Jinu117 SS Worklog (vapoli, UFO, etc)

[elsupremo]

Yes... I officially am done with NF9FX. I will miss the compressor dearly as it was my favorite compressor for last 70+ builds. It has come to time for some more boost in performance without sacrificing on noise front for once
I don't think this is end of how far I will go as I remember my first nf9fx to right now saw about 2-4c gain in average when I actually handed out my last unit to a customer 2 days ago. Let's see how much further I can go now

Oh nooo....this is so sad, a week after I pick up my unit and an upgraded one comes out. lol I didn't think I would have this upgrading problem with phase components!

[damnnit]

Haha same with me, I was waiting on the NF10FX but compressor never came in until work on my unit began. Old compressor is still extremely good, he did mention his tuning got 2-4c higher than when he first began so thats good to know.

[Seufari]

How long is the average build time on one of these units?

[LvMike]

Jinu117 looks like your prepping for 4core chips...

I was wondering if you could pm me with some prices... I am tring to decided if i go it alone, or invest in one of your fantastic looking systems.


LVmike

[Weee]

Jinu117 looks like your prepping for 4core chips...

I was wondering if you could pm me with some prices... I am tring to decided if i go it alone, or invest in one of your fantastic looking systems.


LVmike

All I can say is this, if you want to make something of the same quality as what he can make for you, it's going to cost you more money and SIGNIFICANTLY more time. If you have the budget to buy a unit from Jin, do it! I did, and I was very happy.

[[cTx]Philosophy]

I was also wondering the time it takes for an order to come to light..

Like If I was to order one the begining of the month when could I expect delivery?

Also again Id like to say nice work to Jin, these things are beauties and perform awesome dude, it this full time for you now?

[pablo]

i think its like 3 weeks but your gonna have to talk to him.

[damnnit]

I went over to pick up my unit on the 1st. He seems to be load testing many units with much more orders hanging on the wall. May take a little bit of waiting time but its worth it.

[dominick32]

Guys. Jins time is WELL WORTH THE WAIT... What he is doing is constantly and viciously load cycling and testing your unit before it leaves the facility. I know first hand that Jin is a perfectionist and you guys are going to love your units.

Regards,
Dom

[jinu117]

Well, couple of sad news...
NF10FX has proven itself to be quite capable but falls short of noise level criteria I have set myself for. While better than NF11FX by small margin, it isn't miracle maker.
And my idea of don't sell what I wouldn't use goes straight against this unless I can explain every single customer about the noise level in detail which is extremely hard to quantify as you have noticed on people asking and people answering.

It typically is 3 weeks for me to build, test, test, test, and some more... and pack it and ship the unit out

Guess2098 picked up one of unit tuned for NF10FX with high capacity in mind and also working with noise envelop I have put myself on (well very close to it at least... certain tuning change had to be done) so I will hear back from him on how he feels. But for now, I have returned ALL nf10fx I had in stock, as well as what was in order pipeline (other than already brazed up compressors).

I will go back to NF9FX for vapochill modification unless someone specifically asks for NF10FX, etc.

As for my own creation, some change to plan was in order once I found this out and I do have new compressors showing up tomorrow that was in order pipeline for nearly what seems like 2 months already
I really didn't feel like posting back on some questions here until I got everything settled down with firm idea of what will happen

Thanks for all my current customers who understood the inherent issues with NF10FX, supporting me, etc.

One thing I am 90% sure is, the extra wait you might have due to this compressor mishap or just plain change of compressor would be newer generation of compressor with higher efficiency and better noise/power draw/performance ratio It's been my goal for last one year. (and I will subsidize the cost difference on this new compressor which is tad more expensive than either NF9 or NF10 for current orders)

One bad thing would be my vapochill modification while being superior to premodded system might not be quite as good as what I can custom build as this compressor probably won't fit vapochill LS's case.

[renityren]

Jin,

Have you written up any technical documentation for your units?



Just how busy are and / or backlogged are you?
Would it be better to communicate by way of e-mail? IMing through a forum drives me nuts--I will miss the and the peculiar obsession XS members have with cats.

