Adding cooling and filter wheel to a debayered 450D mono DSLR

[Gina]

Hi Gina,

I've followed most of your threads about your DSLRs transformations (qualified as "butchery" at the begining), being busy myself with the mod of a 1000d to have thermoregulation and as light as possible in a home made enclosure. I have one question regarding this mod : I was wondering myself if I would or not cut the plastic pyramid on which the bayonet is fixed in order to reduce the optical path length. But I ended up deciding not to do it to be sure not to introduce "tilt". I see that in this mod, you decided to cut it. How do you ensure to have the sensor parallel to the focal plane (not having tilt) ?

And a second question also, if I may : is the shutter really needed ? 

I was happy to see you could cut the ribbon from the top, I was wondering if I could do it on the 1000d too, I had the feeling it was really not necessary. I try to eliminate everyting I can : I even removed the battery compartment and the CF reader PCB, the flipping mirror, ... 

Ah, one more question still :  what capacitor to add to the power supply connection to eliminate noise ?

Clear skies, wonderfull nights and alot of fun while busy on your projects !

Mike

Thank you Mike

I haven't decided quite how to adjust the sensor alignment as yet.

I've wondered if the shutter is really necessary too.  I wonder if anyone has tried it.  I seem to remember that the camera doesn't work with the shutter data ribbon disconnected - I haven't tried disconnecting the power.  I feel an experiment coming on

As for the top ribbon cable - I think these models are much the same in this so I would think it would work with the 1000D.  But no guarantees - I really don't know.  Later I might play with the 1000Ds I've got.  I think there are two that just don't work and one that does.  ATM I'm concentrating on the 450D.

The capacitor I'm using across the power supply is 10nF.  The value isn't critical but ceramic type are recommended for this purpose.  I suggest any ceramic capacitor in the region of 10-100nF.  The voltage dropping/regulator circuit will be within centimetres so an electrolytic isn't required but I do use a few hundred microFarads on long runs.

HTH

[TeBilePas]

Hi Gina,

thx for the reply, I also feel some experiment coming ! ;-) Thx for the capa, I'll manage to get the necessary. I got alot of material today, so I can go further with the temperature sensors, humidity sensor et even modulation of the speed of the cooling fan. I'm still waiting for the piece CNC'ed to make the adapmtation between the EOS bayonet and T2 and to close with joint to close the box and the MC clear filter also... but still alot of things to do. I took as your latest project the way of not having display etc, but just to control as much as possible via the computer... simpler and lighter... and maybe later will I try to play with debayering, but, there's a long road before !

Beautiful nights,

Mike

[gromain]

Hi!

First of all, awesome work. I'm myself on this kind of ultra mod for either my 450D, or a 1100D/600D off ebay.

As for the shutter, I believe it's highly necessary.

In fact, if I remember well, if it's not installed, some problems will appears. Brightest stars in a given field will create trails on the picture as the readout of the sensor pixels value goes further.

I'll try to find some schematics explaining this if needed, or will do them myself.

There are some beginning of explanation here: http://www.steves-digicams.com/knowledge-center/why-digital-cameras-have-mechanical-shutters.html

The problem is just named here at the beginning of the page: http://astropix.nl/astro/shutter.htm

I can't find (yet) the nice explanation I once read, but the bottom line is: the shutter is highly needed, or bright stars will leave streaks on the final picture.

Cheers,

[Gina]

Hi!

First of all, awesome work. I'm myself on this kind of ultra mod for either my 450D, or a 1100D/600D off ebay.

As for the shutter, I believe it's highly necessary.

In fact, if I remember well, if it's not installed, some problems will appears. Brightest stars in a given field will create trails on the picture as the readout of the sensor pixels value goes further.

I'll try to find some schematics explaining this if needed, or will do them myself.

There are some beginning of explanation here: http://www.steves-digicams.com/knowledge-center/why-digital-cameras-have-mechanical-shutters.html

The problem is just named here at the beginning of the page: http://astropix.nl/astro/shutter.htm

I can't find (yet) the nice explanation I once read, but the bottom line is: the shutter is highly needed, or bright stars will leave streaks on the final picture.

