What causes dotted diffraction spikes?

[tooth_dr]

See photo below - Ha filter

Is a diffraction spike like a spectrum/rainbow, and only parts of it pass through the filter, giving the dashed line appearance?

 

 

[andrew s]

That is the correct diffraction pattern for almost monochromatic light. Look at HST images and you can see coloured "dots". In RGB they get washed out.

Regards Andrew 

Yes it is analogous to how a diffraction grating creates a spectrum. 

Edited November 19, 2019 by andrew s

[Craney]

So with a mono  camera  the top pattern ( Luminosity  sub )  would not have  discernible  gaps .  The bottom three are monochromatic, similar to the Ha.

 

[tooth_dr]

5 minutes ago, Craney said:

So with a mono  camera  the top pattern ( Luminosity  sub )  would not have  discernible  gaps .  The bottom three are monochromatic, similar to the Ha.

 

Thanks Sean! Makes sense! This is a luminance diffraction spike from another image i took

 

[ollypenrice]

I never knew this. It's also very visible when focusing in NB using a Bahtinov mask but I never thought about why.

Olly

[Craney]

That epsilon sure does produce some impressive spikes....

Another thing to notice is that the angular spread of the colour blocks is less (tighter packed)  with the blue than the red  because of blue's  smaller wavelength.

[glowingturnip]

On 19/11/2019 at 09:50, ollypenrice said:

I never knew this. It's also very visible when focusing in NB using a Bahtinov mask but I never thought about why.

Olly

ahaaa - yes, I've seen that and wondered about it too

Edited November 20, 2019 by glowingturnip

[JamesF]

4 minutes ago, glowingturnip said:

ahaaa - yes, I've seen that and wondered about it too

Me three.

James

[carastro]

Quote

I've seen that and wondered about it too

Me four.

Carole 

[tooth_dr]

Here is a comparison of RGB spikes from the other night:

 

 

[DaveS]

Yup, diffraction around spider vanes will be wavelength dependent, while a B-mask is, of course, a coarse diffraction grating. Most obvious with NB filters, but also seen in RGB. A mono camera will show the L as a single spike, but I guess a colour camera (Which I haven't used in years) will show a rainbow of multiple diffraction orders. Each dot ina B-Mask pattern is a different diffraction order.

[sharkmelley]

It's a good way of telling if someone has added diffraction spikes to their colour image during post-processing.  Such diffraction spikes are typically uniform in colour and won't have the colour banding effect.

Mark

Edited November 21, 2019 by sharkmelley

[wimvb]

On 20/11/2019 at 20:46, sharkmelley said:

It's a good way of telling if someone has added diffraction spikes to their colour image during post-processing.  Such diffraction spikes are typically uniform in colour and won't have the colour banding effect.

Mark

I agree, Mark. But your post needs an addendum, imo. Artificial spikes won't show banding, but absence of banding doesn't mean that the spikes are therefore artificial. With my 150pds I've never managed to get banding in the spikes. My suspicion is that the "quality" (ie, sharpness, banding, length) of spikes in a stacked image is determined by several factors: focus, polar misalignment and collimation, and maybe others. Focus is obvious because it blurrs any detail. Polar misalignment introduces field rotation which, with stacking and pixel rejection, results in shorter spikes. Collimation is somewhat uncertain, but probably similar to poor and uneven focus. 

[tooth_dr]

1 hour ago, wimvb said:

I agree, Mark. But your post needs an addendum, imo. Artificial spikes won't show banding, but absence of banding doesn't mean that the spikes are therefore artificial. With my 150pds I've never managed to get banding in the spikes. My suspicion is that the "quality" (ie, sharpness, banding, length) of spikes in a stacked image is determined by several factors: focus, polar misalignment and collimation, and maybe others. Focus is obvious because it blurrs any detail. Polar misalignment introduces field rotation which, with stacking and pixel rejection, results in shorter spikes. Collimation is somewhat uncertain, but probably similar to poor and uneven focus. 

Also Wim, Ive combined reflector lum with refractor RGB, and obviously the spikes arent coloured.  This is kinda visible in the image below cropped from my M45 image

Edited November 22, 2019 by tooth_dr

[wimvb]

Why the double spikes? The minor spikes do have coloured bands inside the halo, are these from a refractor? 

