Convert Lux meter to PAR meter?

[sowNreap]

Let me see if I have this straight. I think you sent me these 6 filters: Roscolux #17, 36, 55, 363, 4430, and 3304 none of which are the ones you used in this last mod. But if I was to use those filters it would still read in PAR just not as accurate as this new better way. ???

If so, and you have too many requests for the new filter stack, I can just use what you sent. But IF you have any extras after that, I'd appreciate if you could send me a new stack. Just in case I can't make it to the Rosco dealer in Knoxville .. which is about 70-100 miles away.

[Hoppy]

Originally Posted by sowNreap

Let me see if I have this straight. I think you sent me these 6 filters: Roscolux #17, 36, 55, 363, 4430, and 3304 none of which are the ones you used in this last mod. But if I was to use those filters it would still read in PAR just not as accurate as this new better way. ???

If so, and you have too many requests for the new filter stack, I can just use what you sent. But IF you have any extras after that, I'd appreciate if you could send me a new stack. Just in case I can't make it to the Rosco dealer in Knoxville .. which is about 70-100 miles away.

Yes, you have it right - but one big difference between the two filter stacks is that the stack of 6 you have makes the meter read PAR directly, while the new stack makes it read 10X PAR, so if you use the 20,000 lux scale, it reads PAR directly, and more accurately. I mailed you a set of the 4 new ones.

[sowNreap]

Thanks Hoppy for the clarification and for mailing the new filter stack.

I just wanted to make sure everyone that might want to do a mod on their lux meter had at least 1 of the filter stacks before I asked you for the new one. The more people we have that are willing to take lux/par readings on their light fixtures the more everyone on TPT will benefit.

[shd17]

So if I use the 4 filters in your final versions, how would I tell that the lux meter is reading par correctly (or close to it)?

[Hoppy]

Originally Posted by shd17

So if I use the 4 filters in your final versions, how would I tell that the lux meter is reading par correctly (or close to it)?

You would have to borrow a Quantum PAR meter and verify the accuracy. However, each of the Rosco filters has a percent transmission number, which is shown on the Rosco website. If you multiply those 4 numbers together, .88 x .64 x .40 x .61 = .1374 which is 1 divided by 7.28. If there were no other factors involved, it should be 1 divided by 7.6, or .1316. However, since the lux meter photodiode isn't equally sensitive to all wave lengths, it worked out that the .1374 was a near perfect total transmissivity.

The only things this modification does are:
Filter the light striking the photodiode to reduce it by a factor of .1374.
Adjust the spectral response curve to make is a bit more consistent with what a PAR meter response curve should look like.
Using the same optical path, make a waterproof sensor, so it can be used in the aquarium.

The resulting "PAR meter" will be as good as the Mastech lux meter is, but it will just read differently. And, as far as I can tell so far, the Mastech lux meter is all that it needs to be to be a good meter for our uses.

[shd17]

So i tried measuring a 14w cfl in 10in dome light and i got 77-80 par at 12", 40-42 par at 19". Does that sound reasonable? Or is it off?

[Hoppy]

Those PAR readings are very consistent with the reading in Diagram 7 of http://www.plantedtank.net/forums/sh...2&postcount=21

[Hoppy]

I have been working steadily, day after day, trying to make this DIY PAR meter work better, and be easy to duplicate. I think I now have it about as good as it is going to get. After a lot of wheel spinning, and wasting parts, I finally sat down and figured out what it should take to get where I want to be. First, the cosine diffuser, I figured out, is just a way to put a light source for the sensor's photodiode at a fixed distance from the diode, and with a brightness that is proportional to the brightness of the light being measured. The amount of light this provides to the diode for measuring depends on how far it is from the diode, with the brightness dropping with the square of that distance, making the distance a critical dimension.

Next, I figured out that filtering the light to reduce its intensity at the diode, and modify the spectral sensitivity of the diode, isn't a good idea. This is because the diode does have some small sensitivity to infrared and UV, even though there is a glass filter mounted on it to greatly reduce its sensitivity to the non-visible light. But, the Rosco gel filters are not able to block the infrared light at all, or at least they only block a small portion of the infrared. That means as I filter out visible light to flatten the spectral response, and reduce the intensity, I am also making the IR light sensitivity go up as a percentage of the total sensitivity - not good at all. So, I have to use a minimum of filtering, preferably only one relatively weak filter.

But, you can "filter" the light by reducing the aperture the light shines through. The amount of light going through is roughly proportional to the area of the aperture - the square of the aperture diameter. So, the aperture is another critical dimension.

The best way to be able to reproduce both the aperture and the distance between the photodiode and the cosine diffuser is to make each of those parameters be set by an individual part, with one easy to control dimension. Along with that, the design has to position the photodiode in a fixed location.

