After some discussion on SIMUL8R's thread for precondenser tests and some reading through other threads - I tired to do a spreadsheet that will calculate ideal condenser placement, while accounting for arc size.

I believe Rox and DAZ have already worked out a solution to this problem, but a spreadsheet will make it easier for everyone. And if it works (hopefully), I may be able to add extra things to it later - like determining the %gain in brightness from a particular condenser. It might also make it possible to play around with different fresnel/condenser combinations to see which combo gives the best results for a particular setup.

At the moment I'm still trying to test this, and not really sure if my initial "on paper" solution is correct... all I know is that it broke my b@ll$ to work through the maths, so I'm hoping its right

Right now I need numbers.... If anyone has the figures for a working model (ie. is sure that all lens placements are optimal... ie. Rox or DAZ ), could you please tell me the following specs for your model setup:

* Diameter and focal length of the fresnel
* Diameter and focal length of the condenser
* Length of the arc

I will run these values through the spreadsheet and see if I get the same lens placements as you.

Of course, everyone is welcome, and thanks in advance for your assistance


[Edit]: Included updated version of spreadsheet. All the fields are open so you can view the formals below

lets say some random numbers

diameter fresnell; 14"
focal fresnell; 220mm
diameter condenser: 100mm
focal condenser; 117mm
arc; 30mm

let me know your distances, good luck .

Rox, thanks for the reply.

For you example above, I get the following:

Arc to condenser = 30.62mm
Condenser to fresenl = 178.53mm

Is this correct?

If you guys get this figured out that would be cool because I have a lot of lenses that I want to try out. Lets see... at least seven different precondensors between the sizes of 5 and 8 inches.

Hopefully it works... would at least be nice for comparing how different components work together.

About your lenses - did you use some of those in your DIY triplet ?
Nice work in your PLOG btw

Sounds correct to me . First I used Rox’s values and my graphical method and I got 30.1mm to the arc and 179.6mm to the fresnel but these values don’t used the entire pre-condenser. Then I used your values, 30.62mm and 178.53mm, and ray traced it to see if the virtual arc passed through the pre-condenser. It did and is closer than my first attempt. Just a note that my graphical method is note perfect, it gets the arc close then I usually tweak the arc slightly until it works on paper. I also checked it using the thin lens equation and your values seem to work with it as well so I'd say it's a winner, in theory anyway.

DJ

No I haven't - well I have tried them, the five inchers but I think they are too fast of a lens or my problems are from them being uncoated, I seem to get really bad colour abberations with the condensor lenses. It's too early for me to test though, as you can see I broke a lens tonight. I need to slow down and stick with the program. Anyways for the self made triplets I am using different lenses than my condensor lenses.

P.S. This thread is a great idea if you can get it going. Thumbs up.

Dang it, this puts my final build of my light box on hold now. Looking forward to what this thread offers.
sim

Thanks DAZ So it looks OK then.

How do I post the spread sheet, so you guys can play around with it and let me know if you find errors?

SIM... Your thread is what got me thinking about an arc... before that my only thoughts and cacls were beased on a "point source", even though there is no such thing

Hopefully this works and you don't have to wait too long. But please remember... this is still just theory. There will be adjustments to apply it practically, but it should give you a good starting point.

I'm suppose to be in bed but I am too hyped up. I need to sleep. Anyways if you email it to me dberladyn@shaw.ca I will host in on my site to be downloaded.

If it's to big to directly upload here as an attachment then try compressing it, 100K limit.
And don't forget rox still hasn't dissected it yet.

DJ

Hi Lucky... thanks for the offer - I've emailed you the file.

Is there no way to provide it here also?

OK, here it is...did some adjustments... now it's only 23K.

I look forward to Rox dissecting it

[Edit] still doesn't work. I zipped it - hopefully it works now.

it works

Too soon for that claim... remember Rox hasn't had a chance to dissect it yet

Gonna be out for the next 6 hours or so. If there are any questions, will try to answer when I get back.

