined
Aug 3, 2019
Messages
97
Likes
58
Problems with, and failure of, the Topping PA5 amplifier are well documented in another thread.

A Topping PA3s purchased in the last month has completely failed and now has been returned to the retailer. It is fair to say the output levels never stressed the amplifier and the case remained cool at all times.

Is this an isolated failure or an indication of reliability problems with Topping PA series speaker amplifiers?
 

deadwood83

Member
Joined
Sep 1, 2022
Messages
91
Likes
160
I don't see much to go very wrong in the PA3s. Aside from the amp chips and a single UC there's just a few transistors, I think 3 LDO's and a single opamp. I think fewer than 5 SOT-23 transistors, 3 LDOs, one STM8, 2x MA12070 chips, and the rest are SMD passives. There might be 2-4 through hole caps. Not a whole lot to go wrong. Though both of my amps have a terminal block which is soldered very crooked. If subsequent units had a slightly worse assembly then the rear terminals might break from fatigue.

Just not a whole lot to go wrong.
 

dsnyder0cnn

Addicted to Fun and Learning
Joined
Dec 14, 2019
Messages
540
Likes
800
I don't see much to go very wrong in the PA3s. Aside from the amp chips and a single UC there's just a few transistors, I think 3 LDO's and a single opamp. I think fewer than 5 SOT-23 transistors, 3 LDOs, one STM8, 2x MA12070 chips, and the rest are SMD passives. There might be 2-4 through hole caps. Not a whole lot to go wrong. Though both of my amps have a terminal block which is soldered very crooked. If subsequent units had a slightly worse assembly then the rear terminals might break from fatigue.

Just not a whole lot to go wrong.
I was thinking the same thing. Possibly just a bad SMPS?
 

Fleuch

Member
Joined
Aug 3, 2019
Messages
97
Likes
58
Both posts from Deadwood suggest a reasonable prediction for failure.
One other observation where topping dropped the ball. Heatsink is mounted on the plastic case via what feels like a low W/mK thermal pad. The MA12070 is not designed to dissipate heat this way. It is meant to dissipate via epad on the PCB. On both my amps, the epad was not even connected. The heatsink does basically nothing and in operation it is maybe detectably not cold; but the PCB gets quite warm by comparison. The chips are likely running quite hot and only contact (w/o solder) thermal dissipation is happening on the PCB side. Since silicon dielectric breakdown resistance decreases with heat (hence why NAND lives longer / has more write cycles when run a little warm) this is sort of really foolish error.

Though both of my amps have a terminal block which is soldered very crooked

When the case remains cool and the PCB warms up, with the obvious lack of thermal dissipation. gives a reasonable diagnosis for failure, as does imperfect soldering. At best these issues demonstrate a lack of attention to detail on the production line.
 
Joined
Aug 25, 2022
Messages
14
Likes
3
Does anyone have a channel misbalance under 1/4 volume. It seems that right channel is very slow to come up to the match volume of the left channel.

I worked around it by lowering the volume of the DAC but i was wondering if there is a way to fix this permanently?

Another thing i noticed with this amp is even when the volume knob is set to Min I can still hear the music through my speakers albeit at a very low volume. is this normal as well?
 
Last edited:
Top Bottom