R-cores are quite popular for audio. Perhaps its for their look because as regards parasitic capacitance, a split-bobbin EI usually wins.
I've seen some 10v dc, 10...50w switching supplies at LHC at Harwell which had up to 10uv noise, but they were state of the art by all means...and they said that they had litteraly no budget constraits for making them as well as any other electronics sustaining the hadron collider....
Yeah I think it has totally to do with their looks 🙂
It is the way they are wound and their absurdly small stray field IMHO. They are not as broadband as toroids which helps. I have yet to see one that hums.
I tend to believe that 50% of possible PSU trouble can be avoided by chosing a better transformer.
Last edited:
I've seen once a higher spec sony dvd player which placed the r-core transformer literally within 5mm...1cm to the lowest noise regulator chip and no audible hum got into the circuit.
Yes. Please note that cheap R-cores sometimes are not wound the way they should be wound.
Last edited:
A split bobbin EI transformer is one where the primary is wound on a physically separate bobbin from the secondary. Most of the small non-encapsulated EI transformers I see on Taobao are in fact wound this way. It gives a higher leakage inductance than putting the two windings on the same bobbin but better isolation between windings.
Attachments
OK, i have tons of them.I didn't know what you mean by split bobbin...they also used to have a third winding between primary and secondary connected to the iron core and ground, but none of them behave like an r-core honestly.
That is why for ultra-low noise I recommend feeding the SilentSwitcher with a Powerbank. Nothing connected to the mains comes close.
Jan
R-cores of course win on radiated field and beauty. EIs win on cost and isolation.
What's the correct way to wind an R-core @J-P? The ones I've examined are split bobbin on the surface but turn out to put half of each winding on one bobbin and the other half on the other. Which presumably is to minimize leakage.
Thanks a lot again. If i understand well if let's say the power amp is grounded then no problems ? I have to check that.
I would like to ask if you have any device that can work as a spectrum analyzer. I start to think that this is a mandatory equipment for testing.
I have looked at the price ... they are very expensive even refurbished.
As an alternative there are some usb analyzers some of which said to be very good like this one
QA401 Audio Analyzer
– QuantAsylum
as i have no pretension of high accuracy i wonder what would be just a decent starting point. I really feel the need to see some noise on a video
I see it the only way to check the quality of some psu circuits ...
Looks like you haven't understood yet. A grounded poweramp when fed by a balanced cable would be fine as then the noise current flows via pin1. But with an unbalanced IC you've got ultrasonic noise going through the screen of your cable. Which then gets summed with your wanted signal as the amp amplifies the difference between the core and the screen at its input.
I do have one, yes it was fairly expensive but I haven't used it much. So I wouldn't consider it an essential piece of kit for most audio work. With digital technology being used more and more in SAs they are definitely getting cheaper and far, far less bulky. When I was at uni I had a summer job working in the lab that at that time had just designed the Marconi Instruments TF2370. The thing was huge (two stacked boxes) and probably a whole year's salary to purchase.
If you have no need to examine RF signals then a soundcard would be a much better starting point than a full-blown SA. I use a USB ADC for low frequency work.
I see. Pin 1 is the pin connected to the ground
An externally hosted image should be here but it was not working when we last tested it.
why the noise current flows only through pin 1 ?
i see again. And there is no way to avoid this issue ? i mean ... is the balanced way the only very good one ? 😕
i ask this because most of the audio equipment around is unbalanced and in some cases noise is minimal i guess.
very interesting and my sincere congratulations for your work. I like this kind of instruments ... everything is quite evident through it. Noise, distortion, ... very nice.
if to examine RF signals gets expensive than i have no need for that 🙁
Seriously ... i would love to get a very nice equipment like everyone else
But money is indeed an issue. Audio is not the source of my salary but only of my pleasure ... lucky who get both at the same time 😱
I don't think that is correct. The current runs from mains phase through transformer and load back to mains neutral. You may have a fault current going to earth safety but that is what it is called, a fault current. There's no reason for the primary or secondary current to flow to safety earth because it is just as you say a reference, not a current source/sink.
The situation with the XLR pin1 is a bit more subtle. This pin should NOT be connected to signal ground, only to chassis. It is in a sense extending the cable screen to the enclosure but should not carry a signal ground. The beauty of balanced interconnects is that ground does not come into the picture; the signal is the voltage between the two signal pins (here pin 2 and 3). The only requirement is that these two signals remain within the power supply range of the receiving circuit, but they have no reference to ground, and should not have.
Jan
Last edited:
Because pin1 should be bonded to chassis and in turn chassis is bonded to safety earth. So it goes the lowest impedance route.
I think its the only way that completely solves the issue. Other ways are going to reduce it, even substantially but not completely resolve it. Those other ways are common-mode chokes and isolation transformers. They're band-aids on the 'wound' which is unbalanced ICs.
How do you know its minimal? If you listen with no music playing you'll hear nothing because the noise in question is ultrasonic (typically 40-100kHz fundamental, plus harmonics).
I don't think you're missing very much by not going there.
If the impedance to safety ground is lower then the current will predominantly take that route. Going back to mains neutral entails traversing a transformer's inter-winding capacitance which would normally be higher impedance (though of course depends on the size of the trafo).
If you are talking about the mains frequency currents then yeah hardly anything goes to safety earth because at 50Hz the impedances are too high via the parasitics of a trafo. But above 40kHz things change somewhat.
Normally circuit ground goes to chassis too, just in another place. But basically no disagreement with what you've said here.
Last edited:
Hi thanks a lot for your kind and helpful advice.
But i am ready to some compromise in order to stay simple
This is the smps i have got and i would like to test maybe with some LC filter downstream ...
https://www.mouser.it/ProductDetail/MEAN-WELL/LRS-35-48?qs=vDxCgdWo2h9Xg/U5xGVofA==
i do not need to kick down noise to the uV ...
by the way i am attaching the pic of a dac i have. The noise is very low indeed and it uses what looks like an off-the-shelf smps Maybe they have added some passive filtering i do not know.
Passive filtering i understand can be very effective at suppressing some residual ripple out of these smps
If they are good for this professional unit they could be good also for my silly prototypes i guess
But without measurements it is very hard to say
What i am thinking to do is to get a mic preamp battery powered, crank its gain to maximun and listen to the sound of the ripple out of this bloody smps ... maybe i will cook my brain
Attachments
Last edited:
Yes. There are many people that still think the word 'balanced' in balanced connections stands for balanced signal. Which is wrong of course; if you require that the two signals on balanced lines are equal in level with respect to ground you again introduce ground as a reference and that was what you wanted to get away from in the first place.
'Balanced' stands for equal source and load impedance. That way, any EMI, or other magnetically or electrically noise coupled to the cable will be equal in amplitude in both wires, and thus cancelled by the differential circuit at the receiving side.
Jan
'Balanced' stands for equal source and load impedance. That way, any EMI, or other magnetically or electrically noise coupled to the cable will be equal in amplitude in both wires, and thus cancelled by the differential circuit at the receiving side.
Jan
well ... i do not know Can i check that without open it just with a tester ?
I am not using it by long time. Also because it is balanced and i want to stay unbalanced ... more in line with my nature 😱
Actually this balanced thing i do not understand well. Let's take a preamp for instance ... i should see 4 independent identical preamp lines starting from the volume attenuator with 4 sections.
Instead i see balanced preamps that are stereo inside ... just with an op-amp at the input and maybe another one at the output. Is this not cheating ?
Balanced on the outside and unbalanced in the inside
Last edited: