Sonance DSP 2-150 Amplifier Review

[amirm]

At least you got it for a very nice price!

 

[ZolaIII]

Knowing for similar performing class AB with similar performance and functionality, with great build quality and twice the power at half the price I had to vote pore (all do it's definitely in not bad category). Check out Dynacord LXXXXFD lineups, I whose alluding on L1300FD starting model which is capable to deliver 220W @ 8 Ohms relatively clean and can be found for 600~700€.

Cryptocurrency Prices, Market Cap, Trading Charts: Bitcoin, Ethereum and more at PiedPiper

Live cryptocurrency prices, market caps, charts & daily insights for a wide range of coins on PiedPiper. Compare trends, top gainers and trading volume 24/7.

docplayer.net

Unfortunately to my better knowledge they are EU only.

 

[hyfynut]

That's not too bad for a DSP amp. Sadly the DSP itself is very limited, making the number of use cases limited.

The dsp included with these is completely acceptable for it's use case and product category. What did you find it lacking?

 

[Toslink]

I'm the original DSP 2-150 amplifier owner who sent in the amp that was tested and posted here. I also work in the custom installation field and can attest to one of the previous posters who mentioned that Sonance quickly replaced this original model with the MKII version that addressed all of the failure modes of the original design. The new version has a solid top and cooling fans on the sides of the case that actively draw air through the chassis. We saw a number of failures due to heating on the original design and these issues were fully resolved in the MKII.

Additionally, when I sent the amplifier in for testing, it had the digital input card installed and the analog card in the shipping box. I wanted him to have both for testing, but I didn't state that in writing and Amir didn't see the analog card until after he'd completed testing and writing up the review. This test likely shows what we'd have seen with my amp had the analog card been used.

Despite not reaching the ranks of the better amplifiers tested on the site, I use the Sonance DSP line of amplifiers regularly and can say they offer excellent performance in distributed audio and 70v (using their DSP 2-750 MKII amp) applications. My hope is when the amplifiers are updated, we'll see a few newer features, such as firmware updating through the UI, but, again, the amplifiers are quite solid and even offer IP and IR control, which allows for some very creative installation possibilities that are not needed in standard residential 2-channel applications.

 

[Inte Grator Dominus Rex]

Why would turning the trim pots get it back to where it was then?

@restorer-john I know I'm beyond late to this party, but I think I've cracked the answer. In the IN/OUT tab of the Sonance settings, there are options for both "Output Volume" and "Turn On Volume". According to Sonance (in this part of their tutorial video), the unit will default to the latter anytime it is turned off and on again.

Perhaps when Amir was testing the unit these numbers weren't both changed to the same value?

 

[kawal]

I have a MK 2 that I am fixing due to blow amp module. So far this is the info I have:

The DSP used is a Cirrus Logic CS48520-CQZ 300M instructions per second.
ADC ES9840Q is 32bit very respectable performance. This is a 4 channel ADC
+116dB DNR 4 channel mode
+118dB DNR in mono mode
-108dB THD+N 4 channel mod
Bias voltage is 1.65v and based on some Motu forum post that is not optimal. (https://www.audiosciencereview.com/forum/index.php?threads/fixing-the-motu-m4-audio-interface.65701/)
Input buffer are NE5532 and SGM8270-2 - could have been better.

The NE5532 are driving the ADC - maybe a good upgrade would be the OPA1602 as per ESS recommendations. Not sure what the SGM op amp is doing as its powered by 5V while the NE5532 are powered by +/-12v and are biased to 1.65V. maybe the SGM is providing reference voltage?

Amps are hypex UDC250

Power supply is +/- 60v starts up at 0V than ramps to +/-30v and when there is no issues goes up to +/-60v rails.
Caps are rated 100V on the rails 3x220uf on each rail. in the PS. Not that much .
There are unpopulated capacitor positions on the UDC modules and on the interceptor boards. There is room for max 220uf caps because of the 10mm diameter restriction. Might be able to squeeze in 12.5mm caps.

 

[kawal]

The main board has a lot going on too. There is a DSP and CS4398 DAC . Outputs from the DAC drive the amp modules via filters made using NE5532 filter buffers to drive the inputs of the amp.s

 

[kawal]

I started testing the DSP2-150 and ran into an issue. The hybrid analog/digital inputs that came installed on my unit did not work. I had a DSP-2 750 that had an analog only input so i put that into the 150 and it worked.
Not sure why the hybrid ones are not working - also swapped the DSP boards to make sure.
I recorded a video showing THD at 1kHz . The amp produced over 180W /CH both driven. The power supply got very hot.

The ADC DAC chain is probably not contributing too too much distortion as the amp seems to have a lot more distortion than the rest of the system. The lowest THD from the digital chain is seen up to -6dBV (0.5V RMS)input as measured on the buffer amp output post DSP. Max power is at 8dBV (2.5V RMS).

I will test the Signal processing THD alone to make sure the THD of the power amp is truly making the the input stage THD.

 

[kawal]

Testing from analog input card to power output.
Noise dominates the upper frequencies - most likely from power supply.

