Scythe Kaze master Mofsets? reaching 91+C with 2 fans attached within specs??

I have 3 fan controllers and need to condense to 2 for space, so need to run 2 fans per controller on some.

According to scythe fan controller, specs are:
Maximum Fan Ampere per Channel 1 Ampere = 12 W max.

Fans I tried running 2 of are Scythe GT 1850 (<0.1 amps per fan) and Scythe S flex F (0.2 amps so 2 is 0.4 amps).

With 1 fan, temp on ?mofsets was 50's C when running fan full speed, 70C when running fan low speed.
With 2nd fan hooked up, temp was 70C when running full speed and 90-93C (SEE PIC) when running 2 fans at low speed.

Anyone have a guess if 90C within specs of that component? Tried it with more than one temp probe and all are calibrated accurate, and it will burn your finger quickly to touch it (which is why I measured I was testing rpm readouts with 2 fans hooked up an accidentally touched it). I guess if it dies no loss, going to replace soon anyways, just as long as it doesnt smoke and smell like electric fire like my video card did when gpu seals leaked, resulting in a long conversation with wife about watercooling...

http://img51.imageshack.us/img51/2329/scythetemp.jpg

I know you are asking if 90C is within specs, but I have to say that the bigger issue is that MOSFETs running so hot with such a seemingly light load does not seem normal. It seems defective.

That was my first guess, but so far I tested 2 of the 4 on that one controller and both ran ~same temps with 2 fans hooked up, and just testing 1 on my second scythe controller, it is up to 88C just after 2 minutes. :shrug:

Maybe I need a fan to cool my fan controller, until I replace them :p:

No it's not defective - they run as hot as that :)

I will probably put on some ramsinks and use a 40 mm fan for better cooling since I don't feel comfortable with those temperatures.

Quote:

---

Originally Posted by rge

That was my first guess, but so far I tested 2 of the 4 on that one controller and both ran ~same temps with 2 fans hooked up, and just testing 1 on my second scythe controller, it is up to 88C just after 2 minutes. :shrug:

Maybe I need a fan to cool my fan controller, until I replace them :p:

---

Quote:

---

Originally Posted by Main

No it's not defective - they run as hot as that :)

I will probably put on some ramsinks and use a 40 mm fan for better cooling since I don't feel comfortable with those temperatures.

---

That's pretty messed up :shocked:

90C seems awfully hot. What about a different fan controller: the Sunbeam Rheobus Extreme is good, and has six inputs, for possibly up to 12 fans. I personally use one Sunbeam for only six fans, to be safe though.

Second topic: Watercooling is the best option for efficient cooling, however, you still have fans. Maybe fewer fans than now, but maybe more, depending on your set-up.

Actually does look like the 2 scythe F fans are about 10-13C worse than 2 gentle typhoons when tested on same channel, guess since F's have higher amp rating.

ethos: yep I am going to get a different controller. I will get the Lamptron FC-5 (has readouts which I like) if they get in stock soon, if not will get sunbeam...and I like the 6 controls on sunbeam. Or may get one of each of those...only cpu rad fans I am interested in seeing rpms readouts. I am just hoping for no electric smell burnouts until I get new ones.

Main: thanks for info

EDIT: holy crap, dont run 2 led fans on one channel, a few of my 15 fans are .3 amp rated led fans (x 2 still only .6 amps), temps were 103C-109C. I am thinking they should change the max to 0.1 amps on specs instead of 1 amp. Anything above .1 amps nets mofsets temps above 80C when fan controller rpm is turned to lower rpms.

First pic, cant see temp probe, it is measuring temp mofset on second scythe controller still attached to case, in second pic, pulled it out.
http://img237.imageshack.us/img237/2552/109led.jpg

http://img192.imageshack.us/img192/4849/103led.jpg

Quote:

---

Originally Posted by Main

No it's not defective - they run as hot as that :)

I will probably put on some ramsinks and use a 40 mm fan for better cooling since I don't feel comfortable with those temperatures.

