Dual Merus MA12070 Amplifier Sabaj A20 fully balanced : pictures inside )

[Moosi]

To be more precise: The noise from the speakers was analysed and compared without a DAC plugged. So it's not a question of RCA or XLR (tho the XLR has been reported to be silent).

 

[Moosi]

Just saying. Dual chip won't give you much more power. 10% probably.

Wait a second. From the Infineon manual it says: "For providing additional power the MA12070 can be configured for mono operation using a parallel BTL mode (PBTL), as shown in Figure 9-5. In this fashion the two BTL output stages are combined to be able to deliver twice the current. This makes high output power sub-woofer application possible."

By the specs one MA12070 chip can deliver 8A@26V=208W. Doubling that would give 16A@26V=416W total max. in theory, wouldn't it? Now, since volume is related to gain and the speakers sensitivity, the PBTL amp is not playing louder by definition, for the gain does not change, but is in theory able to drive speakers twice the sensitivity rating. Isn't that what you would expect an amp twice the power to do anyway? I've had the BTL configured SMSL SA300 before and there is without a doubt quite a lot more juice to the bone with the A20a on my 3.5Ohm/83db KEF LS50 Metas.

 

[JohnYang1997]

Wait a second. From the Infineon manual it says: "For providing additional power the MA12070 can be configured for mono operation using a parallel BTL mode (PBTL), as shown in Figure 9-5. In this fashion the two BTL output stages are combined to be able to deliver twice the current. This makes high output power sub-woofer application possible."

By the specs one MA12070 chip can deliver 8A@26V=208W. Doubling that would give 16A@26V=416W total max. in theory, wouldn't it? Now, since volume is related to gain and the speakers sensitivity, the PBTL amp is not playing louder by definition, for the gain does not change, but is in theory able to drive speakers twice the sensitivity rating. Isn't that what you would expect an amp twice the power to do anyway? I've had the BTL configured SMSL SA300 before and there is without a doubt quite a lot more juice to the bone with the A20a on my 3.5Ohm/83db KEF LS50 Metas.

Calculate what ohms at which you could achieve that power.

 

[JohnYang1997]

Will using balanced input give us more power? I know for headphones balanced outputs give more power. What about speaker amps?

BTL bridge tied load configuration is basically balanced output. It does provide more power when it's voltage limited.

Ma12070 amps are mostly in this configuration already.

Balanced input doesn't give more power.

 

[Toku]

PBTL connects two BTLs in parallel. It doubles the output current capacity but does not change the output voltage.
Therefore, even if the amplifier changes from BTL to PBTL, the output voltage does not change, so the power that can be supplied to the speaker currently in use does not change.
The power increased by PBTL can be effectively utilized by replacing the speaker with a lower impedance. In addition, some speakers that are said to be difficult to drive have an extremely low impedance at a certain frequency. Such speakers can be easily driven by using PBTL.
To put the relationship between PBTL and BTL more clearly, it is the same as the relationship between parallel connection and series connection of batteries, and PBTL is a state in which batteries are connected in parallel.

 

[bravomail]

Here's my short review of Sabaj A20a.
First, setup - PC, Foobar 2000 WASAPI, with volume at 98% and FLAC, MP3, AAC files (44.1Khz, no DSD, no upsampling, 48KHz for regular audio like youtube) connected to Topping E30 via USB, Topping E30 in DAC mode (which means max volume, or 2V) via RCA cables to Sabaj A20a, speakers are JBL A170 (towers). All cables are generic(!). Search Monoprice....

Power limits - A170 are more sensitive and easier to drive than my 1st gen ELAC B6. But I think even ELAC B6 wouldn't be a problematic load. With A170 I set the volume at -34dB - normal listening. At -22dB - very loud. I haven't tried more, but u get the picture. I don't think power will be a problem for in-room listening.

