Re: DIY UV Spec seeking photodiode circuit help

[Josiah Zayner]

I think the simplest way to make a UV spec would be using filters and a very strong light source. Lots of microplate readers do this.

Also, you say you have high dark voltage readings. Normally one takes a background reading then subtracts out background readings when doing an actual measurement.

Have you tried something different to block out the UV light? I don't know what you are currently using.

Are you using Beers law to calculate your absorbance values?

Also, try other diodes. Would be a waste to spend all this time on this one to find out that someone on Ali Baba sent you a piece of shit.

On Sunday, October 7, 2012 5:25:11 PM UTC-5, Ben Hunt wrote:

Hi!

I have been interested in building a homebrew UV spectrophotometer in order to know the purity and concentration of DNA samples for a while. Mostly I've been discouraged by the high price of UV LEDs on thorlabs and other reputable optical outlets.

Recently I've found some lower cost (80-120) 260 and 280 nm LEDs on ali-baba, which has reinvigorated my quest. I am just trying to build a device which can measure concentration of DNA (at 260) and purity (by 260/280 ratio).

I bought a reasonably useful silicon UV detector (photodiode) from Edmund Optics and hooked it up to an arduino in open voltage, you can see my results at diyscibuildjournal.wordpress.com. As you may have guessed my signal is extremely noisy, even using the large amount of light I get from a normal LED. The UV LEDs will give off much less and the detector will be less sensitive.

The literature tells me to try operating the photodiode in reverse-bias and also try a half-dozen other circuits (this is a good group of them, for example). I have bought most of the op amps and transimpedance amps listed in the paper, but I have very little idea of what is going on on the ground floor of the circuit.

What have read is that the linearity of the photocurrent to incident light is highest around zero volts, but I am mystified by the idea of a circuit that holds the photodiode at that voltage while still being able to get that current and transform it into a (non-zero) voltage.

I am hoping someone out there in the DIY BIO community might be able to help me out with some of this basic electronics knowledge so I don't end up just trying a half dozen circuits until I get something that has low enough noise to be usable or worse, fail and not know why I am failing.

Thanks!

Ben

[Nathan McCorkle]

You need to compensate for line noise (noise in power/ground line) by doing corellated double sampling (cds), you can implement this yourself or buy a video ADC for a few dollars that already incorporates it, check openspectrometer.com for the wolfson part number I am using.

On Oct 8, 2012 5:34 AM, "Ben Hunt" <ben.g...@gmail.com> wrote:

Hi!

I have been interested in building a homebrew UV spectrophotometer in order to know the purity and concentration of DNA samples for a while. Mostly I've been discouraged by the high price of UV LEDs on thorlabs and other reputable optical outlets.

Recently I've found some lower cost (80-120) 260 and 280 nm LEDs on ali-baba, which has reinvigorated my quest. I am just trying to build a device which can measure concentration of DNA (at 260) and purity (by 260/280 ratio).

I bought a reasonably useful silicon UV detector (photodiode) from Edmund Optics and hooked it up to an arduino in open voltage, you can see my results at diyscibuildjournal.wordpress.com. As you may have guessed my signal is extremely noisy, even using the large amount of light I get from a normal LED. The UV LEDs will give off much less and the detector will be less sensitive.

The literature tells me to try operating the photodiode in reverse-bias and also try a half-dozen other circuits (this is a good group of them, for example). I have bought most of the op amps and transimpedance amps listed in the paper, but I have very little idea of what is going on on the ground floor of the circuit.

What have read is that the linearity of the photocurrent to incident light is highest around zero volts, but I am mystified by the idea of a circuit that holds the photodiode at that voltage while still being able to get that current and transform it into a (non-zero) voltage.

I am hoping someone out there in the DIY BIO community might be able to help me out with some of this basic electronics knowledge so I don't end up just trying a half dozen circuits until I get something that has low enough noise to be usable or worse, fail and not know why I am failing.

Thanks!

Ben

--
You received this message because you are subscribed to the Google Groups "DIYbio" group.
To post to this group, send email to diy...@googlegroups.com.
To unsubscribe from this group, send email to diybio+un...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msg/diybio/-/uWgYQOilm8YJ.
For more options, visit https://groups.google.com/groups/opt_out.
 
 

[Simon Quellen Field]

Photodiodes are fairly easy to use.

You just use a JFET-input op-amp which can easily measure microamps

of input and give you volts of output.

Your noise problem also has a nifty easy solution.

