Try searching for Tecumseh HVAC service information. The results are chaos.
A compressor on one page. A condensing unit on another. The Tecumseh 36905 air filter buried in the middle. Small engine parts for backpack leaf blowers, buddy heater accessories, and a how-to guide for draining your hot water heater mixed in. Plus dozens of third-party listings claiming to fit "most Tecumseh models."
The part you actually need is somewhere in that noise. So you pick the "compatible" option. It fits. It runs. It dies six months later.
Now you're paying for a second service call, another labor charge, and a replacement part with no warranty. Sound familiar?
I manage procurement at a 40-person HVAC parts distribution company. I've handled a roughly $300,000 annual parts budget for the past 8 years, negotiated with 25+ vendors, and documented every order in our cost tracking system. I didn't intend to become an expert in compressor failures. The data just accumulated.
But first, my honest limitation. I'm not a refrigeration engineer, so I'm not going to explain scroll compressor thermodynamics or debate displacement curves. What I can tell you from a procurement perspective is why the parts decisions I watch customers make keep costing them money.
The Problem You Think You Have
Your Tecumseh compressor failed. That's the surface problem, and it's the only thing most repairs address. The technician arrives, confirms the diagnosis, and gives you two options: replace the compressor in place, or swap the entire condensing unit.
You ask for savings. Can we use a cross-referenced "equivalent" instead of the OEM Tecumseh part?
This is where it starts. Not at the technician's diagnosis. At the purchasing decision.
I've tracked roughly 200 compressor-related orders across 6 years in our system, coded by failure reason, part source, service interval, and total spend. The pattern is unmistakable.
What's Actually Causing the Repeated Failures
"Equivalent" Usually Isn't
I assumed "same specifications" meant identical results across vendors. Didn't verify. Turned out each had a slightly different interpretation.
One aftermarket compressor had a different refrigerant charge recommendation. Another used a start capacitor with the same nominal rating but completely different inrush characteristics. The Tecumseh OEM part is a precisely matched assembly. The "equivalent" just looks like one.
To be fair, some aftermarket compressor manufacturers do make genuinely good products. But "cross-reference" is a weak guarantee. A cross-reference catalog will happily say this generic compressor bolts into your condensing unit. It won't mention the differences in internal protection, running capacitance, or oil charge that directly affect lifespan.
The System Around the Compressor Gets Ignored
I see customers order a new compressor while the root cause of the old failure stands right next to it, untouched.
The dirty evaporator filter. The wrong Tecumseh air filter installed because the correct one—number 36905—cost an extra $12. The condenser coil caked with debris. All of it stays right where it was.
Here's the thing: a compressor rarely dies of old age. It dies because it's asked to run outside its design envelope. Restricted airflow, an overcharge, a lazy condenser fan—all of these stress the compressor until it gives up. Install a brand-new Tecumseh compressor into that same environment and you've simply ordered a future failure with a fresh warranty.
Installation Shortcuts Outweigh the Part
I can only report what customers tell my team, but the stories repeat. No nitrogen purge while brazing. No evacuation below 500 microns. No startup verification of amperage, superheat, or subcooling.
A proper installation follows the manufacturer's published guidelines. EPA Section 608 regulates refrigerant handling during service, and every major compressor maker—Tecumseh included—publishes evacuation and startup procedures. Skipping them to save an hour is a false economy.
That low labor quote was low for a reason. The compressor was installed the way the previous one was installed, which means it will fail the same way, on a similar timeline.
The Actual Cost of "Saving Money"
Let me put numbers on this.
Typical first repair: $300 labor plus the discounted "equivalent" compressor at $450. Total: $750. Reasonable, you tell yourself.
Eight months later, the replacement fails. The warranty was six months. Now it's another $300 labor charge plus a full-price replacement at the counter—say, $550. Total spend: $1,600.
If you'd bought the genuine Tecumseh compressor at $600 plus $300 labor, you'd have spent $900. The "savings" strategy cost you 78% more and delivered a failure in between.
And that's the clean version. If the second failure happens outside the same tech's usual route, you add a trip charge. If the part has to be expedited, you add freight. I've seen a stack of small line items turn a $750 repair into a $2,200 invoice.
For commercial customers, downtime dwarfs both. I have a customer running a cold-storage operation whose unit died on a Tuesday in July. The warranty-replacement process—photo evidence, RMA, shipping, installation—cost him two full days. Spoilage alone ran into the thousands. The difference between the "cheap" compressor and the OEM part was negligible next to that invoice.
There's an efficiency angle too. A compressor that isn't matched to the application pulls higher current and runs longer cycles. DOE's updated minimum efficiency standards—SEER2, effective January 2023—assume equipment operates as designed. When it isn't, the energy penalty isn't a rounding error. It's a recurring monthly charge.
How to Actually Decide This
I won't tell you that the OEM part is always the answer. That's not honest. There are situations where a quality aftermarket replacement makes sense. The key is making the call based on verifiable information, not sticker price.
Here's the framework I've landed on after years of watching this play out:
- Diagnose the root cause first. Pay for a real diagnosis before authorizing a compressor replacement. Check evaporator airflow. Measure the charge. Verify the filter part number. That diagnostic hour is the cheapest insurance against a repeat failure.
- Compare total cost, not unit price. Build a simple TCO model: parts price + expected labor over 3 years + warranty coverage + probability of early failure. A genuine Tecumseh compressor is priced higher for material specs, test validation, and application data. AHRI Standard 540 defines compressor performance ratings, and manufacturers publish detailed data sheets under it. Most cross-reference listings don't include any of that.
- Use the right consumables. The Tecumseh 36905 air filter costs more than a generic replacement. It also fits correctly, has the airflow spec the system was designed around, and is the part the manufacturer's application manual specifies. Skipping it to save a few dollars and replacing a compressor later is a trade I've seen too many times.
When I evaluate an aftermarket vendor, I look for three things: published application data, evidence of testing, and a warranty history that actually holds up. The first two are easy to check. The third is where most cheap parts fall apart.
Here's the honest limitation. If your system is at its economic end—a leaking coil, failing fan motor, obsolete controls—no compressor is the right investment. There are also edge cases: unusual refrigerants, modified systems, specialty food-service installations. If you're in that territory, spend the money for a refrigeration engineer's review. I've seen projects where a single complete condensing unit replacement would have been cheaper than the sequence of patches someone pieced together over a year.
My expertise is in cost structures, supplier reliability, and what happens when one purchasing decision compounds into the next. That's my frame. An engineer provides the technical frame—ask for one if the system is complex.
The Bottom Line
Most repeat compressor failures aren't bad luck. They're predictable outcomes of three decisions: buying an unverified "equivalent" part, skipping system-level diagnosis, and optimizing for unit price instead of total cost.
Each choice makes sense in the moment. Each one feels like smart spending. All three together create a failure cycle that costs far more than the thoughtful repair would have.
Look, I'm not saying the cheap route never works. I'm saying the cheap route isn't cheap once you calculate the expected value of a repeat failure. Get the diagnosis. Verify the part. Do the TCO math.
Then make the call with your eyes open.