Cheers.

[guess2098]

Well, couple of sad news...
NF10FX has proven itself to be quite capable but falls short of noise level criteria I have set myself for. While better than NF11FX by small margin, it isn't miracle maker.
And my idea of don't sell what I wouldn't use goes straight against this unless I can explain every single customer about the noise level in detail which is extremely hard to quantify as you have noticed on people asking and people answering.

It typically is 3 weeks for me to build, test, test, test, and some more... and pack it and ship the unit out

Guess2098 picked up one of unit tuned for NF10FX with high capacity in mind and also working with noise envelop I have put myself on (well very close to it at least... certain tuning change had to be done) so I will hear back from him on how he feels. But for now, I have returned ALL nf10fx I had in stock, as well as what was in order pipeline (other than already brazed up compressors).

I will go back to NF9FX for vapochill modification unless someone specifically asks for NF10FX, etc.

As for my own creation, some change to plan was in order once I found this out and I do have new compressors showing up tomorrow that was in order pipeline for nearly what seems like 2 months already
I really didn't feel like posting back on some questions here until I got everything settled down with firm idea of what will happen

Thanks for all my current customers who understood the inherent issues with NF10FX, supporting me, etc.

One thing I am 90% sure is, the extra wait you might have due to this compressor mishap or just plain change of compressor would be newer generation of compressor with higher efficiency and better noise/power draw/performance ratio It's been my goal for last one year. (and I will subsidize the cost difference on this new compressor which is tad more expensive than either NF9 or NF10 for current orders)

One bad thing would be my vapochill modification while being superior to premodded system might not be quite as good as what I can custom build as this compressor probably won't fit vapochill LS's case.

had it on for few hrs. 1st hr is ok and after 2 hrs i am geting really tired of the noise of the SS....
however it is pretty good SS, i will do more testing tonight
thank you jin

[jinu117]

What kind of technical documentation? I am sure it is placed all over the place on information that is necessary but organizing it in one place probably is good idea and about time to do it
How busy I am? Well, 3-4 hours of sleep a day busy? Backlogged? Normally not until incidents like this happens. The time frame is setup so that I can do enough testing for each unit and it isn't like I only work on one unit at a time. Usually things happen in parallel here
If you have really problem using PM to communicate to me feel free to drop me an e-mail. It really isn't much different. Sometimes, MSN messanger might be better choice or even phone call to discuss more throughly as well as having no lag time between communication. And no I have no obsession with Cats... one of few who doesn't it seems here.
Now, Denny... please let me know of your finding and noise level. I would appreciate your feed back for sure. It is possible to turn the fan down with simple in line 12v fan controller which will reduce noise greatly for sure at cost of temperature on evap under heavy load. However, I am more worried about inherent compressor noise on that unit even after that major surgery I did on it

Sincerely, Jin

[wdrzal]

use a sound absorbing material to line your cases, too quieten down things even further use fiberglass duct board and turning vanes to channel air flow thru case, mount fan & compressor inboard of the ducts bends that need to be about 90 degrees so the sound is absorbed and does not continue on.2 -90 degree bends together is even better.in a square 90 degree turn, air turning vanes need to be installed or it will be like the air hitting a wall and cfm flow will suffer.

don't worry guys this board is made for A/C ducts and the fiberglass is coated so none will get airborne.It may not offer enough room left as it is 1" thick. there is thinner material but ducts or diversion plates would need to be made of tin then sound asorbing material glued on is another solution.

The Fan doesn't have to be mounted Right on the condenser, as long as the case stays closed. I would draw the air in the bottom,turn it pass over compressor and then thru condenser which can be angled then exhaust out the back,somehow. you will need to put unit on 3/4 to 1 inch legs or make bends in the case floor to hold up off floor so air can be drawn in bottom. you may need to think about provisions to raise another 3/4" if you sell one to a hippy who has long mat shag rugs.

