DIY PAR meter

[mcgyvr]

Anyone made one? I'm about to start and was just looking to see if anyone has already been down the road and might let me know any issues/roadblocks/bad sensors/issues they have had.
Right now with my current plan I'm looking at about $5 in electronics parts and a waterproof housing of some sort.

I will need someone with a real PAR meter to either let me borrow or I will send my sensor to them when its complete just to take readings and see how accurate and if any additional scaling is needed.
My plans are to make the sensor just hook up to a typical multimeter and the sensor would have a 0-5V output that can easily be read by the meter.
I'm looking at a 2000 micromoles per second per square meter upper limit so .0025V=1 micromoles per second per square meter

[BeanAnimal]

You are going to have a hard time getting predictable results. While you can cobble together a sensor and diffuser, their interaction is not going to be anywhere near linear. Building a "correction" look-up table to translate the meter results is already a given, but the problem is that your DIY sensor is will likely need a diffrent table for each different bulb spectrum and will still have a lot of error in it.

The diffuser is very important for correcting cosine error and without a lot of trial and error, you going to need a lot of correction. Correction could be done via software loaded into a PC or microcontroller, but the DIY sensor is going to still be the gotcha.

Even buying an OEM sensor (what I have done for my project) has its hurdles. I still have use software to correct the raw sensor values before displaying useable results.

[mcgyvr]

Points noted... Thanks.
I plan to use GaAsP photodiodes which require far less cosine correction as their angular response is far more linear than Si photodiodes. Not to mention I'm not using this in an external environment with dawn to dusk light angles nor do I need lab grade results.

I just can't force myself to fork over $299 bucks or whatever it is when I know darn well its only a $2 or less photodiode with an op amps in those Apogee units (which are far from lab grade anyways). Not to mention I think alot of users would be interested in an easy/cheap DIY solution that gives you usable PAR readings.

[shlap]

Originally Posted by mcgyvr

Points noted... Thanks.
I plan to use GaAsP photodiodes which require far less cosine correction as their angular response is far more linear than Si photodiodes. Not to mention I'm not using this in an external environment with dawn to dusk light angles nor do I need lab grade results.

I just can't force myself to fork over $299 bucks or whatever it is when I know darn well its only a $2 or less photodiode with an op amps in those Apogee units (which are far from lab grade anyways). Not to mention I think alot of users would be interested in an easy/cheap DIY solution that gives you usable PAR readings.

Heck yeah we'd be interested! A cheap reliable PAR meter would be the holy grail of this hobby (along with cheap lighting).

[redfishblewfish]

I’m here…..I’m watching…..I’m waiting, with soldering iron in hand.

[Sharpie_]

Originally Posted by redfishblewfish

I’m here…..I’m watching…..I’m waiting, with soldering iron in hand.

Don't burn yourself.

I can't wait to see how this turns out, even if it means having to input our readings into an executable or table on the desktop to get the real data, i'm okay with that.

[kbb0118]

The DIYers on RC are da bomb. Seriously I didn't know this underworld of reef tinkerers existed...building controllers and stuff. Its awesome! Someone smarter than me will have to figure this one out but I will support in any way I can. I don't understand circuits fully (yet), but it is my new hobby to learn.

[widmer]

Originally Posted by BeanAnimal

You are going to have a hard time getting predictable results. While you can cobble together a sensor and diffuser, their interaction is not going to be anywhere near linear. Building a "correction" look-up table to translate the meter results is already a given, but the problem is that your DIY sensor is will likely need a diffrent table for each different bulb spectrum and will still have a lot of error in it.

The diffuser is very important for correcting cosine error and without a lot of trial and error, you going to need a lot of correction. Correction could be done via software loaded into a PC or microcontroller, but the DIY sensor is going to still be the gotcha.

Even buying an OEM sensor (what I have done for my project) has its hurdles. I still have use software to correct the raw sensor values before displaying useable results.

So how do store bought PAR meters not need a table or software to correct for their results?

[mike_cmu04]

I am sure it is doable i know i would try....I just dont have a spare par meter laying around to tear apart to see what makes it tick.

[mcgyvr]

Electronic parts have been ordered ($8 total)..should be here in a day or so
Just need to order the parts for the mechanical enclosure which should also arrive in the next day or so..
Hopefully I will have a working prototype this weekend to start playing with.

[williah]

Interested, watching and reading carefully...

aside: I've got to say, from an electronics pov, RC is better than College. I have more practical examples here than I can follow. Awesome.

[XSiVE]

Originally Posted by widmer

So how do store bought PAR meters not need a table or software to correct for their results?

they actually do have correction built into their onboard software, for example the commonly used apogee mq-200 only measures between 360 and 670 nm spectrum, and has a couple different settings based on "expected" type of light. The closest thing to our aquarium light is the "sun" setting, which is good, but it does not lend enough internal correction to the blue spike in most of our lighting thus the advanced LED guys suggesting they should add at least 15% to their PAR readings based on the intense blue spike they give.

[widmer]

Not knowing a thing about how they actually work, I would assume that a PAR meter cannot actually distinguish between wavelengths of light that hits it, and instead employs a light filter on top of the sensor that will block out light that is outside of the 360 to 670 region. Does that sound about right?