Sound - like ppl described above, is "musical", compared to say AIYIMA A07, and more congested/muddy with multiple instruments (say some electronic mix with drums and voice) - both of which I think are signs of some harmonic distortions. A07 is cleaner, resolves all the instruments, and A07 is much more powerful, I listen to 9-o-clock for normal in-room volume. So, is the sound of A20a objectionably muddy? No. Some ppl might just prefer it. "Musicality".

Now, I have to add a comparison to my Yamaha RX-683 AVR. It's underperformance brought me to this search of an amp. With YPAO EQ set to Flat it became better, than YPAO Natural, which was absolutely abysmal and muddy. But even with YPAO FLAT it's still very muddy. Maybe it's me, I just like Class-D sound more. I still have Alientek D8, and it runs circles around Yamaha. Aiyima A07 and Sabaj A20a both do the same - outperform Yamaha in clean sound. Yamaha has many useful features, like NetRadio - it can stream directly some nifty internet radio stations. Yamaha has sub-out with lowpass for sub and highpass for speakers, parametric equalizer, YPAO room correction, speaker distance delays, speaker and sub equalization, DD5.1 and other multichannel formats decoder with subsampling it back to my 2.1 etc etc. But with all that goodness the end result is abysmal. So here we are.

Which one will you choose - if A07 had a remote volume control and dedicated sub-out - I'd embrace A07 with open arms. It's just no contest. But, I need remote control, I need sub-out with volume control - balanced with speakers volume. So, I will keep testing more, if I find some issues with my music (and movies) - I might send it back and wait for "next big amp".

 

[Moosi]

Did you break the A20a in for a minimum of two weeks of playtime with turning the AC power on/off a couple of times (to deload/load the caps)? How was your E30 powered? With isolation or with a mains-loop?

I remember there was a very notable change in sound with this amp which some people might describe as "congested" at first. Mine really opened up after about two weeks of daily play time. It was very notable to a point where I thought my speakers changed. Unless you push it really hard (75%+), there cannot possibly be higher distortion compared to TPA3255 for the MA12070 is superior in about every aspect (reflecting ripple, efficiency, distortion, slew rate), except for total power (it was pretty much designed to beat the TPA chips series in the low-mid power region). Most known amplifier manufacturers break their amps in before they ship it. I seriously doubt Sabaj does it with this ~195$ amp.

 

[bravomail]

Did you break the A20a in for a minimum of two weeks of playtime with turning the AC power on/off a couple of times (to deload/load the caps)? How was your E30 powered? With isolation or with a mains-loop?

I remember there was a very notable change in sound with this amp which some people might describe as "congested" at first. Mine really opened up after about two weeks of daily play time. It was very notable to a point where I thought my speakers changed. Unless you push it really hard (75%+), there cannot possibly be higher distortion compared to TPA3255 for the MA12070 is superior in about every aspect (reflecting ripple, efficiency, distortion, slew rate), except for total power (it was pretty much designed to beat the TPA chips series in the low-mid power region). Most known amplifier manufacturers break their amps in before they ship it. I seriously doubt Sabaj does it with this ~195$ amp.

I don't believe in break-in/burn-in with solid-state amps. There is nothing to break-in or burn-in.
E30 power and isolation is irrelevant as I don't hear any noise or hiss. It's just the way music rendered, and I suspect there will be harmonic distortions. Look at SA300 review by Archimago - it didn't get good SINAD numbers. 60 to 70 db SINAD.
A7 is really fully optimized, they squeezed every last bit of juice out of TPA chip. I also was sceptical when I read Amir review and his recommendation. Some russian guy Stalker tested numerous DIY TPA32xx boards, and they were all crappy. To the point that he gave on a chip itself and its performance. Like that Frequency graph dependent on load. Good amps don't do that. A07 does that but still somehow manages to sound good.
So, maybe there is a hope with Dual Merus, look at Loxjie A30 - it measured fine, and I haven't heard about congested sound in any of A30 reviews.
A20a sound is not objectionable. It is rather pleasant. It is when u try to listen closely, and discern the layers/instruments - u might get disappointed. I'm running it for many hours daily. We'll see if my ears will "break-in"