---

what a difference heatsinks make...this is the sony regular controller not extreme (both it and extreme have similar large heatsinks on mofsets and this is without any fan), max temps are ~45C no matter where I measure, and this is with same 2 led fans at same ~rpms that was at 109C on mofsets on scythe that does not have heatsinks.
http://img693.imageshack.us/img693/2488/sonytemps.jpg

YES the mosfets on the kaze is HOT as hell.

I actually burn't my finger trying to get a fan header back in once, and learned never to try that again.

The kaze master is SUCH CRAP CRAP CRAP!!!!

Thank god i have a lamptron on its way as a replacement.

MOSFETS can take this kind of heat, it's within their usual specs. Nothing wrong here, but still it's not far from the limit (which is around 120C usually if i remember)

i have kaze server and i run 2 fans on one channel that are below specs and the mos gets really hot like burn my finger hot :shocked:... figured its fine lol.. :confused:

EDIT : and i use the Triebwerk TK-122 fans

Quote:

---

Originally Posted by NaeKuh

YES the mosfets on the kaze is HOT as hell.

I actually burn't my finger trying to get a fan header back in once, and learned never to try that again.

The kaze master is SUCH CRAP CRAP CRAP!!!!

Thank god i have a lamptron on its way as a replacement.

---

crap my aZZ - Kazemaster Server is the only real fancontroller available, the others are just fanspeed adjusters ( that goes for the Kazemaster ACE for that matter)

Furthermore we have Kazemaster Server installed in very hot enviroments - standalone colocation servers - I have never measured the temperature of the mosfets - but I know they most be glowing hot.

Quote:

---

Originally Posted by NKrader

i have kaze server and i run 2 fans on one channel that are below specs and the mos gets really hot like burn my finger hot :shocked:... figured its fine lol.. :confused:

EDIT : and i use the Triebwerk TK-122 fans

---

yeah you will be fine - but it doesn't hurt to take measures against it of course. I run 3x140mm LianLi fans on one channel - no problem.

If the kazemasters started to die we would have seen it by now - they have been in the market quite a while.

Quote:

---

Originally Posted by Main

crap my aZZ - Kazemaster Server is the only real fancontroller available, the others are just fanspeed adjusters ( that goes for the Kazemaster ACE for that matter)

---

Thank goodness you're here. We really do need someone who absolutely knows everything about anything anyone posts about.

It's probably a good idea to put the probe on the body of the part, and not the metal. The metal is the collector of the transistor, and it could mess up your readings. Most heatsinks that are mounted to transistors use thermal grease that is either non-conductive, or use a non-conductive insert between the metal tab and heatsink.

Here's the data sheet of the transistors used in my old scythe fan controller http://pdf1.alldatasheet.com/datashe...NICS/D882.html

Tj (die temp) is rated for up to 150C

This might explain why my perfectly in spec fans managed to kill one of the channels on mine. They really need to revise these things.

Quote:

---

Originally Posted by Humminn55

Thank goodness you're here. We really do need someone who absolutely knows everything about anything anyone posts about.

---

:rofl:

i use mine for 3k ultra kazes and deltas.

you just need a heatsink on that bad boy haha!

http://img.techpowerup.org/100124/IMG_0038158.jpg

Yeah they are crap, had one myself and returned it within a day. With only a few fans hooked up it got so hot that it needed it own cooling, which is a bit retarded if you ask me.

I remember going to scythe forums to see if it was just me having the problem and I can remember the first page being filled with posts like: Kaze Master Burning?! :rofl:

Paid a bit extra for a TMS200 and never regretted it.

Just have checked my Kaze temps out. 40C

Quote:

---

Originally Posted by WaterFlex

Just have checked my Kaze temps out. 40C

---

Is that temp reading with fan speed down turned? If the controller is not reducing the speed significantly, ie fan on full blast or near full blast, the temps wont be bad, mine were ~50'sC for scythe F at or near full rpms. The temp increases rapidly and progressively as rpms are turned down, especially in 800-900 rpm range. But interesting if your temps are much better.

