SylphAudio's High-Performance Amplifier Modules (2024 and Beyond)

[daniboun]

We will revive the production of D100 configured as non-PFFB, since there are some people who still prefers non-PFFB configuration for TPA325x.
Sample feedback below.

View attachment 355844

My answer is put into perspective. There are two worlds, that of auditory perception and that of scientific interpretation. But there is something for all tastes and so much the better))

Now to stay concrete :
PFFB offers many benefits including lower output noise, improved THD+N performance, improved IMD performance, lower output impedance, frequency response less affected by load impedance, and suppression of nonlinearities of the LC filter.

All measurements of PFFB vs non PFFB are here

https://www.ti.com/lit/pdf/slaa788

 

[SMen]

We will revive the production of D100 configured as non-PFFB, since there are some people who still prefers non-PFFB configuration for TPA325x.
Sample feedback below.

View attachment 355844

Would letting users' have a switch to go between PFFB and non-PFFB be the answer (unless I have miss-read the spec. I understand these are fixed, or need to be soldered to change in some way)?

I have the impression you pushed the D100 as far as you could performance wise with a DIP8 option and PFFB. Am I right? This might be a better reason to do another run of the D100's I think. Having the DIP8 functionality.

Just my thought. I agree with @daniboun 's observations about PFFB

 

[daniboun]

Would letting users' have a switch to go between PFFB and non-PFFB be the answer (unless I have miss-read the spec. I understand these are fixed, or need to be soldered to change in some way)?

I have the impression you pushed the D100 as far as you could performance wise with a DIP8 option and PFFB. Am I right? This might be a better reason to do another run of the D100's I think. Having the DIP8 functionality.

Just my thought. I agree with @daniboun 's observations about PFFB

I think that the FB100 is already a more accomplished evolution of the D100. Finally, the D100 measurements showed that the addition of the 1656 op amps brought a very slight improvement. For the rest, I am impatiently awaiting the preliminary measurements of the FB100. In the meantime here is the synopsis of the D100 measurements in the attached file

 

[SylphAudio]

Let us know the pre-measurements please.

Will be buying E1DA Cosmos Scaler as well. FB100 THD is better than -116dB, but I can't get the THD+N past -106.5dB.

 

[SylphAudio]

I understand these are fixed, or need to be soldered to change in some way)?

D100 PFFB can be disabled through solder jumpers on the bottom PCB.

 

[daniboun]

Will be buying E1DA Cosmos Scaler as well. FB100 THD is better than -116dB, but I can't get the THD+N past -106.5dB.

Hi,

Any news ? when do you expect to share Cosmos measurements ?

 

[SylphAudio]

Hi,

Any news ? when do you expect to share Cosmos measurements ?

Cosmos measurement is not yet ready. Getting mains harmonics when LPF is added on the class-D output with the cosmos APU.
Using LPF on this site: https://www.openaudiolab.com/measuring-amps/
Still figuring out the correct wiring to eliminate the mains harmonics.
I ordered Cosmos scaler, will arrive after holy week.

 

[daniboun]

Cosmos measurement is not yet ready. Getting mains harmonics when LPF is added on the class-D output with the cosmos APU.
Using LPF on this site: https://www.openaudiolab.com/measuring-amps/
Still figuring out the correct wiring to eliminate the mains harmonics.
I ordered Cosmos scaler, will arrive after holy week.

I read this article a few days ago and I was aware of the problem...
This is what motivated me to entrust my “Micro-Audio” project to a laboratory equipped with an APx555 instead
Another thing, I think it is important with the Cosmos ADC to carry out the tests with both channels drived at the same time

 

[SylphAudio]

Here's with LPF straight to Cosmos ADCiso.

Setup is SMSL DL200 -> FB100 -> LPF without output coupling caps -> Cosmos ADCiso

Can't get lower noise measurements, though the module already has lower noise than my sample PA5 here. So, I believe I can get better noise measurements. It's just some setup problems.

Power supply is MW LRS-350-36 adjusted to 38V DC

left ch 5W 4 ohms

right ch 5W 4 ohms

 

[SylphAudio]

Same setup 5W 4ohm but with FB360. Sample only, this is on open air, no casing
Will make repeat measurements later

 

[Guddu]

@SylphAudio
This looks amazing, congrats.

One question though - are you only after slightly better noise/distortion measurements than an amp like PA5? SNR/THD/SINAD are, sort of, resolved problems already.
How about other aspects like behavior & stability with a complex load, multitone performance w.r.t. power, IMD, etc?

PE. I know that we will find out most of those on Amir's review, whenever that would happen, but do you have any stats on those?

 

[daniboun]

Here's with LPF straight to Cosmos ADCiso.

Setup is SMSL DL200 -> FB100 -> LPF without output coupling caps -> Cosmos ADCiso

Can't get lower noise measurements, though the module already has lower noise than my sample PA5 here. So, I believe I can get better noise measurements. It's just some setup problems.