Air in a home that is not blown about by fans or A/c will stratisfy about 1 degree per foot.So if you have 8 foot ceilings and no fans running the air @ 70F @ floor level will be approximately 8 degree's warmer @ the ceiling. So it makes sense to pull air from the floor. and the exhaust should be out the back ,to exhaust heat and noise away from user.

when I get time I'll draw and post a possible design to reduce noise greatly

If someone gives me case dimensions and compressor and condenser dimensions I'll see what will fit best. oh give fan dimmensions and cfm so a proper static pressure can be calculated for proper condensation of refrigerant and cooling. may take me a little while till I find time..............

[jinu117]

Thanks Walt
I am quite well versed on sound insulation believe it or not. However, what i believe in is not to mask it but to kill it from source as much as you can. In fact, you can find great deal of adventures and extremeties of how far people will go on silentpcreview.com... It is probably one most authorative place for noise related issues in computing world just like XS is for overclocking. Of course basic concept applies either way.
Increased efficiency = less wast such as sound energy/heat/etc.
Maybe you are talking about some ideas like this...

High NRC foam lined up, MDF to kill any vibrational and low frequency noise, Air intake right from bottom moving over to top back and it has shrouded exhaust with more noise reduction path. This one is about 2 year old contracept I did when I was playing VERY heavy watercooling for highest level of watercooling with lowest noise possible... (this is when I got tired of phase and peltier, etc)

[wdrzal]

actually you can't do much about compressor and fan noise unless you use the bends in duct to absorb it and keep it from being exhausted. The bends will do way more than simple adding sound absorbing insulation to the case interior.

The only thing more advanced is ANC (active noise cancellation) where a microphone listens to the frequency of the noise and a controller produces a wave 180 degrees out of phase and sends it to a speaker so the waves cancel each other out.This must all happens incredibly fast.

Bose was the inventor of this technology about 16 years ago iirc for there aviation headsets. they now are on there second generation,generation X ,I have a set and they work incredibly well but a little pricey at 1000.00 a set,but they would fit well over piston compressors to reduce noise http://www.bose.com/controller?event...set_index&ck=0

This technology in now available for industrial applications but only works in enclosed spaces such as a case.I seen the technology used in gen-sets on yachts & aircraft to quieten them down.It would work in phase cases if you want to pony up some big bucks.

[mele]

Nice work!!!

I wish one of these units would fit in my SSTJ07!!!

Mel

[jinu117]

Well new finding for the day. I need you guys with quad core in hand. Apparently I finally found CPU that can dump more than what I consider 200w!!!!
A local customer told me unit not holding load his quad core. Think he was trying for 4.2 or 4.3 ghz or something. He came in and I load tested it on tester at 200w... holding absolutely fine the way it was tuned from start. It seems that these CPUs with heavy voltage are capable of dumping well over real 200w. I've retuned his unit to hold 220-230w range losing huge amount on idle temp, etc. He will report back to me once he gets back home to see how much voltage he can put in, etc.
I've also heard report of same quad core cpu (well diff cpu but same criteria of chip) dumping hugely different amount of heat... So, all bets are off right now on how I should tune my future unit. Get great temp with current dual core with some room for quad core? Get good temp for HUGE load capacity sacrificing some nice temps we are used to seeing up to 170w?
This is going to be interesting development.
Do we just prepare for much more heat than reasonable? Do we stop at certain point saying, don't put too much voltage on it for 24x7 type use?

[elsupremo]

Jin - I suspected this would be the case. My friend (who was just at your house) is currently driving over to my house and we will probably do some testing on his unit together. Initially, we were both surprised to see how easily the quad was overcoming the unit as previously tuned. We are both excited to see how the unit performs now.

He gave me some details of his conversation with you about temps vs. load capacity, and we were happy that our assumptions prior to today's meeting with you were correct. Neither of us could care less about temps, as long as it is not maxing out the compressor and we are still keeping the cpu nice and cold.

We were getting a little concerned about the value of the unit when used with quad, as some people with big air coolers (which are not precisely tuned at all) were matching or surpassing his overclock. Hopefully we'll be able to meet what we think are reasonable goals for a quad under expensive phase cooling. We aren't trying for benchmark records or anything. We'll let you know how it goes.

