LSA Voyager GaN 350 Teardown (Class D Amp)

[respice finem]

Who remember the reply from manufacturer? This is going to get awkward.

During my decades with Hi-Fi stuff, I've developed a "run like hell" reflex whenever I see "audiophile lyrics" from a manufacturer.
I ended up prefering pro gear, where nobody would even try.

 

[JSmith]

reply from manufacturer

Now the product makes complete sense... it reflects their poor attitude to audio reproduction and their customers.



JSmith

 

[Yevhen]

When you take the layout into account, the ESR is rarely much better (ditto ESL), but you can get a boost in ripple current rating. However, for a Class-D audio amplifier, ripple current is practically inconsequential with reasonable quality caps for switch modes. It will be pretty easy to hit the peak required ripple current rating, and the average ripple current will be a small fraction of the rating having little impact on life.

Since they are "glueing them to the board", the same can be done for larger capacitors eliminating mechanical stress. There is a happy medium. These look like 12.5x20mm, which are the highest volume/lowest cost for power supplies, at least for lower voltages (<=100, and I would say <=63). On the primary side, not so much for high voltage.

With larger capacitors, they would not fit in the 1U height.

 

[pma]

What is fascinating is that each amplifier module is stereo! They are wired in bridge mode, doubling their power and likely doubling their noise+distortion.

Bridging suppresses even harmonics distortion products. You can see it in your measurement.

It is stupid that they use PCB XLR connectors as panel connectors, thus the soldering troubles. Amateurism of worst kind.

 

[antcollinet]

I meant to mention this. There is no connection anywhere to the chassis. Whether it is floating or not, is something I should have tested but forgot.

Bloody hell - if that is true, it's a serious electrical safety problem. No way this can be class 2 insulation with all those flying leads. I also doubt there is the required 5.5mm between the underside of the PSU pcb and the metal plate attached to the chassis.

 

[antcollinet]

- Elegant Audio Systems designed the GAn Systems reference design if I am not mistaken

- Are they achieving chassis grounding via the ground connector to the power supply? It suspect that may connect to the PCB under the screw and then the screw to the chassis. Locking connector, so likely would meet UL, and should meet the <0.1 ohm requirement
.....

But standards generally don't allow you to dual purpose a mechanical mounting screw also to fix the safety earth - and certainly not if there is not flat metal to metal contact (IE screw threads are not allowed to be the connection - although this could be achieved here if there is a metal pillar between the PCB and the chassis)

 

[antcollinet]

And even where the soldering is good - for a quality job I'd expect to see heat shrink sleeve or similar over the joints.

 

[Koeitje]

Who remember the reply from manufacturer? This is going to get awkward.

This sounds like a manufacturer that has no understanding of what is required for high fidelity sound reproduction. Can you imagine a reply like this coming from RME, Genelec or Orchard Audio?

 

[mightycicadalord]

The most offensive part of this amplifier for me are the use of the same color wire for pretty much everything. XLR being all silver makes me crazy.

 

[Dennis_FL]

You will never see a review like this anywhere else. The uniqueness of Audio Science is the reviewer can tear down a device and know what he is looking at.

 

[sarumbear]

@amirm It looks like the combination IEC/Switch module also has a fuse on it. I also see a fuse on the AC end of the PSU PCU. Are they in series, do you think?

 

[sarumbear]

The most offensive part of this amplifier for me are the use of the same color wire for pretty much everything. XLR being all silver makes me crazy.

So easy to make a mistake.

 

[Poseidons Voice]

See my post above. Likely through the AC ground connection to the power supply, then through the PCB via the screw just right on the AC connector. You can see the PCB is plated under the screw there.

Now I see it, and I agree. Sharp Eye!

That being said, it is still...unsafe. The chassis ground connection should be a DEDICATED connection to earth ground without having any other responsibility. Like holding down a PCB! Oy vey.

Oh, and how about the UNTWISTED AC wire coursing itself near the RCA input? Is it too hard to twist? Will fingers fall off? Or did we run out of wire on a $3K product? Flip that PS board 180 degrees, twist the AC wiring and minimize the distance!

