Author Topic: DC load using a CPU cooler  (Read 57237 times)

0 Members and 1 Guest are viewing this topic.

Offline Mr.B

  • Supporter
  • ****
  • Posts: 1249
  • Country: nz
Re: DC load using a CPU cooler
« Reply #150 on: October 13, 2014, 06:55:35 pm »
I am going to use IXTH30N50L2.
They are a couple of dollars dearer than IXTH75N10L2, but at least they are in stock at Mouser.
Where are we going, and why are we in a handbasket?
 

Offline microbugTopic starter

  • Frequent Contributor
  • **
  • Posts: 568
  • Country: gb
  • Electronics Enthusiast
Re: DC load using a CPU cooler
« Reply #151 on: October 13, 2014, 07:25:21 pm »
Yes, those would be OK. I'll have to do some searching to see if there are some available online which have similarly good specs to the IXTH75N10L2s (low RDS(on)).
 

Offline Mr.B

  • Supporter
  • ****
  • Posts: 1249
  • Country: nz
Re: DC load using a CPU cooler
« Reply #152 on: October 13, 2014, 07:34:27 pm »
The IXTH75N10L2 has about 20mOhms.
The IXTH30N50L2 has about 215mOhms.
In the bigger scheme of things, is it going to make a lot of difference?
Where are we going, and why are we in a handbasket?
 

Offline microbugTopic starter

  • Frequent Contributor
  • **
  • Posts: 568
  • Country: gb
  • Electronics Enthusiast
Re: DC load using a CPU cooler
« Reply #153 on: October 13, 2014, 07:35:18 pm »
Not really, but if they're around the same price I might as well.
 

Offline Mr.B

  • Supporter
  • ****
  • Posts: 1249
  • Country: nz
Re: DC load using a CPU cooler
« Reply #154 on: October 13, 2014, 08:57:35 pm »
For me they are about NZD 3 each more expensive.
I ordered 16 from Mouser late last week.
Arrived this morning.
BIG DC load.  >:D

Where are we going, and why are we in a handbasket?
 

Offline rob77

  • Super Contributor
  • ***
  • Posts: 2087
  • Country: sk
Re: DC load using a CPU cooler
« Reply #155 on: October 13, 2014, 10:04:20 pm »
The IXTH75N10L2 has about 20mOhms.
The IXTH30N50L2 has about 215mOhms.
In the bigger scheme of things, is it going to make a lot of difference?

Rdson is unimportant unless you're planning to build a dummy load capable of tens of amperes at 1 V or less ;)
 

Offline Mr.B

  • Supporter
  • ****
  • Posts: 1249
  • Country: nz
Re: DC load using a CPU cooler
« Reply #156 on: October 13, 2014, 10:11:50 pm »
Exactly rob77.
In my design I am not wanting to do anything under about 3v3 anyway.
Where are we going, and why are we in a handbasket?
 

Offline Mr.B

  • Supporter
  • ****
  • Posts: 1249
  • Country: nz
Re: DC load using a CPU cooler
« Reply #157 on: October 14, 2014, 01:04:57 am »
@AcHmed99 - Yes, they have exceptional thermal characteristics.
However, I have already spent NZD 250 on the IXTH30N50L2.

On the subject of thermal...
The IXTH30N50L2 has a Tjmax of 150 deg C.
Datasheet attached.
Expert experience question: How hard can you safely push it?
140 deg C?
6.66% de-rated?
Where are we going, and why are we in a handbasket?
 

Offline rob77

  • Super Contributor
  • ***
  • Posts: 2087
  • Country: sk
Re: DC load using a CPU cooler
« Reply #158 on: October 14, 2014, 06:36:02 am »
@AcHmed99 - Yes, they have exceptional thermal characteristics.
However, I have already spent NZD 250 on the IXTH30N50L2.

On the subject of thermal...
The IXTH30N50L2 has a Tjmax of 150 deg C.
Datasheet attached.
Expert experience question: How hard can you safely push it?
140 deg C?
6.66% de-rated?

check the datasheet for thermal de-rating factor - it might be a surprisingly high number (a 100W mosfet can easilly become 40W max at high temperatures)
 

Online Jeroen3

  • Super Contributor
  • ***
  • Posts: 4209
  • Country: nl
  • Embedded Engineer
    • jeroen3.nl
Re: DC load using a CPU cooler
« Reply #159 on: October 14, 2014, 10:09:17 am »
I was looking at the linearity of the Id/Vgs curve. It's not ideal if the fet goes from 0 to Imax in a window less than 1 Volts.
The IXTP or IXTH15N50L2 seems nice for this purpose, you have less risk of instability when you're running milliamps.
And it's affordable on digikey and an ebay seller.

