15k Seagate & Cooling

[jen4950]

I'm in the process of getting everything pieced together- and will be using the 73GB 15000rpm U320 rev.3 Seagate drive for my C:- I was wandering if anyone sees any problem with placing a thermisistor directly on the drive 'spindle' as shown below. I'm concerned with interference caused by the close proximity of the sensor wire to the IC's on the PCB.

Hi Res: http://neubauer.tv/images/Computer/P4_2/PICT7869.JPG

Hi Res: http://neubauer.tv/images/Computer/P4_2/PICT7876.JPG

Let me know if you have any experience with a similar setup...

Thanks!

(PS- I've got a black lab if you even think of making a nasty comment about the hair.. )

 

[Snugglebear]

That's really overdone. SCSI units are built to be run in cramped cases with little airflow. Cooling like that is excessive.

 

[Ice Czar]

I was wandering if anyone sees any problem with placing a thermisistor directly on the drive 'spindle' as shown below. I'm concerned with interference caused by the close proximity of the sensor wire to the IC's on the PCB.

shouldnt be a problem

but does look like your "cooling" your thermistor

 

[jen4950]

I had a RAID 10 setup with 6 15k - 18GB drives; trust me- the cooling is neccessary- these drives get really hot.

Infact the blue fan frame is from that old setup, I just put new fans in it. I've just never actually monitored the temps and have a Digital Doc which makes it really easy to monitor this stuff.

but does look like your "cooling" your thermistor

Well, yes, it is being cooled; But if you wanna get technical, the thermisistor will be drawing heat from the hard drive- the source of the heat, and cools the harddrive. The whole laws of thermodynamics thing. So the fact that the thermisistor is there really makes no difference in cooling, but I am worried about having wires close to sensitive IC's-

 

[DeepFreeze]

Since I have snuggle AND czar in the same plae at one time, I have a problem.

remember the 15K rpm fujitsu drive and Newegg purchased Tekram DC-390U 320Scis card i bought, well it worked fine initially, and winXP ran lovely.

i went to format one night, wasnt sure if i got a virus from girlfriend.or not.

went to install winxp. hit f6, recognized enough to format the drive. even saw controller driver. when it came time to install the files. windows just causes an error.
now, it wont install windows on the controller.

drive is fujitsu mas 3184
host bus adapter, tekram dc390U4 U320 controller.
terminated at the end, drive is right before terminator.
sig stuff is accurate.

 

[Ice Czar]

LOL Its Snugglebear and UICompE02 that are the SCSI Gods not me

Id Start a thread that would catch their eye or drop them a link via PM to it
but test your memory first, try a different slot, check cables
rule out the usual obvious suspects

my point about the thermistor location is that since its exposed on the one side directly to the air flow from the fan, its accuracy is goin to be seriously compromised
and will be measure the fans air exit temperature as much as the motors temperature

Insulating the sensor would solve that, but it would also impede convection on the spindle hub, might be a good compromise if you can get a small enough "shield for just the one side of the thermistor leaving the majority of the hub exposed

 

[jen4950]

my point about the thermistor location is that since its exposed on the one side directly to the air flow from the fan, its accuracy is goin to be seriously compromised
and will be measure the fans air exit temperature as much as the motors temperature

Insulating the sensor would solve that, but it would also impede convection on the spindle hub, might be a good compromise if you can get a small enough "shield for just the one side of the thermistor leaving the majority of the hub exposed

I see your point- but.. I don't agree.

The conduction from the thermisitor being in contact with the spindle outweighs the effect of the forced convection from the fan and the solid/air interface. Case in point - the thermal conductivity of air is 0.026 W/(m.k) whereas aluminum is 237 and copper is 401. (Gold = 317, Silver = 429, Diamond = 2300). So the thermisitor/spindle interface is on the order of 10000 times more effective than the thermisistor/air or spindle/air interface.

In a nutshell- the thermisistor will be essentially the same temperature as the spindle due to their contact and lack of an air void. And the whole point is to remove heat from the system via air whether it's the sensor or harddrive case. I left out alot of thermo and diffeq but you get the idea- that's been 5 years or so ago- don't grill me on the details.


But I'm still worried about having current carrying wires near those lower power IC's-

 

[Ice Czar]

you right it wouldnt be effected by much (pretty good at thermodynamics myself )
but then I also noted that isnt a flat sensor and doesnt really have that great a contact patch from what I can tell, (that could be enhanced though?) the tape does reduce the amount of direct "in air" exposure it wouldnt have even been a consideration in my mind if it wasnt that its directly below the fan
and its hard to tell exactly how large the thermistor is so I dont have all that good of an idea to the conduction interface vs the convection interface, but assumed you want it to be as accurate as possible

looking back on my statement of "seriously" compromised
well guess that relative
been playing with a few Platinum RTDs