how'bout this compressor? [Archive]

Gary Lloyd

01-22-2004, 10:46 AM

It doesn't look too bad, it's rating says to -35c at normal ambient temps using 134a. Would work pretty good on a water chiller.

No, it's rating says it will move 32.6 watts of heat @ -35C or 40.1 watts of heat @ -35C, depending on whose testing standards are used (CECOMAF or ASHRAE). If you move more heat, the temperature goes up.

With an 80W heat load, we are looking at about -20C, but the CPU is not the only heat load. There is the heat that needs to be removed from the water to get it down to temperature, and there is the heat that is absorbed through the insulation. The better the insulation, the lower the temperature can go.

Heat is the enemy. Insulation is our friend. :D

Miles_Teg

01-22-2004, 04:26 PM

So Gary,

As you and DaBit are pretty much on the same line as far as the fine art of CPU cooling & refrigeration go, you'd prob'ly expect around the same temps right?

I've had a not-so-good waterchiller before. Made it down to whopping 11°C. Loaded temps would rise 'till around 15...16°C. Like I said not so good. But every now and then, when I felt the need to go all out, I used to stuff 3L worth of water in icecubes in the coolant container. Made the coolant temps settled for quite some time around 0.8°C. BUT, delta-T between coolant temp and CPU temp made a huge leap at those temps. Upto 17°C difference. Reason for me to search for something better then the Maze2 I used at the time.

Is it normal for Delta-T to increase fast at lower coolant/block temps? With DaBit's predicted liquid temps of around -5...-10°C I think I can forget about sub-0 at the CPU right?

Miles_Teg

01-22-2004, 04:44 PM

Thing is...

I got one of these:
http://members.home.nl/jan.barelds/Miles/Chiller/waterchiller.jpg
And it has a compressor with around the same specs (as far as I can tell) as the danfoss link in the opening post.

I'll strip the blue body off, yank off the thermostat, attach a nice fan to blow at the condenser and make it a new sheetmetal body with a control panel looking somewhat like this:
http://members.home.nl/jan.barelds/Miles/Full-size_Pics/Control_Panel.JPG
Yet, this time I'll pay some more attention to detail.

Basically it's just attaching a hose to the tap and to the blue-bottle side with a pump and a block inline. But I want it to look good. (prob'ly better then it performs ;) )

I reackon the small reservoir inside with the evap is already insulated pretty well. Insulating the hose & block will be on me, but I'll do the best I can.

Let me know what ya think.
...And please don't burn me down, I paid a hundred bucks for the cool-water-thingy... :slap:

Gary Lloyd

01-22-2004, 05:10 PM

As you and DaBit are pretty much on the same line as far as the fine art of CPU cooling & refrigeration go, you'd prob'ly expect around the same temps right?

Probably so. Keep in mind that I work on refrigeration systems. I do not design them. Where others calculate and predict proposed systems, I measure and analyze existing systems.

They design them. I fix their designs... LOL

I've had a not-so-good waterchiller before. Made it down to whopping 11°C. Loaded temps would rise 'till around 15...16°C. Like I said not so good.

What refrigerant? What size/length cap tube? The cap tube is the key component.

Is it normal for Delta-T to increase fast at lower coolant/block temps? With DaBit's predicted liquid temps of around -5...-10°C I think I can forget about sub-0 at the CPU right?

Generally, I view evaporator delta-T as an indicator of heat transfer, but then I am working on systems with a given coolant (or refrigerant) temperature. Increased dT means reduced heat transfer from the product to the coolant.

Keeping in mind that this is not my area of expertise, I would say that yes, a drop in coolant temperature will cause an increase in dT, and you should not expect subzero CPU temps.

By conventional standards, the evaporator dT on these systems is huge, but that is understandable given the tiny surface area we have to work with. The block is the major chokepoint in these systems. We are taking systems capable of running a large chest type freezer, and reducing the evaporator to the size of a postage stamp.

Gary Lloyd

01-22-2004, 05:43 PM

In choosing the cap tube, we can go for cooling the water faster to a higher temperature (less restrictive cap tube), or cooling the water slower to a lower temperature (more restrictive cap tube). Since the manufacturer had no desire to freeze the water, but wanted to provide you with a fast cup of cold drinking water, it is a safe bet that the cap tube is less restrictive.

If you are going to use it as is, it will pull the coolant down from room temperature fairly quickly, but it will not go much below 0C.

We can change it to a higher pressure refrigerant, such as R404A which will pull it down to a lower temperature, and/or we can replace/extend the cap tube to go even lower.

And then if we want to go all out, we can put in a bigger compressor.

Miles_Teg

01-23-2004, 06:09 AM

Gary,

The first waterchiller I built was based on a small absorbtion fridge. No good for our purposes.

I intend to leave the new system as it is, as it's a pain in the butt to get refrigeration parts out here (the Netherlands). It might be possible to get a R404A refill, but I don't think the condensor is up to the task.