I'll look into the ones named above tonight. (EU time)

 

Offline microbugTopic starter

  • Frequent Contributor
  • **
  • Posts: 568
  • Country: gb
  • Electronics Enthusiast
Re: DC load using a CPU cooler
« Reply #160 on: October 14, 2014, 02:36:46 pm »
I got a reply from Delta Components in the UK. They have a sister company, Neutron LLP, that will provide me with two IXTH75N10L2 samples free of charge :)
 

Offline Mr.B

  • Supporter
  • ****
  • Posts: 1249
  • Country: nz
Re: DC load using a CPU cooler
« Reply #161 on: October 14, 2014, 07:28:31 pm »
check the datasheet for thermal de-rating factor - it might be a surprisingly high number (a 100W mosfet can easilly become 40W max at high temperatures)

Checking the datasheet, I cannot see anything directly called thermal de-rating factor - unless I am overlooking something.
The IXTH30N50L2 is quoted as a 400W device (when Tc = 25 deg C).
If I take the quoted SOA: 200W @ Tc = 75 deg C.
RthJC is 0.31 deg C/W.
200 * 0.31 = 62 deg.
Therefore: 75 + 62 = 137 deg C for the junction temperature when the case is 75 deg C and I am driving it at 200W.

I plan to drive the device at a maximum of 150W.
Assuming 137 deg C is within the SOA and working backwards...
150 * 0.31 = 46.5 deg C.
137 - 46.5 = 90.5 deg C.
To remain within SOA, I can allow the case temperature to be as high as 90.5 deg C @ 150W.

Now add to this the RthCS of 0.25 deg C/W --> 150 * 0.25 = 37.5 deg.
90.5 deg (case) - 37.5 deg = 53 deg.

If my math above is correct...
I can allow my heatsink face to get to 53 deg C if I am driving the FET at 150W.

Have I understood this correctly?
« Last Edit: October 14, 2014, 07:31:37 pm by Mr.B »
Where are we going, and why are we in a handbasket?
 

Online Jeroen3

  • Super Contributor
  • ***
  • Posts: 4209
  • Country: nl
  • Embedded Engineer
    • jeroen3.nl
Re: DC load using a CPU cooler
« Reply #162 on: October 14, 2014, 07:44:26 pm »
IXYS is very honest and they specify their SOA at Tc 75.
They do not specify a dereting factor. But this fet is designed for linear use.
If you stay inside the soa drawn in Fig. 14 you should be fine.
 

Offline Mr.B

  • Supporter
  • ****
  • Posts: 1249
  • Country: nz
Re: DC load using a CPU cooler
« Reply #163 on: October 14, 2014, 07:46:29 pm »
I was looking at the linearity of the Id/Vgs curve. It's not ideal if the fet goes from 0 to Imax in a window less than 1 Volts.
The IXTP or IXTH15N50L2 seems nice for this purpose, you have less risk of instability when you're running milliamps.
And it's affordable on digikey and an ebay seller.

I'll look into the ones named above tonight. (EU time)

Sorry Jeroen3, I dont quite follow you.
The graph for the IXTH30N50L2 looks to me to be about 3.5v for the 10A Ids range I want to run mine in.
Am I missing something here?

Where are we going, and why are we in a handbasket?
 

Offline Mr.B

  • Supporter
  • ****
  • Posts: 1249
  • Country: nz
Re: DC load using a CPU cooler
« Reply #164 on: October 14, 2014, 07:50:41 pm »
IXYS is very honest and they specify their SOA at Tc 75.
They do not specify a dereting factor. But this fet is designed for linear use.
If you stay inside the soa drawn in Fig. 14 you should be fine.

Thanks, will do.
Where are we going, and why are we in a handbasket?
 

Offline Mr.B

  • Supporter
  • ****
  • Posts: 1249
  • Country: nz
Re: DC load using a CPU cooler
« Reply #165 on: October 14, 2014, 07:52:23 pm »
Apologies @microbug.
I am not wanting it to appear that I am hijacking your thread.
It is just that we are both designing a DC load and using similar or the same components.
I figured a shared knowledge thread was better than starting another one called "yet another DC load".  :D
Where are we going, and why are we in a handbasket?
 