[tooth_dr]

39 minutes ago, wimvb said:

Why the double spikes? The minor spikes do have coloured bands inside the halo, are these from a refractor? 

I used data from two nights, and there was a different rotation of the tube relative to the sky.  So that produced two sets of spikes.  The coloured spikes correspond to RGB data from the reflector combined with lum data from the same night, the non coloured spikes are just from lum on that different night.

Below is reflector lum (epsilon) and refractor RGB (ED80).  There are double spikes here due to dodgy collimation (when I bought the scope)

[wimvb]

On 20/11/2019 at 16:00, DaveS said:

while a B-mask is, of course, a coarse diffraction grating.

Not quite. A yrue diffraction grating must be much, much finer. Diffraction occurs at the edge of every line in a Bahtinov mask. Having more lines only gives a better definition and a stronger pattern. You can make a Bahtinov mask with only one Y. The edges of the lines in a Bahtinov have to be very straight to get a well defined diffraction pattern. 

Here's a simulation using Maskulator. The dotted pattern is from a true Bahtinov mask, while the continuous pattern is from a Y-mask. Only three wavelengths were used (blue, green, red).

Edited November 23, 2019 by wimvb

[Robny]

Sorry.....

I have nothing to.add to this thread but found it very interesting reading indeed and something I'd never considered

[Firas]

On 19/11/2019 at 15:52, Craney said:

That epsilon sure does produce some impressive spikes....

True! I always wondered, why only few other newtonians are capable of producing similar fine, long diffractions, like the Epsilon. Is it the coma corrector that Takahashi uses? Are there any information on what type of corrector they built into the focuser? The secondary vanes looks thick like most other newtonians. 

Edited December 3, 2019 by Firas

[andrew s]

9 hours ago, Firas said:

True! I always wondered, why only few other newtonians are capable of producing similar fine, long diffractions, like the Epsilon. Is it the coma corrector that Takahashi uses? Are there any information on what type of corrector they built into the focuser? The secondary vines looks thick like most other newtonians. 

It is the relatively fat spider vanes that do it. This puts more light into the spikes.

Regards Andrew 

 

[JamesF]

44 minutes ago, andrew s said:

It is the relatively fat spider vanes that do it. This puts more light into the spikes.

Really?  If the spikes are a diffraction effect along the edges of the vanes, surely those are the same regardless of whether the vanes are thick or thin?  How would thick vanes cause more light to be diffracted?

James

[andrew s]

1 hour ago, JamesF said:

Really?  If the spikes are a diffraction effect along the edges of the vanes, surely those are the same regardless of whether the vanes are thick or thin?  How would thick vanes cause more light to be diffracted?

James

No the whole obscured area contributes contributes to the diffraction pattern.

Regards Andrew 

[JamesF]

3 minutes ago, andrew s said:

No the whole obscured area contributes contributes to the diffraction pattern.

Confused :(  Surely any photon that actually hits the vane contributes nothing?  I could get my head around the idea that it's the spacing between the edges of the vane that changes the form of the diffraction pattern, kind of like changing the slit spacing in the double slit experiment.  I just can't explain to myself how making thicker vanes causes more photons to end up in the diffraction spikes.

James

[andrew s]

5 minutes ago, JamesF said:

Confused   Surely any photon that actually hits the vane contributes nothing?  I could get my head around the idea that it's the spacing between the edges of the vane that changes the form of the diffraction pattern, kind of like changing the slit spacing in the double slit experiment.  I just can't explain to myself how making thicker vanes causes more photons to end up in the diffraction spikes.

James

To really understand this you need to look at the maths. However, maybe this will help.

You can work out what an obstruction will do by treating it as a hole and then inverting it by subtracting the intensity from the rest of the calculation.

In your edge model the wider the veins the further apart the edges are, this alone has an effect.

Regards  Andrew 

[wimvb]

I did a simulation in Maskulator with thin and very thick vanes and got these results:

Thin vanes to the left, and thick vanes to the right. Brightness scaled down to 30%. I'm not completely convinced that Maskulator is always right.

These vanes: