Cheap loadtester materials

[Thrilla]

I used one of these, 120vAC 150w heat firerod
http://www.watlow.com/products/heaters/ht_cart.cfm

I also saw they carry strip heaters, that may be easier to mount
http://www.watlow.com/products/heaters/ht_strip.cfm

[Carlz0r]

No it wouldnt.

All you would do is measure the delta T, it doesnt matter in what range.

So you're telling me that it would work in some strange way contrary to all the resistor load testers I've seen? Because they're all frosted over, and I've never seen frost at high temperatures.

[Thrilla]

They will get frosted if you use phase or LN2 or DICE, unless you insulate the heatsource well.

[jinu117]

http://www.harborfreight.com/cpi/cta...emnumber=43060

Perfect little thing... Couldn't find the link for longest time

[sson74]

Hello all, Hey Jin i'll try posting more, anyway thought I throw this in http://stores.ebay.com/Tibetwalk-Electronics This guy sales all sorts of panel meters, maybe you could use a Ammeter he sells for your load tester. I picked up some Blue Led volt meters and a PID controller for the cascade.

[SexyMF]

If you are going to hook up anything to main please add a fuse to the line.

Also, those of you who have a ghetto resistor/mains loadtester and a clamp meter please measure the current draw at ambient and then lowest temp. Post the result.

[jinu117]

Okay, here finally is guide I am going to write up.

Parts you need to acquire:



1 x Heating element. Cartridge heater. I am using 3/8" 250w high temperature one in this picture with stainless steel sleeving. McMaster.com part number 3618K236. About $20.
1 x Copper block. Using 2"x2"x3/4" on this one. I got mine from Dangerden.com as I only very recently found local wholesaler for metals that is reasonable... -_-. Somewhere from $10 to few bucks if that
1 x Thermal probe. In the picture I use K-type probe. You can use something else as you would like.
1 x Voltage variac, etc - Something to change power output. I recently found great deal on rotary speed controller which works just like variac in our situation. http://www.harborfreight.com/cpi/cta...emnumber=43060 about $10+shipping if you don't have local harborfreight store.
1 x Way to measure amperage and voltage. You could calculate of power draw by voltage * amperage. However, Kill-a-watt or equivalent tools make reading wattage and changing it respectably much easier as well as having some other cool uses such as figuring out total power draw of your system, etc more easily. (like avg power consumption). Search e-bay or froogle for Kill-a-watt. In the picture shown is seasonic powerangel which is same thing.
About $30-40 shipped.

Tools used: Drill press You can get fancy on how to mount evaporator to load block, etc. I just used really simple method on mine... you will see it later. So only required tool is drill press really. Or steady hand, lot of elbow grease and hand drill... Never tried it myself and probably won't.
Lapping glass. (any flat surfaces really).
Sand papers for lapping.

Misc: Socket mounting back from motherboard. 2 x 3 inch 6/32" machine screw, 4 washers for the screw and 2 nuts (or wing nuts)

Now to construction.

[jinu117]

Mount it securely. If you already don't have vise for drill press... go get one. Better safe than sorry. Use all manufacturer's precaution when using tools.




Since I do have 3/8 cartridge, I am drilling 3/8 hole in the copper block. Enough depth so that heating element can get all the way in...
Use some cutting oil, etc as you will burn through your drill bit quite fast in copper otherwise and those larger bits are not too cheap.
Also if copper gets too hot, as well as drill bit... just let it cool down a little and resume. Take it slow.



Now that you are done with 3/8 hole, let's work on hole that will get your thermal reading probe in... Since I am using k-type probe with very small tip, I am going to drill small hole. Right in between the 3/8" hole and surface. So that it reads more consistant value in between. For this one I used 5/64 for about inch deep. Don't go too deep in or you will surely break your drill bit.



It should look like above once all cuts are done. Clean it off, clean off any cutting oils.

[jinu117]

Now it is time to prepare surface of load tester.
The side where it is closer to your probe is going to be the evap contact surface which need to be lapped.


This is grits I go through... sometimes I do use 2000 grit... but since load tester will constantly see deformation over time with many mounts... I skip 2000 grit on this one.



Circular motion is key on this. I've tried varios methods and always comes back to circular motion on lapping for better results. (other than some air HS which will cut through sandpaper instead of getting lapped). I lap mine on flat wood panel attached with glass pane :P lap it on flat surface please. (glass being the best).



Just for the luster, I buffed the thing. Now when buffing if you plan to, be careful to keep all area has been covered evenly. You might start undoing the fine lapping job you did if you go too far... Mirror shine eh?

[jinu117]

Now you fill some portion of thermal paste into 3/8 hole and the 5/64 hole so heat transfer will be good.