Offline Joenuh

  • Newbie
  • Posts: 8
Re: DC load using a CPU cooler
« Reply #166 on: October 15, 2014, 04:05:54 pm »
Just to let you guys know, I just got a reply that I will be receiving 5 samples of the IXTH75N10L2 tomorrow. I only asked once.. :-//
 

Offline microbugTopic starter

  • Frequent Contributor
  • **
  • Posts: 568
  • Country: gb
  • Electronics Enthusiast
Re: DC load using a CPU cooler
« Reply #167 on: October 15, 2014, 04:19:17 pm »
@Mr.B, no worries! That makes sense. I haven't been working on my design for the past week or two; I need to get on with the schematic.

@Joenuh I don't know what country you live in, but here in the UK the official IXYS office appears to be unreachable. The only way I could get samples (which arrived today, BTW) was through a distributor (Neutron LTD). I've only got two of them, so I'll be using cheaper ones until I have to test the power dissipation.

EDIT: Offset voltage for the AD633 (jellybean multiplier) sucks, plus it's really expensive, so I'm going to do constant-power and constant-resistance in software.
« Last Edit: October 15, 2014, 04:38:22 pm by microbug »
 

Offline macboy

  • Super Contributor
  • ***
  • Posts: 2309
  • Country: ca
Re: DC load using a CPU cooler
« Reply #168 on: October 15, 2014, 05:03:42 pm »
Offset voltage for the AD633 (jellybean multiplier) sucks, plus it's really expensive, so I'm going to do constant-power and constant-resistance in software.
Constant resistance operation does not require a multiply operation. Set the current to be proportional to terminal voltage, rather than setting it to a constant value. Constant power operation does require a multiply.
 

Offline microbugTopic starter

  • Frequent Contributor
  • **
  • Posts: 568
  • Country: gb
  • Electronics Enthusiast
Re: DC load using a CPU cooler
« Reply #169 on: October 15, 2014, 05:31:32 pm »
OK... so I could switch the external reference on the DAC to a divided-down version of the input voltage, and compare the DAC output with that same divided down voltage.
 

Offline Joenuh

  • Newbie
  • Posts: 8
Re: DC load using a CPU cooler
« Reply #170 on: October 15, 2014, 06:46:24 pm »
@Microbug I live in The Netherlands. I used the tech support (emailadress), they were happy to forward me to the Europe section of IXYS. Today I got an email from a IXYS dealer.. Anyway you still got a couple :)
 

Online Jeroen3

  • Super Contributor
  • ***
  • Posts: 4209
  • Country: nl
  • Embedded Engineer
    • jeroen3.nl
Re: DC load using a CPU cooler
« Reply #171 on: October 16, 2014, 09:27:00 am »
I was looking at the linearity of the Id/Vgs curve. It's not ideal if the fet goes from 0 to Imax in a window less than 1 Volts.
The IXTP or IXTH15N50L2 seems nice for this purpose, you have less risk of instability when you're running milliamps.
And it's affordable on digikey and an ebay seller.

I'll look into the ones named above tonight. (EU time)

Sorry Jeroen3, I dont quite follow you.
The graph for the IXTH30N50L2 looks to me to be about 3.5v for the 10A Ids range I want to run mine in.
Am I missing something here?
Precisely. With a "normal" switching optimized fet this isn't 3.5 Volts.
 

Offline microbugTopic starter

  • Frequent Contributor
  • **
  • Posts: 568
  • Country: gb
  • Electronics Enthusiast
Re: DC load using a CPU cooler
« Reply #172 on: October 16, 2014, 04:03:27 pm »
I've just ordered the Arctic Cooling Freezer i30 CPU cooler for this project. It's the same as the previously planned A30, but it's for an Intel socket and is slightly cheaper.

Would it be best to solder the two TO-247s directly to the copper heat pipes (using low-temperature bismuth-tin solder)? Solder appears to be better than thermal epoxy, and if I had the two packages opposite ways up then the heat transfer would be about the same for each heat pipe.
 

Offline mrflibble

  • Super Contributor
  • ***
  • Posts: 2051
  • Country: nl
Re: DC load using a CPU cooler
« Reply #173 on: October 16, 2014, 04:11:01 pm »
Would it be best to solder the two TO-247s directly to the copper heat pipes (using low-temperature bismuth-tin solder)?
No. What would be best is to not solder it. Best would be to use a thermal paste of your choosing. Then take measurements of thermal performance. Then remove paste, clean it up, and solder it.  ;) Then measure thermal performance again, and compare findings.
 

Offline microbugTopic starter

  • Frequent Contributor
  • **
  • Posts: 568
  • Country: gb
  • Electronics Enthusiast
Re: DC load using a CPU cooler
« Reply #174 on: October 16, 2014, 04:14:17 pm »
OK. I'll get on with the rest of the project, and do the thermal stuff last. That way, I can put exactly 300W (or whatever) into the cooler and measure its temperature.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf