FV SPIN - NOISE

[pruttelherrie]

when your signal reaches that stage you already mixed the "noisy" signal from the FV1 with the clean in the mix control. You want to do all the reconstruction filtering right after the DAC and then keep processing the signal, mixing, etc.

Ah right! Thanks for the insight.

[AndyF]

ok I read everything, now as I wrote at the beginning of the thread, the noise can be eliminated, but there is no concrete clarity on how to do it, they are all said assumptions, nobody published anything that says this has no noise, I guarantee it! And the one who has it, I don't know if he will share it clearly because it is a job and it is a challenge to do it. I tried many pedals that use the fv spin, and I can say in particular that there are important brands that release their product with a lot of noise in my opinion and others that have almost zero noise, none have zero noise! Therefore, it can always be improved, but the noise is there and it is not something easy to eliminate.
Of course, using a poor quality power supply will not help, they are things that are discarded. A good pcb design? And yes, I think it makes a difference, as well as the algorithms will ...
Now if all this is known and it is understood that it is not very easy to do something that does not produce and amplify the noises, because xnoize does not give parameters of things that should or should not be done, and things that should or should be done, You have to look a lot to be able to do something almost decent, without obtaining totally satisfactory results. I believe that many are satisfied with the result and accept it and relax, but the noise is there and it tortures me to have it there ...
For example, in a reverb or a delay, you hardly use the mix to the maximum, so the noise you will not notice, so you use it without problems until the day you decide to try new things and find a hurricane of background noise and it is impossible to record in a pc with tremendous noise. In a chorus, in a pahser, that all the mixing is needed it is a big problem to do something good without noise.

It must be me, I'm very bad at all this and all the rest must be great experts, but as you know, we all start one day, and my intention is not to criticize the product because for me a great solution is great, but I do not understand how everything This noise issue passes by and nobody says or does anything to improve it, as there is no basic list of use and solutions to possible problems, it would be less frustrating.
Of course when you sell a lot you don't need to do anything else, leave it like that, it totally works, right?

Thanks to all who have responded, I hope someone contributes something concrete and tested, I would pay what I do not have to eliminate the damn noise, my message box is open to receive all possible help

[ElectricDruid]

You keep telling us how bad it is, but you haven't put any numbers on it. What is this "unacceptable" noise level you're seeing coming out of this chip? How many mV P-P on average? Or RMS, or whatever? I don't really care *what* measure you use as long as it's something we can relate to.

If we have a number on the noise that you've got, we might be able to get a figure on the S/N that you might hope for, and that might give us some idea whether you're being realistic or aiming for the moon in a flying bucket!

[Sweetalk]

You keep telling us how bad it is, but you haven't put any numbers on it. What is this "unacceptable" noise level you're seeing coming out of this chip? How many mV P-P on average? Or RMS, or whatever? I don't really care *what* measure you use as long as it's something we can relate to.

If we have a number on the noise that you've got, we might be able to get a figure on the S/N that you might hope for, and that might give us some idea whether you're being realistic or aiming for the moon in a flying bucket!

That's what I was thinking!. There's a lot of answers with things you can try to lower the noise. Dead silent is impossible! but you haven't show us the layout or measurements or anything to get some info about. If you're using the circuit that you showed before try to improve the filtering and the signal level. If you don't want to share the layout you can send it via PM and I have a look at it.

ok I read everything, now as I wrote at the beginning of the thread, the noise can be eliminated, but there is no concrete clarity on how to do it, they are all said assumptions, nobody published anything that says this has no noise, I guarantee it! And the one who has it, I don't know if he will share it clearly because it is a job and it is a challenge to do it.

No one can tell you "do this and the noise will go away" if we don't know where the noise is really came from and how big it is. As I tell you, there's a lot of options stated in the thread you can (and have!) to try to lower the noise.

The SN ratio is really good in the FV1 if you follow those steps. I have a bunch of old PCB's (professionaly manufactured) that have mistakes on the layout and get ground loops so there's a lot of noise, and other designs that are really silent with the propper techniques applied, same circuit in both cases.