PS. if you wanna unlock the spreadsheet, you'll need a password.
In the spirit of a good riddle, I'll give you a clue - It's the name of a place. That's probably too big a clue

I'm liking this RymGB, trying it out but need some help on a couple of things:
1. Whats the diameter of 317mm fresnel?
2. The diameter of the condenser is 114mm but my mount allows open of 98mm so I should use 98, correct?

sim

the example on 3post is fine for me. I used to work with 0.1234 decimals but this is stupid. Your ideal placement is close enough (less than 0.1 mm error).

good work.

will have a look to that downloadable file now...

This is great work RymGB but it seems like I can't get it to work can someone put these figures in for me

Frensel Diameter--------355.6mm
Frensel Focal ---------220mm

Codenser Dia ---------114.3mm
Condenser Focal---------165.1mm

Lamp s400dd arc--------27mm or 30mm

I had my lamp for a month I'm not sure if it's the one where they changed the ark size or not can you give me both

27 arc case;

arc to condenser; 40.68mm
condenser to fresnell; 166.04mm

30 arc case;

arc to condenser; 39.49mm
condenser to fresnell; 167.92mm

I think you were using "." instead of ",".

Thanks ROX

Easy Q: For the distances, does this refer to the middle of the condensor lens ?

The calced values look good in my case:

Diameter 200
Focal Length 160

Diameter 90
Focal Length 240

Length 24

Arc to Condenser 45.85
measured 37 from flat of lens or 42 to center, I think I could move it out a bit further

Condenser to Fresnel 103.33
measured 108, or 103.3 to center - WOW

Can you model the light increase? That would be sweet

I believe the principal plane is some distance forward like you have observed. These models assume a perfect thin lens were in reality they have two refractive surfaces. Also spherical aberration hasn’t been factored in which could account for some of this forward adjustment. Your results do look encouraging to the accuracy of these models though. What I mean is the 5mm error is common to both sides of the pre-condenser.

DJ

yea, the model is based on ideal thin lenses, i bet your condensor is fater than 0mm

assume the center of the lens as the distance references.

mmm, about the light increase I havn´t found yet a way to cuantify a non point source (arc source) light... Im all ears though

How innaccurate would it be to use a point source?
this assumption seems to work quite nicely when using just a fresnel, based on the light cone solid angle...

Anyhow, something is bugging me with the calculation of the fresnel to condenser distance.
In practice, this distance depends on the focals of both fresnels and the distance to the projection lens.
Eg. if the triplet is moved to adjust focus, ideally the lamp should be moved also to ensure that the light cone converges in the right place (somewhere in the triplet)
I think the fact that my measured distances coincides with the calculated value is just (a freaky) coincidence, and should not be taken too seriously

whats considered the diameter of a rectangular fresnel? its diagonal? the shortest length?(9 12/16" for a LL 15" fresnel) its active length(ie 9" for a 15" lcd)

good point michaelJ, this model is just assuming we want the lamp at rear fresnell's focal. But if ou want you can just change 220 value by any almp to fresnell distance you have alredy worked out. The procondensor lens idela placement does not care on the lamp to fresnell not being the fresnell focal distance.

Just fill the rear fresnell box term by your lamp to fresnell distance.

about how accurate the 0 arc is... you mean the precondensor lens placement or the light cuantification?

Available here uncompressed:

http://dan.berladyn.net/projector/Condense...tCalc-0.52b.xls

1. Diameter of fresnel corresponds to the diagonal of LCD you will be using. If it's a standard 15" LCD, then input 381mm into that field.
2. yes, use 98mm.


I see Rox gave you the values (thanx Rox). I tried it on my end and it's working ok. Is anyone else having trouble getting it to work?? please let me know.


Yes.


Yep, assume distance is to centre of lens as you said.

All the calculations are based on thin lens theory. I realise this is not "perfect" - but... what is? Even thick lens calcs are just theory, and they would be very hard to incorporate into the spreadsheet. Besides, we're using thin lens theory to model all the other optics of the pj anyway .

I can try a very simplistic way to model the light increase (ie. from a point source) but I'm also stuck on that one... How can it be done with an arc source? - I'm not sure.

But a simple point source assumption to model light increase should allow you to compare different setups (ie. using different fresnel/condenser combos). A comparison between different setups "should" be accurate, because the difference is relative - and error is the same in both cases.


Use the diagonal size (in mm) of your LCD.