Thanks for your great support Jin. To any buyers in waiting, rest assured that Jin sticks with you after the sale. Great service and support.

[hdinmotion]

Hey Jin Thanks for your awesome support!

I did just a little testing on the quad, but still need more time to do more.

It looks like I can set 1.60 volts (1.56 actual) and be close to the max I can go for the compressor.

At 1.63 (1.59 actual) volts the heat takes over the unit. It takes about 10 min in prime for it to happen.

Oh and my clock speed for these tests were at 3.9ghz.

If it were me I would sacrifice temps for higher load capacity.

[jinu117]

This bugged my mind so much that I thought over this:
Here is quick ideas I drew up.
Core Duo Xtreme (assuming 1.75v high overclock vs 1.35v stock) and overclocked: Approximately 170-180w range.
Kentsfield (1.6v high overclock vs 1.35v stock)... based on calculation... about 250w range -_-;
Considering programs probably don't stress quad core efficiently enough (even for orthos, etc), I am guessing maybe even 220-230w range we saw might explain some idea on quad core's evil heat dump when overclocked.
Basically, same overclock with same voltage on quad will be approximately twice the heatload vs core duo. This is the kind of heatload we didn't deal with before at all.
I am almost tempted to make 2 lines one for dual core one for quad core as the nice -40c loaded temp will not be practical shortly for 24x7 setup.
I can get all the help from what you guys feel on which direction to go.

[MrDeeds]

Thats actually a very good idea jin. Theres no reason to sacrifice temps for customers that have no intention of running quads. Im just curious as to how much different idle and load temps would be on the unit tuned for a quadcore.

[rjw]

Here is what I found with my quad cores and vapochill units (stock)

On an evga 680i board, running 4.3Ghz @1.8 volts (which is reading as 1.74) the unit holds ok, with normal rise in temps when running benchmarks.

same cpu on asus p5b32-e sli, anything over 1.62 volts and the cpu drives the vapochill up to extreme plus side temps.

Lower the voltage just a little and vapochill keeps the quad core about on par with the evga 680i setup.

something doesn't seem to add up. either the evga or the asu mb is reporting a wrong voltage.

Possible theory:

around 1.6v on a quad core is probably going to get you as high an overclock as possible with a ss unit.

If you need more, then you probably need a cascade or ln2, etc.

I have run my quads between 1.5 and 1.74 and at 4.3Ghz it's a crap shoot as to whether it will lock up or not.

vapochill seems to hold the quad at around -24C (temp probe at end of evap touching cpu). Under fairly heavy load, between 6 and 12C. Severe load = 22C.

So as you can see, the vapo isn't making it and is marginal at best

Hoping Jin can keep me on the minus side with the unit he is building me!!!

I think that with a stronger ss unit, several things may happen

1) I will be able to stay at or slightly higher then 4.3Ghz and

2) I will be able to do this with less voltage, which will help the ss unit keep the quad under control a little better.

On a side note: If I do not wait for the vapochill to get the quad well below 0c before powering up, the quad overpowers the unit and temps are in the 40's (positive), even on the vapochill unit.

[elsupremo]

Rjw, I have a hard time believing you are able to put more than 1.6v (ACTUAL) through the quad CPU and have it be stable while Orthos priming (100% load over a considerable time period) with a stock Vapochill unit. It seems that you would be overpowering the Vapochill almost immediately...no?

The reason for my skepticism is that on hdinmotion's unit (using evga 680i), which has been proven up to Jin's "heavy-load" tuning, anything above 1.56v (ACTUAL) overwhelms the unit during Orthos testing within about 10 minutes.

[rjw]

I don't think that the voltage readings on MY evga are correct

I hit 22099 in 3dm06 in sli with a quadcore on an evga 680i running a 15 x 1147/4 or 4.3 Hgz

orthos runs for about 30 minutes at up to 22 C, which to me is scary.

Not all cpu's are created equal

I bought 2 at the same time, same batch. The second one would send temps through the roof in orthos.

Is there a quad core version of orthos?Or maybe I can run 2 sessions from sifferent folders, in order to load all 4 cores?