Cheers,

Thank you for that - very interesting and entirely answers my question

[TeBilePas]

Thx ! Seems like there's an expirement less to do ! ;-)

Clear skies and beautifull nights

[Gina]

This project is back on but with the idea of trying it on the Esprit 80ED rather than an SLR camera lens.  This means some changes to the design because I would like to try LRGB as well as NB so I would want 7 filter positions as a minimum.  The alternative would be interchangeable carousels which makes things more complicated to produce as well as more fiddling about in use.  With the bigger wheel I will be able to drive it from the centre directly from a stepper motor which means it would only need a home position sensor and the individual filter positions obtained by counting steps.

While a 7 hole wheel would by quite practical the geometry is pretty awkward and 8 would be easy - 45 degrees per filter change.  Another point is the possibility of adding an extra filter though I don't know what.  OTOH the fewer filters catered for, the smaller the carousel for a given spoke width.  Also, filling all the holes would mean perfect balance.  I will be using Baader 36mm unmounted filters (big enough to cover the APS-C size sensor) - I already have Ha, OIII and SII and will be getting LRGB.  With them all being Baader, they will be parfocal.

[Gina]

[Gina]

Here is a scale drawing of an 8 hole carousel - 150mm OD.  8 36mm unmounted filters will just fit in - barely.

[Gina]

Another question is what to make the carousel out of.  The usual material is aluminium but in the other thread I'd said acrylic.  Two problems with that - reflections from the surface (could roughen it) and lack of strength and rigidity particularly with so little between the filters.  I think aluminium wins the day.  In fact I already have some aluminium discs of 145mm diameter and 3mm thick.  This size however, is just on the edge of the filter holes so they would have to be moved in a bit further reducing the metal between. 

I think as a result of the above I might try for 7 holes which will be further apart and further in from the edge.  That will mean either awkard geometry or maybe I could mount the wheel on a stepper motor and drive it 1/7 of a revolution and mark the position.  Of course the steps per revolution of 4096 will not divide by 7 exactly - being 585.142 but the error turns out to be only one count.  585 x 7 = 4095 that's accurate enough - I couldn't get that sort of accuracy with the machining anyway, I don't have the equipment.

One thought...  Has anyone got a clever trick to divide a circle into 7 parts?  Many other numbers have these tricks.

[Gina]

In my lathe and turning investigations I've been looking at acetal as an engineering material and now I'm considering the possibility of using acetal for the carousel - black, of course.  I could attach a sheet to the face plate on the lathe and round off the edge etc.  Alternatively, I have just ordered a set of 4 black acetal billets - 154mm diameter and 22mm long.  I already have some other ideas for these anyway.  They're rather thick/long for this purpose but I could turn a boss into one to attach to the stepper motor.  Just an idea really and I think I'll probably keep them for thicker items and use sheet for the FW.

[Gina]

Well, I can't find any black acetal sheet at a sensible price so it looks like it's back to aluminium - far far cheaper   Unless I use ond of thoe 6" billets...

[vulkan]

For purchasing plastics in the past I have used theplasticshop.co.uk, the key is to use the 'custom dimensions in mm' option in the sheet size drop down.

They aren't the cheapest and watch out for their shipping costs (a flat £10) but for getting exactly the material you want in the right thickness and size they are pretty good.

Nick

[Gina]

Thank you Nick

[StuartJPP]

I like working with Aectal/Delrin, very easy to shape and work with, much better than acrylic.

[pixueto]

Dear all,

I've been considering modding my camera adding cooling. One thing that worries me a bit is adding a fan. I would have thought this will create vibration in the system affecting all the pictures. Isn't that the case? If not, how is this possible if just by touching the scope a photo can be wasted.

Thanks for your help

[Gina]

I wondered about this before I tried it but the fan hasn't spoilt any images so far.  Of course, this could be "famous last words" and next model could suffer from it - fingers crossed.  If this happened I could always reinstate water cooling - I still have the parts

[Gina]

I have now found a source of black acetal sheet in various sizes and thicknesses from 8mm up though nothing thinner.  Anyway, I think the 8mm acetal sheet would be suitable for the carousel.  Not sure about the sides but I have ordered 2 pieces 150mm x 200mm x 8mm to see what's what   I can see how stiff this stuff is in a largish sheet and whether it would be rigid enough for the sides.  I'm also thinking that I may be able to cut teeth into the edge of the carousel and rim drive it from a stepper motor.

