Choosing the Optimal Product: 7 Questions on Thermostats, Intercoolers, and Bearing Pullers

Every year, a few buyers send me a part print and ask, “Is this an optimal product?” Usually they mean, “Is this the best looking option we can put on a purchase order?” That’s not the right question, and I’ll explain why.

I’m a quality and compliance manager at Optimal, an automotive parts manufacturer with stamping, die, CNC machining, forging, and aluminum extrusion capabilities. I review roughly 250 part lots a year before they go out the door. In 2024, I rejected around 10% of first submissions—not for cosmetics, but for dimensional results that did not match the PPAP record.

If you’ve ever asked any of the following questions, this is for you.

1. What makes a part an optimal product?

The term optimal product gets treated like highest specification. That’s wrong. The optimal product for an automotive application is the one that does the job with controlled variation, not the one with the tightest tolerance you can afford.

It took me four years and a couple of awkward supplier audits to learn that. We once quoted a CNC bracket with ±0.05 mm on a critical hole because the customer asked for high precision. Every first article passed. Then their assembly fixture couldn’t pick up one in ten brackets. The gap between individual parts was fine, but the process capability was not.

So when someone asks for an optimal product, I ask: Which failure mode matters most? What data do they have? The best part geometry is often a stamped feature with a wider, controllable tolerance instead of a machined feature with a beautiful drawing number.

2. Are “optimal thermostat settings winter” a real thing?

This one shows up more than I expected. For a vehicle, there usually isn’t a thermostat setting you adjust for winter. A mechanical engine thermostat is a temperature-controlled valve with an opening point set by the manufacturer, typically around 190–207°F.

The phrase “optimal thermostat settings winter” often comes from someone who wants the engine to warm up faster in cold weather. A lower-temperature thermostat doesn’t do that. It opens earlier and sends coolant to the radiator before the block is up to temperature. That can keep the engine colder, increase fuel dilution, and make the cabin heater less effective.

The most useful winter thermostat setting is still the one the OEM specified—unless you have a recalibrated ECU and know why you’re deviating.

3. How do you use a thermostat for cooling?

Honestly, you don’t use it for cooling. The thermostat’s job is to block coolant from the radiator until the engine reaches operating temperature. The radiator, cooling fans, and coolant do the actual cooling. The thermostat only prevents overcooling.

The most frustrating part of this question is that the word cooling makes people think colder is safer. I’ve seen vehicles where someone removed the thermostat to fight overheating. The engine ran cold, the computer kept adding fuel, and fuel economy got worse. Warm-up is part of the cooling system design.

If you want a practical answer for how to use a thermostat for cooling: install the OEM-rated opening temperature, position the bleed hole correctly, and pressure-test the system after installation.

I should add one clarification: a stuck-open thermostat causes cool running and poor fuel economy; a stuck-closed one can produce an overheating condition. Both failures tell you the thermostat, not the radiator, is the problem.

4. What should I inspect before buying a Perrin front mount intercooler?

Brand name is not a quality spec. I’m not going to attack Perrin—I don’t have a reason to. But the question should be about the whole system around the Perrin front mount intercooler: mounting brackets, end tanks, tube routing, and seals.

I once had a vendor tell me a bracket face was flat. We both said flat. My meaning was within 0.3 mm across the mounting face. Their meaning was visually flat. That communication failure let the end cap seal shift after the first pressure cycle. It looked like an intercooler defect, but the bracket caused it.

Check three things: weld quality on end tanks, core density suited to your boost level, and bracket material with aligned holes. A front mount intercooler that moves 3 mm under load can pull on charge pipe couplers and create leaks that look like tune problems. The core can be great; the hardware around it has to be equally good.

5. How do I remove a differential bearing without wrecking the housing?

If you’ve ever watched someone try to pull a differential side bearing with a pry bar, you know why this belongs in a quality discussion. A proper differential bearing puller is the start, but technique matters.

Position the puller arms or bearing separator behind the inner race. Pulling on the outer race bends it because the outer race is unsupported as the bearing begins to move. A bent race can score the bore in the carrier. Tighten the puller center screw gradually. If it binds, stop and check for a snap ring, burr, or corrosion. Don’t hit the arms with a hammer and hope.

At Optimal, we use the same discipline when removing dies and tooling components. The right puller setup takes ten minutes and can save a $2,000 housing. The wrong one costs more than the tool you were trying to save. Trust me on this one.

6. When is paying extra for guaranteed delivery smart?

When the cost of being late is higher than the rush fee. That sounds obvious, but I’ve seen buyers avoid a $400 expedite fee and then accept a $22,000 delay.

In early 2024, we needed a set of aluminum extrusions for a prototype fixture. The lower-priced vendor said “probably” three weeks. They missed by four days. The line trial delay was budgeted at roughly $6,000 per day. We then paid a different supplier $410 for guaranteed delivery and had parts in two days. The $410 didn’t just buy speed. It bought a written schedule commitment.

The red flag is “we think it will ship Friday.” Certainty costs something, and in a deadline situation it is usually worth it.

7. How do you know a supplier can make the optimal product consistently?

Ask to see their control plan before you ask for a quote. A good quality engineer is happy to show how a critical dimension is monitored. A vendor who only has a CMM report on a first article is showing you one good moment, not a process.

At Optimal, we release a new die only after a 30-piece trial run and SPC review. We look for drift, not just whether every part is inside print. A part can be inside tolerance and still not be optimal because the process is wandering from batch to batch.

If a supplier can’t explain what they do when a process goes out of control, take that as the warning sign it is. That is where “optimal” actually gets decided.