DIY mini-split install — 3 zones, 2 years in, honest assessment
By GridFreeGuy·6 replies·803 views
Installed a Mitsubishi MXZ-3C24NAHZ2 multi-zone system (24,000 BTU outdoor unit) with three 9,000 BTU indoor units myself two years ago. This is not a quick project — I logged about 40 hours of work over three weekends — but it's been completely reliable and the performance has exceeded what I expected.
Why mini-splits made sense for my situation
My house has no ductwork. Previous heat source was propane forced-air (with its own air handler in the crawlspace). Adding central ducting to the house would have cost $12,000–18,000 according to three HVAC quotes. The mini-split system I installed cost $3,800 in equipment and maybe $400 in materials (line sets, disconnect boxes, conduit, mounting hardware). It heats better, cools better, and cost $14,000 less.
What DIY means for a mini-split
The compressor and refrigerant lines in a pre-charged DIY system (like the Mr. Cool DIY series) are factory pre-charged — no EPA 608 certification needed, no recovery equipment, no vacuum pump for the refrigerant lines if you use a pre-charged unit. However, Mitsubishi, Daikin, and other commercial-grade units are NOT pre-charged. They require:
- A vacuum pump to evacuate the line set before opening the refrigerant valves
- A micron gauge to verify the line set holds vacuum
- The refrigerant is already in the outdoor unit; you're pulling vacuum on the lines and indoor coils, then opening valves to flood them
I used a two-stage vacuum pump ($120 refurbished) and a digital micron gauge ($65). The vacuum procedure takes about 45 minutes per zone. If the lines don't pull to below 500 microns and hold for 30 minutes, there's a leak to find and fix before opening the refrigerant valves.
Performance at cold temperatures
The Hyper-Heat designation on Mitsubishi's H2i series means rated heating down to -13°F (-25°C). My actual experience: at -10°F the system heats adequately but you can feel it working harder — the defrost cycles are more frequent (maybe every 40 minutes versus every 90 minutes at 20°F) and the discharge temperature is lower. I added baseboard electric resistance heaters in two rooms as emergency backup, which have run a total of about 8 hours in two Idaho winters. The mini-splits handle 95% of heating demand.
At summer temperatures my garage zone cools to 68°F from 95°F outdoor in about 35 minutes. The SEER ratings are real in my experience — my annual electricity for HVAC (heat + cool) is about 2,400 kWh for a 1,600 sq ft house, versus ~900 therms of propane previously ($1,800/year at $2.00/therm). My HVAC electricity cost is about $264/year at my rate. Net savings: ~$1,536/year.
What I'd flag for first-timers
The line set penetration through exterior walls is the fussiest part. You're drilling a 3-inch core hole, routing refrigerant lines + communication wire + condensate drain + electrical in a single sleeve, sealing around it against air infiltration and pests, and keeping the line set slope consistent so condensate drains properly. Plan this carefully before cutting. A slightly wrong hole location is painful to fix after the fact.
The communication wire (the multi-conductor signal cable between indoor and outdoor units) is often the source of commissioning errors. Mitsubishi uses a 3-wire S-Bus communication protocol. Polarity matters. If the indoor unit displays a communication fault, this wire is the first thing to check.
Permits: mini-split installation requires a permit in most jurisdictions. The electrical connection (240V 20A or 30A circuit per outdoor unit) definitely requires a permit. Some jurisdictions also require a refrigeration mechanic license for the refrigerant handling even on pre-charged lines. Check locally before you start.
The vacuum procedure note is important and often glossed over in DIY tutorials. 500 microns is the target, but best practice is to pull to 300 microns or below and hold for 30 minutes after closing the vacuum pump valve. If the micron gauge climbs from 300 to 500+ after you close the valve, you have a leak — find it before opening refrigerant valves. Leak check with nitrogen pressure before pulling vacuum if you want to be thorough: 150 PSI nitrogen, 15 minutes, then release and vacuum. This adds an hour but eliminates any question about leak-free lines.
The Mr. Cool DIY units are genuinely good for simpler single-zone applications. I installed a 24,000 BTU DIY unit in my detached garage/workshop and it was the easiest HVAC install I've done — literally plug the pre-charged lines together, mount the units, wire the disconnect. For a single zone with modest performance requirements they're hard to beat. For a whole-house multi-zone application with cold-climate performance needs, the Mitsubishi Hyper-Heat equipment is the better choice even though the install is more complex.
The permit and licensed contractor question for refrigerant handling is one I run into constantly in California. California requires an EPA 608 certification AND a California-specific contractor license (C-20 HVAC) for any work involving refrigerant. This means a purely DIY refrigerant installation — even on a pre-charged line set — is technically unpermitted if you're not licensed. Mr. Cool's pre-charged system is designed to be installer-friendly but in CA you still need a C-20 licensed contractor to pull the permit and sign off. For multi-zone commercial-grade equipment, the permit process reliably requires licensed installation.
The 2,400 kWh/year for heating + cooling a 1,600 sq ft house is a data point I'm going to save. I'm planning a similar system for a 1,400 sq ft house in Colorado and every estimate I see online varies wildly. What's your insulation situation — attic R-value, wall construction?
@PacketDriver 2x6 wall construction, R-21 in walls, R-49 attic, double-pane windows throughout. House was built in 1994 to good standards. If your house is older with 2x4 walls and R-11 insulation, you'd see higher energy consumption — I'd estimate 35–50% higher for the same climate zone. If you have a blower door test done (sometimes free through utilities) you'll know your actual infiltration rate and can model more accurately.
Two years is a solid data set — thanks for posting this. I just crossed 14 months on a similar setup (single 12k BTU Senville Leto in a 400 sq ft workshop, 24V/400Ah LFP bank + 800W solar). The inverter sizing question you raised is real. I originally had a 2000W pure sine and the startup surge was tripping it on cold days — compressors need headroom. Moved up to a 3000W Victron and haven't had a single trip since. One thing I'd add: if you're in a climate with hard winters, a low-ambient kit is worth the extra $150. Mine handles down to -15°F without issue.