Rotary CPEV bench unit with 'custom' case

[Gray Mole]

Ended up throwing together a Bench unit yesterday and today, along with 'making' a case for it. The case is pretty ugly, but not 'the ugliest case ever' since that might give it some redeeming qualities





Don't know if you can see it, but I used a wire type condensor as the rear fan guard and it worked well. Looks a bit funny but it's really strong and let's good airflow through.

Actually the case works out really well, since it's only about 8 screws at the bottom and it slides off (other than undoing and reconnecting the temp readout) and can just be used 'open' if needed.

I built the case just for a shipping idea really, since it's harder to pack and open unit. It got to be fun after a while, and for shipping it's great the suction line fits nicely into the hole along with the power cord.

So the specs...

Toshiba 13cc or 3/4+hp rotary compressor
Chilly1 spiral evap and Duniek mount (had one around)
9000btu condensor with 200cfm fan (lowest speed) barrel type
Fachs CPEV with 1/8" 'feed' tube setup
30" flex line HD flex
LAE temp meter K type probe
R290 charge





Didn't bother making the fan adjustable, though it does have higher speeds, they're not necessary for this setup. Tested up to the 520w max I have and it held no problems, condensing at ambient.

The idea initially though, was to see where a person stood with tuning for 300w or so. This is 260w tested though I found a little more than req'd gave a better plate temp, so it's really a 300w setting at the cpev for best results.

I was just curious, since it's easy with a cpev to set the load/temp and then go straight to 'no load' to see what the comparative captube tuning gives. A cpev at a given setting isn't really any different to a length of captube tuned for a particular load.

This one, when tuned to the usual 300w mark, gave a no load temp reading around -53c or so. Even with a big rotary, the high load tuning won't give you a really low no load temp to brag about. It's just the way life is, a trade off.

The cpev does have that little advantage of the 'money shot' of a lower temp if you dial it right down, but that's a pretty meaningless number, since all you're doing is running the evap almost dry, and it wouldn't hold a load when you do.

This unit dialed right down hit -65c. Not bad with just R290 I guess. Not much point in taking a picture at all of it, since it takes a while to retune it back to load and it's got a really nice setting where it is. Whoever ends up with it can play around at super low temps if they really want to.

The K probe LAE display is reasonably accurate. I was surprised really. I might just do a zero calibration to make it match on lower temps, but it's only like 1 degree off on load temp, so it's not bad at all. I blame the probe, since it's a -50 accuracy probe it tends to slow down at low temps. The display is really only for reference for startup, and works well enough for retuning to a specific cpu and load, but the K probe moves a lot faster than the PT100 probes I've been using.

The head is a chilly1 spiral, so the head size is a little smaller than some of the other heads I've got. Great for being load sensitive though, the temp moves quickly to the load, and back down. Very responsive.


It held 500+ watts at about -25c, which is cool. The plate temp read about 24c so close to 50 degrees of difference, and that's a few more than the larger head. I figure the surface area does help a lot with really high load. Still not bad though, nice to know that the smaller heads should be able to hold up to 500w on the right unit. Don't know if any cpu's will get that power hungry but you never know, and it's good to be able to be sure of the evaps we're using being able to hold that.

Went with a length of the heavy duty flex I have. Not much left now. It's pretty stiff as the metal is a touch thicker than some I've used, but you can manhandle it a little more without issue. Still recommend never to bend when frozen but this stuff is pretty forgiving, doesn't seem to get so brittle. Holds a shape very well but you also have to 'preshape' it a little to get it the way you need. I like it well enough I suppose, and I've never had a unit come back with a cracked flex with this stuff. Seems like decent quality. I've had really 'light' flexlines that cracked way too easy, so if I had to choose, the heavier stainless flex would win even though it's got it's own niggles.

The case, welll...it's not totally pretty but it was kind of fun to chop up an old AC unit to fit. Worked out well enough I guess. I had it and it was just going into a recycling bin anyways, so I re-recyled it

I cut a hole in the side where the CPEV is for access to the adjustment, the cap normally goes back on when you've adjusted but I took pics with the 'dial' adjuster on. Pretty nifty. Guess you could glue it on and insulate up to it and it would be ok, but seems silly. Better to just cap it until needed but to each their own. Just didn't want to have to remove the case every time an adjustment was needed in case whoever get's it wants to keep it on.


