I don't understand sensor sizes

[StuartT]

The inch designation on sensors baffles me. For example, the ASI485MC has a 1/1.2" sensor which is 11.1x6.2mm

So what does the 1/1.2" mean?

It obviously doesn't mean 1" x 1.2" (because that would be much bigger). So then I thought maybe it's an odd way of expressing the sensor diagonal as a fraction of an inch, so 1/1.2 would be 0.833", but it can't mean that either because that would give a diagonal of 21.16mm (whereas the diagonal of 11.1x6.2 is 12.7mm)

Can anyone explain how this works?

 

[vlaiv]

https://en.wikipedia.org/wiki/Optical_format

Quote

Optical format is a hypothetical measurement approximately 50% larger than the true diagonal size of a solid-state photo sensor. The use of the optical format means that a lens used with a particular size sensor will have approximately the same angle of view as if it were to be used with an equivalent-sized video camera tube (an "old-fashioned" TV camera). In a video camera tube the diagonal of the actual light-sensitive target was about two-thirds the outside diameter, which was the measure used.

The optical format is approximately the diagonal length of the sensor multiplied by 3/2. The result is expressed in inches and is usually (but not always) rounded to a convenient fraction. For instance, a 6.4x4.8 mm sensor has a diagonal of 8.0 mm and therefore an optical format of 8.0*3/2 = 12 mm, which is expressed as the convenient 1/2 inch in imperial units. The reason it is expressed in inches is historical, dating back to the early days of television

1/1.2" means that sensor is roughly 1.2 smaller than 1" sensor.

Problem is that 1" sensor does not correspond to anything physically 1" long on that sensor. Above wiki article explains that it has to do with early days of television and diameter of tube used in camera rather than sensor itself. It also explains that you get rough size of sensor if you multiply with 2/3 (or divide with 3/2).

1/1.2" sensor is therefore roughly (1" * 2/3) / 1.2 = (25.4 * 2/3) / 1.2 = 14.11mm diagonal

11,1 x 6.2 give diagonal of 12.71mm

As you see - it is only a rough "class" of sensor - rather than precise measurement.

For another example, both ASI533 and ASI183 are classed as 1" sensors and should have 25.4 * 2/3 = 16.9333mm diagonal.

First has 11.31 x 11.31 = 16mm while second has 13.2 x 8.8 = 15.86mm

[Stuart1971]

44 minutes ago, StuartT said:

The inch designation on sensors baffles me. For example, the ASI485MC has a 1/1.2" sensor which is 11.1x6.2mm

So what does the 1/1.2" mean?

It obviously doesn't mean 1" x 1.2" (because that would be much bigger). So then I thought maybe it's an odd way of expressing the sensor diagonal as a fraction of an inch, so 1/1.2 would be 0.833", but it can't mean that either because that would give a diagonal of 21.16mm (whereas the diagonal of 11.1x6.2 is 12.7mm)

Can anyone explain how this works?

 

You are correct and it’s totally confusing to anyone who does not understand, of which I am one of, I have had this discussion recently with QHY as they are advertising there new 533 mono camera with a 1” square sensor, which it a total misrepresentation, why can’t they just just put 16mm diagonal, it’s far easier to understand, it’s no where near 1” square…but to a newcomer or even someone who has been in the hobby for years, sees this and thinks wow 25mm x 25mm sensor, only to receive and find its approx 11mm x 12mm 🤔🤔

[Mr Spock]

I don't know why they don't quote standard sizes like 'proper' cameras. My D500, for example, is 23.5 x 15.7mm with a diagonal of 28.26mm and 5568 x 3712 pixels, effectively 20.9 megapixels.

[DaveS]

Damn right. To me it just smacks of obfuscation  with a distinct whiff of snake oil.

The Vidicon tube sizes might have made sense in the days of Vidicon tubes, but those days have long gone, unless you run a TV museum.

[vlaiv]

I guess it is just a convention that is still used.

They don't actually try to fool anyone - for example ASI labels their sensors like this:

It gives exact measurement in millimeters as well as "inch class" - it should really be thought of as "class" of sensor rather than actual size.

We also say for example 5" telescope for 120mm to 130mm apertures although strictly speaking 5" telescope is precisely 127mm (some might be precise by calling 120mm - 4.7" instead ).

 

[StuartT]

Thanks everyone! So it really is a mess then, and not just me being dumb 🤣

As for the ZWO diagram @vlaiv it's crazy! Writing 1" on a diagonal that clearly is nowhere near 1"

Just give the mm dimensions and throw away the crazy "inch class"

 

1 hour ago, vlaiv said:

for example ASI labels their sensors like this:

[Stuart1971]

1 hour ago, vlaiv said:

I guess it is just a convention that is still used.

They don't actually try to fool anyone - for example ASI labels their sensors like this:

It gives exact measurement in millimeters as well as "inch class" - it should really be thought of as "class" of sensor rather than actual size.

We also say for example 5" telescope for 120mm to 130mm apertures although strictly speaking 5" telescope is precisely 127mm (some might be precise by calling 120mm - 4.7" instead ).

 

And that’s not confusing, 11.31 x 11.31 is not 1” diagonal….it's 16mm🤔🤔, so yes that is totally confusing…..

I mean c’mon 1” class, what does that mean, many people have no idea….

Edited April 27, 2022 by Stuart1971

[Adam J]

3 hours ago, StuartT said:

The inch designation on sensors baffles me. For example, the ASI485MC has a 1/1.2" sensor which is 11.1x6.2mm

So what does the 1/1.2" mean?

It obviously doesn't mean 1" x 1.2" (because that would be much bigger). So then I thought maybe it's an odd way of expressing the sensor diagonal as a fraction of an inch, so 1/1.2 would be 0.833", but it can't mean that either because that would give a diagonal of 21.16mm (whereas the diagonal of 11.1x6.2 is 12.7mm)

Can anyone explain how this works?

 

Honestly I would not think about that, just take the number of pixels and multiply by the pixel size if you want to understand relative sensor size. 

Adam 

[Toltec]

I've been wondering this for a while, I did think it was maybe the physical chip size rather than the sensor area, but even that seemed odd. Not unlike the scaling from a 35mm film format for dlsrs I guess. I have looked at the actual sensor dimensions as even the diagonal is dependent on the shape anyway.