NOTE: The views expressed herein are those of the author and not necessarily those of Overclockers.com.
So you just cracked your piggy bank after years of contemplating and saving up for phase
change cooling – the ultimate in extreme 24/7 overclocking. You splurge and buy that
gorgeous Vapochill Lightspeed you have been eyeing in the classifieds. You turn on your rig
for the first time and clock up that brand spankin’ new Conroe to 4 GHz without breaking a
sweat. And to top it all off, you’re dual prime stable for 24 hours at over 4 GHz. You then
kick back your feet and stare at the clear Liquid Crystal Display on the Vapochill reading
-47ºC and say to yourself:
That’s where Mr. Jin Park of Jin CPU Cooling based in Riverside, California, comes into
play. Jin is a master of phase change cooling and if you ask him, he is still only an
apprentice. There is much knowledge to be gained and experience to be had in this game. Jin
offers a service modifying your Vapochill LS or Mach units under his part-time full
production line in California.
The service consists of taking a stock Vapochill LS and
simply gutting and replacing the internals with more efficient and capable parts. This is
finalized with a professional re-gas with “r402a” refrigerant and a custom tune based on
the customer’s goals for the unit. He will follow the re-gas with extensive leak testing
and load cycling while making sure the unit is tuned for maximum usage of the compressor’s
Jin will start his conversation with potential clients by asking them simply: “What are your
goals with this unit? What kind of system will you be cooling?” He than compiles that data
and custom tunes the gas in the new compressor based on that spec.
For this unique and “one
of kind” service, be prepared to pay $375 for parts and labor. He is extremely honest and a
downright pleasure to deal with. His claims are not exaggerated and right away he explains
to potential clients that this mod will give you two benefits:
- It will decrease noise substantially over the stock Vapochill LS;
- It will decrease all around temperature by 5 – 15ºC depending on how the new
Before we get into details about our Vapochill LS modification, I would like to disclose something that you do not see too often in business today – extra special care and tedious attention to detail when packaging your product. From start to finish, the entire Vapochill LS modification takes approximately 2 to 3 weeks to complete.
Jin ships DHL Insured and shipped out our unit from California on Friday. We received the unit surprisingly fast on the following Monday morning. Upon receipt of your modded Vapochill, you will be shocked at how great it is packed. Jin uses a unique (and expensive) system called Instapak to ship all of his products. It is a form fit packing system that fills every box with the exact amount of Styrofoam. Not only does he use Instapak, but he double boxes and includes Instapak in both of them.
Here is our Vapochill upon receipt Original large box with Instapak material:
Here you can see Instapak within the first box!
And finally, after more Instapak material we can finally start to see our shiny new modded Vapo!
Our stock Vapochill LS was a tad less efficient compared to Asetek’s
general performance spec measured at the evaporator, so a “Jin Mod” overhaul was definitely
in line for this unit. To be fair to the readers and show a more realistic performance gain
from Jin’s mod, I chose to use Asetek’s evaporator-probed data on a stock Vapochill LS Heat
Capacity in wattage.
This data is ultimately what every Vapochill LS should produce
straight from the manufacturer in Denmark. Just for the record, our Vapochill test unit
performed a few degrees worse across the board compared to Asetek’s spec, so releasing data
from our test unit would give skewed “over promising” results. The results in this article will give you a real idea of what to expect from a “Jin Mod” based on Asetek’s ideal specification with close to
perfect ambient temperature.
The stock Vapochill LS was also a tad noisy. This is due to the stock compressor
that was actually louder than the two 120 mm fans inside the unit. Jin claims the new
compressor will reduce noise substantially over the original unit. He was right – there is
no longer a buzzing in my ears long after leaving the room where the stock Vapochill
was located. Now the fans actually drown out the noise from the compressor.
** Just a
little side note for Performance Junkies: During a Jin Custom Build on a new unit, he offers
the ability to use an even larger compressor for more performance and colder temps.
However, this will put your unit right back to the same decibel level, if not louder than
the stock unit. ** The Vapochill LS “Jin Mod” will decrease noise level while increasing
performance and capacity.