Perfect answer

And this is exactly what I'll be doing with my setup ... I'm not going to put the lamp at the fresnel's focal. If you don't plan on having the lamp at the fresnel's focal just input your own preffered value into the "focal length" field of the fresnel - instead of the actual focal length of the fresnel. As Rox correctly explained, all the calcs will still be accurate in terms of placement distances.


By the way... did anyone figure out the password yet??

Password to what?

Rox, maybe you should look into a spheroid for the equations?

DJ

Sorry, was just trying to have some fun with the password thing If you want to unlock all the fields on the spreadsheet, and have a look at the equations below, go to: Tools > Protection > Unprotect sheet... then type "lumenlab" as the password.

Excuse my ignorance, but what's does this "spheroid for the equations" mean?

I think I've got a relatively accurate way to determine the %gain from the condenser. It's not perfect... it uses a point... but think it should be good enough. I'll try to explain my proposal tonight and see if you guys agree.

A sphere is a 3d circle and a spheroid is a 3d ellipse.

about how accurate the 0 arc is... you mean the precondensor lens placement or the light cuantification?

I meant the light quantification

Rym, look forward to seeing your proposal, I have a least one datum to check with.
Oh, and a short arc lamp and new condenser on its way...

Great!! I was hoping someone can back these calcs up by real life measured results. There'd be no point if the models don't approximate reality.

mmm, yea, I played with condenser increases calcs some time back as you are (will) sugesting, point assumtion. But this is not propper cuantifcation... I mean, considering the rear fresnell, yes, we can assume a point source because arc/rear fresnell distance is more than 10 times ratio (enginers are teached something 10 times bigger is omitable the smaller one).

But the condensor is so close from the arc that assuming a point source is not valid in my opinion... well, if you post an image i will try to draw on it...

I agree, It's not sensible to assume a point for the condenser, and to make things even worse, the fresnel is now looking at a magnified image of the arc... so the ratio becomes worse for the fresnel also.

I've been playing with different ideas all day... now I've got a headache, and still no solution for an arc source :angry:. I really can't see how it could be done if we don't assume a point source .

A little later, I'll post a very basic method of what I feel is probably the best way to do it (with diagrams), then you can let me know what you think.

Warning: long post. Rox - I think you can probably skip to the third paragraph.

Refer to pics...
Pic1 shows an ideal point on the arc (rays emanating from central point) and a point at the tip of the arc (rays emanating from the arrow head). You
can see that the point on the arc, which corresponds to the exact optical centre of the lenses, will behave like the "ideal" point source. This setup is optimised for the ideal point. You can see that all the rays emanating from this point are passed through the condenser spread over the fresnel. BUT rays from the arrow head are focused to a point that is only seen by the middle of the fresnel. This is a direct result of the virtual arc being too magnified - its extreme points cannot pass through the condenser to the edges of the fresnel... Sorry, you'll have to imagine where the image of the arrow head is (I couldn't get the applet to put it in view .

Imagine also the arc isn't there, and we have just the ideal point. %gain would be easy... It's just a ratio of the maximum (useful) light cone captured by the condenser to the max light cone captured from the fresnel's focal point. Indicated by the green and red angles in the pic. But with an arc, the whole setup changes.

Click to view attachment

Now look at the second pic. This is the setup is optimised for the whole arc, not just the ideal central point. Now the rays from the extreme point on the arc can be spread to the edges of the fresnel, because the virtual arc is smaller. All we did to achieve this is move the condenser closer to the lamp (to decrease Mag) and move the fresnel further (so it can still see the virtual arc at it's focal). But in doing this, we've lost some of the rays from the "ideal" point on the arc. Now only a few of those rays are useful, and the rest are missing the fresnel. The max useful light cone captured by the condenser (again in green) is now only little bigger than light cone from the fresnel's focal (in red). This is the price we pay for the large arc... But remember the condenser has still resulted in an INCREASE - just maybe not significant.

Click to view attachment

I feel using this method (ie. comparing max useful cones) to work out the %gain in an arc system, is probably as accurate as we're going to get. And the comparison will be based on a 3D solid angle, not 2D as shown in these pics.