@fitseries3, nice job on heatsink, how did you attach with thermally conductive adhesive?

edit: @diverge, thanks for thermal specs. The black mold is roughly 5C cooler than metal (see pic), here metal was 110C, mold 105C. btw, all measurements are made with thermocouple coated with thermally conductive, electrically nonconductive thermal grease in which this thermocouple was calibrated for surface readings. Back in realtemp thread when milling holes in IHS for cpu IHS readings, and then reading how intel did it, then using calibration plates to check + true surface probe, thermocouples wont accurately read surface temps unless coated in thermal grease, since they will average air temp + surface temp, unless a true flat temp probe like wife had access to at her lab, which is also where got calibration plates. And even mx2 coating the thermocouple tip will read within about 1C of thermal grease, so works in a pinch.

http://img513.imageshack.us/img513/3857/105cmold.jpg

The reason they get hotter when the fan is speed is turned down is cause the voltage is dropped across the transistor. So if your fans are running at lets say 7V, the other 5V is pretty much being converted to heat by the transistor. When it is at fullspeed, the only drop across the transistor is Vcesat, which is the saturation - when the transistor is fully on.

Sigh.. So I might as well throw away my kaze server even though I havent even used it yet... Gonna have 7 GT 1450rpm's so... Any suggestions for some good controller? Would prefer speed control based on temperature... So aquaero?

i used 2 small zip ties to affix the HS to the mosfets. use whichever TIM you would like.

Quote:

---

Originally Posted by Elpy

Sigh.. So I might as well throw away my kaze server even though I havent even used it yet... Gonna have 7 GT 1450rpm's so... Any suggestions for some good controller? Would prefer speed control based on temperature... So aquaero?

---

GT's only draw a fraction of current that a normal fan does, you should be fine.

edit: they only draw 0.049A, typical 120mm fan draws like 0.3A. you could put them all on one channel and still only draw 0.343A... a little more then typical fan.

Quote:

---

Originally Posted by Diverge

GT's only draw a fraction of current that a normal fan does, you should be fine.

edit: they only draw 0.049A, typical 120mm fan draws like 0.3A. you could put them all on one channel and still only draw 0.343A... a little more then typical fan.

---

Thats what the specs GT says. But I don't believe it. Time to make a special cable and check amps myself.............

Grr my multimeter fuse is blown from another past experiment, more tomorrow once the fuse is replaced.........

Quote:

---

Originally Posted by Diverge

GT's only draw a fraction of current that a normal fan does, you should be fine.

edit: they only draw 0.049A, typical 120mm fan draws like 0.3A. you could put them all on one channel and still only draw 0.343A... a little more then typical fan.

---

Thats the normal current. However just like other fans, they have a higher starting current which is somewhere around same numbers as that 0.3A. Depending of which GT ofc. For some reasons some sites just write the running current of the fan and some write the starting current of the fan... :shrug:

And its pretty easy to figure out since if those currents would really be like 0.049A... You could run like a dozen of them with any controller. But anyway... They use just as much power as normal fans.

Quote:

---

Originally Posted by Elpy

Thats the normal current. However just like other fans, they have a higher starting current which is somewhere around same numbers as that 0.3A. Depending of which GT ofc. For some reasons some sites just write the running current of the fan and some write the starting current of the fan... :shrug:

And its pretty easy to figure out since if those currents would really be like 0.049A... You could run like a dozen of them with any controller. But anyway... They use just as much power as normal fans.

---

GT's are the leader of the pack when it comes to efficiency. Just check out this chart.

http://www.xbitlabs.com/images/coole...agr_pw_big.png

Here's the whole review of many different fans: http://www.xbitlabs.com/articles/coo...s-roundup.html

And here is data taken for all the different GT's, including their peak current:
http://www.xbitlabs.com/articles/coo...oundup_18.html

Check out the peak current (which would include startup) for some other fans :p:

Quote:

---

Originally Posted by Conumdrum

Thats what the specs GT says. But I don't believe it. Time to make a special cable and check amps myself.............

Grr my multimeter fuse is blown from another past experiment, more tomorrow once the fuse is replaced.........

---

Go for it. I'd check mine with my multimeter, but they are all soldered in parallel, heat shrinked, zip tied, and in my case.

