PDA

View Full Version : mitsubishi 1.5 ton split system install







zeke2u
21-10-2012, 12:45 PM
I recently purchased an 18,000 BTU Mitsubishi heat pump which I'm about to install. I've never installed any AC or refrigeration before, but I did once work as a design engineer for a commercial HVAC firm. Although I teach mathematics at a local university, I've also worked in manufacturing where we built very large coil winding machines with complicated hydraulics and consider myself to have excellent mechanical/machining capabilities.

My question is with the purging of my system. My piping (3/8 & 5/8") only has connections at the evaporator and condenser and the pipe length is relatively short (25') with only 5 curves (5" radius). I picked-up a relatively small vacuum pump (Welch 8890) on eBay, which will need to have its' 1/4" NPT fitting changed. I think it is rated at a little over 1 CFM, but is a very good quality pump.

Seeing as the pump size is relatively small, should I change the NPT tapered hose fitting that is on the Welch to a 1/4" SAE, which would match the hoses with my new manifold/gauges, or should I go larger, say 3/8" for the hose from vacuum to the manifold? Also, if I do go with a larger diameter hose from the pump to the manifold, should I also use a larger (3/8") hose from the manifold to the service port on the gas side?

SeanB
21-10-2012, 01:10 PM
You can change to the1/4 SAE, and it will take longer to get down to a good vacuum. Change the oil in the pump before you start, and if it is dark then run it closed for a half hour and change it again, so that you start with clean oil and a clean pump inside. Using a larger pipe will make it faster to vacuum down, but with a small pump a short pipe connecting it to the gauges and the outdoor unit will work. Just make sure you get a good vacuum, and you probably will need to add refrigerant to the unit as the pipe run is longer than what the prefilled charge is meant to do. You will need a canister of the right refrigerant to do the top up, and the installation manual will tell how much mass wise you will need to add.

What refrigerant is it designed to use?

install monkey
21-10-2012, 01:34 PM
dont bother changing the size of the fitting on the unit, as the schraeder core will hinder the vac, with a 25ft piperun then theres no point in modifying anything, i assume this has been pressure tested with nitrogen, as although ur pipework has no brazed connections the flared connections are prone to leak

zeke2u
21-10-2012, 05:13 PM
Well SeanB and install monkey, I was hoping that I wouldn't have to change the oil in the pump as the oil that is recommended is very expensive ($90/l). And regarding the charging, the Mitsubishi installation manual says that the charged condenser/compressor is good for a pipe run of 33'. I was hoping that with very careful flaring technique and proper torquing, I could be fairly confident of the fittings not leaking - provided that it held a vacuum for a decent amount of time. However, I'm aware that the R410A operates under higher pressures than the older refrigerants. I haven't priced them, but would it be feasible and safer to use the Serto SO 40221 compression fittings?

install monkey
21-10-2012, 05:18 PM
r410a flares should be at an angle of 45deg formed with a rolling flare tool, if your vac pump is new then theres no need to change the oil, as for any compression type fittings, thats a no no unless they are specified fit for purpose on ac equipment- 410a kit can run upto 35bar in heating mode.- im not familiar with the serco fittings.
a system that holds a 30 inch vac is equivelent to a 2 bar pressure test, so youve got a 50/50 chance of whether theyre ok.
might be worth getting a bottle of oxygen free nitrogen to prove all is ok, due to if ur flares leak then how much is a bottle of 410a? - will upload some tips

zeke2u
21-10-2012, 05:26 PM
Here's a link to the Serto compressions (http://www.serto.net/SAEserto1.htm) Thanks in advance for any tips!

install monkey
21-10-2012, 05:27 PM
http://www.mediafire.com/view/?6mvriuboj1cc4ob
http://www.mediafire.com/view/?9waxuln28ca1vl5

zeke2u
22-10-2012, 02:35 PM
Thanks for those Daikin downloads. I'll probably have more questions after I finish reading them! My vacuum pump is not new, so I'll buy a liter of oil and change (DirectTorr Premium or DirectTorr Gold). If I go with the Premium grade oil, the pump (it's a direct drive, rotary vaned, and what's really important to me, made in the US!) is guaranteed to achieve .1 Torr (1.3 mbar). The DirectTorr Gold oil, which has even lower vapor pressure, will give lower Torr values but I'm not sure how much lower. The DirectTorr Gold is twice the cost of the DirectTorr. Anyway, I have not installed the piping yet, and will need to devote some time to my "regular" job, but hope to have things ready for testing (I'll try to find someone who has a tank of N2) by next weekend.

