Fosi Audio V3 op-amp rolling, has anyone tried it? Snake oil? Or are there actual differences?

[WhyLee]

my two cents regarding "opamp sound":
we are dealing here with an extrem noisy power rail.
1. we get the power from the switching power supply
2. we get noise on the power rail from the TPA3255

when i switched from my (already pimped) 90's Rotel Pre and Amp combination to a Fosi Audio BT30D Pro including a Fosi Audio 32V 5.0A power supply I instantly noticed the noise in the high frequencies.
the noise in not alwas present. it is a kind of modulation between noise and the audio signal.
i don't know if the same phenomenon is there with the Fosi Audio V3 and the power supplies that are coming with them.

For me, it is absolutely clear that the switching of the power supply and the TPA3255 creates a lot of noise on the power rail.
The question is, regarding the Opamps, how good is their Power supply rejection ratio in the higher frequencies. On the datasheet they usually have a very nice number, but that number cannot be independent from the frequency of the power supply distortion.

A lot comes down to the circuit design and if good L-C filters has been added to make two different power rails:
* one filter for the power rail for the high power supply for the TPA3255
* one filter for the power rail for the low power signal path that supplies the Opamps etc. maybe even two filters to make two power supply lines for the left and right channel

all this has a strong effect on the performance of the opamps as well. and different opamps have different power supply rejection ratios.

and we know that these filters come at a cost. i mean, it is very nice to being able to buy a Fosi Audio V3 including a switching power supply for 100$ but hey, how is this price even possible without cutting at least some corners !!!!!!

Comparing my Fosi Audio BT30D Pro with my Rotel Setup, i am not disapointed, because of the huge difference in price. But the difference in audio quality is a bit annoying.
I just use it for now because my Rotel Pre has a noisy balance-poti that needs to be repaired (or bridged completely). I might try to make a two stage LC filter to be placed in between the Fosi power supply and the BT30D Pro just to check out if it improves the quality. And if I buy a Fosi V3 - just because it is cheap - i could try this filter as well.
I already looked at mouser to find the parts for a two stage filter, first with a corner frequency of about 1000Hz and second LC-stage with a corner frequency with about 150Hz.
That should give a nice flat and stable power supply and makes sure, that there is at least no noise coming into the amp.
But that cannot not solve any potential problems inside the Fosi Amp regarding the separation of high burrent and low current power supply pathways that should be separated with filters as well.
The parts costs for this filter (two inductors and two capacitors) are about 15€ excluding VAT at mouser. part costs for filters easy add up. maybe it is a bit over engineering, who knows. But stable noise-free power rails are the base for good audio circuits.

Any thoughts regarding this input?

 

[fpitas]

There are not many right things in this post..

You get points for tact and restraint.

 

[HarmonicTHD]

my two cents regarding "opamp sound":
we are dealing here with an extrem noisy power rail.
1. we get the power from the switching power supply
2. we get noise on the power rail from the TPA3255

when i switched from my (already pimped) 90's Rotel Pre and Amp combination to a Fosi Audio BT30D Pro including a Fosi Audio 32V 5.0A power supply I instantly noticed the noise in the high frequencies.
the noise in not alwas present. it is a kind of modulation between noise and the audio signal.
i don't know if the same phenomenon is there with the Fosi Audio V3 and the power supplies that are coming with them.

For me, it is absolutely clear that the switching of the power supply and the TPA3255 creates a lot of noise on the power rail.
The question is, regarding the Opamps, how good is their Power supply rejection ratio in the higher frequencies. On the datasheet they usually have a very nice number, but that number cannot be independent from the frequency of the power supply distortion.

A lot comes down to the circuit design and if good L-C filters has been added to make two different power rails:
* one filter for the power rail for the high power supply for the TPA3255
* one filter for the power rail for the low power signal path that supplies the Opamps etc. maybe even two filters to make two power supply lines for the left and right channel

all this has a strong effect on the performance of the opamps as well. and different opamps have different power supply rejection ratios.

and we know that these filters come at a cost. i mean, it is very nice to being able to buy a Fosi Audio V3 including a switching power supply for 100$ but hey, how is this price even possible without cutting at least some corners !!!!!!