These aren't your typical computer fan, with all marketing gimmicks, ect. They are pretty much precision made industrial fans made by a fan manufacturer ( www.nidec-servo.com ). They are just marketed by scythe.

:O. I've been lied to. read somewhere about their starting current being higher. Oh well thanks! I wish I would had found that chart earlier. Although even so... Just put my kaze server for sale :rolleyes:. I want something more reliable. I've heard of them being bad few times before but after seeing this.. Yeah dont want :S

Quote:

---

Originally Posted by Elpy

:O. I've been lied to. read somewhere about their starting current being higher. Oh well thanks! I wish I would had found that chart earlier. Although even so... Just put my kaze server for sale :rolleyes:. I want something more reliable. I've heard of them being bad few times before but after seeing this.. Yeah dont want :S

---

AFAIK, There have been no reported cases of the server version dying. They have all been the original and ace versions.

If there has been reported deaths of "server's", can someone please point me to a link or two?


Also, on the topic of the GT's, Nidec-Servo's own PDF shows a 83mA "rating" but it also shows a 360mA "starting" for the 1850.

Hmm... Well then I guess I'll see how it works and if it gets toasted, then buy something else.. did some quick googling and apparently the server versio has some heatsinks unlike the other versions.

the 2 led fans (0.3 amps per fan) at 800 rpms on Scythe mosfets that were 105-110C, on sony only 45C, and on Lamptron 35C (just got FC-5 today after ordering last 2 from frozen cpu saturday). Even hooking 3 of those LED fans, only 40C on lamptron reading multiple places. After I bought it, was worried lamptron FC-5 was going to have same issue with temps, since has small copper heatsinks?, but obviously not an issue. I think I will be sticking with it or sony, though like lamptron better since has rpm and voltage readout.

http://img192.imageshack.us/img192/6371/lamptron35c.jpg

*Grrr* I meant to ask this in my last post but forgot. :mad:

Is there something that I can pick up at my local Radio Shack that will replace the D882 for more "reliability" or are heat sinks the best option?

http://www.radioshack.com/family/ind...goryId=2032279

Well, as you can tell from this datasheet, the D882 is rated for 3A, so I doubt it's unreliable. I'd suggest getting (or making) some heatsinks.
Now what I am wondering, if the D882 is rated for 3A, what component is bottlenecking the fancontroller, not being able to handle more than 1A per channel?

Quote:

---

Originally Posted by Waterlogged

*Grrr* I meant to ask this in my last post but forgot. :mad:

Is there something that I can pick up at my local Radio Shack that will replace the D882 for more "reliability" or are heat sinks the best option?

http://www.radioshack.com/family/ind...goryId=2032279

---

Find out what the transistors used in rge's pic above are. They are in a TO-220 package, which should handle more current then TO-126. Look up the datasheet of the 2 parts and compare them to see if they are similiar.

If I recall correctly though, the D882's are lined up side to side, in very tight area... so the larger transistors wouldn't fit there. you'd have to remotely mount them to a heatsink, and use decent gauge wire. Might as well just buy a new fan controller though.

Pretty much the reason I don't recommend these controllers. No heat sinks at all = doom. There is at least one person here that had an experience with the controller burning up and wrecking havoc on her loop. I'm not gonna mention names though ;)

Just mount a heat sink to it and you'll be all set. Old video card and NB heatsinks can be easily cut and used for this. :up:

You should just get a real fan controller and quit toying with the kids stuff. It will run every fan you have and them some. Mine is setup to run 1 outta my 6 fans until the digital sensors on the radiator go above 33C, then they will start to kick on as needed to keep the computer cool and quiet. http://www.sidewindercomputers.com/tbalancerbigng.html

Quote:

---

Originally Posted by Alexandr0s

. . .
Now what I am wondering, if the D882 is rated for 3A, what component is bottlenecking the fancontroller, not being able to handle more than 1A per channel?

---

One of my main suspects is the rotary pot. It just doesn't strike me as a well built or designed.

