Noob trying to gether some info.

[mike8913]

OK fellas. Right now I'm at school so I won't be building my first unit until may/june, but right now I'm trying to gather as much information as I can.

In the past couple of months I've done my research and I have the basic ideas(and maybe even slightly advanced) on how to assemble my unit.

First item: temperature probe. I looked through all the stickies and couldn't find anything on what type of probe is used on the head of the Evap or how it is mounted there. I've seen K-Types used but I think different type is used for controllers. Right? If some one could illustrate this for me it would be great.

Secondly: Wiring. I'm so confused on this because I've seen different routes taken by everyone. It seems to me that LittleDevil uses a system where power is delivered through one wall cord that powers on the compressor and then controller. Then some how i think it allows power through the secondary cord to which the computer can then power on. But I feel like I'm missing something. How des it all get wired together and then I seem to remember some sort of 24-pin power connection adapter coming from one of his controllers?

anyway, multiple insights would be appreciated and pictures would be amazing. I just want to get a solid base to work from come summer time.

Oh, and lastly. I keep finding different "formulas" for captube sizing and length, but they all seem to differentiate. Is there a definitive formula out their that will give me an accurate estimate of what size and length I should be using.

I will be using a Danfoss NF11Fx and I'm going to be using the water-cooled desuperheater application using a PA120.2 and PA120.1 for dissipation. If this helps with any of the above mentioned "fuzzies", if not than oh well.

[mike8913]

I'm looking for insight from some of the SS builders out there. Off the top of my head I can think of Jinu, Pete, LittleDevil, Tim, Johann, SSilencer, and Dimas.

Just looking for of the nitpicky stuff; if you guys can help me here that would be great.

[n00b 0f l337]

Go take a look at the gallery I'd say. Your looking to watercool the unit then? Or have a condenser as well? Captube depends on your load though so we need some more information.

[mike8913]

Well, I suppose it's all speculation for now because I'm planning this for future chips (e.g. Nehalem) which will most likely run at lower TDP's. I figure a 200w load is appropriate and future proof. I was planning on a 2 foot suction line (please feel free to suggest otherwise). I'm planning on not using a condenser, only the water cooled desuperheater.

Anything else that might help? I have looked throughly through the gallery and SS build catalogs. It's more of the stuff you really can't "see" that I'm confused on; specifically the power wiring for the unit itself.

Hope that helps.

[SexyMF]

Do you intend to purchase a premade controller for your unit. If so, which one. As this will determine the electrical setup.

200W load is not enough for future proofing. Wattage goes up incredibly when you overclock. The current lot of quad-cores CPUs present greater than 200W loading. You are either rich, brave or other, if your first use of a direct-die cooling setup is to be on the latest expensive CPU.

A desuperheater is not needed in a single stage. You just need an appropriatly sized condensor. Water cooled condensor is fancy but is overly complicated to achieve a mediocre result when using a standard watercooling setup. This has been brought up in recent threads.

[[XC] gomeler]

Well, I suppose it's all speculation for now because I'm planning this for future chips (e.g. Nehalem) which will most likely run at lower TDP's.

I would rethink this statement considerably. Just factor Nehalem will be 8 core on 45nm process and while current 45nm quads are maybe pulling 225w or so, do the math, carry the one, looks like we might have another B3 quad situation on our hands. I'd aim higher more like 275w, if anything the extra capacity will be useful when overclocking to the limits of the chip rather than the cooling.

If this is going to be your first build just do the standard compressor, condenser, suction, evap. Unless you want to deal with getting the waterloop correctly tuned, then go for it but I may point and laugh down the road :D Wiring the unit itself depends on what's going into the unit, I use about 8 feet of wiring in each of my units to wire everything up properly including grounds and relays. In the end it all matters how you want to build the unit and what goes into it. For a first build I'd stick with a simple on/off switch and wire the fans and compressor into the single switch.