Comparing my Fosi Audio BT30D Pro with my Rotel Setup, i am not disapointed, because of the huge difference in price. But the difference in audio quality is a bit annoying.
I just use it for now because my Rotel Pre has a noisy balance-poti that needs to be repaired (or bridged completely). I might try to make a two stage LC filter to be placed in between the Fosi power supply and the BT30D Pro just to check out if it improves the quality. And if I buy a Fosi V3 - just because it is cheap - i could try this filter as well.
I already looked at mouser to find the parts for a two stage filter, first with a corner frequency of about 1000Hz and second LC-stage with a corner frequency with about 150Hz.
That should give a nice flat and stable power supply and makes sure, that there is at least no noise coming into the amp.
But that cannot not solve any potential problems inside the Fosi Amp regarding the separation of high burrent and low current power supply pathways that should be separated with filters as well.
The parts costs for this filter (two inductors and two capacitors) are about 15€ excluding VAT at mouser. part costs for filters easy add up. maybe it is a bit over engineering, who knows. But stable noise-free power rails are the base for good audio circuits.

Any thoughts regarding this input?

Is there a problem of PS noise reaching the output in this amp? If yes, can you quantify it (frequency and magnitude)?

Before you go off proposing a solution to an undefined problem.

 

[fpitas]

If only op-amps had complete graphs that describe PSRR. But, that's just a dream

 

[bgravato]

Is there a problem of PS noise reaching the output in this amp? If yes, can you quantify it (frequency and magnitude)?

Before you go off proposing a solution to an undefined problem.

I'm not who made that comment about PS noise (I'm the OP), but here's my opinion on the Fosi V3.

I ordered it with the 48V PSU option.

My only source of reference to compare it to is a fully working vintage Philips FA950 (which I think was supposed to be quite a decent one back then).

Subjective listening I can't hear any noticeable difference between the two.

In terms of noise floor, the Fosi V3 has actually a lower noise floor than the vintage Philips (both my ears and measurements in REW confirm this). In either case I have to crank the volume up beyond my normal listening levels to be able to hear any noise, so not a problem I need to worry about.

The noise floor I talk about is just white/pink noise, I don't hear any high frequency hiss or anything similar. My aged ears can't hear anything above 13-14 kHz, so if there's any noise above that I wouldn't hear it anyway.

 

[bgravato]

and we know that these filters come at a cost. i mean, it is very nice to being able to buy a Fosi Audio V3 including a switching power supply for 100$ but hey, how is this price even possible without cutting at least some corners !!!!!!

I like to believe the corners they cut were on the "extra" features. V3 has no tone controls, no remote, no bluetooth, etc... It has just 1 stereo RCA input, L/R speakers out and a volume knob.

Before the V3, I tried an Aiyima T9 Pro, which has a lot in one small package, and it wasn't good at any of it... (I believe the cheap tubes where probably a big part of the problem). I sent it back to Amazon for full refund.

I later ordered the Fosi V3 (directly from Fosi, because I wanted it with the 48V PSU) and so far I haven't been disappointed... I had to buy a separate DAC and I still don't have all the features the T9 Pro had (such as bluetooth, which I don't care much for anyway). DAC+amp I spent $250. If I wanted to add more inputs, maybe a phone preamp or a headphone amp/output, I'd probably end up spending $500 or more. Which would be the price for an (entry level?) integrated amp, with all those features, from a reputable brand (Yamaha, Marantz, Rotel, etc).

Bottom line, all in all, I don't thing the Fosi V3 is "that cheap", considering what it offers (which is very little feature wise) and I think $100, just for a basic power amp + volume knob, should be able to pay for the same components (regarding that task alone) as the ones in a $400-500 integrated amp with a lot more features (and components).

PS: one other thing the V3 probably lacks in comparison to a more "traditional" amp is max power... Even with the 48V PSU it doesn't get anywhere near those advertised 300W per channel... I have a vintage Philips amp (FA950) rated at 100W per channel, that can a output quite more power than the V3, but for my normal listening levels I can't hear a difference between the two. I can't hear any noise either unless I really crank up the volume and get my ear close to the speaker... In that matter, the Fosi actually has a lower noise floor than the vintage Philips. I don't hear any high frequency noise/hiss either (but my aged ears can't hear anything above 13-14kHz). After my disappointing experience with the Aiyima T9 Pro, I wasn't expecting much from the Fosi V3, but it actually surprised me positively.

Also if it sounds as good as I was hoping for with the stock opamps, I see no reason to experiment with different opamps... It already does what I wanted it to...

 

[WhyLee]

Is there a problem of PS noise reaching the output in this amp? If yes, can you quantify it (frequency and magnitude)?

Before you go off proposing a solution to an undefined problem.

no, i didn't do any measurements. not sure if I have the time soon to do real measurements.
the noise seems to be a modulation product of power supply noise and switching noise from the TPA3255.
And this noise is audible. It has nothing to do with a noise floor. There is no noise floor audible for me.
I directly plugged the speakers from the output of my Rotel Amp to the Fosi BT30 Pro. So, only a few seconds in between. And the difference in the high frequencies is very easy audible. But again, this is with the BT30 Pro. Maybe the V3 has better power supply filtering.

