SMPS _ shielding options.

[ginetto61]

I think your ARTA software should be able to measure the noise from the power supply. Put a small capacitor (say 1 uF, film cap) in series, so that DC voltage from the power supply is blocked by the capacitor, and only the AC (noise) gets through.

Hi ! Thanks again for the helpful advice.
But i was thinking ... if my need is to check the all ensemble, i.e. psu + usb interface, maybe the measurement more telling is the one on the whole set up ? in this way i will have an idead of the real noise of the system.
Of course if i test the psu i could screen out the best one in terms of noise.
I will try the procedure you recommend.

This is normal and expected, and there is something to be learned from this observation.
The smallest audio signal is at the (mic) inputs.
Because the signal is very small, this is the place where it is hardest to get a really good signal to noise ratio. The electronics has to have really low noise to keep it well below the really small audio signal.
That means the input stage is always the noisiest stage in a well-designed preamp.
That in turn, means that the total noise from any properly designed audio amplifier (including your TEAC interface) will be dominated by the noise from the input stage (mic input stage).
With mic gain at zero, you are measuring only the noise from the later stages of the preamp. This looks impressively low, but it's not going to be the major source of noise.
It does not really reflect the performance of the unit, since you cannot actually use this device with the mic gains turned down to zero!

Thank you very much for the very helpful explanation. I will carry out the measurement again but this time with gain to max and see the result.
I understand the huge challenge of designing a very low noise and high gain unit. I have always been fascinated by phono preamps for moving coil cartridges. A MC cartridge gives a very low signal

MC 20 Super re-introduction
Output voltage at 1000 Hz, 5cm/sec. - 0,25 mV

When (if) you actually use this circuit to record audio, the electronic internal noise will be dominated by the mic input stage.
That is the electronic signal-to-noise ratio that actually matters.
I keep using the word "electronic", because even the mic input isn't the biggest source of noise in the recording. There is almost always a much bigger source of noise: the acoustic noise in the room or environment where you are making the recording.
When you connect a microphone to the input and make a recording of the "silence" in the room, it's amazing how much acoustic noise you will typically find. For example, you may hear the refrigerator humming in the kitchen, two or three rooms away. And the cooling fan inside the PC connected to the TEAC interface. And the room air-conditioning or heating. And your neighbour's cats saying "miaow!"
My experience is that with home recordings, this acoustic background noise is always much, much, worse than the electronic noise from any microphone preamp, even a cheap one. Because of this, I hardly care what the advertised signal to noise ratio of my mic inputs is.
It really doesn't matter, because the electronic noise will be inaudible, buried in the background acoustic noise around the microphone

Very interesting. What i find also amazing that in reality it seems that the brain filters some of this noise. Like a neural noise cancelling process.
However i am more on the party "i want to hear it all".
I think that in order to get a strong sense of ambience also this minute details are very important. For the live effect i mean.
I still prefer a "silent" recording system, that adds very little.

" Honestly, I don't think I'm the right person to give you that advice - because I get all my SMPS from the local thrift-stores, and I usually pay only about $2 for each one! "

I see that normal figures for ripple for smps are 100 mV peak to peak.
I guess it is also important the spectrum of this noise.
While i understand very well the importance of having very low noise in the audio band i am still confused about the need to have also very low noise above the audio band with digital equipment. This is a big and fundamental doubt i have.

The fact is, these days our electronics is usually not the limiting factor in the quality of our music. So I simply don't care about advertised things like 24-bit A/D converters and 192 kHz clock rates and -115 dB signal-to-noise ratios. As far as my ears can tell, none of it matters at all! (And it's not that I have bad ears, it's just that a lot of people hear things that aren't actually there, or simply believe "more is better" when it comes to paper specifications.)
-Gnobuddy

Thanks a lot again for the very interesting and precious advice.
Personally i have been impressed lately by the extreme quality of recorded special effects in movies. I mean real sound that so realistic in an even scary way. They are just so real.
My reasoning is that if they are so good at recording natural sound they must be similarly good at recording music.
I guess that they do not use computers but just standalone and portable digital recorders.
Unbelievable quality indeed.
And actually i would like to start some recording one day. I have to find a music group willing to be recorded
Thanks again a lot for the very helpful advice.
Kind regards, gino

 

[Gnobuddy]

...if my need is to check the all ensemble, i.e. psu + usb interface, maybe the measurement more telling is the one on the whole set up?

