[cubalis2]

Like many here I use the popular inuke amps for cheap sub and transducer power.

I am currently using a pair of nu6000DSP's, and a single nu1000DSP. I have done fan mods on each of them, as they are quite annoying with the stock fans.

I decided to do some 'empirical' testing on the effect of reversing the fans and how it impacted internal temperatures and stability.

First of all - Although I tested the nu1000DSP, the results were not significant for either direction of fan so it is not included. I am using a dead silent Arctic cooling F8 like this and it has been rock solid. This is powering 2 crowson transducers in bridged mode, and runs pretty close to clipping with the right material (ULF heavy movies).


As for the nu6000DSP's - I was using a mixture of 1 Scythe Flow2 (40CFM - discontinued) and 1 Noctua NF-R8 in each amp. The Scythe's are noticeably louder as they blow more air, and I was getting thermal shutdowns with just the Noctua's installed in a single amp. The combination of both fans has been solid for over a year now. Although not completely quiet, it was much better vs. stock.

Each channel of each amp is powering a sealed LMS-R 15".

I installed both Scythe fans in one 6000DSP, and both Noctua fans in the other for a test that would best represent common fan modifications. I fully expected to run into some power cycling with the Noctua cooled amp, as I had before (and wasn't wrong - but more on that later).


For the test I did 3 conditions with both directions of airflow - fans blowing front to back, and back to front:

1) Amps out of the rack in 'open air' idling with no signal for 30 min

2) Amps in the rack, stacked on top of each other normally, idling for 30 min. In my rack the amps are at the bottom.

3) Amps in the rack, 13 minutes of very heavy bass music played to just below clipping (the song was a compilation from a sub meetup a few years ago). I measured peaks of 128dB at the listening position.


I tested the temperature in 6 locations per amp after each test using a Raytech minitemp laser IR device:

1-4 I believe these are the backs of the transistors, where they are attached to the PCB. These were easy spots to hit accurately.
6 was the hottest spot I could find on the board.

- - - If anyone has a better idea of where to probe for temperatures I could do this again quite easily - - -

I wasn't looking to get perfect readings of the absolute temperature of certain components; I just wanted something to repeatedly compare the two fan directions.

Here are the results:

First - the difference between the Scythe and Noctua fans is evident right away, especially in the number 6 spot.

100°c is pretty hot for any piece of electronics, no wonder the Noctua cooling resulted in thermal cycling.

Spots 1-4 seem to be dependent on distance to the cool air coming in. 1 is best cooled back to front as it's nearer the fan, 4 is the reverse of that.

*No thermal/power cycling happened in this test.

In the rack idling is more of the same, just elevated temperatures in general.

*The Noctua cooled amp cycled twice during the 30 minutes - BUT only in the Back to Front fan position. This is Behringer's default direction.

This was a fun test to run twice. I made sure I beat on them pretty good, constantly lighting up the third signal light and flickering the clip signal at the heaviest times. I cut the back to front test short as the Noctua amp was cycling every 30 seconds or so.

I was actually pretty surprised how similar the temps are to the idle conditions, which tells me my probe spots weren't the best.

*As mentioned the Noctua amp cycled 5 times until I cut the test short, BUT only in the back to front position again.


Concluding remarks -
- I would like to run this again some time with better spots to record more meaningful and significant temps.
- My ambient temperature was quite low today, 17-18°c in the basement.
- 120°c seems pretty damn hot.
- I measured ~63dB with the Scythe fans, and ~55dB with the Noctua, although the Scythe has a metallic noise that is annoying.
- I am going to order another pair of Noctua's now that I found reversing them stops the cycling. I have been running both amps with some decent volume music for the last 3 hrs and no cycling (I kept both Noctua's in one amp now).


TL/DR:
Reverse your Fans in the 6000DSP! Pull the cool air in from the front, instead of pushing warm air from the rest of your equipment through them.

 

[vitod]

Interesting! I'll try your method of pulling air in.

 

[vitod]

Interesting! I'll try your method of reversing the fans.