Use the same type of diode for the detector as for the emitter.

An LED is a photodiode. It just also has the ability to produce light when

you put current through it.

Actually, all diodes are LEDs. Silicon diodes emit in the infrared.

LEDs have this nice feature that they emit a rather narrow spectrum.

Not as good as a laser, but narrow enough that we see the light as a

pure color. You can see the narrowness in the graphs of light output vs

frequency that come with your diode.

This narrowness also applies to the light they can detect. And it is the

same wavelength as the light they emit. So it acts as a narrow band 

detector for the color it is a narrow band emitter for.

Here is a simple microamp to single-digit volt conversion circuit using

an LED as a color sensor: "http://www.robotroom.com/ReversedLED.html".

You would build two of them, one for 260 and one for 280 nm.

You would illuminate them with two more of the same diodes.

You put your sample in-between the emitting diodes and the detecting

diodes, and record your ratio.

-----

Get a free science project every week! "http://scitoys.com/newsletter.html"

[Simon Quellen Field]

...and run the circuit from a nine volt battery -- no line noise, and it will

last a month or more before you need to replace the battery.

-----

Get a free science project every week! "http://scitoys.com/newsletter.html"

[Simon Quellen Field]

Also, I would not bother with a quartz window.

I would dunk the sending and receiving diodes directly into the solution

I was analyzing. Just make the diodes face one another, and use some

glue to make the leads watertight and dishwasher safe. The rest of the

circuit you can pot in epoxy and it will go through the dishwasher just

fine. Disinfect with bleach.

-----

Get a free science project every week! "http://scitoys.com/newsletter.html"

[Nathan McCorkle]

On Mon, Oct 8, 2012 at 9:00 PM, Simon Quellen Field <sfi...@scitoys.com> wrote:
> ...and run the circuit from a nine volt battery -- no line noise, and it
> will
> last a month or more before you need to replace the battery.
>

Doesn't digital logic noise creep in from the processor itself?

--
Nathan McCorkle
Rochester Institute of Technology
College of Science, Biotechnology/Bioinformatics

[ByoWired]

On Sunday, October 7, 2012 6:25:11 PM UTC-4, Ben Hunt wrote:

Hi!

I have been interested in building a homebrew UV spectrophotometer ...

Be especially careful about the materials from which you build your unit.  As you know, UV can cause a lot of things to fluoresce, so even small bits of dirt or lint in view of your sensor can give it unwanted signals.  You might also be surprised at how many electronic materials will fluoresce in UV, including labels on items.  Glass and plastic that looks unresponsive might also emit light, so be sure you measure your background carefully.  All the little things add up after a while.  

[Simon Quellen Field]

The high impedance analog inputs are separate from the ground.

The processor sees the low impedance output from the op-amp, and

the op-amp sees nothing from the processor. But if you are worried,

you can connect the analog ground to the digital ground using a 1 Meg

ohm resistor, so that any switching noise on the digital side is reduced

by a factor of 1 megohm/vcc. A fat electrolytic capacitor between vcc

and ground on both sides will also help, but I don't think either will be

needed.

-----

Get a free science project every week! "http://scitoys.com/newsletter.html"

--
You received this message because you are subscribed to the Google Groups "DIYbio" group.
To post to this group, send email to diy...@googlegroups.com.
To unsubscribe from this group, send email to diybio+un...@googlegroups.com.

[Josiah Zayner]

Pretty awesome how much monies, time and effort you are investing for the community. Keep trying hard.

On Tuesday, October 9, 2012 6:21:06 PM UTC-5, Ben Hunt wrote:

Hey guys!

Thanks for the huge response! Some of this is super helpful.  I want to clear up a few things . . . 

-The LEDs I'll be using are actually the most expensive part of the device ($80-120). I may end up reconfiguring the device to use them as detectors, but I have read that LEDs are also sensitive to wavelengths below their wavelength so I will have to do some testing.

-There are also 260 nm filters which are similarly expensive, and I will probably use them if I cannot get enough light from the (very low power, 0.1 mW) UV LEDS. 

-There is a quartz window on both the LED and the photodiode, and quartz cuvettes are not difficult to come by, so I don't have to worry about dipping the parts into the solution when I have to take a measurement :).

-The dark (or reference) voltage of my photodiode is problematic now not just because of its size but because of the amount of noise in it: it has a standard deviation of ~15-20% of the signal (compared to ~1 percent at high light signal), which makes it impossible to use itself as a reference in this current circuit configuration.