Our stock Vapochill LS was able to hold load on an E6600 Conroe processor for about 15
hours at 4.14 GHz and 1.62 vcore. But it was apparent that more MHz could be had on a 24/7
overclock because at 4.2 GHz, Ortho’s failed on Core 0 at 3½ hours. With a max suicide clock
of 4.41 GHz, we definitely blame the 4.2 GHz 24/7 Ortho’s failure on heat, which should
easily be cured with our Vapo “Jin Mod”.
With probed readings from the evaporator, our Vapochill “Jin Mod” decreased
temperatures unloaded @ 0 watts by 13ºC:
At 70 watts, the modded unit held 13 more degrees over the stock Vapochill LS:
At 150 watts, of load the “Jin Mod” unit held 10 more degrees over the stock
And finally at 180 watts of load, the modded unit held 7 more degrees over the
Upon initial testing of our modded unit, a
vcore selection of 1.4 and 3.0 GHz on the E6700 Conroe processor produced BIOS
CPU Idle temperatures of -44ºC, so expect to subtract on average about 10ºC to
calculate your evaporator temp:
As you can see in the following chart, the “Jin
Mod” Vapochill LS decreases temps across the board by an average of almost 11 degrees over the stock Vapochill LS. Considering Jin’s very honest statement of a “5 to 15 degree decrease” in temperature, I think it’s safe to say that we are very happy with our modded unit, especially when temps on our
unit were even higher than the Asetek spec.
- E6700 Conroe Core 2 Duo Processor (L627A) Week 27
- OCZ Titanium Alpha VX2 DDR2-1000 (2 x 1 GB) 4-4-4-12 @ 2.35v
- P5B-Deluxe Wifi w/711 Bios and Vcore and Vmch volt mods
- Silverstone ST56ZF PSU
Now that we are aware temperatures are much
better and the unit is much more efficient over the stock Vapochill LS, we
decided to give our Conroe E6700 a shot for the gold 4 Ghz mark. This was easily
attained without even breaking a sweat. We ran SuperPi32m for just about 3
minutes to see how the compressor held load at 1.65 vcore before really trying to
bench the processor for high numbers:
Our next step was to try for a maximum
overclock on 1.65 vcore. This vcore is the so called "under the table" maximum
recommended voltage on these processors if you are extreme cooling with phase
change, dry ice, or liquid nitrogen. The chip maxed out at 4.49 GHz at
1.65 vcore, just a hair away from 4.5 GHz.
Being the overclockers at heart that we really
are, we decided to push this chip to its absolute maximum and see theoretically
how well it scaled on voltage. Luckily our P5B-Deluxe motherboard was already
modded for a little extra juice on Vcore. We slowly turned the resistance on the
potentiometer until we approached an exact 1.85 volts on the CPU. We than booted
Windows at a staggering 4.61 GHz, an almost 75% overclock.
Extreme overclocking results like this just
continue to confirm how happy we are with the modded unit. The unit is now
substantially less noisy and much more efficient.
Jin Park is a 33 year old full time DBA/Systems Integrator and part time phase change builder. We went ahead and asked Jin the following questions:
- Where are you currently?
- Where do you plan to be in 5 years?
- What do you do exactly in your product/service orders?
- Who are you and how did you get started in Phase Change cooling?
I will give Jin’s unedited response below:
“I have been using single stage units for over 3 years now. The first Mach I unit was repaired and regassed by a local builder (I eventually learned not to buy used phase units over the internet) who doesn’t build anymore. My second unit was a modified Mach II unit by Baker (whom is a very well known historic figure on phase change building) but rather recently has disappeared from the scene.
A year and half ago, I have decided to embark on phase change cooling again. I asked a local builder to make a custom single stage unit inside my Lian-Li V2000 and the end result was not satisfactory. The noise level and cooling performance as well as the customer service just wasn’t what I was expecting. I felt stumped, frustrated and tried to look for some other options. Being rather stranded (no, I didn’t want a regular phase change unit that sells out there when I know there can be a colder/quieter alternative), I decided to make one myself.
Eventually, after spending quite a bit of time and money, I achieved my original goal of making a single stage unit that is colder and quieter than units that are out there. At this point, I got an idea that there are people like me out there stranded. People who want colder temperatures, people who want a quieter phase change cooling unit, and people who want someone reliable to do it for them. I carefully looked into my flexible savings I had, wrote out a business plan and than decided to offer the service to people who are just like me.
