Am I off base on these temps?

[sdumper]

I recently built a custom bencher for someone in the US (Not from this Forum) and felt the temps were respectable and on par with what I was expecting.

Compressor was a Danfoss SC18 3/4hp LT compressor and no load was -47c and eventually -50c.

240 watts -42c
280 watts -39c
310 watts between -35c to -37c (after recover from -35c).

The gentleman I built this for is less than happpy because he was expecting -56c to -60c unloaded since he was told that even Cryoz's were doing -58c or colder.

I tried to explain that adding more refrigerant to increase load handling would result in the temps under no load being warmer than they would with a lower charge.

Anyway maybe im the one off base here so I wanted to ask you guys if these temp seemed reasonable or if maybe I was the one short changing the guy.:shrug:

[wdrzal]

Sdumber you should tell him no load temps mean absoultly nothing,and are not worth the forum space to publish.

As Far as actual temp you posted they are in line with the best builds we ever seen.

So greart work. :up:

Your customer needs to come back to reality........there no magic in THERMODYNAMICS........just way to many people exaggerate ,read temps before the system is in steady state or their testing equiptment is out of calibration or they use bios temps that are not accurate.

[wdrzal]

One more thing iirc Cryo-z was meant to be a low cost introduction into the phase world from water cooling. They were never meant meant to even compete with the coldest custom built units. Just a first step into phase.

[One_Hertz]

Sounds like he was just buying this for bragging rights. Anyone buying a unit for a purpose would not care about zero load temps. In fact, warmer is better for idle due to condensation and such...

[Gray Mole]

Sure sounds familiar :(

Doesn't really help that you get so many people giving glowing feedback on the units you make, sometimes all it takes is one guy to talk you down and it breaks your spirit just a little huh?

Hope you don't let it get to you, your work is and has been really good.

I remember arguing a similar point with someone about the old Mach2's, and how they were reaching below -60 unloaded :p even though their temperature monitoring was a mile off.

Anything better than -30 at 300w is excellent, and like Walt said, unloaded temps mean very little in a unit. A long time ago there was a discussion about that very issue, and the consensus was just as One Hertz stated, that the best unit would have a really small difference between max load and no load. The best way to overclock a chip would be in a tight range of low temps for consistency.

I personally found that there were a few people that could never be pleased by the work you do, and I had to try to get past that. I didn't, really. Part of why I disappeared for a while. Hope you don't let that happen to you.

One note, though, and the only way I can see to get super low unloaded temps in a unit with high load capacity is with a CPEV. A captube system is going to allow a stream of refrigerant that's fairly consistent, and the more it's charged the higher the backpressure, so the higher the temps, loaded or not.

With a CPEV you can tune it for 300+ watts and record the temps, then dial it down and record a stupidly low unloaded temp. Means the end user has to keep retuning for both of those results but if that's what the guy really wants, is bragging rights on a cold sounding unit, then that's about the only way to get both from it. Adding a CPEV just so you can brag about your unit seems frivolous but some guys just want what they want.

So yeah, the temps you're getting are pretty much in line with the best I've seen, and the best I'm personally getting on the 3/4hp rotary SS unit I threw together. If he wants something different then I guess it's a matter of finding out what that is, but maybe 'selling' him on a cpev mod to the unit you made is the way to go?

Don't let him get you down though :) Your work is good. Those temps are great.



Gray

[sdumper]

Thanks Guys...

The buyer actually wanted a CPEV but I have had bad results when using a CPEV although those were on cascades.

Anyway when I get back I will send him an email. I have my password stored in software cache for OCN and cant log in from my laptop while i am out of time for the holidays...

[Gray Mole]

Never liked cpev's myself :p though they have their place I guess. Inevitably they get used when cold, which is both awkward, and kind of asking for a leak in the valve later.

That said, I'm throwing together a Rotary Bench SS with a CPEV, just so I can play with one again. Been a while, but like I said, never much liked 'em.

When I'm done though, I'll be sure to 'brag' about how cold it is unloaded :D :D


Gray

[kitfit1]

I'm thinking of fitting a CPEV in my next build as well, in fact i thought long and hard about fitting one in my first build.
But i have to agree that unloaded temps mean nothing, always have and always will.
But the nearer you can get one to the other the better me thinks.
The pics below give you an idea of why i think "the closer the better"

This was taken a few seconds before the end of a WPrime 1024 run:

http://s5.tinypic.com/2zsr128_th.jpg (http://oi39.tinypic.com/2zsr128.jpg)

And this was taken a min or so after the same WPrime 1024 run:

http://s5.tinypic.com/119q6w9_th.jpg (http://oi41.tinypic.com/119q6w9.jpg)

As you can see, not a lot of difference between loaded and unloaded temps.
To me, that's 100% more important than having a massive unloaded temp.

[Gray Mole]

I guess if you wanted to 'reach' for a reason for lower temps at low load (cpu idle) is for the 'suicide screenshot' of the hgihest cpu clock at idle.

Those, while also relatively meaningless, can be fun. A cpev could be retuned while the cpu is at idle to the loest possible temp, and the cpu clocked higher while the cpev monitored and changed for the best high clock screenshot possible.

I think that's really the only place where a cpev would be any kind of benefit over captube, but as soon as the screenshot is done, the cpev would need to be retuned for max OC at full load (cpu actually working)

So I suppose lowest LOW LOAD temp can have a purpose even though NO LOAD does not.

Ah well, there's a rotary out in the Garage waiting to be butchered into a system. And it's callin' me...

:D

[sdumper]

Walt always good to get a touch of reality :)

[Stewie007]

It sounds like his customer is just inexperienced and doesn't really know a lot about refrigeration.

[runmc]

I'm thinking of fitting a CPEV in my next build as well, in fact i thought long and hard about fitting one in my first build.
But i have to agree that unloaded temps mean nothing, always have and always will.
But the nearer you can get one to the other the better me thinks.
The pics below give you an idea of why i think "the closer the better"

This was taken a few seconds before the end of a WPrime 1024 run:

http://s5.tinypic.com/2zsr128_th.jpg (http://oi39.tinypic.com/2zsr128.jpg)

And this was taken a min or so after the same WPrime 1024 run:

http://s5.tinypic.com/119q6w9_th.jpg (http://oi41.tinypic.com/119q6w9.jpg)

As you can see, not a lot of difference between loaded and unloaded temps.
To me, that's 100% more important than having a massive unloaded temp.

Interesting anology but a bit off topic. I'll move this thread to customer/builder section.

As you tune for higher loads the distance between loaded and noloaded temps do decrease.