890nm LED is Looking Red at 100mA current

I have an IR LED with 890nm peak wavelength emission (part number is VSMF3710 the datasheet is attached). When I drive the LED with 100mA current the LED is looking red in color. For my application LED must be invisible to human eyes and only visible to the camera sensor, for eye detection.

Thanks in advance

You should take a look at some of the references (2-5) here.

Seems to indicate you can see infrared under certain conditions, where intensity is high.

Maybe you need to pulse the led instead of using a constant drive.

Also note that 100mA is the absolute maximum rating for continuous operation. It isn't a good idea to push it to it's limit, use more LEDs at lower current to achieve the radiance you need.

Brian.

It is not healthy to look directly into 100mA of IR LED. 95% is invisible 5% is visible due to bandwidth of the spectrum. The invisible part is caterogenic.

- - - Updated - - -

Any LED or photo diode , reverse biased to a -5V MAX should pick up a pulsed current. Bandwidth depends on Transimpedance amplifier.

however a photodiode will be more sensitive ~ 0.5mA/mW

Very Thank for your valuable suggestion and good try to resolve the problem.

I tried to drive the LEd with 20mA and it is still visible, then I tried much low current 2mA and LED is still visible.

I know It's not good to drive the LED at such a high current, but a 100mA current it given to LED to check the Intensity.
I will lead to more bright LED, If the red visibility problem is removed.

I am working on it to give the LED pulsed current. I will update soon the results.

Thanks again.

Test results are negative,

I have tried frequency from 5Khz to 44Khz signal, with duty cycle ranging from .1 to .6, but the led is stil visible.

It's in fact unlikely that pulsed operation changes the emitted spectrum more than slightly. Small wavelength shifts according to current and ambient temperature are expectable.

We can just assume that the spectrum specification in the VSMF3710 datsheet isn't exact and the device emitting more visible light than expectable.

I know that standard remote control 950 nm LEDs are invisible, I would expect the same for 880 or 890 nm types, but there are apparently differences.

I would guess these are similar in spectrum to the IR LEDs used for security floodlighting where the camera sees a brighly illuminated area but a person only sees a dim red glow from the light source. I've also seen it in IR data applications where vocal translations are sent to IR receivers on headsets in multi-lingual meetings. In a country where public meetings have to be bi-lingual there are a lot around here. I agree with FvM, pulsing may reduce the average power but it would have minimal effect on wavelength.

What exactly is your application?

Brian.

Thank Brian,

The exact application of this IR is to detect the eye movement of a patient having vestibular disorder. In complete dark when the patient is said to hold his eyes on a single point, it will get nystegmus.

Now If IR led of eye tracking system is glowing in red color then the patient will get visual fixation and the test is go wrong.

I have tried 850nm, 890nm, 940nm and 860nm, all the IR LED are glowing red in complete dark.

I have tried to lowering the emitted power of LED's but then the brightness is too low that we can't detect the eye pupil movement.

That is unusual. Obviously there is some spread of spectral power from the peak but I wouldn't expect it to reach from even 850nm into the visible range although I confess I've never tried running such LEDs at maximum current before.

The amount of IR power you are emitting is quite considerable, I wonder if the solution isn't more light but to use better sensitivity. I'm not a medical expert but it occurs to me that the red glow may not be the IR light itself but an effect of the light stimulating something else along the path to produce light. I have chemical IR detectors here that glow dark orange in the presence of IR radiation, perhaps you are seeing something similar.

I don't know what kind of sensors you are using but putting the issue in perspective, I have a CCD camera here that sees a picture as good as in daylight (except in monochrome) over a distance of 10 metres while in total darkness and it uses 20 x IR LEDs each drawing 20mA @ 1.5V. I appreciate you probably want a point source of light so one LED would be preferred but over a short distance you should be able to use very low power. Running your LED at full power and aiming at the eye could actually be harmful in itself.

Brian.

I can just believe your observations that you see even 950 nm LEDs glowing.

It would be helpful to quantify the brightness by comparing it with a light source of known intensity, e.g. a red LED operated at a very low current.

I guess, the problem may arise because you are shutting off any ambient. I'm unable to determine if it may be a residual IR eye sensitivity or residual visible emission. You may want to use additional daylight filters, e.g. IR LEDs with black package to further reduce possible visible emissions. Worst a case a long-pass interference filter must be used.

I would also suggest that you check the eye safety of your appartus thoroughly. You surely have read the warnings in LED datasheets, there are strict regulations for lasers and high intensity LEDs with power limits for IR radiation.

After thinking again about the problem (and a cursory review of IEC 60825-1), I'm rather sure that you need some kind of indirect diffuse lighting. The LED chip must not be directly visible, otherwise you'll never comply with safety rules.

All the IR LED's below the 940nm are visible.

Please update if anyone find any other solution.

I believe the question has been answered before. Part of the answer is that it's absolutely prohibited to have a person look directly into an IR LED, diffuse and indirect lighting is required.

Hii FvM,

I know very well that diffuse or indirect lighting is required, but if the led is visible at very low intensity, what is the matter to diffuse it. The eye pupil tracking system will not work at such a low intensity. And as we increase the intensity of light it is visible which leads to visual fixation of subject.

The invisible IR LED which can be used is 940nm and above, although my pupil tracking system will suffer a lot by doing that, but it is fairly good.

Thanks

I don't think that the diffused IR LED light will be still visible, but I may be wrong. The direct LED beam is visible due to it's extreme high (and not at all eye-safe) intensity.

Look close to these two led's the only difference is the radiation pattern. Both LED's are 850nm from OSRAM.



If the LED is visible due to high IR intensity (as I thought so a long time ago), the LED must be invisible at the higher viewing angle. But both the LED's are visible at 80 to 85 degrees of viewing angle.

Industrial security cameras, which use an infra red source have a proper filter. From the ground it looks black, i.e. you can not see what is behind it.
Now my physics is not very good, but I regards infra red as "heat waves", as pitch is virtually transparent to them. I wonder if a thin black plastic film would work much the same? Try a bit of bin liner.
Frank

I still think the issue here is camera sensitivity, not the brightness of the LED. An IR filter in front of the LED will block all visible light but I can't understand why high power is needed. I'm assuming a CMOS/CCD camera is being used and not photodiodes or phototransistors, is the camera a 'standard' one or is it specifically an IR camera. The reason I ask is most normal cameras have a filter in them to reject IR light so to get them to work you need a bright source. Ones without the filter are many times more sensitive.

Brian.

Hi,

I agree with FvM.

You mixing up "viewing angle" with "diffused" light.

With varying viewing angle the size of the light source stays the same.
But with a diffused light source, it may be achieved with an opaque glass or a diffusing screen, you have a big area of light source.
With a big area of light source, there is not a "point" you can fix, if it is recognised...

Also the safety issue. If the light comes from a big area, then the light energy is spread in the eye also. A small spot of light with the same energy may destroy the eye at those points.

The camera doesn't care if the light source is a small point or a big area.

Klaus