Btw, we have one of them cool-water-machines at work, and pulldown is really fast. In the summer it was like 30°C and dozens of guys were constantly getting cool water from the machine. The watertemp remained the same for as far as I can tell. I believe the thermostat is set to 8°C. So yes, probly a less restrictive captube.

Gary Lloyd

01-23-2004, 10:34 AM

Windshield washer fluid should work nicely as your coolant.

It will be interesting to watch your progress on this. Post lots of pics. :D

Miles_Teg

01-25-2004, 12:40 PM

Gary,

I've used windshield washer before with my previous chiller, but I did not like it. My previous chiller wasn't an open designs, but not exactly air-tight at the reservoir either. the windshield washer smelled terribly, and bubbled like h3ll 'cause of the soapy additive.

Anyway, I'm thinking about using methanol, like DaBit proposes on his pages. Should also be a tad better for the thermal conductivity of the mixture (Not that I'd even notice it I guess...). Anyway, I think I'll start by removing the thermostat and monitoring watertemps unloaded. If it does dip below zero I'll add some methanol, 'till the point it'll stay liquid.

As soon as the cool-water-thingy arrives I'll start taking pictures of the whole progress. Promiss. :P

As you might notice, the pic of the cool-water-thingy changed. This is because I decided to get this one, saves me a few bucks. It's a tad older, but I think it does work with R134a.

Russell_hq

01-26-2004, 01:56 AM

You guys ever tried using propanol/isopropanol, I dont think its as toxic as methanol and should work at low temps. It should be easy to get hold of as its used as a cleaner for many applications. It be nice to see someone trying out this solvent.

Miles_Teg

01-26-2004, 07:36 AM

Dunno if there's a difference between isopropyl and isopropanol, but isopropyl does have a nasty viscosity in relation to temperature. Slightly less then twice the viscosity of water, whereas methanol is 'thinner' then water, lower viscosity. Freezing point do not differ that much between the two though.

Still, I dunno about the difference between the -pyl and -anol, and neither do I know anything about propanol (without the iso).

Gonna try to dig up some more info on this, and check availability too...

Russell_hq

01-26-2004, 08:36 AM

Ok, here's the deal, you get 2 types of propanol (or n-propanol);

1-propanol
2-propanol

They have the same chemical formula but different chemical structure.

2-propanol is also known as isopropanol as its an isomer of propanol, so my mistake in thinking isopropanol was propanol, its not.

The -yl alcohol is just another naming convention for the group of alcohols, thus propyl alcohol is propanol and isopropyl alcohol is isopropanol (or properly 2-propanol)

The properties as far as melting and boiling points are different as far as:

Methanol 2-propanol 1-propanol
B.p. 65ºC 82ºC 97ºC
M.p -98ºC -89ºC -126ºC

I got this information from Database of Organic Compounds (http://129.219.157.125/)

I had a quick look and it looks like 2-propanol is most common in cleaning solvents and is also used as rubbing alcohol. Maybe someone else knows a source of propanol. You might ge it in antifreeze.

Miles_Teg

01-26-2004, 09:14 AM

Russell,

Thanx for the info! That's some pretty tough scientific stuff on that site :P
I'm pretty sure I can get propanol at work, probly for free. I'm gonna check out tomorrow what kind they have exactly. Antifreeze like windshield washer often contains a soapy additive. Easy to recognise though, I always shake the bottle or can violently if I come across one (will bubble). You should see the look on them salesmen's faces :D Never came across one that wouldn't bubble so far.

Edit: Been checking out some fancy professional thermal conductors. 3M has some fluorinert liquids that would serve us extremely well, yet cost up to $3000 per litre. You can find a link here (http://products3.3m.com/catalog/us/en001/oil_gas/specialty_materials/node_HX0DNRHXKWge/root_GST1T4S9TCgv/vroot_G1F6DNZDBVge/command_AbcPageHandler/theme_us_oilgas_3_0).

Miles_Teg

02-05-2004, 02:23 PM

Lill' update on the project...

Things are NOT looking good. The compressor is only 1/20hp (!), so performance will probably not be very good. Hopefully it'll be able to maintain below ambient, otherwise this project is just wasted time. The compressor used in my particular chiller is the Danfoss TL2.5F (http://www.danfoss.com/compressors/pdf/datasheets/r134a_220v_50_60hz/t-series/TL25F_R134a_220V_50Hz_06-01_Cd42s402.pdf).

Edit: Oh yeah, sorry I'm not posting any pics. Progress is actually pretty decent, but I wanna save all the pics/info up to do a lill' article if thing happen to end up just fine after all.

Gary Lloyd

02-05-2004, 03:07 PM

Looking at the capacity charts on page 2 of the link, according to the ASHRAE chart, if your heat load is 169 watts, then you can expect a water temperature under load, of about 0C. But that is with R134A. With a higher pressure refrigerant, such as R404A, and the same heat load, your temp will be about 10C cooler (-10C).

The question then is, what is your heat load? The heat load is the heat output of the devices you are cooling, plus the heat that comes through the insulation (or lack thereof).