If/when I get a milling machine I may machine some of the camera and filter wheel parts from either acetal or aluminium.

I have decided to go for a 7 hole carousel and fill these with Baader 36mm unmounted filters - Ha, OIII, SII and LRGB.  Using 7 holes rather than 8 means I can reduce the size of the carousel and the FW overall.  Those are the filters I will be using and see no benefit in allowing for any more.

This camera/FW will be primarily made to use with the SW Esprit 80ED with field flattener and have the appropriate threaded attachment.  I think I will probably use a separate piece for the scope connection - my lathe is not big enough to swing the FW casing to cut a thread directly into it.

[Jooshs]

Hi Gina, I've been following your projects for a while from afar, but am finally about to jump in and further modify my 1100D with cooling. You seem like you frequently need custom parts and or shapes... I am curious if you have considered using 3d printed parts. They are extremely accurate, strong, and you can insert whatever details you need. I'm fortunate and have access to several at work and they encourage us to experiment and play to make things with our free time (robotics company)... However, there are quite a few places to order parts of a wide range of different materials and sizes. Just a suggestion that have made all of my DIY projects significantly easier and better performing.

[Gina]

Yes, I've looked at 3D printers and would cetrainly like to have one but they are still very expensive.  A mill is cheaper and capable of more jobs.  I think a lathe and a mill will probably do most of what I want.

[Gina]

Been working out measurements to see if there's any chance of including an OAG in this design and it seems that there is

Here's a preliminary diagram with some of the filter wheel casing omitted for clarity.  (Not to scale.)

[Horwig]

I'm probably a bit late with this, but this is where I get my plastics from:

http://www.directplasticsonline.co.uk/

there are probably cheaper places out there, but their stock range is huge, and the website is informative and well laid out.

BTW, how's the lathe going Gina, and when will the mill arrive, (I find that I do far more work on the mill than the lathe)

Huw

[Gina]

Thank you Huw   I have looked on DirectPlastics a few times.  Their prices seem reasonable but delivery charges are quite high if you only have a small order.

I now have the two pieces of black acetal sheet and it seems very rigid so I think I'll use it for the filter wheel casing on the scope side and probably the carousel.

I'm still doing improvements to the lathe but have done some turning - see my lathe saga thread for more info

The mill is out of stock ATM...

[Gina]

A more realistic diagram.  The plate between the camera and filter wheel is 3mm aluminium and conducts heat from the copper box to the optical window to counteract condensation.  The scope attachment side is 8mm thick acetal and the carousel the same.  I plan to use a surface silvered mirror salvaged from the original camera viewfinder instead of a prism for the OAG.  I can see no advantage in a prism.

I propose to cut gear teeth in the edge of the carousel and edge drive it by a pinion on the stepper motor - I wasn't happy to direct drive the carousel from the stepper motor for fear of stripping the gears in the motor gearbox,  The drive gear ratio will be 7:1 making one revolution of the stepper motor per filter position.

There are more details to sort but I'm making progress

[Gina]

Another change of plan for the carousel drive   Now that the FW casing is containing an OAG there is plenty of space between carousel and scope side plate to get timing pulleys and belt in.  This is a much easier method than cutting gear teeth in the rim and edge driving it.  The timing belt has a small amount of give to protect the stepper motor.  Alternatively, I may be able to incorporate a bit of give in a centre direct drive - the enlarged FW case would give room for a "soft" coupling.  These stepper motors have 4096 steps per revolution - 4096 / 7 = 585.14285714285714285714285714286 so 585 steps per filter.  585 x 7 = 4095 so an accurate sensor for the home position would be required and any request for position 1 would need to use this rather than using n x 585 steps each time otherwise the error would accumulate in time.  I guess software correction could be used to adjust the step count by 1 each time the wheel "passed go".   A 7:1 reduction ratio makes the "maths" easier but OTOH the stepper takes 10s per revolution at maximum step rate - thus 10s per filter and 70s per revolution of the carousel - rather slow IMO.

[Gina]

Furthering the design in SketchUp has shown that there is no room for the stepper motor to drive the centre of the carousel - the camera is in the way one side and the field flattener for the scope on the other side.  So it's timing pulleys and belt then - probably a 60t on the wheel and 20t on the stepper motor.  That would mean 585 x 3 = 1755 steps per filter - just over 4s at full speed.