56K Warning!
http://img706.imageshack.us/img706/8...tarypot.th.jpg

Quote:

---

Originally Posted by Elpy

Hmm... Well then I guess I'll see how it works and if it gets toasted, then buy something else.. did some quick googling and apparently the server versio has some heatsinks unlike the other versions.

---

You are correct there. The server version has had heatsinks added to it. Also, I'm running 6x Scythe GT's 1 channel, 3x S-FLEX's and another channel, and some random case fans on the other channels. All running at lower than standard RPM's, with ambient temps of 30 to 40 degrees C during the day (australian summer with no aircon). This machine has been running 24/7 for at least a couple months now with now problems.

Conclusion, I think you will be fine with the server version. :)

Quote:

---

Originally Posted by Waterlogged

One of my main suspects is the rotary pot. It just doesn't strike me as a well built or designed.

---

I could be wrong here, but I don think the rotary pots are actually passing the current that goes to the fan. I'm pretty sure they just send a control voltage to the mosfets. I will be adding a few things to my loop in a couple days do I will have a closer look while everythings apart.

darn, just ordered one. thankfully i have a ton of ram sinks lying around :p

Quote:

---

Originally Posted by R3alsp33dy0ne

I could be wrong here, but I don think the rotary pots are actually passing the current that goes to the fan. I'm pretty sure they just send a control voltage to the mosfets. I will be adding a few things to my loop in a couple days do I will have a closer look while everythings apart.

---

There was quite a rash of ppl that lost the use of the pots (as well as the display) but the fans kept running at full speed. This was one of the things that gave me the idea that it was possible that I could modify the Kaze for use with thermo controllers which I pointed out in my video

i would think a sink is the best option.

In almost everything that runs current though them, if you can keep the temps in check, it has better reliability.

However i will NEVER EVER EVER BUY A KAZE ANYTHING.

ITS CRAP CRAP CRAP.. i dont care what you guys say about the server, i still think its CRAP.

If a 13 dollar rheobus can use big alu sinks.. why the hell cant a 40 dollar kaze controller use the same?
It tells me the company still has a lot to learn in fan controller department.

Their design and idea is great, however the implementation and correct use of parts could be better.

Would it be possible to replace the potmeters with ones that are more reliable/can handle more amps? If so, what specifications would they need?

Quote:

---

Originally Posted by NaeKuh

It tells me the company still has a lot to learn in fan controller department.

---

Worse, they could learn from their own heatsink department since they make some of the best stuff out there. Case of left arm ignoring what right arm does.

Quote:

---

Originally Posted by Alexandr0s

Would it be possible to replace the potmeters with ones that are more reliable/can handle more amps

---

The pots only control the grid/base of the FET so they dont see any current go through, it's the transistor itself that overheats. Just glue heatsinks on them.

Quote:

---

Originally Posted by Alexandr0s

Their design and idea is great, however the implementation and correct use of parts could be better.

Would it be possible to replace the potmeters with ones that are more reliable/can handle more amps? If so, what specifications would they need?

---

The pots aren't usually part of the output stage. They control input to a much smaller transistor, which drives the output transistor.

Edit: woops Gmat beat me to it

Quote:

---

Originally Posted by gmat

Worse, they could learn from their own heatsink department since they make some of the best stuff out there. Case of left arm ignoring what right arm does.



The pots only control the grid/base of the FET so they dont see any current go through, it's the transistor itself that overheats. Just glue heatsinks on them.

---

Quote:

---

Originally Posted by Diverge

The pots aren't usually part of the output stage. They control input to a much smaller transistor, which drives the output transistor.

Edit: woops Gmat beat me to it

---

question still not answered.. could somone solder a better higher amp one onto the board to replace current ones?

They get that hot because they're used in a rather crude manner (directly dropping the voltage from a +12v reference) - a multi stage solution would reduce the heat impact on each transistor. One could browse electronics retailers to see if they carry higher rated MOSFETs but in any case with such a brutal, single-stage circuit, wou'd have to use proper heatsinks anyway. The ones on Kaze controllers arent bad by themselves, they just need proper cooling.