Yes, absolutely. But you wanted to also test power supplies, to compare them to each other, right?

Output voltage at 1000 Hz, 5cm/sec. - 0,25 mV

I think those moving coil cartridges were quite challenging to design engineers when they were first introduced, many decades ago.

But consumer electronics never was the cutting edge - the most advanced technology is usually secret military stuff.

Today's most challenging signals are probably the ones JPL/NASA gets from their most distant satellites. Voyager 1 has a 20-watt transmitter, and it is now about 20 billion kilometres away from earth. Imagine how dim a 20-watt light bulb 20 billion kilometres away would be? Yet JPL now has receivers with such low noise that they are *still* in communication with that ancient satellite!

I know next to nothing about how it is done - it would be fascinating to find out. I do know that cooling the front-end electronics down to incredibly low temperatures is a must (to reduce thermal noise in the electronics).

What i find also amazing that in reality it seems that the brain filters some of this noise.

I agree, absolutely amazing. Imagine if you strap a video camera to your forehead, and run a 100 metre sprint. If you looked at the video, it would be so shaky and blurry that it would be unbearable.

But your *eyes* were experiencing the same shaking and movement as that video camera - yet, you saw everything around you clearly, without blur or shake, as you ran. Your brain magically filtered out all that shaking, jerking, and erratic movement!

A good thing, otherwise our distant ancestors on the African Savanna would have become dizzy and fallen down as soon as they tried to run away from a hungry lion. That would have made things very easy for the lions!

i am still confused about the need to have also very low noise above the audio band with digital equipment.

All sorts of "experts" will tell you that it is terribly important to have tweeters that work till 40kHz, and so on.

But, when was the last time a bat flying outside your house woke you up? How about a mouse? These creatures make loud ultrasonic noises, but we humans never hear them!

The necessary research was done by Bell Labs scientists (as well as others) many decades ago. 20 kHz was a nice generous upper limit with a nice safety margin - most normal people actually cannot hear that high. These days, growing up in noisy cities, many adults can't hear anything above 10 or 12 kHz.

Many years ago, I was very interested in Hi-Fi. I built a simple passive low-pass filter that cut off frequencies above 10 kHz. While my friends were listening to my music system, I would turn on the filter, without telling them. I found that if I turned it on during the silence between two songs, nobody would ever notice. If I turned it on in the middle of a song, once in a rare while, someone would notice.

So, most of the time, even young people listening intently to music did not notice when I removed everything above 10 kHz!

I have to find a music group willing to be recorded

If you are willing to do it for free, I think it won't be too hard to find someone. There are usually plenty of musicians with no money and big egos, just hoping someone will record them!

There is also the possibility of recording natural sounds - birds, insects, animals, flowing water, etc. A long time ago I remember reading about a fellow who won a contest for the most interesting audio recording - he recorded the sound of a snail eating a lettuce leaf!

-Gnobuddy

 

[AndrewT]

an SMPS emits emi.
That is why the regulators require them to pass emi tests before they can be retailed.

Most are inside faraday cages to help with emi. The PC smps is always inside a faraday cage even though they are also inside the PC case.
The wires into and out of the smps also emit emi. This is why we pair them off to attenuate that emi. It is usual to fit filters to both the input wiring and the output wiring of smps.

I opened up an very old Alcatel mobile phone. It had 5 or 6 little soldered closed boxes and when I opened one of these it had a second even smaller soldered box inside.
This is Acatel attenuating emi.

 

[ginetto61]

Yes, absolutely. But you wanted to also test power supplies, to compare them to each other, right?

