KRONOS Fan Information

[curvebender]

I was pleasantly surprised when I turned on the Kronos, and noticed that it was approximately half as noisy when compared to how it was when the fan was mounted with metal screws.

It's not silent, but it's a big step in the right direction.

[dfahrner]

Have to say, I'm tempted to drill vent holes in the K! :D

I thought about this too...slots on each end cover / plate (with appropriate location, filtration, etc.) would solve any cooling problem...doesn't the OASYS have vents on the end like this?

But it's hard to believe that the Kronos need this much cooling: it's pretty much a sealed box (is there any air moving out the slider slots on the top, or out from under the keyboard?), it's been run for several weeks without the internal fan, etc...the fan must be there to just move the air around inside, so that there aren't any hot spots; as long as there's even a little air movement, then the entire area of the case can dissipate the heat (that warm spot on the upper right, isn't that directly over the power supply? - that would be why it's warm)...but I'm no thermodynamic engineer, and may be completely wrong here...

My K61 is very quiet with the Silenx IXP-34-12 (14 CFM) installed, and the upper right front panel is only slightly warmer, but I think I'll install an IXP-34-16 (18 CFM) or one of the Noiseblockers, for something a little closer to the stock fan's airflow - the IXP-34-12 just doesn't move much air...

df

[aribo61]

I was pleasantly surprised when I turned on the Kronos, and noticed that it was approximately half as noisy when compared to how it was when the fan was mounted with metal screws.

It's not silent, but it's a big step in the right direction.

this would have been even better, if you had used another fan (e.g. Noiseblocker), just don't use the Silenx (which seems to have max. 1/10 of the desired throughput, more likely 1/24!). The Noiseblocker is *far* more silent, than the stock fan. I only had higher expectations and thought I could go even further, that's why I wasn't satisfied, when I tried it.

I have a new way of installing both 60mm and 80mm fans interchangeably with little effort and with good noise decoupling.

I ordered two additional 80mm fans which look promising (and have at least three more models in the queue, if those don't fit). I'll tell more about those fans when I receive them or on request (they are on the way, so they should arrive tomorrow).

Also, I am preparing photos of the 80mm mounting to add them to my kronos-album.

And, I got an infrared thermometer (as maphill urged me to ), so from now on I can deliver more exact measurements.
But I didn't order a sound level meter, as I think those values don't say much about how annoying the noise is, which is what I am only interested in.
I see (=hear) this from a musicians point: I have a big problem with all kinds of tones (hum etc.), which badly influence my musical feel, even relatively silent ones. I have a far smaller problem with "windy" noises, even if they are not that silent.

[aribo61]

Have to say, I'm tempted to drill vent holes in the K!

I think this could work. The motherboard is specified to be used fanless, if a certain air flow is guarantied.
But I am unsure about the PSU. Did anyone identify the PSU and knows how it's specified?

Acoording to air flow, I think there should be a grid of holes in the bottom cover just below the chip set cooler, so the cool air directly gets to the cooler. Because hot air flows to the top and the hot spot was already identified in the right back corner of the Kronos case, this would be the best place for a second grid. A better place from a design point of view would probably be at the back right top of the case.

My new (and still unpacked) infrared thermometer also should have an external sensor with a cable. With that I want to find out how hot the cooler is in different situations. I think as a good base for good cooling I can use an open Kronos, which stands on it's back side (so with the keyboard at the top). We could compare the according cooler temperature with any other solution to decide if it's working enough.
However, the best way for measuring would be with the motherboard sensor (which we could eventually get from the Kronos OS, as it should be a kind of linux). But I still didn't find a way to connect a VGA cable. If I would remove the side bar of the case (and I don't see how I could do this), the Kronos hardware would become very unstable (mechanically), I think.

btw. @dfahrner: the hot spot is mainly above the chipset, for the K61 this is where the most right switch panel is. The PSU has it's own hot spot a bit more on the left at the second switch panel.
And I think the Silenx ixp-34-16 is also *far* away from enough throughput. If we would very graciously grant the ixp-34-12 (14cfm) a factor of 1/10 of the desired throughput, it would be about 1/8 for the ixp-34-16 (18cfm). I think you could better go without any fan, without seeing a noticable difference.

[curvebender]

this would have been even better, if you had used another fan (e.g. Noiseblocker)

But does the Noiseblocker move enough air to cool the system?..

I ordered two additional 80mm fans which look promising (and have at least three more models in the queue, if those don't fit). I'll tell more about those fans when I receive them or on request (they are on the way, so they should arrive tomorrow).

Looking very much forward to your findings with the 80mm fans!!

And, I got an infrared thermometer (as maphill urged me to ), so from now on I can deliver more exact measurements.

Me too!