I know we'd be doing this from the "ideal point" on the arc... but where else can we do it from ?? I'm open to suggestions if anyone has any. Besides, wherever we do it from, it will always be based on a POINT, not an area. Even the calculation we did for condenser placements only dealt with points (we simply picked the point at the extreme end of the arc, and made sure it can pass through the condenser, to the edges of the fresnel).

My only assumption here is that %gain for the ideal point on the arc, is equivalent to the %gain for all the other points on the arc, and things average out... this is probably not true, but it's an assumption I HAD to make... if only for the sake of my mental health .

Wow RymGB! Those are interesting plots. I see your software is performing some ray tracings for you, but apparently still using lens formulas to do so.

Your spreadsheet is really nice. I wanted to use it to compare your calculated values for what my own fundamental ray tracings (using no lens formulas or assumptions) seem to recommend for an arc to condenser spacing.

Therefore I used the following values for the Surplus Shed condenser:

Fresnel: 381mm dia x 317mm FL
Condenser: 114mm dia x 165mm FL
Arc: 27mm

Your spreadsheet calculates an arc to condenser spacing of 49.1mm, compared to my CAD plot of 35.8mm. However, I realize that your spacing is intended to be to the center of the lens. The actual Surplus Shed lens is 22.5mm thick, so if we take one-half that amount, then your value reduces to 37.9mm - only a 2mm difference.

Edit: Although your condenser to fresnel spacing differs a lot more - about 30mm - to the center of the lens.

So our methods might be converging in some ways after all.

Hi Dajyn,
Good to see the methods are converging. To be honest, I think your way is more accurate, but it's very hard to put into a spreadsheet. The more optical components we use, the less accurate thin lens calcs become and tracing through every optical surface becomes necessary. But we're only dealing with 4 optical surfaces here, so I think thin lens calcs should be close enough, as you found comparing with your traces.

Just curious though... what did you calculate for condenser to fresnel spacing using your method for the example above?

oh... btw, the traces above don't require special software, I used an online applet - just search google for "virtual optical bench", there are many around. They're great for quick traces, and simple observations, but they're not exactly feature packed . I usually use this one http://www.hazelwood.k12.mo.us/~grichert/optics/intro.html - but if you manage to find one that has adjustable lens size, please let me know... I'll be forever grateful.

Does anyone have any thoughts on the %gain calcs ??

If not, I'll implement the method I described above, then hopefully someone with a test setup and a lux meter would be willing to help me determine if it's accurate.

I’m still thinking about it.

DJ

223.6 mm to the flat side of the precondenser lens. This is determined purely through geometry by fitting the lens within the focal triangle of the fresnel. That value is independent of any method to ray trace or position the lamp arc.

I tried the simple light cone angle approach and answered my own question about accuracy:
Calculated gain: >100%
Measured gain: 25%
lol
Will be interesting to see what the measured gain for the same pre-condenser, but with a 3 times shorter arc will be

...

But curious... how did you measure gain? and where? (ie. at fresnel, triplet, or pj output).
I"ve got some semi-finished gain calcs in the spreadsheet. I'd be very interested in running your numbers through and see what it calculates, I don't think it'll be over 100%.

[Edit]: Forgot to ask could you give me the lminous flux for your lamp also... So I need arc size, luminous flux, fresnel size and focal, condenser size and focal

Measured at screen, average of ANSI 9 points, but gain at center and corners is roughly the same.
I'm curious to see what you come up with...

MichaelJ - Could you give us the sizes and spec's of all your components again, for reference?

25% is still a pretty significant gain.

Thanks!

Actually... Dajyn's right... please tell us the specs for ALL the components

sure!

HQI 250W 24mm arc
90mm dia 240mm FL pre-condenser
160mm fresnel
8" Hami LCD
305mm fresnel
64mm dia (clear) 300mm FL triplet

hows that

Thanks!

Initial calcs give me 22.7% gain, but if I knew the dimensions of the panel it might be closer - is the Hami 8" 16x9 or 4x3?

Also just to be sure I'm not making any mistakes - back fresnel 160mm fl, not diameter - correct?
And do you know how many lumens your lamp outputs?

Will check back tomorrow... it's 9:30am, and I'm yet to hit the bed