Hi again !
sorry for the slow reply but for this week i am visiting parents and relatives ... lot of food and bla bla but not electronics
Yes ! you are perfectly right. I would like to measure noise/ripple of different psu. But the task is not trivial at all when ripple is low in the mV region.
I have seen a National Instruments video on Youtube and i got sincerely scared. The lab engineer had equipment for thousands of dollars
So i little give up to the idea. Of course this would be the sane approach.
Measuring the noise floor on an interface changing the psu can give some hints about how sensitive is the interface to psu quality.
Until now i have only used cheap smps and more or less the results are the same with all interfaces i have at hand.
One case aside. The stock smps in the Tascam gives bumps in the noise floor that vanish using an external and low price Dell refurbished 12VDC psu.
Just think that i have found this Dell smps that was about to end in a garbage bin ... still is doing very well with the Tascam. Noise floor nicely flat.
Again i should "scope" the original. Maybe just a cap more and the bumps go.
About extremely low noise psu solutions i have already asked here about batteries. And i got mixed answers.
Someone was citing their noise being not that low in reality.
Then a member directed me to this paper

http://tf.nist.gov/general/pdf/1133.pdf

if you look at fig.3 (that i am attaching) you will see that NiCd batteries have a noise just a little below the noise floor of the testing device
Problem with batteries is that they are functionally a pita, especially with high consumption units.
But they can be spectacular and they provide complete isolation from mains, not a bad thing usually.
After all i am just looking for 12VDC/2A with ripple within some mV peak to peak.

I think those moving coil cartridges were quite challenging to design engineers when they were first introduced, many decades ago.
But consumer electronics never was the cutting edge - the most advanced technology is usually secret military stuff.
Today's most challenging signals are probably the ones JPL/NASA gets from their most distant satellites. Voyager 1 has a 20-watt transmitter, and it is now about 20 billion kilometres away from earth. Imagine how dim a 20-watt light bulb 20 billion kilometres away would be?
Yet JPL now has receivers with such low noise that they are *still* in communication with that ancient satellite!
I know next to nothing about how it is done - it would be fascinating to find out. I do know that cooling the front-end electronics down to incredibly low temperatures is a must (to reduce thermal noise in the electronics).
I agree, absolutely amazing. Imagine if you strap a video camera to your forehead, and run a 100 metre sprint. If you looked at the video, it would be so shaky and blurry that it would be unbearable.
But your *eyes* were experiencing the same shaking and movement as that video camera - yet, you saw everything around you clearly, without blur or shake, as you ran. Your brain magically filtered out all that shaking, jerking, and erratic movement!
A good thing, otherwise our distant ancestors on the African Savanna would have become dizzy and fallen down as soon as they tried to run away from a hungry lion. That would have made things very easy for the lions!
All sorts of "experts" will tell you that it is terribly important to have tweeters that work till 40kHz, and so on.
But, when was the last time a bat flying outside your house woke you up? How about a mouse? These creatures make loud ultrasonic noises, but we humans never hear them!
The necessary research was done by Bell Labs scientists (as well as others) many decades ago.
20 kHz was a nice generous upper limit with a nice safety margin - most normal people actually cannot hear that high. These days, growing up in noisy cities, many adults can't hear anything above 10 or 12 kHz.
Many years ago, I was very interested in Hi-Fi. I built a simple passive low-pass filter that cut off frequencies above 10 kHz. While my friends were listening to my music system, I would turn on the filter, without telling them. I found that if I turned it on during the silence between two songs, nobody would ever notice. If I turned it on in the middle of a song, once in a rare while, someone would notice.
So, most of the time, even young people listening intently to music did not notice when I removed everything above 10 kHz!
If you are willing to do it for free, I think it won't be too hard to find someone. There are usually plenty of musicians with no money and big egos, just hoping someone will record them!
There is also the possibility of recording natural sounds - birds, insects, animals, flowing water, etc. A long time ago I remember reading about a fellow who won a contest for the most interesting audio recording - he recorded the sound of a snail eating a lettuce leaf!
-Gnobuddy