zeke2u
04-11-2012, 12:22 PM
Well, my system is up and running. After replacing the oil in the Welch pump I vacuumed my lines for ~ 3 hours (~16C ambient). It took less than a minute for my gauges to hit -30". I was very careful forming and measuring all the flares, and used POE oil on all the joints (4). After opening both (liquid & gas) stop valves, I checked all the flare fittings with soap-water. I noticed no bubbles from any of the four joints. I do not have any sophisticated meters to check for leaks. I'm hoping my less-than-sophisticated "leak-test" was not a false positive result and that I wont have to worry two months down the road about recharging a leaking system! Perhaps I'll hire someone in the trade to check the piping and remove all my doubts - I'm actually rather skeptical, by nature! Nevertheless, for the past 3 days the Mitsubishi seems to work very well.

install monkey
04-11-2012, 02:16 PM
don't bother getting someone to check the piping- soapy water can be alkali based and cause problems- cheap option using aerosol based leak spray- run the system on full heat, full cooling and record running pressures, coil temps,air on/off ambient temps inside /outside- running currents- will give you an idea if it loses performance

zeke2u
04-11-2012, 05:57 PM
Thanks again, install monkey - I'll rinse-off any soap/water residue, and if performance problems do arise, I'll heed your advice, measure pressures, temperatures, and currents and report back here.

hkj
17-11-2012, 12:11 PM
don't bother getting someone to check the piping- soapy water can be alkali based and cause problems- cheap option using aerosol based leak spray- run the system on full heat, full cooling and record running pressures, coil temps,air on/off ambient temps inside /outside- running currents- will give you an idea if it loses performance

There is no flare joint ,which will not leak!

Acc- to EU-directives max. 5g/annum is the allowed leak rate for F-gas systems (minisplits).

Then,how does your system "tolerate" annual leak ? Capillary systems are very sensitive ,TEV/EEV -systems
do not show the COP/Output deteriation that easily! Even 50% of the refrigerant could be lost without serious symptoms in the performance ... (originally ca. 1000g r410a is a common charge in the beginning).

http://ecaaser3.ecaa.ntu.edu.tw/weifang/hp/The%20effects%20of%20improper%20refrigerant%20charge%20on%20the%20performance%20of%20a%20heat%20pump %20with%20an%20electronic%20expansion%20valve%20and%20capillary%20tube.pdf (http://ecaaser3.ecaa.ntu.edu.tw/weifang/hp/The%20effects%20of%20improper%20refrigerant%20charge%20on%20the%20performance%20of%20a%20heat%20pump %20with%20an%20electronic%20expansion%20valve%20and%20capillary%20tube.pdf)

The basic question remains:

How much is much and whether small (leak) is tiny enough and worth to be accepted?

A leak of max. 10g/a is a rate of utmost challenge.
But you would not ever find that out by pressure/soap test .


How many "bubbles" is that 10g/annum?

Lets assume a H2O-droplet is 50mg=> 20 droplets is 1g (= 1ml by volume).
How many bubbles ( sized water droplet volume) your system will tolerate during the common
pressure-testing (N2,max. 30bar/30 min.) procedure?

=> 10g/annum as a leak corresponds 0,15ml F-gas/ per.30minutes r410a/ 4x flare joints.
=> 0,15ml F-gas leak shows as 7 bubbles sized a H2O-droplet in a Fairy-test provided the bubbles will not explode,dry out by their surface or dissolve into water as a slightly polaric medium.

Does somebody really expext to discover a leak rate of 10g/a by pressure/Fairy-test ?

I would not!

I would notice 50g/annum ! (= tiny leak).
=> 500g /10 year as a leak=> disaster for capillary system but not that serious for TEV/EEV-system!

I would personally rely more on the SERTO=Armacell connectors in this context (leak).

http://www.armacell.com/www/armacell/INETFAQ.nsf/ID/0DF09634C55E97EF80257117005730FB?OpenDocument

zeke2u
30-11-2012, 03:02 PM
Thanks for that post hkj - and your point about there being no flare-joints without leaks (it's more a question of the RATE of leak) is well-taken. My Mitsubishi seems to be performing well so far. I should be receiving my electric bill any day and will have a better idea of the power it consumed over the past 30 days. Ambient temperatures have been in the 40's-50's during the days, 20's-30's at night, and my living area (~1400 sq. ft) has been kept ~21C. However, do you think it would make sense to replace the flares with the Serto - perhaps next spring while my system will be no longer heating?

install monkey
30-11-2012, 08:19 PM
a 1.5 ton split will contain less than 3kg of refrigerant so its exempt from f gas
There is no flare joint ,which will not leak!