Power supply is MW LRS-350-36 adjusted to 38V DC

left ch 5W 4 ohms

right ch 5W 4 ohms

Very nice !
Input sensitivity of Sylph-FB100 is 5 Vrms, I assume you opted for the DL200 to match, right ? )
I'm almost sure that the APx555 measurement should honor this FB100 by a few more db....
Openaudiolab highlighted a difference between the SINAD measured via the Cosmos ADC and the APx555.
But it gives us a preview of the beast's potential.

Could please you push the tests with your Cosmos > Multitone, Frequency response, THD vs Level @4R / 8R, Driven channel VS Undriven channel....


PS : What PSU did you use for the FB360 for this test ?

 

[SylphAudio]

@SylphAudio
This looks amazing, congrats.

One question though - are you only after slightly better noise/distortion measurements than an amp like PA5? SNR/THD/SINAD are, sort of, resolved problems already.
How about other aspects like behavior & stability with a complex load, multitone performance w.r.t. power, IMD, etc?

PE. I know that we will find out most of those on Amir's review, whenever that would happen, but do you have any stats on those?

No stats yet since were using the Cosmos ADC as unbuffered, i.e. directly connected to a passive low pass filter after the DUT's speaker out. There's no protection to transient and large voltage swings. So, we can only measure the amp at 4.5Vrms output (with 4R load). And the passive LPF has a high output impedance driving the low impedance input of Cosmos ADCiso, hence measurements will not be reliable. Once we get the Cosmos Scaler and use it as a buffer, then we can provide the other measurements, and can correlate the measurements with Archimago's.

 

[daniboun]

No stats yet since were using the Cosmos ADC as unbuffered, i.e. directly connected to a passive low pass filter after the DUT's speaker out. There's no protection to transient and large voltage swings. So, we can only measure the amp at 4.5Vrms output (with 4R load). And the passive LPF has a high output impedance driving the low impedance input of Cosmos ADCiso, hence measurements will not be reliable. Once we get the Cosmos Scaler and use it as a buffer, then we can provide the other measurements, and can correlate the measurements with Archimago's.

Did you use the same PSU to measure both FB360 / FB100 modules ?

 

[SylphAudio]

Did you use the same PSU to measure both FB360 / FB100 modules ?

Yes, same setup, LRS-350-36 bumped to 38V.
Cosmos ADCiso and DL200 used the same laptop without USB isolator on the DL200, So, I will try again some other time with separate laptops, might fix the mains harmonics when using the Cosmos APU.

 

[daniboun]

Yes, same setup, LRS-350-36 bumped to 38V.
Cosmos ADCiso and DL200 uses the same laptop without USB isolator on the DL200, So, I will try again some other time with separate laptops, might fix the mains harmonics when using the Cosmos APU.

Thanks. Should be nice to see the FB360 with a 48 or 51v PSU ))) TPA3255 has a better THD+N figure with a lower PSU.
Many thanks for all those sharings

 

[MCH]

Yes, same setup, LRS-350-36 bumped to 38V.
Cosmos ADCiso and DL200 uses the same laptop without USB isolator on the DL200, So, I will try again some other time with separate laptops, might fix the mains harmonics when using the Cosmos APU.

104dB sinad with a humble LRS? Next time I read discussions about fancy PSU I will roll eyes even more than I do today
Thanks for sharing.

 

[daniboun]

104dB sinad with a humble LRS? Next time I read discussions about fancy PSU I will roll eyes even more than I do today
Thanks for sharing.

The fact is that these modules have such a good PSRR that using a "modest / humble" PSU will do the trick. Even if I am convinced that quality is a guarantee of durability)

In any case, I intend to remove the mystery by soon measuring my "Micro Bamboo" project (3E Audio + Micro Audio PSU) with an APx555.

 

[daniboun]

@SylphAudio > when you have time could you please make those same measurements with the stock PA5 brick PSU since I know you have it )))))
I am pretty sure to get the same results as the LRS )

 

[Roland68]

104dB sinad with a humble LRS? Next time I read discussions about fancy PSU I will roll eyes even more than I do today
Thanks for sharing.

Maybe you should first take a look at the data sheets for the Mean Well LRS power supplies.
The LRS series also meets higher industry standards, including all 4 protection standards, 5G vibration tests, operating temperature up to 70 degrees, a 100% full load burn in test, 88.5-89% efficiency at 36-48v and a very low ripple & noise value. In addition, these power supplies have all relevant and verifiable certificates and the CE declaration of conformity required in the EU.
This power supply series in particular is proven and is likely to be used in 100,000 or even millions of industrial devices.
I would prefer an LRS power supply to any supplied power supply.

If you call such a power supply a "humble", then what is a non-humble power supply?