Thank you very much indeed for the very interesting stories and valuable advice.
This month i will be testing what i have already at hand. I have at least two decent interfaces. Then i need two mics i guess. Both can provide +48V for 2 mics.
But going back to noise meausurement i bought one year ago one of these (as soon as i will go back home i will check the model)

Qingdao Hantek Electronic Co., Ltd.

just for noise measurement on psu. I am very obsessed.
But it missed i think the cap coupling feature strange ... maybe i am the only one interested in this kind of tests.
For sure checking the noise floor of interfaces is much much more immediate. If it is easy for me it means that it is really easy
So i will think about direct ripple measurements for the future
One thing interesting are the preamps to increase low ripple level in order to measure it more easily.
They must be very silent preamps.
Thanks a lot again, gino

 

[ginetto61]

an SMPS emits emi.
That is why the regulators require them to pass emi tests before they can be retailed.
Most are inside faraday cages to help with emi. The PC smps is always inside a faraday cage even though they are also inside the PC case.
The wires into and out of the smps also emit emi. This is why we pair them off to attenuate that emi. It is usual to fit filters to both the input wiring and the output wiring of smps.
I opened up an very old Alcatel mobile phone. It had 5 or 6 little soldered closed boxes and when I opened one of these it had a second even smaller soldered box inside.
This is Acatel attenuating emi.

Hi ! thanks a lot for the very valuable advice and precious information.
The fact that a very very common solution is an external power adapter i guess it to avoid this challenge.
Nevertheless i know for sure that there are equipment with internal smps and fantastic figures of noise.
The troubleshooting on these units is really a task for expert people.
I see 220VAC input and a 400V cap
A very unelegant solution could be to place the Dell adapter inside in the place of the stock one ...
I cannot check precisely now (i am far from home) but it looks extremely similar to this one here (look at the price also ... )

Dell PSCV360104A 12VDC 3A AC Adapter 6mm Barrel with Centre Pin (1503FP) | eBay

I could just use some bi-adhesive industrial tape to stick it to the case, connect in and out and be done with that.
As i said the noise floor now is very much flatter than with the stock internal open smps (not shielded at all).
But first i will try the effect of some nice linear psu.
If the noise graph will not change i will go with the smps.
I see many choice here

http://www.ebay.co.uk/sch/i.html?_f...stpos=&gbr=1&_nkw=12VDC+3A+AC+Adapter&_sop=15

In this way i will use the unit as intended, with mains voltage
Thanks a lot again, gino

P.S. or something like this one here

http://www.ebay.com/itm/12Vdc-3A-ac...604854?hash=item3a8f9e13f6:g:6FgAAMXQeW5TXn1F

 

[AndrewT]

I see 220VAC input and a 400V cap .

Why eek?

253Vac is equivalent to 357.8Vpk.
That leaves 42V for spikes on the mains.

 

[ginetto61]

Why eek?
253Vac is equivalent to 357.8Vpk.
That leaves 42V for spikes on the mains.

Hi ! I see a very high risk for a not experienced guy.
When i see tube equipment and the voltage involved i am scared
Actually i should use one of this

Powering Your Radio Safely With a Dim-bulb Tester

I wonder if the bulb must be on the LIVE wire or also one the NEUTRAL will be just fine. I mean, is the orientation of the mains plug important ?
I will built one 1st thing in the next days.
Better safe than sorry.
Thanks again, gino

 

[AndrewT]

the Mains Bulb Tester will work both ways, the Bulb resistance in the Live line or in the Neutral line.
The effect is the same.

The difference is inside the equipment being tested.
With the Bulb resistance in the Live line, the voltage at the equipment is lower relative to Earth.
If the bulb goes to "bright" the voltage inside the equipment is probably around 5Vac to 10Vac relative to Earth.

 

[Gnobuddy]

this week i am visiting parents and relatives ... lot of food and bla bla but not electronics

Remember, electronics will never love you! Family is far more important!