Acc- to EU-directives max. 5g/annum is the allowed leak rate for F-gas systems (minisplits).

Then,how does your system "tolerate" annual leak ? Capillary systems are very sensitive ,TEV/EEV -systems
do not show the COP/Output deteriation that easily! Even 50% of the refrigerant could be lost without serious symptoms in the performance ... (originally ca. 1000g r410a is a common charge in the beginning).

http://ecaaser3.ecaa.ntu.edu.tw/weifang/hp/The%20effects%20of%20improper%20refrigerant%20charge%20on%20the%20performance%20of%20a%20heat%20pump %20with%20an%20electronic%20expansion%20valve%20and%20capillary%20tube.pdf (http://ecaaser3.ecaa.ntu.edu.tw/weifang/hp/The%20effects%20of%20improper%20refrigerant%20charge%20on%20the%20performance%20of%20a%20heat%20pump %20with%20an%20electronic%20expansion%20valve%20and%20capillary%20tube.pdf)

The basic question remains:

How much is much and whether small (leak) is tiny enough and worth to be accepted?

A leak of max. 10g/a is a rate of utmost challenge.
But you would not ever find that out by pressure/soap test .


How many "bubbles" is that 10g/annum?

Lets assume a H2O-droplet is 50mg=> 20 droplets is 1g (= 1ml by volume).
How many bubbles ( sized water droplet volume) your system will tolerate during the common
pressure-testing (N2,max. 30bar/30 min.) procedure?

=> 10g/annum as a leak corresponds 0,15ml F-gas/ per.30minutes r410a/ 4x flare joints.
=> 0,15ml F-gas leak shows as 7 bubbles sized a H2O-droplet in a Fairy-test provided the bubbles will not explode,dry out by their surface or dissolve into water as a slightly polaric medium.

Does somebody really expext to discover a leak rate of 10g/a by pressure/Fairy-test ?

I would not!

I would notice 50g/annum ! (= tiny leak).
=> 500g /10 year as a leak=> disaster for capillary system but not that serious for TEV/EEV-system!

I would personally rely more on the SERTO=Armacell connectors in this context (leak).

http://www.armacell.com/www/armacell/INETFAQ.nsf/ID/0DF09634C55E97EF80257117005730FB?OpenDocument

zeke2u
18-12-2012, 03:38 PM
OK - I'm not sure about the implications of this. My question remains, should I replace the flares w/Sertos this spring? And, what is f gas? Anyway, my 30-day electric bill arrived and it appears the pumps energy usage was ~700 kWh, which added about $100 to my monthly bill. Average heating degree days for the month were ~850.
If I average a little more through December and January, say $150/month, then I'll be very happy. In the same amount of time I would probably use ~$700 of oil, @$3.50/gallon. What I am even more pleased with is my house is not only warmer (I used to set the thermostat at 60 F, it's now at 70 F) but much cleaner/less dusty. The oil-fired hot-air system made my house more like a coal pit.

install monkey
18-12-2012, 07:40 PM
to satisfy the f gas police- 130 should get you an electronic refrigerant sniffer- to detect any signs of oil from your unpressured tested flares- if theres no leaks then don't bother replacing your flares. with a flare joint its 50/50 chance of being ok- not bubbling when sprayed with leak detector spray- then you have a 50/50 chance of it failing over time- as for being energy efficient then its going to lower your heating costs- also keep the cond coil clean from debris.

zeke2u
06-03-2013, 12:44 PM
A quick update - My 1.5 ton Mitsubishi has performed flawlessly, since being installed last November. I'm not sure of how many degree days we've had, here in central Massachusetts, and I can't quantify exactly how many kilowatt hours the pump contributed to my electric bill. My bills for January and February ran about $100-$110 more than they did last year. With heating oil prices close to $4/gal, I figure that the Mitsubishi will fully amortize in about 4 years.

I can't get over the fact that this thing has been running, continuously, for four months without a hiccup. Well, a few nights in January, when the temperatures went below -10 F, the indoor exchanger fan was working noticeably harder, but my oil-furnace only kicked-in if the wind was really stiff, where the convection losses of the house were probably too great for the heat-pump to overcome. I can't say enough about how impressed I am with the new compressor and control technology that is integrated into these units. I was producing heat in temperatures I would have never expected 10 or 20 years ago....

install monkey
06-03-2013, 07:39 PM
nice 1- bout time you cleaned the filters,cleaned outdoor coil with a hose pipe- not pressure washer, buy some evaporator coil cleaner and spray the indoor coil,clean any dust off return air sensor-take fornt cover off indoor carefully and look for the thermistor near the pcb- maybe check electrical connections are still tight :D