Capillary tube tuning is a pain in the butt, I myself spent a long while just hacking it up on a unit till I mapped it out for R507, soon I'll be doing the same for R290 to see how cap tube lengths affect performance. Like many members would say, go grab a book and read a bit. Just get a basic build down then worry about fancy things like controllers and water loop-de-loops and flashy lights.

[mike8913]

I understand that wattage increases considerable when I over clocking and that's why I asked for insight though. So 275w is more appropriate you'd say. Is that load attainable with a NF11FX?

The only reason I want to do the water cooling bit right off the bat is because I'm very familiar with the premise of water cooling because I have experience in that field (not to mention I love to experiment). And to SexyMF: brave(maybe arrogant to a degree as well) is probably the right word for it, my bank account doesn't allow me t claim riches.

Gomeler, I do have to say that everything in the microprocessor world is going towards low power draw. So even though over clocking drives up wattage; I don't think it's unfair to say that 200w load capacitance will be the norm for future over clocking.

I'd still like input on the fancy power setups though. I know I've never done a build before but I am a perfectionist and take my time, so I'm willing to learn the intricacies right away. If anyone can offer some more complex input, I'd love to hear it.

Oh and I suppose I'd use the G-Froster controller. It seems to be the most established one.

[n00b 0f l337]

Attainable? Easily, I've got a 1/6hp 3cc compressor holding 250 easily.

[jinu117]

Water loop isn't advisalbe for first build on my end as well. Too many things factors in for condensing temperature and add hugely to cost. (Greatest benefit i found was ability to relocate the condenser, and be able to reduce noise). Performance wise, it's not much different at all.
Capillary tubing depends on compressor you are using and size of condenser as well as evaporator you are using. Probably why you are getting so many different cap tubing rules as there really isn't a rule.
Hope this helps.

[mike8913]

Thanks Guys.

The only reason I'm going to try using the water cooled desuperheater idea is because I already own the water equipment including the two PA's. Quiet is important to me too, so there's another reason. So that's really not an added cost. I suppose that, theoretically, if my coil were cooled to ambient by the water (say 22C) then I'd have to factor that temperature with the charge of my unit. Is that correct?

So if that's the relative temperature I will achieve and I'm using a NF11FX, can you give me some sort of rough way to figure out a captube length. Some sort of way to guestimate.

[SexyMF]

Attainable? Easily, I've got a 1/6hp 3cc compressor holding 250
easily.

Read into this comment. The question to ask is 'what load at what temperature'.

Start 0.031" capillary at 3 metres. Nobody can be more specific because it is a matter of the components used. ie condensor size, volumn, flow rates etc

If you want some cool graphs and the ability to work things out more scientifically I suggest a look at cool-pack (http://www.et.web.mek.dtu.dk/Coolpack/UK/download.html)

[boshuter]

Attainable? Easily, I've got a 1/6hp 3cc compressor holding 250 easily.

At a usable temp?

[mike8913]

what about adding a CPEV. Since its a SS would I still need to add a receiver in conjunction with it.

[teyber]

If someoen could answer the probe question and the electrical question that would be great, as i also have the same questions

[n00b 0f l337]

Yes Bo.

[mike8913]

If someoen could answer the probe question and the electrical question that would be great, as i also have the same questions

actually thanks for saying that. Fundamentally, this is more important for now, I'll worry about captube later down the road.

[jinu117]

Before getting into CPEV or TXV, I believe you need to find out how they work and are getting used first. I suggest looking into parker/sporlan websites on how correctly to install that. Electrical part for controller, forget about this till you are done with cooling part. Trying to kill two bird in one shot just copmlicates clear view of project.
Once you understand how it should really work, than you can think about how you would like things to be wired and controlled. But cooling comes first.

[mike8913]

I understand that I should get the unit powered on and working before getting into complicated wiring, but I would still like some suggestions on how to address the latter. I'm not looking for simplicity, I'm looking fr performance and function. Don't worry though because I have 6 months to think about everything and get m plans mapped out.

Thanks for any tips.