 

[G-Money1572]

Very interesting. I don't pretend to understand amps, but it would be nice to figure out what component's temperature the software is monitoring when it decides to thermal out. If those components are on the front of the board, front to back makes sense based on your results. If they are on the back, they might run a little cooler using the stock configuration.

 

[BrutalBodyShots]

Good point above.

Thanks for taking the time to do this test with the fans.

You would think that Behringer would have taken the time during the R&D/testing process to do a similar test and based on their results design accordingly. I suppose with them assuming buyers would stick with the noisy factory fans that move a lot of air it really wouldn't matter which direction they blow, but when doing the fan mod which results in less air movement perhaps direction becomes more important.

 

[Cryptic1911]

Well, in behringer's defense, they didn't design these to be used at home.. they are supposed to be in a noisy environment, in a rack, so they put the best performing fans in that they could, and didn't care about the db levels

 

[G-Money1572]

I'll agree that fan noise likely didn't even play into the engineer's design considerations. In their world, "best performing" is code for cheapest fans that met minimum specs. I'd assume that would be cfm, static pressure, hours of expected life, and possibly others? They probably didn't care about ball vs. sleeve bearings, fan blade geometry, motor struts and other factors that can make identically performing fans run much quieter.

The fact that they have the fans blowing forward would lead me to believe that their analysis indicated the most critical parts for cooling are at the back of the case. So they'd want to hit them with cool outside air first, before it had traveled through the case and warmed up. But I'm no engineer, just someone who has built enough computers to have a very rough grasp on the concepts of cooling computer components.

One thing I've always wondered with these. Evidently, the case is the heatsink based on what I've read. Somewhere I read that the bottom gets warm on heavy usage. I wonder if providing a larger air gap under the amp might help?

 

[cubalis2]

This is taken directly from Behringer's user manual of the inukes:

"Make sure enough cool air reaches the rack, especially when other rack
equipment emanates a lot of heat. The iNUKE amplifiers circulate heat from the
rear to the front vents to relieve heat inside the rack enclosure.
Fan speed adjusts automatically to assure safe operation. Never block ventilation
openings. Should internal temperature reach extreme values, the unit will shut
down automatically"

Pg. 17,

PDF

I also know from researching the fans, it seems to be a minor selling point that the fans blow back to front. I don't have the sources saved, but a few product info blurbs on random sites listed things like 'Back-to-front cooling prevents thermal buildup' and such. It makes me question if it was an engineering move or a marketing move... with enough static pressure airflow it wouldn't really make a difference which way the fan is blowing (such as the stock fans) as these units have so much empty room inside them for air to circulate from either side.

 

[audio0947]

They are made to be mounted in a rack and as such it usually makes more sense to blow hot air out the front to the outside of the rack which is almost always a large open airspace rather than have it build up inside the rear of the rack quickly raising temps for all the equipment located within. In a home environment this may or may not make sense depending on where the amp is located.

 

[audio0947]

Now that I think about it, my ep4000 has a filter on the front of the unit so it would make more sense to pull from the front through the filter to keep the interior and my fan blades cleaner, gonna reverse it this evening.

 

[asarose247]

thanks for taking the time and adding some scientific/investigative method to helping some us understand the cooling problems/solutions
I have 2 6000's mooded with the Noctua's
After reading about other threads/ success wrt to cooling an Inuke3000 I bought the tiny heat sinks for internal placement ( .5 x .5 x .2 size)
only to eventually learn that the 6000 uses the bottom plate as its primary sink and no particulars wrt to any recommended internal placement options . .
Of course using the detector to test for / find hot spots on the bottom could be useful but difficult as it introduces "not real world" positioning (I can't picture how, right now)
so
I see a mod for taller feet is a possibility and reversing factory fan direction and putting the sinks sorta under the base wrt to your photo of test spots
yet
Using your heat detector data and the picture of your 6 test spots. IF you had the opportunity to place some heats sinks internally,
Would we think that the backs of those transistors could be the first likely targets, the sinks are small and have the thermal adhesive stuff
Is there any evidence wrt circuit board connections/configuration or whatever that advises against that, electrically speaking (ELECTRICITY!?!-I know how it works . . just pay the bill!)
suggestions/pictures appreciated
this could be the thread breaking the "Inuke6000 cooling code" , not discounting the numerous other external fan mods
thanks