[octfrank]

This is at least the third thread he has started on noise, he created two on the Spin forum last year. He has been provided solutions which he constantly rejects and never provides any measurements, PCB layouts, etc. I consider him a troll on the Spin forum now and while I have not blocked him there I really don't respond to him either.

I've only commented on the thread here because there are some really good suggestions that may help others that will listen and provide the necessary info to get help. And there is a post on the Spin forum http://www.spinsemi.com/forum/viewtopic.php?f=2&t=868 that does a nice job summarizing some great suggestions.

[AndyF]

This is at least the third thread he has started on noise, he created two on the Spin forum last year. He has been provided solutions which he constantly rejects and never provides any measurements, PCB layouts, etc. I consider him a troll on the Spin forum now and while I have not blocked him there I really don't respond to him either.

I've only commented on the thread here because there are some really good suggestions that may help others that will listen and provide the necessary info to get help. And there is a post on the Spin forum http://www.spinsemi.com/forum/viewtopic.php?f=2&t=868 that does a nice job summarizing some great suggestions.

Look frank wrote here, because in your forum, you close the topics and you can't talk anymore, what you do is not very democratic.

I have always said that the fv spin is great, however it is very easy to have noise with it, I am not talking about my designs, I am talking about "professional" and recognized brands that have their product for sale, and when you turn it on you have a lot of noise there, then it is not a problem of my designs, the fv spin has noise and that's it.
It does not mean that it is bad, I have never said that, I said that the possible solutions you give, most do not change anything. Ldo regulator does nothing, ground on one side, does nothing, and so several other things that have been said, as they said above, the design of the pcb can surely make a big difference, but that is why it said to show with images, giving tips how to do it, it would be great for the product and that's what I mean, you say I refuse the indications? Why would I refuse if they are helping me? I tried everything and that's why I ask again, I do not demand anything, I only ask and if someone helps me great, that is only the end of this.
Anyone who starts working with this, has more breath than katrina, and when you start looking for possible solutions, most of them are not effective, then there are many technicalities that require much more knowledge, so if you know what I'm talking about, I was wondering because initially there are no schematics, or help to avoid all those hours of work with noise filtering.
But for you to understand once again, first I tell you that troll will be your sister ...  I'm not! I have only told my position and my problems with all this, I am sorry if it bothers you, it is not the purpose of generating a problem for you, I only want to achieve something good and decent. Now what is a decent sound? There is nothing written about tastes, so I can be me, too demanding with something that I don't know if he ever gave me what I'm looking for.

Look, finally I tell you that it is full of pedals with the pt2399 and I would never use it, so that responds to many things.

If the administrator prefers it, delete the post and that's it, it is not the idea to generate any conflict or damage the image of the product, because once again I say, that for me it is EXCELLENT, but complicated to achieve something of a good level, at the Except that happens to me and more than 10 people that I have spoken with and do things with this, the difference is that they relax and do not give it importance, I can not. sorry!

[niektb]

Well, that's not really constructive... Calling each other names...
Still I tend to agree with Octfrank, you're mainly reiterating your own words without giving any kind of factual info that we can go on. What are you listening to? How bad is it really? How did you decide that it's the pedal with the FV-1 that's too noisy. Can you give audio examples? (For the noisy case, but also for the improved case?) What schematics are we currently talking about? (As you said you tried different solutions)

most of us don't recognize your problem but you're not even trying to give a detailed explanation. Please do so we can get to the bottom of this  or, if you're not willing to, otherwise I would suggest to just completely stop this topic because this is not going anywhere in the current psce

[Blackaddr]

AndyF, there are tons of boutique pedals based around the FV-1 that are not unacceptably noisy for pedals costing a few hundred dollars. There are also a lot of amateurish pedals that do have noise and interference issues and it often has nothing to do with the schematic and everything to do with the PCB layout.