And again for the understanding about which noise i am talking here. There is no audible noise floor. It is modulation noise. power supply noise modulated through the switching of the TPA3255. People who understand the two mirrord sidebands of AM modulation will also understand the modulation of two switching frequncies that are both above 20kHz but the modulation product is NOT only above 20kHz but can also reach down to the audible frequency range.

 

[bgravato]

no, i didn't do any measurements. not sure if I have the time soon to do real measurements.
the noise seems to be a modulation product of power supply noise and switching noise from the TPA3255.
And this noise is audible. It has nothing to do with a noise floor. There is no noise floor audible for me.
I directly plugged the speakers from the output of my Rotel Amp to the Fosi BT30 Pro. So, only a few seconds in between. And the difference in the high frequencies is very easy audible. But again, this is with the BT30 Pro. Maybe the V3 has better power supply filtering.

And again for the understanding about which noise i am talking here. There is no audible noise floor. It is modulation noise. power supply noise modulated through the switching of the TPA3255. People who understand the two mirrord sidebands of AM modulation will also understand the modulation of two switching frequncies that are both above 20kHz but the modulation product is NOT only above 20kHz but can also reach down to the audible frequency range.

I see.

I'm guessing for the kind of noise you're mentioning I'd think it should show an increase in SPL measurements in REW for the higher frequencies and perhaps a higher level of distortion in the distortion graph. Is this a correct assumption?

I'm looking at my measurements in REW, comparing between Fosi V3 and that vintage Philips amp. This was with a non-calibrated mic, but for the matter of checking differences between two devices that shouldn't be too relevant. In the SPL measurements, there are some slight differences: most notably above ~7.5kHz (with the biggest difference probably around 11kHz where there's a difference of about 0.5-0.6dB. I will not swear for the accuracy of those measurements though. I'll try to repeat them one of these days...

In the distortion graph (THD) I can't really see any difference between the two, especially in the high frequencies.

I'm not very experienced yet with REW measurements so not exactly sure what conclusions to take from this.

 

[solderdude]

no, i didn't do any measurements. not sure if I have the time soon to do real measurements.
the noise seems to be a modulation product of power supply noise and switching noise from the TPA3255.
And this noise is audible. It has nothing to do with a noise floor. There is no noise floor audible for me.
I directly plugged the speakers from the output of my Rotel Amp to the Fosi BT30 Pro. So, only a few seconds in between. And the difference in the high frequencies is very easy audible. But again, this is with the BT30 Pro. Maybe the V3 has better power supply filtering.

And again for the understanding about which noise i am talking here. There is no audible noise floor. It is modulation noise. power supply noise modulated through the switching of the TPA3255. People who understand the two mirrord sidebands of AM modulation will also understand the modulation of two switching frequncies that are both above 20kHz but the modulation product is NOT only above 20kHz but can also reach down to the audible frequency range.

Is this noise also there when only the speakers are connected and the volume contol is at lowest setting and nothing is connected to the input ?

 

[WhyLee]

Is this noise also there when only the speakers are connected and the volume contol is at lowest setting and nothing is connected to the input ?

no audible noise when the volume is turned down OR the music signal is silent.
that means NO audible noise floor. it is NOT the usual noise that everybody knows. it is just a kind of dirty noisy music in the high frequencies. nothing extreme but the difference to a good class AB Rotel amp is audible and noticeable once you switch over the speaker cables.

 

[solderdude]

Looks like a groudloop or leakage issue.

 

[MaxwellsEq]

no audible noise when the volume is turned down OR the music signal is silent.
that means NO audible noise floor. it is NOT the usual noise that everybody knows. it is just a kind of dirty noisy music in the high frequencies. nothing extreme but the difference to a good class AB Rotel amp is audible and noticeable once you switch over the speaker cables.

Should be visible in IMD, impulse or multitone. Something is faulty.

 

[antcollinet]

no, i didn't do any measurements. not sure if I have the time soon to do real measurements.
the noise seems to be a modulation product of power supply noise and switching noise from the TPA3255.
And this noise is audible. It has nothing to do with a noise floor. There is no noise floor audible for me.
I directly plugged the speakers from the output of my Rotel Amp to the Fosi BT30 Pro. So, only a few seconds in between. And the difference in the high frequencies is very easy audible. But again, this is with the BT30 Pro. Maybe the V3 has better power supply filtering.

And again for the understanding about which noise i am talking here. There is no audible noise floor. It is modulation noise. power supply noise modulated through the switching of the TPA3255. People who understand the two mirrord sidebands of AM modulation will also understand the modulation of two switching frequncies that are both above 20kHz but the modulation product is NOT only above 20kHz but can also reach down to the audible frequency range.