 

[G-Money1572]

FWIW - About 3/4 way down the second page of this thread, a guy took some FLIR pictures of his inuke6000. He found the fan control circuit and FET driver chips as the hot spots, although that was under stock cooling.

 

[asarose247]

^ checking that link, thank you
in the pic, the twin pairs at the middle left side? could a sink be placed on top "those" and would that "help"?
I can see sticking a pair of sinks on the OP test spots 1 thru 4 as the transistors are underneath black screw assembly, an attachment point (?)


and some on the bottom plate spread out around those areas
just asking

 

[cubalis2]

FWIW - About 3/4 way down the second page of this thread, a guy took some FLIR pictures of his inuke6000. He found the fan control circuit and FET driver chips as the hot spots, although that was under stock cooling.

That would have helped me yesterday

Interesting ir pics for sure, now I'm curious about re-running some of the tests with these spots in mind.

Anyone know where exactly to find these spots?


Fan control circuit with dropper resistors:

Output inductors:

Close up of FET driver chip:

Here is a high resolution image of my nu6000DSP board:

 

[G-Money1572]

That would have helped me yesterday

A day late, and a dollar short... The story of my life. I hope this thread leads somewhere good, cause I'd love to buy an iNuke6000dsp - but what seems like a decently high failure rate on fan mods has scared me off. I'm not sure what the actual percentage is, but you hear more about these mods failing than the 3000dsp. I also have an exhaust fan in my entertainment center right behind where the iNuke would go, so with the fan blowing forward I'm afraid I'd have two fans fighting each other. So I think I really need to do the fan mod, with fans reversed...

I'm guessing with locations here, but I think the output transductors are the four coils on the right of your picture. I suspect that the fan control might be the two white boxes in the top middle above the large yellowish box. The driver chips...well my best wild guess is that those are over in the right 1/4 of the picture directly to the right (on the newest picture you supplied) of where you took the temperature readings 1-4. They are surrounded by a lot of smaller things. That's only a guess though.

 

[Augerhandle]

...anyone know where exactly to find these spots?...

 

[cubalis2]

...anyone know where exactly to find these spots?...

Thanks!

I'll try the in-rack tests again using these spots in the near future and update with results. I'm mostly curious to see if the delta is larger from push to pull fan direction at these more critical spots.

 

[asarose247]

the small feet are glued on, into the indents at the bottom for about 5mm clearance
the back ones are under circuit boars so that's the end of dismantling any further
what if . . .?
simply direct if rudimentary . .
2 pieces of 3/4 ply , cut to be trapezoidal,
say 1 1/2 base and 1/2 across the top, slightly shorter than the depth of the amp
painted black (for less show)
place judiciously on shelf, under amp
space under the base now 3/4" or about 18 ish mm
just part (?) of the answer
would using some metal fabrication then become part of the base increasing "heatsink" dissipation capacity


looking forward to some more analysis, the IR stuff will be the most telling where cooling is needed/most effective

 

[asarose247]

this also in . . . from left field


in looking to draw more heat from the base in a 6K
suppose, theoretical and passive cooling . . .
short pieces 8" or so of drywall corner bead
arranged with the "spread" down, the corner contacting the base and conduction spreading any heat into the rest of the piece
given the standard dimensions, each 1" length of a piece will provide about 5 sq. in. potential heat dissipating cooling surface
and raise the amp about 1" or so off the shelf
i.e. the 8" piece has about 40 sq in. of metal surface area, counting both side,


back to the drawing board

 

[NODES]

you can sink those hot spots for dirt cheap. $10 will get you 40 heat-sink on the bay