Fire break?
I only know that as a track in the forest to stop fire from spreading.
A smoke sensor should be more effective as fire detector, because there is usually smoke before flames (heat). That's why we use smoke detectors in buildings.
Don't try to re-invent the wheel.
Leo..
flame sensor ? or lack there of
if so not the best choice
I know that smoke erupts first before fire breaks. So, installing a smoke detector would be a step to stop possible fire break.
Seven topics devoted to determining if your cheap chinese hobby grade sensor is broken or not. If it’s such an important measurement, then just spend some bucks on a reliable industrial grade product and be done with it.
That statement seems to assume that the problem must be caused by a “cheap sensor”. There is no evidence for that in this or other threads.
The DS18B20 itself is widely used and generally reliable when the 1-Wire bus is implemented correctly. In many cases, unexpected readings are not caused by the sensor at all but by bus topology, wiring, timing, or power conditions. That is exactly why my threads focus on measurement, comparison, and practical implementation details.
Simply replacing a device with an “industrial grade” product does not automatically solve those underlying issues. If the bus design or wiring layout and cable routing are problematic, the same electrical effects will still exist.
The intention is to document measurements and provide practical examples that help others build robust and predictable DS18B20 setups. Readers can decide for themselves whether that information is useful for their projects.
It is likely that a large percentage of the so-called “DS18B20’s” purchased by hobbyists are counterfeit. If reliable operation is desired then the first step is to use a genuine device. Without that no amount of “robust and predictable setup” will help.
It is surprisingly difficult, even when buying from reputable vendors, to obtain units that pass the full cpetrich test.
Using genuine parts is certainly good practice, but reliable systems should never rely solely on the assumption that a component will always behave perfectly.
In real-world engineering every component — regardless of price or origin — can occasionally produce incorrect data due to electrical noise, wiring issues, timing effects, or simple device faults. Because of that, robust systems normally include some form of error detection or plausibility checking anyway.
In other words: good hardware helps, but robust system design should not depend on perfect hardware behavior . Both aspects belong together if reliable operation is the goal.
Only buy from distributors who are authorized to sell Analog deices products, otherwise you have no idea what you are getting.
Then your threads should focus on those problems and not blinking an LED.
You say one thing regarding star topology then you contradict yourself in the same thread and in other topics.
It is not at all difficult when you purchase your parts from authorized vendors.
No one suggested that at all, only that it is the first step in a robust system design.
This post serves as the official Solution of DS18B20 Optical Feedback – Visualizing Conversion Time with an LED (former title). For the full content, see my original post at the very beginning of this thread.