That is not how noise works.

You are talking about intermodulation distortion. You can see results of harmonic disotortion and IMD (also from ultrasonics) in the multitone frequncy test. Below audibility for this amp.

And - I assume you've reached. your conclusions from listening. I'll assume no controls were used for this listnening?

Also

the noise seems to be a modulation product of power supply noise and switching noise from the TPA3255

How you think you can determine the electrical mechanism for your sighted listening impressions without any measurement data to even validate there is anything to hear - is a total mystery.

 

[WhyLee]

That is not how noise works.

You are talking about intermodulation distortion. You can see results of harmonic disotortion and IMD (also from ultrasonics) in the multitone frequncy test. Below audibility for this amp.

And - I assume you've reached. your conclusions from listening. I'll assume no controls were used for this listnening?

Also

How you think you can determine the electrical mechanism for your sighted listening impressions without any measurement data to even validate there is anything to hear - is a total mystery.

Do you have a different explanation for dirty/noisy high frequencies that has nothing to do with noise floor?
As I stated, I plugged over the speaker cables directly from my Rotel Amp that I am using for many years now. That was my control.
And the difference was very noticable. It is no typical distortion that I have experienced in the last 40 years before. It has a unique characteristic that I can only explain with an intermodulationproduct of two frequencies - the switching frequency from the power supply and the switching frequency from the TPA3255.

 

[solderdude]

It has a unique characteristic that I can only explain with an intermodulationproduct of two frequencies - the switching frequency from the power supply and the switching frequency from the TPA3255.

If that were the case one would hear 'something' when idle too.
If it only happens with audio a single tone should result in a tone with not related harmonics which should then shift in frequency when the output load was changed.
Do this with a dummy load not a speaker.

That power supply noise would then also need a path into the audio. It is unlikely to come from the voltage rail unless the power supply is seriously broken.
It could make it into the audio path through RCA cables and a groundloop.

And... it has nothing to do with op-amp rolling.

 

[MaxwellsEq]

Do you have a different explanation for dirty/noisy high frequencies that has nothing to do with noise floor?
As I stated, I plugged over the speaker cables directly from my Rotel Amp that I am using for many years now. That was my control.
And the difference was very noticable. It is no typical distortion that I have experienced in the last 40 years before. It has a unique characteristic that I can only explain with an intermodulationproduct of two frequencies - the switching frequency from the power supply and the switching frequency from the TPA3255.

As I explained before. Something like this is absolutely measurable, most easily with IMD, but you might also see it on THD or impulse. This is not an op-amp issue (unless there is a broken device)

 

[antcollinet]

Do you have a different explanation for dirty/noisy high frequencies that has nothing to do with noise floor?
As I stated, I plugged over the speaker cables directly from my Rotel Amp that I am using for many years now. That was my control.
And the difference was very noticable. It is no typical distortion that I have experienced in the last 40 years before. It has a unique characteristic that I can only explain with an intermodulationproduct of two frequencies - the switching frequency from the power supply and the switching frequency from the TPA3255.

Well you haven't tested blind - so we don't know if what you percieved even existed in the sound waves - or was a result of cognitive biases in your (subconcious) brain. And given the measurements of the amp, the latter is by far the more likely - if the amp is not being over driven into clipping.

So yes - until then the different explanation is expectation bias - or similar.

Another possible different explanation - you have simply got used to the old amp freqency response - possibly with some high frequency roll off. If this is the case, the more extended flat response of the Fosi may well sound "wrong" to you after 40 years of a different response.

 

[LeoGB]

Do you have a different explanation for dirty/noisy high frequencies that has nothing to do with noise floor?
As I stated, I plugged over the speaker cables directly from my Rotel Amp that I am using for many years now. That was my control.
And the difference was very noticable. It is no typical distortion that I have experienced in the last 40 years before. It has a unique characteristic that I can only explain with an intermodulationproduct of two frequencies - the switching frequency from the power supply and the switching frequency from the TPA3255.

I also experienced distortion in i think the mid-high frequencies which are only apparent when there is something "going on" in these frequencies and wonder what it could be / how I can solve this / need another amp.

My system:
Denon CD player > smsl su-1 DAC > Fosi Audio V3 (36v/5a psupply)) > Quad 21L floorst.

 

[SIY]

Have you tried borrowing a different amp to see if your amp is broken?

 

[LeoGB]

Have you tried borrowing a different amp to see if your amp is broken?

Good idea. Will try that. Thanks !
(Seems to get better in time / depend quite a bit on recording quality).
No rolling of op-amps advised ?