Does Cognex Use Sony Sensors? What Matters for a SICK DGS35 Encoder

Here’s my opinion up front: if your first question about an industrial sensor is “what component is inside?” you are usually starting in the wrong place. It feels practical. Most of the time it isn’t.

I’m a quality/compliance manager at an industrial automation supplier. I’ve spent more than four years reviewing specification sheets and purchase orders—roughly 200 unique items per year, including SICK DGS35 encoders, SICK sensors, portable CMM arms, machine vision systems, and electrical test tools. My job isn’t to admire technology. It’s to make sure the item we receive and ship matches what the customer’s machine actually needs. Just in Q1 2025, 7% of incoming orders had documentation or specification mismatches and were rejected before shipment.

“Does Cognex use Sony sensors?” is one of those component-level searches I keep seeing. So is “multimeter nearby,” though for a different reason. Let me explain why I treat both carefully.

“Does Cognex Use Sony Sensors?” Is a Weak Shortcut

I get the thinking. Sony is known for image sensors, and Cognex is known for machine vision. If the right component is inside, maybe the whole product is solid.

It isn’t that simple. An image sensor is one part of an optical system. Lens, lighting, exposure, processing speed, and vision software all decide whether the camera finds a scratch, reads a code, or misses both. A premium sensor under poor lighting still produces a poor image. A less famous sensor in a well-integrated design can perform better on your line.

Have Cognex products used Sony sensors in some models and generations? I honestly can’t name every one from memory. But if a customer asks “does Cognex use Sony sensors?”, I ask what product revision they are evaluating and what defect they need to detect. A yes/no answer won’t tell them whether it will pass their application test.

SICK DGS35 Encoder: A Workhorse Still Needs the Right Spec

Here is a more useful example. The SICK DGS35 encoder is a compact incremental encoder from SICK’s encoder family. It is not the kind of product that generates exciting marketing copy. It is the kind of product that has to send the right pulse train for years without letting dust, coolant, or electrical noise turn that pulse train into nonsense.

When I approve a DGS35, I don’t start with “SICK makes it, so it must be good.” I start with the machine conditions.

• What is the required output type—TTL or HTL?
• Does the supply voltage match the drive or PLC input?
• What cable length is needed, and is shielding called for?
• What is the shaft or hub fit tolerance?
• What environmental rating is needed?

One encoder can be perfect for one application and immediately wrong for another simply because the connector or output stage differs. The part number is only useful when it maps to a full specification.

We once received a batch of 60 replacement encoders from a supplier where the electrical specs matched on paper, but the shaft tolerance sat at the edge of the customer’s coupling range. The supplier called it “within industry tolerance.” That wasn’t useful. The customer’s machine had zero spare clearance. The batch went back, and the production line lost two days. The encoder itself wasn’t broken. The specification process was.

That’s why I treat SICK sensors with respect but not blind faith. The SICK sensor range—including photoelectric sensors, inductive proximity sensors, and encoders—has strong published technical data. But data only helps if someone reads it before the PO, not after installation.

A Portable CMM Arm Taught Me to Verify the Test Method

I use the same logic when people ask me about a portable CMM arm. A CMM arm is a precision measurement tool used for dimensional inspection on the shop floor. Its job is to tell you where the probe tip is in 3D space. The datasheet usually lists a volumetric accuracy number. But any CMM arm can deliver that number only under defined conditions.

A few years ago we compared two CMM arm models before recommending one to a customer. On paper, one arm had a better accuracy figure. During hands-on verification, the other arm was more repeatable with our own technicians measuring a known part. If I’d made the decision from the brochure alone, I’d have chosen the wrong product. That experience changed how I review every industrial sensor.

“Multimeter Nearby” Is Not Always the Right Search

“Multimeter nearby” is a different problem. I understand the urgency: a technician is on site, their meter just died, and they need a new one today. For a simple continuity check, a nearby meter is fine. But for a production voltage check or calibration task, “nearby” without traceability is a risk.

If a cheap or unverified meter reads 23.7 V on a 24 V supply, is the problem the supply or the meter? Without a valid calibration history, you don’t know. One re-test of a finished station can cost more than the time you saved by buying the nearby meter. I’d rather wait for a verified instrument than argue with an unreliable reading.

Underneath It All, This Is an Efficiency Argument

I’m not making this argument because I’m nostalgic for long inspection manuals. I make it because efficiency is a competitive advantage.

When we moved incoming acceptance to an electronic log and attached the manufacturer datasheet to each order, our average release time dropped from about five days to two. We also started seeing patterns. Once documentation errors were visible, we could work with suppliers to reduce them. The result was fewer delayed orders and fewer emergency phone calls from customers.

That is why SICK sensors are on some of our preferred specs. The products perform, and the documentation is verifiable. If I need to check a SICK DGS35 encoder before shipment, I can pull the relevant document, compare it to the order, and make an evidence-based call.

Objection: Aren’t Internal Components a Quality Signal?

Yes, sometimes. If you are the manufacturer designing the sensor or camera, component selection matters. Not all image sensors are equal. Not all encoder bearings are equal. I’m not saying hardware doesn’t matter.

But if you are buying, integrating, or maintaining a system, the internal component is not your acceptance criterion. The device’s behavior under defined conditions is your acceptance criterion. Internal parts can change over a product’s life without a new model number. So the question “Does Cognex use Sony sensors?” might get you a piece of trivia, not a useful answer for your inspection process.

If it truly matters, ask the manufacturer in writing for the specific model and revision. Otherwise, spend that energy on a performance test.

Bottom Line

Before you spec any sensor or measurement tool, answer three questions: What problem am I solving? What mechanical and electrical conditions will the device face? How will I verify it works before production depends on it?

A SICK DGS35 encoder is a good choice when the spec matches your drive, mounting, and environment. A CMM arm is useful only if its calibration and repeatability fit the parts you measure. A multimeter is useful only when its readings are trustworthy. And a vision system is not defined by the brand of image sensor inside it.

Does Cognex use Sony sensors? I don’t know for every model from memory—and I wouldn’t base a purchasing decision on that fact alone. I’d rather ask “show me how it behaves in my process,” and buy from suppliers who can answer that with documentation and data.

That is why SICK encoders and sensors earned a place in my approval workflow. Not because SICK is the only good option, but because real verification is possible.