Hello both, Thanks again for your replies. It is very hard to pin point the exact source of the tone. Depending on how I hold the dB meter it might come from inductor L30 or capacitor C427, or something in between. At one point I had the idea that it came from the edge of the cooling element mounted on the CPU, however, that seems illogical. Also, putting some pressure on the cooling element (with something clean) does not influence the tone, which I would expect if that were the case. So, perhaps it is indeed the pcb that is the acoustic source.
Anyways, changing the input voltage of the board seemed like the easiest potential solution. Fortunately I've got an Elabo laboratory power supply, so I did some experimenting with various voltage levels. At 15 volt the noise was virtually gone. Below 14.6V the tone was less stable and at a much lower frequency (more like a hum), regardless of activity on the board. Unfortunately, the power supply does not give enough ampere's below 11.7 volt (I assume) so the board quits below that level. Perhaps the lower frequency hum is the result of not enough ampere's below 14.6V? I do have a simple 5v power supply. That didn't have any impact on the noise. It seems that a higher voltage power supply could be a workable solution, especially since I don't expect noise isolation to improve the cooling of the board. Nonetheless, I will drop a line to Soekris, see if they have any experience with this issue. Perhaps there is a solution C. Kind regards, Maarten On 28/01/16 16:43, ED Fochler wrote: > Sound absorbing mats like dynamat do work. Expensive, of course. Surround > any direct path for sound from the MoBo to the outside world with that and > you won’t hear it. Just allow some angled hallway of airflow to keep the box > cool. It doesn’t have to be close. A big box with indirect airflow lined > with acoustic mat will hide lots of fairly noisy components. This is my > favorite option as dynamat etc. is beneficial to any sound rig. > > If you are going to try something smaller and closer to the source, like goo > on the ferrite block, then thermal properties become as important to acoustic > properties. Go for the thermal pads like Gavrik mentions or black potting > compound (thermal epoxy). I think the acoustic output is from the green > board itself though, no matter what is driving it. So if you’re trying to > isolate the sound close to the source, you’re likely going to be putting pads > on top and bottom of the MoBo to try to catch the vibration. This sounds > problematic / messy in this situation. > > Also, I have used 12v and 19.5v inputs to the barrel connector on my soekris > board. Soekris sells 12v power supplies, but the board is rated to deal with > 6v to ~36v as valid input. The tone of the sound that I could make out > changed a lot with different input voltages. You may want to try a laptop > power input instead of the 12v input to see that moves your noise problem > outside of the human hearing range. > > Good luck. > > ED. > > >> On 2016, Jan 26, at 11:22 PM, gavrik peterson <grassr...@gmail.com> wrote: >> >> It appears that C427 and C418 are 2.5V electrolytic capacitors. Their low >> voltage rating suggests that they are associated with a switch-mode >> regulator that supplies power to the CPU. Notice that the inductors, L30 and >> L31, are very close to the place you identified as the source of sound. I >> suggest that it is far more likely that the noise is coming from one or both >> of these inductors. This because the ferrite that forms the core of these >> inductors is a material that exhibits magnetostriction. That is, the >> material responds to changing magnetic flux by changing its dimensions. >> While the dimensional changes are small, it is enough to produce sensible >> levels of sound. I have never experienced this sort of behavior from >> electrolytic capacitors. >> >> At this point a question remains. Is the regulator circuit associated with >> these capacitors and inductors doing something irregular that is producing >> this noise? Or, is the regulator behaving properly, and this noise is simply >> due to that fact that the CPU current demand is changing (CPU current is >> very dynamic) and this is causing the sound? I am thinking that the second >> possibility is most probable. As I don't have a schematic, or a board to >> examine, I can only guess. But most likely the two inductors and two >> capacitors are the output of a two-phase buck regulator that reduces the >> voltage from the external power adapter to something like 1.5 Volts for the >> CPU. If this is the case, one would expect the variation in CPU supply >> current to cause the inductors to produce sensible sound. >> >> If it is the inductor(s) that is producing the sound, I am not sure what one >> can do about it. Perhaps some non-conductive material can be placed over the >> inductors to absorb the sound. Some rubbery stuff that is acoustically lossy >> might work. One thing that comes to mind is the materials that are used to >> remove heat from electronic components. This material is sort of gooey and >> is very soft and conformal, so it can be applied to irregular surfaces. This >> stuff is referred to as "Thermally Conductive Gap Filler Pads". One supplier >> of the material can be found here. And more specifically, the HCS10 material >> is here. If you are interested in this solution you might call the company >> and ask about acoustical properties. Usually suppliers will send free >> samples for evaluation. Notice that the HSS10 material is very soft and is >> available in up to 0.20 inch thickness. >> >> -- Gavrik >> _______________________________________________ >> Soekris-tech mailing list >> Soekris-tech@lists.soekris.com >> http://lists.soekris.com/mailman/listinfo/soekris-tech > _______________________________________________ > Soekris-tech mailing list > Soekris-tech@lists.soekris.com > http://lists.soekris.com/mailman/listinfo/soekris-tech _______________________________________________ Soekris-tech mailing list Soekris-tech@lists.soekris.com http://lists.soekris.com/mailman/listinfo/soekris-tech