I would like to measure noise/ripple of different psu. But the task is not trivial at all when ripple is low in the mV region.

You are already measuring noise much smaller than that, using your ARTA software. For example, a noise floor at -115 dB is around 2 micro volts. That is a thousand times smaller than 2 millivolts!

Just put a capacitor in series to block DC from the input, as I said earlier. ARTA may not be the ideal equipment to measure power supply noise, but you will get a noise spectrum from it, which is more information than you have now. That will let you compare one power supply to another.

I have seen a National Instruments video on Youtube and i got sincerely scared. The lab engineer had equipment for thousands of dollars

When you work for one of the best electronics manufacturers in the world, top-notch test equipment is part of the job. That equipment is probably used for a lot of different tests, and the accuracy of the results has legal significance - i.e., National Instruments could be sued if their measurements are inaccurate. Their data must be super-accurate, and super-trustworthy.

Your situation is very different. You just want to compare different power supplies - and nobody will sue you if your measurement has an error of 10%!

if you look at fig.3 (that i am attaching) you will see that NiCd batteries have a noise just a little below the noise floor of the testing device

NIST is an even more extreme case than National Semiconductor. Their job is to make measurements so precise that the rest of the world uses them as reference standards!

By the way, the "Cd" in "NiCd" stands for cadmium, which is extremely poisonous. Decades ago, cadmium was used in a lot of consumer products - not only rechargeable batteries, but even to electroplate some automotive parts. But eventually sanity showed up, and this very toxic material is no longer used in consumer products.

-Gnobuddy

 

[ginetto61]

the Mains Bulb Tester will work both ways, the Bulb resistance in the Live line or in the Neutral line.
The effect is the same.
The difference is inside the equipment being tested.
With the Bulb resistance in the Live line, the voltage at the equipment is lower relative to Earth.
If the bulb goes to "bright" the voltage inside the equipment is probably around 5Vac to 10Vac relative to Earth.

Hi ! thanks a lot indeed for the very precious advice
I will keep it in mind. I will check the correct phase when connecting the bulb tester to the mains.
Regards, gino

 

[ginetto61]

Remember, electronics will never love you! Family is far more important!

Hi ! yes you are right.
Unfortunately in my family no one has this interest.

You are already measuring noise much smaller than that, using your ARTA software. For example, a noise floor at -115 dB is around 2 micro volts. That is a thousand times smaller than 2 millivolts!
Just put a capacitor in series to block DC from the input, as I said earlier.

Thanks ! but still i need a dummy load to measure the psu noise at 1-2A.
I will try to measure the current draw but the stock psu is 1.5A.

ARTA may not be the ideal equipment to measure power supply noise, but you will get a noise spectrum from it, which is more information than you have now. That will let you compare one power supply to another.

Yes ! this is exactly what i would like to do. Arta spectrum analyzer goes up to 100kHz.

When you work for one of the best electronics manufacturers in the world, top-notch test equipment is part of the job. That equipment is probably used for a lot of different tests, and the accuracy of the results has legal significance - i.e., National Instruments could be sued if their measurements are inaccurate. Their data must be super-accurate, and super-trustworthy.
Your situation is very different. You just want to compare different power supplies - and nobody will sue you if your measurement has an error of 10%!

Good ! so decent tests can be performed with much more basic equipment.
This is supposed to be my first step in the testing.

NIST is an even more extreme case than National Semiconductor. Their job is to make measurements so precise that the rest of the world uses them as reference standards!
By the way, the "Cd" in "NiCd" stands for cadmium, which is extremely poisonous. Decades ago, cadmium was used in a lot of consumer products - not only rechargeable batteries, but even to electroplate some automotive parts. But eventually sanity showed up, and this very toxic material is no longer used in consumer products.
-Gnobuddy

Thanks a lot again for the extremely interesting and valuable advice.
I understand the health issue. Still the figures are very impressive.
The noise floor of the best batteries is almost equal to the noise floor of the equipment
Who said that the batteries are noisy ?
Thanks a lot again, gino