A lot of time has gone into finding the correct components, packing/shipping procedure, design and testing based on a few ideas I had on such a plan. And even more money went into acquiring proper tools, getting some prototype jobs done, etc.
Now, let’s talk about how my units are made in general (some obvious steps will be skipped):
- Order comes in;
- Evaporator is brazed (this is part where it contacts CPU). It is to be leak tested for at least 2-3 days under 250 psi;
- Evaporator is lapped, then cleaned. Flexible suction line gets brazed;
- Rest of phase change circuit (copper tubing) are prepared and cut, bent (lot of modification of pre-existing models gets desuperheater at this point). In case of modification of the unit, it means reclaiming gas used inside to a proper recovery tank, then filled with nitrogen at this phase to insure that no contaminant can get into unit, as well as making it legally safe;
- Now, all loops are joined in brazed (while purging with nitrogen gas). Purging while brazing is an important process as oxidization inside the circuit can leave carbon deposits which can result in unpredictable behavior;
- Once brazing is complete, it will be charged with 175 psi of nitrogen to be leak tested for 2-3 days minimum;
- When the system passes this test, it will be triple evacuated (getting rid of all nitrogen, moisture and air, which can change performance as well as longevity of the unit) down to below 40 microns, from a micron gauge. Unfortunately, due to time consumed on doing this (takes 4-24 hours depending on condition of the system), I find lot of people don’t enforce doing this;
- Once evacuation is done, I will charge the system with the refrigerant and start tuning the unit for load using my load tester. A typical unit is tuned for 165 – 175 watts best temp with holding capacity into 200 watts. I use a copper block load tester with a heat cartridge inserted which seems to give a very large amount of heat vs. some other designs I have tried. Each wattage will run for few hours to insure that the compressor will reach its final balancing point (sometimes left over night);
- After initial load tuning is done, it is time to go through what I call thermal cycling (sort of like a 72 hour burn that was used for custom made computers).
It consists of turning on the unit, putting various loads to the unit, turning it off and waiting for it to cool down, then repeating the cycle. This helps me making sure that the unit to be shipped will be working when it arrives in the customer’s hand;
- Time for packing and shipping. I use something called Instapak to insure additional safety in a double box. So far, I’ve had great success using this shipping method with no real failure due to shipping related damage.
There seems to be an excessive amount of time spent on tuning, and testing for leaks in the system, but it is my firm belief that spending those extra hours testing prevents any problems leaving my shop. This would ensure that both the customer and myself are very happy as an end result. (As you can imagine, any RMA can be a nightmare in this type of business when you are talking about 45-60 lbs oversized package traveling through courier).
As a business, here are few rules I adhere to:
- Building a great customer relationship is crucial to success of my business
- Never rest on your laurels. The system must keep evolving to get better until it physically is impossible to do so in a manner that isn’t cost prohibitive. For this currently, I reinvest 100% of net income on tools, R&D, and stock of components.
- Always have open mind when dealing with customers. I love getting feedback for some “What if I had this”, “What would have been cool”, which only helps the quality and ease of use of units further down the line.
Where would I be in 5 years? Who really knows? But here is my current goal for year 2007:
- Improve product appearance, improve ease of use of system and reliability of electronics that go with the system, and of course performance of the system while reducing the costs of such improvements. Yes people, there are more new things coming in future, starting the end of this year.
- Once major production design is decided, start marketing the unit. Also comes the need for better CRM tool. I eventually plan to make a backend operated web site for all orders that come in so that customers can log in and see where their build process is going for custom jobs. For production models, I hope to have them pre-produced and ready to be shipped sometime in 2007.
There is more vertical and horizontal expansion of the business that I can see down the line from the year 2008 and on, but for now I would rather get my customer happy, myself happy, build my reputation as a solid business to deal with, and get toward establishing the company as the phase change cooling company that is “sought after” and “exceeding customer’s expectations”.
Sincerely, Jin Park. Jin CPU Cooling. Riverside, CA.”
I would like to thank
Overclockers.com for publishing this article and most importantly Jin Park for
allowing me to take the time to interview him and his company. Jin can be
reached on OCForums.com with his username: Jinu117 or at his company website vapoli.com.
I can be reached on OCForums with my username: Dominick32
I hope everyone enjoyed!