[aribo61]

look here (album: korg-kronos-inside) for some new photos of my fan mounting etc., now with some captions to describe what they should show

[aribo61]

But does the Noiseblocker move enough air to cool the system?..

in my table you see:

Kronos-Standard 23s 60s
Noiseblocker XR2 50s 50s

the first numbers are for 12V, where the stock fan is about twice as fast, but at 9V the Noiseblocker doesn't move less air (I couldn't believe it, may be the speed stays the same?), so at 9V the stock fan and the Noiseblocker have a comparable throughput.
This is backed up by my observations: with the Noiseblocker my Kronos stayed cool.

[aribo61]

I tried some calculations with my measurement table

Code: Select all

Fan                    12V      9V (motherboard)
Kronos-Standard        23s      60s
Silenx IXP-34-12     >240s      --
  about half filled  >120s    >180s
Noiseblocker XR2       50s      50s
Arctic F8 (80mm)       15s      20s 

using a linux calculator called qalc, which can handle units, but it doesn't know cfm, so I wrote ft^3/min instead:

the stock fan at 12V should have 14 cfm:

14 ft^3/min * 23s / liter =~ 151 [liter], so I assume I have measured with a 150 liter sack

(150L/23s / (ft^3/min) ~= 13.8 [cfm])

and at 9V:

150L/60s / (ft^3/min) ~= 5.3 [cfm]

for the Silenx ixp-34-12:

at 9V: 150L/360s / (ft^3/min) ~= 0.88 [cfm]
at 12V: 150L/240s / (ft^3/min) ~= 1.3 [cfm]

with 28% more for the ixp-34-16 it would be:

at 9V: 1.1 cfm
at 12V: 1.7 cfm

and for the Noiseblocker at 9V and 12V:

150L/50s / (ft^3/min) ~= 6.35 [cfm]

[Davidb]

Have to say, I'm tempted to drill vent holes in the K!

I think this could work. The motherboard is specified to be used fanless, if a certain air flow is guarantied.
But I am unsure about the PSU. Did anyone identify the PSU and knows how it's specified?

.

I think this theory is valid for the moteherboard, but you have to asume the fan is there for the SSD and mostly for the PSU.
Both need the fan in order to not overheat the system, the same way a PC or Mac work. So ...Be prepared to live with (more o less) fan noise.

[aribo61]

I think this theory is valid for the moteherboard, but you have to asume the fan is there for the SSD and mostly for the PSU.
Both need the fan in order to not overheat the system, the same way a PC or Mac work. So ...Be prepared to live with (more o less) fan noise.

I'm quite sure the SSD doesn't need cooling.

I just measured it:

after some hours with a closed case without a fan (and no holes in the case) I get the following temperatures (running my worst case test):

With an external sensor (with a cable) directly at the bottom of the motherboard cooler, I get 70°C.

Measured with an infrared thermometer just after opening the case:

67°C motherboard cooler
67°C PSU
27°C SSD

For the chipset I think this will be very near to it's limit, because the chip will be much hotter than the cooler measured from outside. I think it is probably near 90°C (which is the specified limit).

I don't know much about PSUs, but I think it's also near it's limits.

Before the closed test, I ran the Kronos standing on it's back and with an open bottom cover. This should be the best air flow for the cooler without a fan. I never got above 42°C (measured with the external sensor).

Coming from the closed case (70°C) I have now a constant 53°C (external sensor, 50°C with infrared).

I am afraid, that we cannot get an air flow to sufficiently reduce the critical temperatures simply by punching holes into the case.

[dfahrner]

To verify aribo61's test results: I replaced the Silenx IXP-34-12 in my K61 with a Noiseblocker NB-BlackSilentFan XR2 (instead going to a Silenx IXP-34-16)...the Noiseblocker makes maybe a little more noise than the Silenx (it is rated at 15dBA vs. 12dBA for the Silenx) but far less than the stock fan with the screw mounting, and the warm spot on the upper right of the front panel is now much cooler than with the Silenx, in fact barely warmer than the rest of the panel...the Noiseblocker is certainly the best compromise in cooling / fan noise so far......again, thanks to aribo61 for all of his work...

As for drilling holes, that would be a real design change, as opposed to simply replacing the fan with a quieter one with the same airflow...again, I think the Kronos must have been designed so that air circulating inside the case will transfer the heat from internal components to the entire surface of the case, and that the whole surface area of the case will then dissipate the heat fast enough so that no internal component will be outside of its operating temperature specification...

df

[aribo61]

To verify aribo61's test results

thanks for verifying this...

I did a new test with the original fan mounted in the original position (but on rubber mounts, because I didn't want to have this heavy noise for several hours. I think, the rubber mounts shouldn't make a difference for cooling.

I measured the temperature directly at the bottom of the motherboard cooler (which was about 70°C without a fan).
With the stock fan I get about 57-60°C depending on the room temperature (measured from evening to after midnight).

I noticed, that the temperature slowly increases over several hours, so one hour seems to be much too less time to get a stable result.
The internal temperature increases even when the external temperature (at the warm spot at the top right of the case) has already stabilized.