[tim-]

you should skip the tew stuff. I've done that and I'm not satisfy with that, long delays before the valve catch up with the load and that cause liquid flood back in some situations and really hot cpu when loading it 100% from idle.

the cpev isn't the best choise either in my opinion, you need a reciver to make sure you always have liquid to the cpev.

honestly I do like the captube more and more by time. I have over 10cpevs at home and I have use cpev on one unit under this 2years I've build phase units for this perpose and I've never been comfortable with it. go for captube. and something in 2.2-2.7m of 0,8mm ID captube should suit your needs good. I use ~2.4-2.5m of 0,8mm and hold -33dgr @ 230w with nl11f with proper condenser (500w stuff).

about the electrical and such, K-type is most common for measure real temp, I know gefroster x2.0 uses K-type probe as well, but I also found out that lm35 is working pretty good too if you calibrate it proper, for example unseens controller uses it, proper adjusted for our temprange it works good as I said.

otherwise about electrical there is 3 kind of compressor wireings, start capacitor, ptc and 3phase. 3phase is just l1, l2, l3 can't be easier. the startcapacitor and ptc is little more to read the electrical scheme.

I havn't seen many builds for our perpose with 3phase, so there we are left with two kind, ptc and startcapacitor, I've seen more and more people going over for ptc, atleast for small compressors as nl11f. which is good, since it's cheaper :p but if you are using some kind of valve to make the pressuredrop for evaporator you need a compressor with startcapacitor since ptc can't start with a difference of pressure between hi and lo side.

and as I said earlier of own experience I've found that captube is the best way for SS. both for the load that can change pretty fast, and the pressure will always equilize between startups which is good for compressors health.


captube is a pita to calculate on, i've done some studys about that at the university and we have found a formula ( i think it's somewhere here at XS) but it requires 100% liquid true the captube which we don't have (since it evaporates in the end) etc etc.. to many factorys that makes the theoretical formulas to not fit in, as well as different gases have different viscositys on different temperatures etc etc.. you have to take that in your calculations. for example, if I build a unit without slhx I need a longer captube compare to if I use slhx, since the slhx makes the refrigrant give higher viscosity which means it flows slower true the captube. this doesn't mean the slhx will make a advantage or disadvantage in a unit, it depends on alot of stuff. with nl11f I would say a slhx would decrease the capacity instead of increase, atleast I get better temps by skip the slhx, you should always crap the captube arround the suction line thought.

hope this will help you a little on the way.


cheers
Tim

[jinu117]

Sigh... I was hoping to save embarassment...
TXV requires receiver.
AEV(CPEV) requires suction accumulator. Receiver is not a requirement nor desired part of CPEV system. It is another form of critically charged metering device.
That's why I said need to study bit more.

[mike8913]

Tim, that was wonderful. Thank you so much. But thanks to Gary Loyd I am eternally confused. Here's why.


For R404A/R507:

Using a center point of 150 watts load I would guestimate 10 feet (3 M) of .028 inch (.7 mm) cap tube.

For each increase of 15 watts, I would shorten the cap tube by 1 foot (30 cm).

So, by this, if I want my load to be 300w then I have to get rid of 10foot of captube....but that means I have no captube??? Is this why you use .8mm, because .7 would not allow for such a load?

[mike8913]

Jinu, I was just throwing out the CPEV idea. I really had no idea how they function. It was mroe an open ended question, but at least now I know. Thanks!

[teyber]

Tim, that was wonderful. Thank you so much. But thanks to Gary Loyd I am eternally confused. Here's why.



So, by this, if I want my load to be 300w then I have to get rid of 10foot of captube....but that means I have no captube??? Is this why you use .8mm, because .7 would not allow for such a load?
key word:


guestimate....

[tim-]

mike:
forget about that rule, it's pretty old and isnät 100% sure now.