 

[steve_b_in_az]

Trying to decide if should try one of these up or wait for impressions/reviews of the SA400... I have been using a pair of modded Bel Canto evo 200.2 monoblocks for years but after a recent move been playing music through Topping tp32 + JS2 power supply and I've been impressed by just how good that little amp sounds - making me question 2 big amp boxes + matching pre, especially with all the listening impression provided for these newer chips. Not sure if I should expect a20 and/or SA400 to be clearly better than my old bel canto gear or if those are still relatively good performers?

I also have a CI Audio e200s with older hypex modules and my ears prefer the bel canto monos. The hypex is probably superior from measurement standpoint but it certainly doesn't control speakers as well as the bel canto monos, nor is it rated for as much output. Speakers are JM Lab/Focal micro utopias, Clements Rt-7, and Gallo reference (original large aluminum spheres). Really appreciate all the details listening impression you've provided here + references!

I have been a tube amplifier fan for many years, but I find that I really like the definition and instrument separation in good class-D amps as much or perhaps more than in my tube amp.

I am considering buying a used Bel Canto evo 200.2 The reviews were great for it when it first came out. Also, I have listened to much more modest Tripath-based amplifiers, and I have been very impressed by their musicality - especially given their low price. However the EVO design is 20 years old now, and I am sure that time has marched on.

I was thinking about something like the Sabaj A20 (or perhaps the Aiyima a08 or a09), which would be in the $200 range, as opposed to the Bel Canto, which would be in the $800 range. However, despite using older technology, the Bel Canto was designed from the start as a higher-end product. Therefore I am really uncertain about which direction to go.

I'm wondering if you've had a chance to compare any of the newer class D amps with the venerable Bel Canto 200.2

Thanks so much in advance for any advice you can offer!

 

[dr0ss]

I have listened to much more modest Tripath-based amplifiers, and I have been very impressed by their musicality - especially given their low price. However the EVO design is 20 years old now, and I am sure that time has marched on.

This is probably not the right place to discuss this -- not only because it isn't specifically about the Sabaj amp but also because there are no measurements involved -- but if you like the sound of T amps (as I do) have a look at this thread:

Review of Poppulse T180 MD version German Market, dual Tripath

Hi amigos, I don't know if you've heard of the Poppulse T180, in its MD version for the German market. But I was lucky to be able to have one) The MD version was specially made for the German market with best caps and some improvements (wiring, OP amps etc). I discovered this amp on a German...

www.audiosciencereview.com

 

[bravomail]

wanted to add. take my review with a grain of salt as it is sighted comparison. Wait for objective measurement and base your decision on that.

 

[ces04848]

I have been a tube amplifier fan for many years, but I find that I really like the definition and instrument separation in good class-D amps as much or perhaps more than in my tube amp.

I am considering buying a used Bel Canto evo 200.2 The reviews were great for it when it first came out. Also, I have listened to much more modest Tripath-based amplifiers, and I have been very impressed by their musicality - especially given their low price. However the EVO design is 20 years old now, and I am sure that time has marched on.

I was thinking about something like the Sabaj A20 (or perhaps the Aiyima a08 or a09), which would be in the $200 range, as opposed to the Bel Canto, which would be in the $800 range. However, despite using older technology, the Bel Canto was designed from the start as a higher-end product. Therefore I am really uncertain about which direction to go.

I'm wondering if you've had a chance to compare any of the newer class D amps with the venerable Bel Canto 200.2

Thanks so much in advance for any advice you can offer!

DM sent

 

[Moosi]

I have just put the last puzzle piece to this amp. My journey has come to an end. It's as close to perfection as it gets.