-I think my photodiode is appropriate, because it is sold by a company that also builds UV specs using it (Edmund Optical). Of course I may be wrong!

What I think I need to figure out is how to build a different circuit, which keeps the photodiode at or below 0 V while still getting a current out of it. I am not sure how to do this, but I have a few different circuit diagrams that I am going to try out, the parts just came in the mail.

Thanks for everything guys!

On Sunday, October 7, 2012 5:25:11 PM UTC-5, Ben Hunt wrote:

Hi!

[John Griessen]

On 10/09/12 18:21, Ben Hunt wrote:
> it has a standard deviation of ~15-20% of the signal

That along with the 0.1 mWatt might be trouble. You'll need to find out if you
are up on the curve of signal to noise ratio... or else lame results.

[John Griessen]

On 10/09/12 19:23, Josiah Zayner wrote:
> I think I need to figure out is how to build a different circuit, which keeps the photodiode at or below 0 V while still getting a
> current out of it.

Zero volts is just relative to some other voltage. What you need is to use feedback to
control the photodiode volts and then the other stuff after the photodiode can be at other volts
and other currents. The other current can go through a shunt and be measured.

[Nathan McCorkle]

John can you interpret the circuits here:
http://sales.hamamatsu.com/assets/html/ssd/si-photodiode/index.htm

?

[John Griessen]

On 10/09/2012 10:36 PM, Nathan McCorkle wrote:
> John can you interpret the circuits here:
> http://sales.hamamatsu.com/assets/html/ssd/si-photodiode/index.htm
>

Figure 2-1 Photodiode equivalent circuit

Is a model of the impedances of the detector diode to the left of terminals marked Vo.
To the right of Vo is the load impedance of your measuring circuit.

Figure 2-4 Photodiode operational circuits
(a) Load resistance circuit

(a) has a resistor across the detector, then gain.

(b) Op-amp circuit has detector at the + - terminals of an op amp, so they will stay
right at zero plus some microvolts, (as was mentioned being desirable for signal to noise ratio).
When the op amp is on, no current flows into
the inverting input. The gain is the ratio of Rfb to the detector resistance measured around the
zero volts level it's operating at, since they both have the same current flowing through, (since none
goes into inverting input).

[Ben Hunt]

John, that is super helpful. I did not know that putting the detector into the inverting and non inverting inputs of the op amp would hold the voltage at zero.

More about that specific circuit :2-4b: the output voltage looks to me like the negative of the saturation current x the feedback resistance. You said that the gain of the op amp is the ratio between the resistances of Rf / Rd.

Does that mean that V = -Isc x Rf only when Rf  is much greater than Rd?

If that earlier stuff is true I know the shunt resistance of the diode at -10 mV is 500  megaohms and the saturation current is 0.1 milliamps so to get 0.1 V for saturation I would put in a 500 megaohm resistor.

Do I still need to put power on the op amp?

Either way I'm going to go try this out today, thanks a lot.

[Ben Hunt]

T1000,

I didn't really think about the stability of the light source, and I have no specs for it. At this stage of the design process I am using an excess of light.

At this point I have the gain high enough and the noise low enough to start thinking about linearity of signal. The blog is up to date with my current schematic and arduino code, now including sampling and schematics. 

Next phase, however, is the real live light source in the real live package (printed at Ponoko), and I have to start thinking about what to order.  Thanks for the tip.

Ben

On Thursday, October 11, 2012 1:18:14 AM UTC-5, T1000 wrote:

 

Did the spec sheet that came with the LED tell you the stability of the LED?  What is the stability?

 

Thanks,

T1000

 

[Nathan McCorkle]

How are things progressing?

--

-- You received this message because you are subscribed to the Google Groups DIYbio group. To post to this group, send email to diy...@googlegroups.com. To unsubscribe from this group, send email to diybio+un...@googlegroups.com. For more options, visit this group at https://groups.google.com/d/forum/diybio?hl=en
Learn more at www.diybio.org
---

You received this message because you are subscribed to the Google Groups "DIYbio" group.
To post to this group, send email to diy...@googlegroups.com.
To unsubscribe from this group, send email to diybio+un...@googlegroups.com.

Visit this group at http://groups.google.com/group/diybio?hl=en.
To view this discussion on the web visit https://groups.google.com/d/msg/diybio/-/_6-5o0qHlHQJ.

For more options, visit https://groups.google.com/groups/opt_out.
 
 

--
-Nathan