The dB noise spec for the FV-1 is in the 90s. That is very good, but it's very difficult to achieve from end to end across a circuit board without a very careful PCB design. Total SNR in the high 60s end-to-end most people would still find acceptable for a guitar pedal. SNR in the 30 and 40s is where most people would complain about a pedal being unacceptably noisy.

If you are getting unusually poor SNRs the two most likely situations are:

1) The PCB design does not employ sufficient PCB design techniques such as ground and power plane shielding, isolation of analog from digital sections, current steering cutouts, and reduction of cross talk (no parallel aggressor/victim tracks).

2) One or more of the ESD sensitive components on the board have been damaged by static discharge.

My advice:
1) measure your SNR, if you're getting 80 dB and it's still unacceptable to you, you have golden ears. If it's in the 30s or 40s, your PCB layout has issues.
2) Build a second board following strict ESD procedures to determine if you've damaged your parts with static at some point.
2) go find some online resources or courses on 'mixed-signal design' and do an improved PCB design.

[octfrank]

Well, that's not really constructive... Calling each other names...

And you are correct, I just didn't know of another way to describe my opinion as this started a year ago on the Spin forum and in that time he is yet to provide a single measurement or PCB layout. At some point I just determined he is intentionally wasting peoples' time,  I would like to be proved wrong and have him post actual data that may assist others but I don't believe that will ever happen.
 

[AndyF]

AndyF, there are tons of boutique pedals based around the FV-1 that are not unacceptably noisy for pedals costing a few hundred dollars. There are also a lot of amateurish pedals that do have noise and interference issues and it often has nothing to do with the schematic and everything to do with the PCB layout.

The dB noise spec for the FV-1 is in the 90s. That is very good, but it's very difficult to achieve from end to end across a circuit board without a very careful PCB design. Total SNR in the high 60s end-to-end most people would still find acceptable for a guitar pedal. SNR in the 30 and 40s is where most people would complain about a pedal being unacceptably noisy.

If you are getting unusually poor SNRs the two most likely situations are:

1) The PCB design does not employ sufficient PCB design techniques such as ground and power plane shielding, isolation of analog from digital sections, current steering cutouts, and reduction of cross talk (no parallel aggressor/victim tracks).

2) One or more of the ESD sensitive components on the board have been damaged by static discharge.

My advice:
1) measure your SNR, if you're getting 80 dB and it's still unacceptable to you, you have golden ears. If it's in the 30s or 40s, your PCB layout has issues.
2) Build a second board following strict ESD procedures to determine if you've damaged your parts with static at some point.
2) go find some online resources or courses on 'mixed-signal design' and do an improved PCB design.

thanks for your recommendations and suggestions.

[AndyF]

Well, that's not really constructive... Calling each other names...

And you are correct, I just didn't know of another way to describe my opinion as this started a year ago on the Spin forum and in that time he is yet to provide a single measurement or PCB layout. At some point I just determined he is intentionally wasting peoples' time,  I would like to be proved wrong and have him post actual data that may assist others but I don't believe that will ever happen.

Don't worry, frank, do you think it's something to annoy you or that my intention is to speak ill of a product, which I always say is great, why do you only read what you think?
I do not understand how you have me so present ...
As I said before, my doubts were amplified when I tried pedals of recognized brands and they had the same noise and I understood that it could be something more than my pcbs.
Now I am clear that it can be removed or filtered, if someone did, then it can. I was just looking for helpful tips ...
I do not usually write in the forums, I do not have much to contribute unfortunately, I just made a query, thinking that someone else would go through the same thing and give me some good indication, there is nothing more than this.
Sure the problem must be mine, do not worry I will not write more and I will not say anything else, it does not make sense. It works great for all of them and only I have noise, it is clear that I have the problem.

To clear all your doubts, I tell you that my intention is not to bother, my intention is not to speak ill of a product that I see great, my intention is to try to solve this with my little knowledge and with some tip that works, nothing more for say.