An N30 Ferrite ring from Epcos/TDK (B64290L40X830) has been installed to the AC cable from outside. Huge improvement! Not only to white noise, but from to my subjective ears even to the blackness of the sound (subtly). From this last expriment (with the knowledge of all prior) I am certain that there is HF noise leaking from the AC mains to the speakers-out-panel. You might not notice it right away. Putting the lighting match box inlay to the spkout-panel has improved it a bit, but the ferrite ring is far better at it. Thanks to daniboun for recommending ferrite. However, I did not want to put any inside the case to the speaker cables, for some people claim it to "choke the sound". Putting it to the AC cable is a "safer" solution.


Noise reduction roadmap (all RCA, all ear to speaker):

default amp: from -26db (from 57% vol.)

1. insulating SPcables(inside) +lighting match box cover to SPout-panel : from -21db / 65% vol.
2. replacing Mornsun (burst mode) with Meanwell RPS-400-24 PSU (no burst mode): from -17db / 72% vol.
3. DAC running from 5Vout of Meanwell: from -29db / 52% vol. (ground loop)
4. galvanic mains isolation for DAC: from -12db / 80% vol.
5. N30 ferrite ring to AC power cable: from -02db / 97% vol.

*from = noise starting to be noticable

I highly recommend anyone (even if you use XLR) to get such a ferrite ring, for they cost only a few bucks and are probably having the most effect on cleaning the amp up. It's crazy detailed now. I could hear a hamster coughing on most tracks.

 

[Toku]

I have just put the last puzzle piece to this amp. My journey has come to an end. It's as close to perfection as it gets.

An N30 Ferrite ring from Epcos/TDK (B64290L40X830) has been installed to the AC cable from outside. Huge improvement! Not only to white noise, but from to my subjective ears even to the blackness of the sound (subtly). From this last expriment (with the knowledge of all prior) I am certain that there is HF noise leaking from the AC mains to the speakers-out-panel. You might not notice it right away. Putting the lighting match box inlay to the spkout-panel has improved it a bit, but the ferrite ring is far better at it. Thanks to daniboun for recommending ferrite. However, I did not want to put any inside the case to the speaker cables, for some people claim it to "choke the sound". Putting it to the AC cable is a "safer" solution.


Noise reduction roadmap (all RCA, all ear to speaker):

default amp: from -26db

1. insulating SPcables(inside) +lighting match box cover to SPout-panel : from -21db
2. replacing Mornsun (burst mode) with Meanwell RPS-400-24 PSU (no burst mode): from -17db
3. DAC running from 5Vout of Meanwell: from -29db (ground loop)
4. galvanic mains isolation for DAC: from -12db
5. N30 ferrite ring to AC power cable: from -02db

I highly recommend anyone (even if you use XLR) to get such a ferrite ring, for they cost only a few bucks and are probably having the most effect on cleaning the amp up. It's crazy detailed now. I could hear a hamster coughing on most tracks.

Thank you for the report.
I read the content repeatedly with great interest, but I would like to ask you because there are some points that I cannot understand well with my English ability.
What does from -21db or from -17db mean? Is it the volume level where noise starts to appear?
After these measures, did the previous noise disappear even if the volume was -00 at MAX?

 

[Moosi]

I'm sorry for being not precise enough.

-60db = min. volume
-00db = max. volume

"from -21db" for example means that there is noise starting to be notable from -21db (of -00db) on the volume. I will translate to percentages to make things easier. It is now totally dead silent up to 97%. I have never achieved that level of silence before, neither with insulating all cables inside (except AC), nor with swapping the PSU with another one emitting far less EMI (CRISPR32 to CRISPR11) -> leading to the conclusion that the main source of EMI inside this amp is coming from the AC-ins which are in close proximity to the SP-out-panel (panel of where to plug the speakers). The ferrite to the AC cable was the cherry on top eventually and is most probably the biggest (and cheapest) way of improving this amp. Just make sure to leave to room from the ferrite ring to the actual speakers cables for we want it to filter out frequencies from the AC-ins, not from the speakers.