I don't care what the PSRR is of this design, but there are ways to build electronics in a safe and repeatable manner. Moreover, what use is an amp if it isn't reliable? Does it matter what it looks like or costs at that point when its...um....CAPUT?!

Has anybody heard of a bonafide engineer called Leo Ayzenshtat who designs fantastic GaN based designs, called the Starkrimson? Hellooooo?

End rant...

Thanks,
Anand.

 

[JayGilb]

@amirm ,

In this picture, I cannot see exactly what is happening at the balanced connector, but it looks like the balanced input is not wired properly. It looks like IN- and GND are simply connected together:


View attachment 166307

Under rated solder gun was used and a lot of the joints look cold. Those large plated connectors really sink a lot of heat, especially when soldered while attached to the chassis. I always take them apart and tin the connector prior to installation, because it's a fine balance between making a proper joint and melting the plastic insulators. Always use flux on any joint.

Since pcb mount xlr connectors were used, at least purchase some simple, small pcb boards that can be soldered after the connector is mounted. There are a lot of companies that will make even small quantities of simple boards for very low cost.

 

[audio2design]

With larger capacitors, they would not fit in the 1U height.

The connectors are taller than the caps already.

 

[audio2design]

But standards generally don't allow you to dual purpose a mechanical mounting screw also to fix the safety earth - and certainly not if there is not flat metal to metal contact (IE screw threads are not allowed to be the connection - although this could be achieved here if there is a metal pillar between the PCB and the chassis)

An end user ground connection cannot be an mechanical assembly screw. This is not a end user connection.

I assumed there is be a metal standoff. Let's hope the paint/anodizing has been removed and that a proper screw/washer is used that will cut through oxidation.

Class-2 is under 100 watts so this can't be class 2. Let's hope the standoffs are tall enough for isolation. The AC wires look suitable and use insulated terminals. They should be mechanically stabilized.

 

[audio2design]

@amirm It looks like the combination IEC/Switch module also has a fuse on it. I also see a fuse on the AC end of the PSU PCU. Are they in series, do you think?

The one on the PSU is due to the MOV. If you have a MOV you need a fuse. They would be effectively in series.

 

[Tatteredmidnight]

Now I see it, and I agree. Sharp Eye!

That being said, it is still...unsafe. The chassis ground connection should be a DEDICATED connection to earth ground without having any other responsibility. Like holding down a PCB! Oy vey.

Oh, and how about the UNTWISTED AC wire coursing itself near the RCA input? Is it too hard to twist? Will fingers fall off? Or did we run out of wire on a $3K product? Flip that PS board 180 degrees, twist the AC wiring and minimize the distance!

I don't care what the PSRR is of this design, but there are ways to build electronics in a safe and repeatable manner. Moreover, what use is an amp if it isn't reliable? Does it matter what it looks like or costs at that point when its...um....CAPUT?!

Has anybody heard of a bonafide engineer called Leo Ayzenshtat who designs fantastic GaN based designs, called the Starkrimson? Hellooooo?

End rant...

Thanks,
Anand.

Once you see the orientation of the psu board you can’t unsee it. It’s such low hanging fruit! Shorter cables, and no need to route the AC mains next to an am module, unbelievable.

 

[audio2design]

[

Under rated solder gun was used and a lot of the joints look cold. Those large plated connectors really sink a lot of heat, especially when soldered while attached to the chassis. I always take them apart and tin the connector prior to installation, because it's a fine balance between making a proper joint and melting the plastic insulators. Always use flux on any joint.

Since pcb mount xlr connectors were used, at least purchase some simple, small pcb boards that can be soldered after the connector is mounted. There are a lot of companies that will make even small quantities of simple boards for very low cost.

Hard to call the joints cold from the pictures. It looks like the solder has penetrated the strands and you can see flow on the connector itself. Likely lead free so not as shiny and does not flow as well. Ditto high silver. I was surprised by no damage to the wire insulation from soldering, but likely some "audiophile" wire with Teflon insulation.

 

[Billy Budapest]

Who remember the reply from manufacturer? This is going to get awkward.

I had forgotten that Walter Liederman purchased LSA from its former owner (Larry Staples, I think). Walter is a stand up guy. If he were to participate here, I am sure dialogue would be constructive, or at the very least, congenial.