I also didn't want the CPEV right at the evap like I've seen on some. Too much grief and the feed tube being 1/8 is flexible enough to just use the line as normal. Couldn't really go 'too long' though, didn't want the small pipe to start restricting the flow. Rather have the full adjustment of the CPEV.

The one 'real' advantage to CPEV on a unit that I'd forgotten about before is that you can overcharge it just a little. You need the discharge volume to allow for it of course, and a nice big condensor to prevent any high discharge pressures, but that extra refrigerant does a lot for making the unit attain temps at startup MUCH faster than captube, especially at low ambient or after longer periods of disuse. This one is great for that, It cooled down close to max no-load temp within 10 seconds after sitting overnight and I hit it with the 260w tester within 15 seconds of startup to see if it was 'ready' for load. It took it and held the load without issue, even though the compressor was still cold. I really like that feature, since long startups are a pain at times. Having a superfast subzero PC is awesome but having to wait 3-5 minutes to start using it can be irritating. The CPEV setup seems to bypass that issue and that's really nice.

So it answered my questions about cpev vs captube and to be honest, other than swapping to low load tuning when needed it's not absolutely necessary. For those needing the lower low load temps for the suicide screenshots it's handy I suppose, and the fact that on units where floodback is an issue you can dial it down for cpu idle if you needed to. Not a benefit you 'should' need but it's something. The cost of adding a CPEV to a unit is relatively high, since ideally a sightglass is normally something that goes with it. I didn't put one on myself but knowing the behaviour of a unit with CPEV I wasn't concerned about it's 'need'.

If you add both CPEV, Sightglass and a larger filter (if discharge volume is any kind of issue) or accumulator then the extra cost can be significant for the few benefits it gives. I know a lot of us don't pay full retail but looking in the book you could be adding 150 to 200 bucks or more to the cost of a unit just to incorporate a CPEV setup. That's a fair bit, so something to consider.

So the final results...

Tuned for 'best no load temp' was -65c evap

Tuned for ~300w, was -53c no load, and -39c 260w Evap


Plate temp -9c after 2hours load

Tuned for 500+w no load was -45c and -25c 520w

Plate temp 24c after 1 hour load


Plate temps were measured with a K probe drilled into the load tester right over the 2 resistors. The probe is closer to the one resistor, so I use that resistor for the 260w testing to be more realistic. I tried the other resistor for single resistor 260w tests, but it showed about -13 so I know the result isn't as accurately reflecting what the CPU core temp might be due to the placement of the probe.

[Gray Mole]

I didn't get that specific by what I mean about Plate vs Evap temps but The tuning with the cpev made that easier to get the best result.

I tuned the unit a little 'lean' and got -40c on the evap, and it actually rose to around -7 on the plate sensor.

I tuned it 'heavier' and the evap rose to -39c but the plate temp dropped to -10c.

That's one way the cpev can help, though tuning captube you can do that as well during the load testing and tuning. Just found it interesting that a little more refrigerant flowing was able to flood the base of the evap just alittle more, and though the evap temp might be colder, it's ability to tranfer heat is lessened. Something to remember for a builder, but also an end user, that the temp of the evap is great, but the actual cpu core temp is the more important factor.


Gray

[Buckeye]

Don't know if you can see it, but I used a wire type condensor as the rear fan guard and it worked well. Looks a bit funny but it's really strong and let's good airflow through.[/URL]

HAHA at first I thought you used that condensor for the build, I was going man how did you get that thing to hold any load

Do you have any pic's of the insides at all ?

Nice work

[Gray Mole]

Yeah that might have been a little tough.

I don't know how many of those it would take, but I think it'd be a lot.

Like I said in the 'cheap parts' thread, they can make a decent fan guard but that's about it.

I was pretty tired last night, so didn't take the 'lid' off. I'll do that today, but I'm supposed to meet up with the family for breakfast so I don't know if I'll make it before that.

Took me the whole day and putting up with the stink of burning plastic to build that ugly case so I don't know if it was worth it, but I suppose it's my first (and probably last) custom case.

Maybe I should have just called it a custom shipping container. I could feel better about that.

Will get some pics though. The unit is surprisingly simple considering it's a cpev setup, but a valve system of any kind doesn't need to be over complicated. Long as it does the job, it's all good




Gray