You might have hard time filling 5/64 hole... no problem... just put them on the probe. (I do this before I attach mine to evap, etc anyway... )



Now insert the cartridge heater and k-type in...
Cartridge heater might take some effort to put in as it might trap air inside and won't come out... just keep on trying... (also edge of cartridge heater is concaved so putting some thermal paste there beforehand can help)



Time to insulate it I used high temperature fibre based tape in here. Armaflex, styrofoam box, etc will do just fine for this really. Insulate all side excet where the evap will touch.

[jinu117]

Now let's connect cartridge heater to ac line.



This is how I mount it. DFI bracket here (any cpu bracket can do), in this case i use 2 3inch 6/32" machine screw in there.



Well, just sandwich our load tester between bracket and evap and mount it as you would on motherboard. I don't use wing nut for myself... Deep socket wrench with 5/16" does great for saving my hand and getting me good balanced mounting often times. Wingnut is convenient without tools but if you have proper tools, it makes even these little things easier



Now, let's wire them up. Kill-a-watt goes to ac line. Variac or load variance tool goes to kill-a-watt. Load block connects to this load variance device.
Now you are ready to power up your phase, waterblock, air heatsink, etc and see how it handles load

Few things of notes...

1) It is recommended that you install thermal shut off switch. Something that won't burn through enclosure material if you forget to shut off the load tester after you are done testing..... (I've burnt 5 enclosures so far :P)

2) Bottom of load tester insulation... I prefer something stiff so that mounting is more consistant each time. Hard refrigeration foam, the tape I used, etc works great. I suggest armaflex tape type if these are not an option instead of foam... foam tends to be too squishy.

3) Always turn load tester off when you are done.

[epion2985]

sweet jesus!! you motivated me to start on mine

[runmc]

Very nice jinu - thank you
I will give it a try.

Do you have to worry about the heat melting the DFI backing plate?

[jinu117]

Very nice jinu - thank you
I will give it a try.

Do you have to worry about the heat melting the DFI backing plate?

Acutally, unless you forgot to turn off loadtester at the end of test, it should start feeling cold at back. With cascade, the backside start getting water drips... You see, the evap should be cooling well enough to cool down entire copper since there is no other way of heat escaping other than through evap (thoeritcally) from load tester with all the insulation. Some resistor load tester people showed with no frost to me looks like just bad contact with no correct load on evap.
And thanks to nature of cartridge heater, variance with temperature and resistance is very minmal which you could run for hours.
Another tip... never let it run unattended. Your unit might not hold load and collapse and temp will just keep going up.

[Jack]

Nice guide and good pics

But one question, the variac is probably using some power itself so we you connect it the way you did, the Kill-a-Watt reader will give you the powerconsumption of the cartridge+variac.

[twisted5]

Nice work there Jin. You really do take pride in your work and it shows!

[Alex08]

this thread should be stickied or linked in the guide section.

great work like always.

OT: kind of weird seeing you wearing jeans especially in 90f+ these past couple of days

[Weee]

this thread should be stickied or linked in the guide section.

great work like always.

OT: kind of weird seeing you wearing jeans especially in 90f+ these past couple of days

It wasn't terribly hot that day, plus half the photos were taken after 7PM. The longest part of this thing was watching Jin try and drill that hole deep enough to the the cartridge all the way in without punching through.

Okay now everyone start using these!

As far as the comment about the variac using power, transformers shouldn't be terribly lossy, but the key here I think is to have a consistent basis for comparison of temperature, and the power draw numbers can be correlated with real CPU conditions to really get a handle on how phase units being built are handling the load.

[yngndrw]

Can we have an earth wire going to that copper please.

[wdrzal]

Nice work Jinu Looking forward to feedback on testing!

[froudeg]

Nice work Jinu - best designed load tester i've seen so far on XS....some of the one's i've seen are shocking lol.

[martinjon666]

Awesome work, i think i need one

Its almost impossible to find any copper blocks here in idaho though. Still looking for a decent source for copper.

[ilkkahy]

Nice work but the contact surface area of that tester with evaporator is about 25cm^2 at maximum (if evaporator is that large) where normal cpu die is only 1-2cm^2 large. Reducing heatload tester contact area with evaporator bottom is more cruel for the evaporator temps but it think it responds the reality better. I have used a self-made 160w heatload tester in testing co2 containers and its contact surface area was around 2cm^2. Anyway it works very nice indeed.

[Crazystang01]

Hello I got a question about the Cartridge heater. Now Jinu you suggest the 200w. Where as DetroitAc is using a 300w Cartridge heater. Since with 115v=192w wouldnt a 250w Cartridge heater be good. Think it be the sweet spot between the 2. Also I see that the same one you listed from McMaster can go to 250w with same sleeving.

Doh was rereading all of this. Saw that you posted up 200w at start then later 250w. So lol guess just by reading I got the info I was looking for.

[LittleDevil]

jinu thanks for that guide and runmc thanks for cartridge heater , here is mine



regards