With all of that in mind I came up with this: (Using an Excelitas photodiode http://www.newark.com/jsp/displayPro...79268381|plid| instead of the lux meter diode, to get around the filter on the lux meter diode.)



This is made by first making the 6 acrylic pieces. I used plastics from Tap Plastics, http://www.tapplastics.com/product/p...ylic_discs/137 http://www.tapplastics.com/product/p...ylic_tubes/141 http://www.tapplastics.com/product/p...rylic_rods/149 Also, I had a scrap piece of 1/16" thick acrylic I used for one piece.

Starting at the top: I drilled a centered .25" hole through one of the 3/4" diameter acrylic discs. Then I used black nail polish to paint the inside face of that disc black. I poked and tapped in a short piece of .25" dia. frosted acrylic rod in the hole with a bit of Weldon #16 cement on it. When it cured I hack sawed off the excess rod and used fine sandpaper to trim it flush on the top side of the disc, with the rod already flush with the bottom side.

For the aperture, I drilled a .125" diameter hole through a 3/8" square of 1/16" acrylic. and painted the inside face of it with black nail polish. When that cured I cemented it on the bottom of the disc with the frosted rod in it, so the hole was centered over the frosted rod end. This is the cosine diffuser plus aperture.

Next, I cut a piece of 3/4" OD, 1/16" wall acrylic tube about 3/8" long with square ends. (I used a small cheap miter box and hacksaw.) Before I cut it, I drilled a hole centered on one side, with the hole diameter the same as the electric cable diameter.

To mount the diode, I used another of the 3/4" diameter acrylic discs, with a carefully centered .312 dia. hole through it. (Drilling centered holes is easy if you first make a jig from a piece of plywood, drilling a shallow 3/4" dia recess to fit the disc, then using the center of that hole to locate a .25" or .312" hole the rest of the way through the plywood. Push the disc into the hole, clamp the plywood to a workbench with the disc captured between the plywood and the bench, then drill the hole through it, using the hole in the plywood as a centering guide.) The Excelitas diode is .312 in diameter, with two parallel cutoffs making it a rounded rectangle, so it fits snugly into the .312 hole in the disc. With the face of the diode at the top of the disc, I used Weldon #16 to glue the diode into the disc.

Poke the electric cable through the hole in the 3/4" tube piece far enough so you can strip the wire so you have two stripped wires about 1/2" long. Use the plywood drill guide to hold the disc that holds the diode, so you can locate the electric cable in between the two electric leads to the diode, and wrap one wire around one lead and the other around the other lead. The positive lead is identified with a "+" marked on top of the diode. The red wire goes around that lead. Solder the two connections and cut off the excess wire.

Carefully push and pull the wire back to get the diode inside the piece of 3/4 tube, with the disc sitting on top of the tube. Use Weldon #16 to cement it on, being careful not to get it on the diode sensing surface. Apply a filet of Weldon #16 around the cable inside the 3/4 tube, and outside as well. This seals the water out of the inside, and acts as a strain relief to avoid pulling the wire off the diode.

Cut a .25 inch long piece of the 3/4" acrylic tube to be the spacer that positions the diode .25" from the cosine diffuser. Keep the faces parallel and square, and use fine sandpaper to get the length accurate. Use Weldon #16 to glue that on top of the diode mounting disc.

Cut a rough circle of Lusco #3313 Tough Minus Green filter and put that on top of the diode - it can be loose since it can't go anywhere.

Use Weldon #16 to glue the cosine diffuser/aperture to the top of the assembly, with the aperture to the inside. After that cures, glue the assembly to a strip of acrylic about 1.25" x 5" to seal off the bottom and act as a handle for the sensor. Add another coat of Weldon #16 to the outside of all cemented joints to be sure they are waterproof.

Use black nail polish to paint the exterior surfaces, being very careful not to overlap the frosted rod diffuser disc.

Connect the end of the electric cable to the luxmeter readout. I cut the cable to the readout off as close to the luxmeter sensor as I can, to have a straight section of wire to make the soldered connections. Connect red to red. Use shrink tubing to insulate and reinforce the connection. Since the luxmeter cable wires are very thin, and weak, I use the ground wire on the electric cable, not otherwise used for anything, to overlap the insulation of the luxmeter cable by an inch or more, and use black electric tape to tightly hold it to the insulated cable to relieve the load on the soldered wires.

The performance of this PAR meter is:

[mistergreen]

Did you find a big different with and without the minus green filter?

[Hoppy]

Originally Posted by mistergreen

Did you find a big different with and without the minus green filter?

No, that filter is just there to theoretically broaden the spectral response. What I did find, was that using many filters stacked up to get the broad response I wanted, jacked up the IR respons excessively, so hot light sources gave 10X response by the meter compared to cool light sources. So, I went for the minimum of filtering. The 1/8" aperture alone didn't drop the meter reading enough, and 3/32" dropped it too much, so the aperture size to correct that would have required buying a set of drills in very small increments of size around 1/8th inch. So, I looked for a filter to fine tune it, one which would do a minimum of IR emphasis, and not drop the response in the blue area. This one fit that criteria.