I apologize if I disturb you or someone else, anything I said.

[mark2]

The dB noise spec for the FV-1 is in the 90s. That is very good, but it's very difficult to achieve from end to end across a circuit board without a very careful PCB design. Total SNR in the high 60s end-to-end most people would still find acceptable for a guitar pedal. SNR in the 30 and 40s is where most people would complain about a pedal being unacceptably noisy.

...

1) measure your SNR, if you're getting 80 dB and it's still unacceptable to you, you have golden ears. If it's in the 30s or 40s, your PCB layout has issues.

What's an appropriate way to measure SNR?

e.g. I can estimate the standalone noise to be (from memory, maybe not accurate) 140mV peak-to-peak by disconnecting the input signal and looking at the scope. But I'm not sure how to isolate the noise from the signal in the combined output.

[ElectricDruid]

What's an appropriate way to measure SNR?

Yeah, it's sometimes tricky. Perhaps especially when the levels get very low.

I've been working on and off trying to reduce the noise level caused by clock jitter in my Flangelicious design. There's a certain level of "watery" background noise that's caused by that, and I'd like to try and reduce it. But first I have to measure it. That's not that simple, since the actual level of that noise varies depending on the clock frequency. What I'd like is some sort of measure of "overall noise level across the whole clock frequency sweep" but that's not at all an easy thing to get a figure on. And because we're talking about jitter, the amount at any particular frequency can vary enormously from another frequency very close, so measurements at specific points don't tell us too much.
So how would I quantify the background noise level, and furthermore how would I quantify it over a whole range of output clock frequencies?

Going back to the problem at hand, I don't think this is that difficult. The FV-1 can be set up with a straight input-to-output algorithm, which should show only the noise introduced by the ADC-DAC process. You can then ground the input to eliminate noise going in. If you also eliminate the dry signal from the output, then what you're looking at is noise caused by the PCB layout and whatever other circuitry there is on the board. That's a fairly clear measure of the background noise level of the circuit. If the noise goes up horribly when it's running a specific algorithm, then the suspicion clearly has to fall on the algorithm, not the rest of the board.

HTH,
Tom

[octfrank]

What's an appropriate way to measure SNR?

e.g. I can estimate the standalone noise to be (from memory, maybe not accurate) 140mV peak-to-peak by disconnecting the input signal and looking at the scope. But I'm not sure how to isolate the noise from the signal in the combined output.

Honestly you need a spectrum analyzer as you need to add all the noise in the range of interest (20Hz to Fs/2) so an FFT is required. In addition there must be a reference signal at say -6db that is centered in one of the FFT bins. You eliminate this bin in calculating the noise floor for SNR, if you cannot do so in the spectrum analyzer then use a notch filter.

If you have a digital scope it may have some basic FFT capabilities so you can do a simple measurement with a notch filter to eliminate the reference signal.

You need a really pure sine wave as any distortion can raise the noise floor due to the reference signal which means you are not getting a good measure.

What may help more is a digital scope with FFT, put in a reference sine centered on a bin and look for peaks in other bins that indicate issues in the circuit/PCB.

A nice simple example is here https://itectec.com/electrical/electronic-how-to-measure-snr-with-an-oscilloscope/

[mark2]

Thanks Frank and Tom,

This helps a lot. I'll definitely start with the simpler method.
Failing that (or just to explore further) I'll try the FFT method as well. My signal generator is pretty crude, but I can probably figure out something.

[MetalGuy]

Since this thread is still going on I decided to check again my unit for FV-1 noise. Usually if I need to check something for noise I connect it to the Return stage of my 100W tube head and put the Master to max. No noise can escape from that.
With my unit at idle (input0 to gnd, in a quiet room, 1 meter from the speakers) I can't hear any extra noise whatsoever from the unit. With some effects engaged I can barely hear the LFOs. Note that with the Master at max the amp would produce deafening volumes when played so for my purposes the noise from my FV-1 unit is insignificant (if at all).