 

[ModDIY]

I'm sorry for being not precise enough.

-60db = min. volume
-00db = max. volume

"from -21db" for example means that there is noise starting to be notable from -21db (of -00db) on the volume. I will translate to percentages to make things easier. It is now totally dead silent up to 97%. I have never achieved that level of silence before, neither with insulating all cables inside (except AC), nor with swapping the PSU with another one emitting far less EMI (CRISPR32 to CRISPR11) -> leading to the conclusion that the main source of EMI inside this amp is coming from the AC-ins which are in close proximity to the SP-out-panel (panel of where to plug the speakers). The ferrite to the AC cable was the cherry on top eventually and is most probably the biggest (and cheapest) way of improving this amp. Just make sure to leave to room from the ferrite ring to the actual speakers cables for we want it to filter out frequencies from the AC-ins, not from the speakers.

Thanks Moosi for the advice about the ferrite. I have a few, I will definitely do it.

I must say that I was starving myself about this amp versus my 150w/c class AB amp.

But yesterday while listening seriously at higher volume, it's been 2 weeks since it's plugged in ... wow I had the same amplifier, well a powerful bass (not as much impact as my class AB amp) but certainly excellent. The width and depth have also increased, in short this is what was lacking at the beginning.

I finish my listening tests and put my main amp back on and give you a little feedback.

 

[ModDIY]

@Mosi I have a quick question, did you pass the 3 wires of the AC cable through the ferrite?

If so, I will prepare a cable, cut it and resolder.

 

[bravomail]

An N30 Ferrite ring from Epcos/TDK (B64290L40X830) has been installed to the AC cable from outside. Huge improvement!

Hi Moosi. While we all appreciate the work u r doing and sharing - try using some, even rough, measurements before and after your mods.

 

[Moosi]

While we all appreciate the work u r doing and sharing - try using some, even rough, measurements before and after your mods.

No. For various reasons:

a)Measuring "something" with "some cheap/mainstream measurement device" will not contribute anything of value beyond word of mouth.
b)In order to give you real value by measures, I would have to use professional or at least semi-profesionell equipment.
c)Even If I had the equipment, In order to do that I would have to be educated on how to conduct these measures properly.
d)I'm not Amir or Archimago. Seriously. Not even close.
e)From all the things I did to the amp, the ferrite ring is the most effective and also the easiest to do and cheapest to try. If you are hesitant to spend ~8$ and try for yourself, don't. If you demand measures, well, there are plenty for those ferrites. Just look for "N30" because there are different materials for different frequencies.

edit: To clear it up. The improvement described as "huge" is related and compared to those other alterations I made and described above. It's not a "huge improvent" all by itself. Don't expect this ferrite ring to turn it into a new amp. If you don't have white noise from the speakers, it might not actually do anything. If you do have white noise, this would be the very first thing I would "mod" the amp with. This is from the experience of about a dozen times opening it up.

 

[Toku]

Since the noise problem of Sabaj A20a disappears when the volume is lowered in proportion to the volume, it can be 100% concluded that the noise is mixed between the RCA terminal and the electronic volume chip of NJW1194.
Looking at the internal photos you uploaded, the speaker output of the MA12070 chip is connected to the speaker terminal on the rear panel from the connector just behind the front panel via the top of the amplifier board.
The speaker wiring during this time is closest to the OP amplifier and the NJW1194 chip.
High-level high-frequency noise associated with the switching operation of the MA12070 chip is superimposed on this speaker wiring. This high-frequency noise is mixed into the OP amplifier circuit and electronic volume, and this is thought to be the cause of the noise emitted from the speaker.

To improve this, it is effective to install many 2-split type ferrite cores as shown in the photo near the connection on the front panel side of the speaker wiring, or to shield the speaker wiring with aluminum foil. Also, although the work is difficult, it is necessary to bypass the route so that the speaker wiring is as far away from the OP amplifier and NJW1194 chip as possible.