0,8captube is what will suit your needs best, atleast of my experience with 0,7 and 0,8mm captube and big loads.

maybe down to 2 meters if you're aiming for holding 300w with the nl11f. you have to try and fail :p noone can tell you 100% sure what's the best lenght. also it depends on evap, some seems to possible to run "dry" (longer captube) with good capacity and some seems to need a little "wet" condition to perform best (shorter captube).

and ofcourse r507/r404 is your choise of gas if you're going for aircooled condenser, which is the only thing I recommend for first time, at for 200-250w ~500w condenser is good, for 300w load i would go for something little bigger.

regards
Tim

regards
Tim

[mike8913]

Do you recommend r507/r404 because they dissipate load easier? or for other reasons? I tired to study up on the refrigerants, but studying political science is taking up enough time.

[tim-]

well. thats about the pressures and condensing temperatures. nl11f doesn't like much bigger then 15bar delta between hi and lo side, we all want the compressor to last atleast a couple of years. expect that it's about performance. r134a for example need very big vac for same temp which mean the massflow is very low = hold less load at same temp compare to r507 (i'm going to say r507, but it involves r404 since the two refrigrigrants is almost identical, they use same tev etc). and r410 is 4-5K colder @ 0bar(e), but the condensing temperature and pressure makes it hard to do with regular stuff, you need very big condenser or very high pressure to condense it which affect evap temps, yadayada, from experience we have figure out that r404/r507 is the best gas for our perpose. R402 is a better choise but since it's including r22, which is illegal in whole europe i guess? that's not possible to get legally.

this put us in the situation, of the allowed gases r507/r404 is the best choise for regular SS use. I've build some rotary units with 14K btu + compressors and 1kw+ condensers and I've found out that even there the pure r410 isn't the best choise, the best choise (for SS) is a mix of them two. but thats same there, to condense in room temp we need something like 22bar on hi ide but we still want low pressure (equal to low temp) on lo side which means the compressor need to work harder. it's always about calculations and make a balance between temps and performance, here is also the massflow that limit us to how low we can go.

a short answer, the r507/r404 is the best to go 'cause it have low evaporation temp but still have decent condensing temps when we aircool it, also the gas is pretty nice to us since it allows pretty high superheat without overheat the compressor. this is very nice, if we compare with amonia there a very small superheat can cause a overheated compressor. this is also a reason we choose r404/r507, since it's a very easy gas to deal with and doesn't require to much attention to get superheat and subcooling correct, it will run pretty good anyway, thats why a slhx is pretty easy to use in this gas as well, on amonia it would be the same as asking for a overheated compressor. but I know many of you havn't been dealing with amonia units, and be happy, thats an experiance I could live without.

there is alot of experience from both me and other users that have made those statesment and can proof them, well I have proberbly not made to much for the XS on that but i've done it the hard way and figure the stuff out my own to know WHY it's happening and understand it.

another thing I would recommend you is to avoid blending different gases 'cause it's very hard to understand whats happening.

regards
Tim

[mike8913]

Very interesting. So if on my NF11FX I have 22bar on hi side I would want no less than 7bar on low(preferably a smaller a gap right)? Makes sense a lot more now if that assumption is correct.

[n00b 0f l337]

22 Bar is quite high. So is 7bar low.

[tim-]

yes 22bar is high, and thats' why I don't recommend r410 for this kind of applications, if you're able to hold condensing @ ~16*C @ 15bar constantly, then r410 would give you an big advantage 'cause it both holds more capacity for same mass as well as lower evaporation temperature.

borrow some books at the libary and make the reading to get the basic, I think thats the best advice I can get you if you want to have all the answers :)

if you run r404/r507 you won't have more then 14-15bar on highside if the unit is proper tuned :)

[teyber]

yes 22bar is high, and thats' why I don't recommend r410 for this kind of applications, if you're able to hold condensing @ ~16*C @ 15bar constantly, then r410 would give you an big advantage 'cause it both holds more capacity for same mass as well as lower evaporation temperature.

borrow some books at the libary and make the reading to get the basic, I think thats the best advice I can get you if you want to have all the answers :)

if you run r404/r507 you won't have more then 14-15bar on highside if the unit is proper tuned :)

The next wdrzl!!!

(In a good way!)