Insulation is our friend. :D

Miles_Teg

02-06-2004, 02:39 AM

169W of cooling capacity at around zero °C seems quite decent, better then I expected. I'm only using the chiller to cool the CPU, and I'm doing a very thorough job on insulation. This way I intend to minimize additional heatload and maximize cooling capacity for the CPU.

- Can any1 gimme an indication on what heatload to expect from losses given a good insulation job?

DaBit

02-06-2004, 04:13 AM

Download CoolPack. It has a module for estimating heat losses through pipe insulation.

Miles_Teg

02-20-2004, 03:54 PM

DaBit, thanx for the 'CoolPack' recommendation. Though most of it's features are rocketscience to me, I was able to roughly calculate the losses I have to bare in mind.

The project is around half-way now. Most of the technical stuff has been done, and now it's time to make a pleasant enclosure. I took lots of pics in the process, and will continue to do so. I'm intending to do a 'How to build a Chiller guide for phase-change n00bs' article.

kingsack

03-12-2004, 07:14 AM

Originally posted by Miles_Teg
Russell,

Thanx for the info! That's some pretty tough scientific stuff on that site :P
I'm pretty sure I can get propanol at work, probly for free. I'm gonna check out tomorrow what kind they have exactly. Antifreeze like windshield washer often contains a soapy additive. Easy to recognise though, I always shake the bottle or can violently if I come across one (will bubble). You should see the look on them salesmen's faces :D Never came across one that wouldn't bubble so far.

Edit: Been checking out some fancy professional thermal conductors. 3M has some fluorinert liquids that would serve us extremely well, yet cost up to $3000 per litre. You can find a link here (http://products3.3m.com/catalog/us/en001/oil_gas/specialty_materials/node_HX0DNRHXKWge/root_GST1T4S9TCgv/vroot_G1F6DNZDBVge/command_AbcPageHandler/theme_us_oilgas_3_0).

Try Prestone Ice Buster Windshield Washer Fluid.
Running in my system for weeks now with no sign of
any foaming.

kingsack

03-12-2004, 07:30 AM

Originally posted by Russell_hq
You guys ever tried using propanol/isopropanol, I dont think its as toxic as methanol and should work at low temps. It should be easy to get hold of as its used as a cleaner for many applications. It be nice to see someone trying out this solvent.

The viscosity gets high quickly and it tends to absorb moisture from the atmosphere much more quickly than methanol resulting in icebergs in your res. Nothing like a having you pump ingest a good slushball to create a little flow problem :(
I use Prestone Ice Buster Windshield Washer Fluid + 10% Antifreze + Water Wetter. Charcoal lighter fluid (odorless mineral spirits) is a possibility. Good down to < -50C. Viscosity is about twice that of water at ambient and it gets worse at lower temps. 10X at -30C, 14X at -40C and 22X at -50C. In a low restriction system like a water chilling setup this might not be a problem.

As if it weren't obvious I would recommend being careful when using any flammable liquid in this application. If you run it unattended and your chiller fails you could return home to find that you don't have one!:explode:

Miles_Teg

03-12-2004, 07:52 AM

Kingsack,

The matter is solved already, in the end following Gary's recommendation after all. Got concentrated windshield washer fluid without the foamy additive. The stuff will go down to -70°C before it'll 'change phase' ;)
As I'm never ever gonna see -70 with my chiller, I'll be okay me thinks. I have tried wiper fluid / water mixture (50/50), but it didn't seem to mix well. The water sunk to the bottom of the coolant container, freezing over after all. So I'm using 100% windshield now, seems to work fine.

At this very moment I'm working on the unit, it's near completion. Working on it's casing now, when finished all there's left is a spray-paint job.

kingsack

03-12-2004, 10:58 AM

Originally posted by Miles_Teg
Kingsack,

The matter is solved already, in the end following Gary's recommendation after all. Got concentrated windshield washer fluid without the foamy additive. The stuff will go down to -70°C before it'll 'change phase' ;)
.

Must be anhydrous. Be sure to add some watter wetter or Antifreze anyway if you have both alumimum and copper anywhere in your system! My system is still a work in progress as well.

Miles_Teg

03-31-2004, 10:59 PM

Fellas,

The project is almost finished by now, all there's left is to install the waterblock onto the CPU. Even though performance is better then what I expected (-18°C unloaded), I feel like there might be more in it. As Gary mentioned the captube is probly less restrictive (short) so that pulldown is quick, but lowest temp isn't that low. I thought up 2 ways to improve performance, and I'd like some feedback:

1.- To make a couple of fans blow at the condensor and compressor.
2.- To pinch the captube for it to become more restrictive (lower temps, but slower pulldown).

In general I don't like the idea of pinching captubes, cause of the risk to pinch it shut. But I would give it a try if that;s what it takes. Would some fans help the temps, or would they only help preserve the compressor for some more years?

Plz let me know what ya think, any other idea's are more then welcome too ofcourse.