I then measured with the Noiseblocker and got a stable temperature of about 60-61°C. I think this is similar to the stock fan, because during measurement my room was some degrees warmer (around high noon).

I was surprised, that a good cooling (stock fan as reference) is only 10°C lower than without a fan. The PSU is also pretty warm with the stock fan: 55-57°C (infrared). The motherboard cooler also was at 55-57°C (infrared).

Conclusions?

btw. currently I have the Noiseblocker installed with a new kind of mounting (cable ties but different construction with only one loose connection to the SSD mount). This gives me a better result than before, when I connected the fan with four cable ties to the SSD mount).

At the end, these 80mm fans:

• - Arctic F8
  - BeQuiet BL011 Silent Wings USC 80mm
  - Xilence DualWing XQ 80mm

were more or less comparable to the 60mm Noiseblocker (the Xilence had lesser throughput). But it's more difficult to install them with a good decoupling, because there is not much room for a freely swinging construction.

I found, that the main problem is a bad decoupling from the SSD mount, which then transfers the vibration noise to the bottom cover. The second problem is bad decoupling to the LCD display, which also amplifies the vibration noise. With the 80mm fans it was difficult to avoid both.

The noise through the bottom cover can be reduced by dampening the cover (e.g. by slightly lifting it in the middle by some foam pieces).

I think the dampening of the bottom cover should get a bigger focus.
Perhaps we should use some kind of sound dampening mat?

[Davidb]

The noise through the bottom cover can be reduced by dampening the cover (e.g. by slightly lifting it in the middle by some foam pieces).

I think the dampening of the bottom cover should get a bigger focus.
Perhaps we should use some kind of sound dampening mat?

Yes, well thought, probably you are right here as well.

[aribo61]

I have some new measurements:

The Noiseblocker directly controlled by the motherboard was too loud for my taste.
So I tried to insert the Zalman fan mate 2 on maximum speed (which is much slower than without it).

With my worst case test I got a stable temperature after about 6 hours of about 64°C.

After that I switched off any sound and waited again for a stable temperature. After >6hours I had 58.9 °C.
To be sure I switched the Kronos on today (so again without any sound) and got 58.7 °C after 6 hours.

All these temperatures were measured with the external sensor (which I always leave in place since mounting it to the motherboard cooler).

Currently I am running another test, with a new cooling configuration:

I inserted a rubber mount to each hole of my Silenx ixp-34-12 (so I used 8 rubber mounts) and sticked it to the motherboard cooler (inserting the rubber mounts between the fins of the cooler), the other four rubber mounts are used as spacers to the bottom cover. So the fan is jammed(?) between the cooler and the bottom cover.

As a second fan I used my Artic F8 (80mm) and lashed(?) it to the PSU with cable ties (this time I used ball cable ties, which can be reused).
This seemed to be nice but after installing the bottom cover it sounded suboptimal (again!).

But I am currently mostly interested in the resulting temperatures.
I hope, this construction (which btw. is rather standard for CPU coolers etc.) will give much better results, as it directly sucks hopefully cool air from the bottom of the case and pumps it to the cooler.

As I already said before, the Silenx has very very low throughput, I often can't feel it with my hand, then I use my face to detect air flow. But it is very silent, even when directly mounted to the cooler. And I hope the throughput will be enough for this usage.
The F8 has a very good throughput and so I reduced it's speed with the Zalman fan mate 2, this time at minimum.

Unfortunately the F8 is not well decoupled now, this has to be improved.
Also, the fan could be slower for this purpose. Perhaps I can use my Xilence DualWing XQ fan, which is slower.

As the warm air isn't transferred to the cool side of the case, I assume that the right side will warm up. But I hope it won't be too much, and the air on the bottom will also keep cool enough.

I think a very long test is mandatory to detect problems. Perhaps I need to run the test during the night.

[aribo61]

the test was more or less successful...

after 7 hours I had a stable temperature of 55.8°C at the bottom of the motherboard cooler, which is a very good value.

The right top of the case was warm, but not too much.

I also found a reason for having surprising noise increases with an installed bottom cover:

I always thought the bottom cover would be in parallel to the electronic boards, but of course it is not!
The LCD is in the same plane as the boards, and as everyone can easily see, the LCD is not in parallel to the bottom cover. I just don't get why I didn't notice this.

When mounting a fan to the motherboard cooler or to the PSU and also when mounting an 80mm fan similar to the stock fan, they tend to get in touch with the bottom cover, simply because they have less space near the keyboard area than near the back area.

So I needed to take this into account.
For the Silenx, I replaced two rubber mounts against some adhesive foam pads.
For the fan mounted directly to the PSU, there isn't enough space. So I simply placed it more to the back.
This worked.

One problem remains: two fans at different speeds have interfering frequencies, which creates another kind of annoying tonal noise.