Hello ppl,

I'm thinking of building myself a projector using an old HPS armature I have lying around.. HPS also being one of the more efficient lightsources
(150lm/W compared to MH 100lm/W) makes it very nice to use in a projector where every extra lumen is welcome.

Diggin into the matter, I read in a previous post that someone did not recommend HPS as a light source because of the warmish light.
However lights can be filtered and the procedure is really not rocket science.

Turning a 3200K Kelvin Tungsten (normal lightbulb) into 5500Kelvin daylight (pure white) takes only a blue filter (80B it's called).
Filtering a 2000K HPS into 5500K light takes some more blue, but is still not impossible.
Now the downside to this is that you loose some light when filtering, but other than that I see no other downside.

Any thoughts on this?

btw. anyone know response time for the PSone display?

Each lamp has a color spectrum. What matters the most is how much red green and blue [RGB] they make not the color temp or lumenis per say. The only thing we care about is what makes it through the LCD color filters. Look at the spectrograph of each lamp the LOWEST level of RGB is the level that you will have to set the LCD's RGB to have a balanced color spectrum. If you take the lowest value of RGB output of the lamp and add the sum of the colors at that level this true usable output. You will find that any lamp that is deficient in any part of the RGB will have to be balanced somehow. It does not matter if you use external or internal color filters none of them will create light so you always have to balance to the lowest level.
Most of the lamps we use are deficient in red and have a large green spike but the HPS are much worst.

I spent some time a few months ago looking for some spectrographs of lamps and was going to start a thread on it but maybe this a good place to start.

The first image is a "pro short arc" lamp the second one is the best lamp spectrograph I have ever seen from the venture 950 series. Link- http://www.venturelighting.com/LampsDataSh...elamp/95103.pdf

I could use some help if any one can find the spectrographs of the m400sx or the mh400u I lost the one I had.

Click to view attachment



Click to view attachment

Hello Arizona,

thanx for the input. But still I'm not sure of what you're trying to say..

Filtering a lamp changes the colour temperature.
Colour temperature (K) has *very much* to do with how much Red Green and Blue is perceived.

Tuning the LCD to the 'lowest value of RGB output' is doing exactly this - you are adjusting the whitepoint
of your LCD to match the whitepoint of the lightsource. So what you are doing is actually exactly the same thing!

If I remember correctly, HPS more or less fills in all of the spectrum, however it contains less blue than MH
and therefore (as previously stated) needs more blue to reach white.

I went to the Venture Lightning page to find some HPS graphs but accidently found something else.
'White-Lux Plus' HPS compatible lamps with a colour temp. of ~4000K. Interestingly enough this is a MH lamp
and it is also available with integrated UV filter
(http://www.venturelightingeurope.com/catalogue/types.php?spec=10077)

Anyway, to add to all this confusion there is also CRI which measure the purity of light (purity meaning the conformity of the spectrum).

According to the following data, both MH and HPS are almost equal in this aspect.

Incandescent 1A
Metal halide 1A ... 2
Fluorescent 1A ... 3
High pressure sodium 1B ... 4
Low pressure sodium 4

1A 90...100 accurate colour matching Galleries, medical examinations, colour mixing
1B 80...90 accurate colour judgement Home, hotels, offices, schools
2 60...80 moderate colour rendering Industry, offices, schools
3 40...60 accurate colour rendering is of little importance Industry, sports halls
4 20...40 accurate colour rendering is of no importance Traffic lighting


PS. To all you people with itching fingers. Keep this thread uncluttered please unless you have something with value to add. Thanks.

MNA, I think what arizonavideo is getting at is that a typical HPS Spectral Energy Distribution is in a range that will be blocked by the filters in the LCD, converted to heat and not projected. Have a look at the SED of the bulb below. Most of the light produced is in the yellow region. None of the filters in the LCD will pass much of this color. There would be so little light left, you'd be better off using a 100 watt halogen. This bulb is supposedly an enhanced-spectrum type for growing plants. A standard HPS SED would likely be worse.

That graph says it all do you see any green at all? All the filters in the world will not make green. They call them filter for a reason they block light so to make an even color balance with this lamp you will have to block all the yellow down to the same level as the green!

Another way to look at this pretend to have three lamps all running at 100% output one red one green and one blue {this is what a TV is} if the red lamp is 150 watt and the green lamp is 25 watt if you want a neutral color balance you will have to lower the output of the red to 25 watts.

If you look at a white LCD with A 20X lope it is not white at all you will only see three color bars one red one green and one blue. All that maters from this point on is the relative brightness of the three colors.

The whole debate about lumens and efficiency of a given lamp is meaningless if the color spikes are not fairly close to centered on red green or blue and even in level. What maters is the "power under the curve" for each color.

We still need the graphs of the ushio lamps that most people are using. Any one found one yet?

But the amount you lose with a filter can be quite substantial. A Rosco #3202 full blue is only 36% transmissive.

even if the filter only eats about 30% of the light(probably be much more) you'd be down to the efficiency of MH ... in which case you might as well use the s400dd which will run on an HPS ballast..... specs on the psone(the sony or the verge, same sharp panel) is 25 ms and 350:1 contrast, 320x240

Thanx MikyD! Just wondering why nobody told me this from start ..... dooooh! This solves everything..
This info should be stated in the shop department. Btw. just found out Philips also makes something alike. 'MASTER HPI-T Plus' but it has much lower CI value (65 compared to 89).
Still I would love to see spectragram for the S400DD.

Btw. regarding the PSONE display, I thought it was 640x480? 25ms and 350:1 seems acceptable for such a small thing.

well I think its in the reference/tutorial or the wiki...anyway, no its a qvga(320x240) but you can do a mos to it so it will run a 640x480i signal from a pc and it makes a nice little pj, works good up to about 60" diagonal

Found it. The UHI-S400DD USHIO.

Yes the venture natural white 950 makes over twice the red over the USHIO. It is BU bace up but on some of thier lamps they say you can run a BU lamp vertical with a slight increase in color temp.

Click to view attachment

just looked at the venture 950....big arc length 43mm compared to what, 24-27 on the usual MH bulbs?

I think it will still fit but it will be close. For a reg setup 330/220=1.5 mag 1.5 x 44mm arc=66mm a little big. My 80mm triplet has 62mm clear.

The pro lens is no better. 650/220=2.95 mag 44 x 2.9=129mm!
The pro lens with a 550mm fresnel looks a little better
550/220= 2.5 x 44mm=110mm arc still a little large.

So it looks like the standard lens is better?

yeah, I found that out in a discussion on lens shifting(upshot: pro is worse for lens shifting also, for the same reason)

What was the arc length of the venture lamp that everyone was using?

I've got the 250watt version(and I think the 400watt jas the same dimensions) and I think its the same arc length as the ushio,24-27mm

Btw. what is the expected light output for a typical LL projector?
Does the LCD really take ~60% of the light? 40% of 40k = ~16000lm which is unbelievably good or is this wrong? Anyone know how ANSI lumen is calculated ?

Also, what are the differences between the pro lens and the regular?

estimated is 200-250 lumens , the average lcd is 6% transmissive(yes, thats right the lcd eats about 94% of the light)..... the fresnels another 20% take a look at the excel file in this thread for theoretical calcs:
http://www.lumenlab.com/forums/index.php?showtopic=8594&hl=

I believe ansi lumen is calculated using a light meter(CIE compliant I think is the qualification on the light meter) and tested at 9 points on the screen and lumens are based on a square meter(ie 200 lux average on a one square meter screen = 200 lumens) although I do tend to get lux(which is what the light meter reads) and lumens confused