[XSiVE]

Originally Posted by widmer

Not knowing a thing about how they actually work, I would assume that a PAR meter cannot actually distinguish between wavelengths of light that hits it, and instead employs a light filter on top of the sensor that will block out light that is outside of the 360 to 670 region. Does that sound about right?

I dont know exactly either, though I would assume its more on what the sensor can "see".. to expand on that, it would be a certain subset of radiation(nm) excites certain somethings within the sensor which generate voltage, that is translated through a table then some sort of weight-based calculation of known photosynthetic wavelength ranges.

....but I could be WAAAAAYYYY off

edit: here's some basic info on photodiodes http://en.wikipedia.org/wiki/Photodiode

[mcgyvr]

The photodiodes typically have a cover/BG filter that blocks UV light.
The sensor I'm using has a 400nm to 700nm spectral range with max sensitivity at 550nm.
I just finished the diffuser. It's a 1mm thick 3/8" dia teflon disc. Turned it on the lathe at work. Ran it really slow so I have very good surface finish.

[theatrus]

Originally Posted by widmer

Not knowing a thing about how they actually work, I would assume that a PAR meter cannot actually distinguish between wavelengths of light that hits it, and instead employs a light filter on top of the sensor that will block out light that is outside of the 360 to 670 region. Does that sound about right?

Pretty much, coupled with tuned sensitivity on the substrate its self.

To the person building it, which part number diode did you use? Whats your amplification strategy?

I have an Apogee sensor here (not with the overpriced multimeter though), so we can do some comparison readings. I have some MH, LED and T5s amongst various setups.

[mcgyvr]

I'll publish the parts list/schematic information when it is completed... (once I'm done and satisfied with the results).

I will take you up on your offer about comparing it with the Apogee sensor though. Once its completed.. (tracking info now says I won't have the electrical parts till monday..boo hoo)

[theatrus]

Once you post the parts/are relatively happy with it, let me know

[mcgyvr]

Teaser pic.. body and diffuser. Just gonna paint it black.

[ERICinFL]

This is gonna be SWEET! I can't wait to see what you come up with and the final results. Pics pics and more pics please, as you go along. I'm just a dumb fireman, but I can follow pics and a good description.

[BeanAnimal]

Originally Posted by theatrus

Pretty much, coupled with tuned sensitivity on the substrate its self.

To the person building it, which part number diode did you use? Whats your amplification strategy?

I have an Apogee sensor here (not with the overpriced multimeter though), so we can do some comparison readings. I have some MH, LED and T5s amongst various setups.

The "overpriced" multimer has filtering and amplification circuits that correct the sensor readings and fit them to the PAR curve. That is, the op-amps transfer functions are tuned to profile of the sensor and its physical filters and designed to correct the non-linearities and weight the response according to the proper curve and the defects in the sensor and filter/optics. In theory it is a very simple device, in practice building it so that the results are accurate and meaningful is no as easy.

I have a lot of experience with tri-color devices (used for display calibration) and even though many use the SAME sensor, the supporting hardware (filters, amplificiation and/or software correct) make the difference between a useless toy and a precision instrument.

Also consider, that many of the sensor manufacturers will tweak the specs of the sensor based on the needs of an OEM

[theatrus]

Bean, I completely agree with you on the non linearity part. Its going to need some tweaking and lab testing.

The newer apogee sq110/120 deliver a compensated signal direct from the sensor; they sell it for use with external data recorders. In fact their handheld now has the sensor stuck inside of it (which lets them change calibration with code) .

[BeanAnimal]

I am using the sq110 in my lighting controller... It is gathering dust on a shelf waiting for a motivated rainy day (along with 1000 other dusty project components waiting for their rainy days).

[mcgyvr]

Parts came in last night.. Got the sensor soldered to the wires and installed in the housing and all sealed up so its waterproof.. Got to complete the external circuitry (amplifier/9Vto5V convertor,etc.. tonight, should have the unit completed by tomorrow ready for testing.. I want to set up initial calibration/output based on the noon sun to be midway in the output. So far so good. Sensor output seemed really linear and the teflon diffuser seemed to be the perfect thickness (1mm).
I'll finish the unit and dunk it in the tank and take some measurements to see if its actually giving a reasonable output.

Still looking to see if anyone will let me borrow their PAR meter to compare readings.. I think the LFS has one that they will let me borrow but if anyone else wants to offer that would be great too. IT should only take me a day to compare readings and be able to scale the sensors output to PAR readings.

[mcgyvr]

Sensor is finished/working as designed... (need a trip to radio shack for an enclosure for the battery and circuit board but thats just the frills to make it look pretty)
Need to get out in the sun tomorrow at noon (yeah for the weekend) for baseline calibration. Found a great site from Apogee that lets me calibrate the sensor (close enough for now) based on the noon sun in my area. http://clearskycalculator.com/

I will then be able to take some test measurements in my tank to see where I'm at.. Overall I'm very happy with this first go at it. I will post information (schematic/parts list,etc..) in a few days or so when its finalized. Still need to figure out a good way that anyone can create the housing/diffuser (I'm lucky and have full access to a machine shop so mine is a little over typical DIY right now).

But the response/linearity of this sensor I used is GREAT.. Seems VERY accurate to even the slightest changes in light levels. Even a .005" change in height from the light source registers a change in readings. (slid some shim stock under the sensor)

Found a local guy with a PAR meter that is willing to let me compare my readings to the Apogee unit too.