 

[kawal]

I tried Analyzing the system from a THD perspective. To get full power one needs to provide about 8dBV input. Here I tested the Analog input to Anlog output buffer right before the amplifier section. The amplifier modules where removed from the chassis for the test. At 2dBV input the system starts clipping and distortion is crazy high . This is inline with my previous testing where -6dBV produced the best THD. Basically the fronted needs more gain to drive the power amp module. The gain should be set to +10...+12dB on the DSP to get best results.

I will test higher gain later but for now the results at the default 0dB gain in the DSP.

I took this opportunity to test a different input buffer op amps. Original NE5532 in input 2 were replaced with OPA1602. It seems the result was a waste of time. The OPA1602 from Arrow.com produced worse results than the original NE5532.

Original NE5532

OPA1602 - signs of instability at higher frequencies. Disappointing results.

 

[thcdru2k]

I run a DSP 2-150 MKII in an installed system and this thread matches something I learned during my setup: this amp is very sensitive to gain structure.
It wants a hot input and level managed inside the DSP. When I have time I'll see if I can do some REW measurements with my speakers.

 

[kawal]

I run a DSP 2-150 MKII in an installed system and this thread matches something I learned during my setup: this amp is very sensitive to gain structure.
It wants a hot input and level managed inside the DSP. When I have time I'll see if I can do some REW measurements with my speakers.

I had the same issue with the DSP2-750 and the amp already had the DSP turned up to 12 but that was not enough to avoid clipping in the ADC/DAC while getting full power. In this situation I changed the gain of the system by changing the gain post DAC buffer/filter all that was needed was 4 1% resistors. Before the change the amp would clip at ~300W

I had a chance to listen to the amp last night and I was not expecting much other than a decent sounding amp. I was actually very surprised what I heard from this amp and some underwhelming mission speakers. The base became very well controlled and mids and highs where very clear / transparent. I thought to myself it must be the class D sound so i switched the setup to a TPA3255 based amp and after the the UcD based amp it sound very dull and boomy - almost lacking in the midrange and highs . There was a night a day difference in the perceived sound quality. The DSP2-750 was running on average 15W of power more from the outlet (80...85W) than at idle but was playing the speakers at almost max excursion. very pleasant sounding amplifier.

I wonder if bypassing the DSP completely in these would bring out even better results.
One could design a balanced input card that feeds the DAC balanced output buffer . The 1500 resistors are best replaced with 3.9k if you intend to use the system with the DSP. otherwise you can leave the 1.5k and add extra gain in the infeed section that replaces the DAC block.

 

[Scrivs]

Thanks for sharing your analysis and pictures, Kawal, very interesting!

 

[kawal]

Thanks for sharing your analysis and pictures, Kawal, very interesting!

For now I changed the gain on the buffer stage of the 150 .

Before modification notice the rise in distortion at 50W
After the modification the amp seems to play cleaner up to 100W

This is with the increased gain on the buffer stage. This make it clear that the distortion is from the ADC/DAC chain working to close to 0dB. Right channel seems to perform better than Left. Gains are modest at best

 

[thcdru2k]

I am thinking since Sonance has IP control it may be possible to create a Pi script to control sonance volume more easily. Can be automated / or custom GUI. That way you can keep your pre-amp source at 100% output and maintain proper gain structure.

 

[kawal]

I am thinking since Sonance has IP control it may be possible to create a Pi script to control sonance volume more easily. Can be automated / or custom GUI. That way you can keep your pre-amp source at 100% output and maintain proper gain structure.

Not sure that will help .The issue is lack of gain on the analog side . There is no control in sonance over he analog side its all in the DSP.

 

[thcdru2k]

Not sure that will help .The issue is lack of gain on the analog side . There is no control in sonance over he analog side its all in the DSP.

I got things mixed up. I swapped my unit to the digital input card, so I’m no longer using the analog inputs.

My idea does not increase headroom or fix the underlying analog gain issue. It's more for operability and living with the issue: if you use a preamp/source as the volume control, you’re constantly changing ADC drive level and may away from the ADC’s cleaner operating region at times.

A more stable approach (when using analog in) is to set the source to a known safe fixed output level that does not overdrive the ADC, then use DSP output attenuation inside the Sonance for day-to-day listening level. A Pi-hosted slider/UI over Sonance IP control would be more convenient

 

[thcdru2k]

I inspected the Sonance DSP web UI and found that the control layer is essentially a thin frontend over simple HTTP handler calls. After tracing the handler endpoints and validating them with curl, I was able to implement direct control of output volume. Technically this could be expanded upon to have full control over anything that is available on the web UI. I wrapped this into a small Python-based control card that interfaces with the amp the same way SonARC does, but without requiring the full UI. Sharing here in case it helps anyone integrating these into their system.

 

[thcdru2k]

I think I still have my analog input module somewhere around I can send to you for free if you want to try to modify it.

 

[kawal]

I think I still have my analog input module somewhere around I can send to you for free if you want to try to modify it.

I am actually looking for one.