Before I can do actual calibrating of the sensor I have to glue it all together and paint it black, making it a permanent assembly, so I was very reluctant to do that kind of testing. I just used black electrical tape to hold it together and block light leakage as I experimented, with just a 10000K PC bulb. Even with this I used up 3 of the photodiodes in unsuccessful assemblies. But, I plan to try to saw them apart to salvage the diodes. This has been a very interesting project, but also extremely frustrating.

[Hoppy]

I think I am finally satisfied with my design, after a lot more work. I finally thought to look up the spectral response for the Apogee Quantum PAR meter, to see what they use and how good it is. I was surprised to find that they also have a far from perfect match with the 400-700 nM PAR spectral range, and, as a result their meter is only good to about +/- 4% for typical aquarium lights, and less accurate for incandescent lights. So, I tried to match their spectral response as best I could with the photodiode and Roscolux filter gels that I have. That proved to be relatively easy, once I worked out the details of what the filter response should look like. I settled on using the Roscolux # 4815 filter gel, giving this spectral response vs the Qualtum meter spectral response:


That left the problem of how to adjust the calibration of the meter without changing the spectral response. So, I noticed that in the Roscolux sample book are 19 different diffuser filters, white diffusers which would not alter the color of the light passing through them. I set up a PAR meter and checked the reduction in the reading that each of those diffusers gave me when I put one on top of the PAR meter sensor. The transmissivities ranged from 45% to 90%, so by using different combinations of those diffuser filters I could adjust the sensitivity of the meter over a very wide range.

With that problem solved, I made some modifications to the last design of the sensor, all aimed at making it easier to make the parts and to assemble them with minimal variation from assembly to assembly. My final design, which I prototyped, was:


With this design I can make 5+ of each part in about 30 minutes, with the aid of a few very simple fixtures. Assembling the parts was very easy, even with the tiny parts.

Today I received a batch of 5 more lux meters, so I was able to assemble a complete PAR meter for testing and to determine what diffuser filters I needed. It took just 2 diffuser filters, Roscolux # 102 and #114, to get a near perfect match with a Quantum meter. This PAR meter is now well within my goal of +/- 10% accuracy. I have two more just about assembled now.

If you have reasonable mechanical skills you should be able to easily duplicate this and make them to sell, if you wish to do so. The Roscolux filter sample books are free at a dealer, or you can buy one at Amazon. The acrylic tubes are available on EBay. You should have little difficulty finding usable flexible 2 conductor cable at a local electronics supply store, although Radioshack doesn't seem to have any.

I have a list of 7 people who have asked that I make them one, so I will finish off that list, and maybe make a few more, then I hope someone else picks up this relatively easy, high demand "product" for others to buy.

Here is the assembly sequence I am following:
Assemble Photodiode to item 5.
Apply acrylic cement to holes in item 5 to attach diode.
Thread cable through hole in item 6, and solder cable wires to leads.
Assemble above assembly to item 6, with acrylic cement.
Paint outside cylindrical surface of above assembly with black nail polish.
Glue item 4 to item 1 with acrylic cement.
Paint outside cylindrical surface of that assembly with black nail polish.
Load filters into item 4-1 assembly and push item 3 in to retain.
Use masking tape to assemble above assemblies for testing.
Adjust diffuser filters to get correct PAR reading.
Apply 2 drops of acrylic cement to hold item 3 in place.
Glue the two sub-assemblies together with acrylic cement.
Glue above assembly to item 7, and paint bottom of item 7 under item 6 with black
nail polish. Glue on a “socket” for holding wand made of piece of acrylic tube to
above assembly.

This modified meter works on the maximum scale of the lux meter, the 50,000lux scale. I just use a couple of drops of acrylic cement to lock the range selector slide switch in the right position.

[sowNreap]

Great work Hoppy!! I figured that you wouldn't quit until you had it perfect in your eyes.

[jerrytheplater]

Is this the meter you are offering for sale here?:
http://www.plantedtank.net/forums/sh...d.php?t=308809

I just bought one from you and did not see this thread till now.

[Hoppy]

Originally Posted by jerrytheplater

Is this the meter you are offering for sale here?:
http://www.plantedtank.net/forums/sh...d.php?t=308809

I just bought one from you and did not see this thread till now.

Yes, this is the PAR meter you bought. As I make more of these the mechanics of doing it get a bit easier so they tend to be more nearly exact duplicates. It is still difficult to get the photodiode mounted at exactly the same distance from the diffuser, so I am finding that I have to tweak the diffuser filter selections a bit at times. When all goes well I use the same 2 diffuser filters in all of them. (The meter calibration is done by choosing diffuser filters which also change the total amount of light passing through them.)