IR LED to repair remote control

Measure it. Put a couple 1n4001 rectifiers in series with a 0.1 ohm resistor, then operate the remote. View waveforms with an oscilloscope.

Or just get DigiKey's hardiest IRLEDs (one of each wavelength), salvage some LEDs from another cast-off remote.

Might help us to know if the failed LEDs failed open, shorted, etc.

Cheers, James Arthur

If you could ever see them light up, they were almost certainly 880 nm LEDS, not 940 nm types. A few IR receivers are also narrow band filtered for 880 nm sources.

(snip)

The filter serves no function in emitters, except for visual identification. Some manufacturers tint 940 nm units but put 880 nm units in clear packages (or vice versa), to distinguish them. I don't know how standardized this practice is.

(snip)

Most receivers work with either wavelength. Only a few based on narrow band gallium arsenide aluminide detectors work only with the 880 nm types.

The great majority of receivers use silicon detectors with long pass optical filters that block most visible wavelengths that the silicon is also sensitive to.

I don't think the package color matters, but high peak current rating and high milliwatts per steradian output intensity are good. Though, the highest intensity units are also the narrowest beam width units, meaning you have to aim them pretty accurately.

A good choice might be the TSAL6100 (950 nm) or its high output cousins (assuming you can use the 5mm diameter package).

But if I was going to order something, I would get both 940 and 880 nm variations. If the control has more than one emitter, I might use some of both.

I didn't look at the remote ever in the digital camera before hand, but I tested another remote just to be sure I could see IR, and the tech support were the people who asked me to try it to look for the IR lighting up. I presume they expected me to see results this way. I guess that means that at least one of the two transmitters was 880nm?

need, it seems that all the 940nm that digikey carries are clear package. But that radioshack emitter at 940nm was tinted, and the radioshack matching emitter was clear. Go figure?

So it seems that a more "universal" wavelength of this purpose would be 880nm?

package).

It seems that the LITE-ON 940nm I listed has similar, if not a little "better" specs in all categories than the 950nm TSAL6100? And a little wider viewing angle to boot. Was there something about the LITE-ON LTE-5228A that you didn't like?

Yeah, there are two emitters, so I think I might try an 880 and a

940. Likely the 880 from Vishay and the 940 from lite-on.

As an additional question. Does anyone have any clue why these diodes would both go out completely randomly in the original unit? It was about a year and a half old, and I found a number of posts of the exact same thing happening to other people after almost the same amount of time. I could see one of the diodes going out, but both simultaneously? Maybe they are connected in series? Or maybe one went out and I never noticed, then the other went? I wouldn't mind examining it a little further to find the root cause, or at least hopefully the new diode I put in exceeds the original diode's specifications such that the problem isn't repeated.

package).

You could have an intermittent bad connection in the remote, which would explain the replacements failing soon. You might check for voltage with a meter, or preferably a scope, at the exact moment when you know there is no light coming out.

-- John

880 nm LEDs give off a faint bit of visible red light. If you look into the LED in total darkness you will see a faint red glow, if the LED is on long enough for your eye to respond. 940/950 nm LEDs are generally totally invisible to the eye.

I think so.

package).

An overload is the only reason they should fail. Their driver may have failed and put DC across them, instead of a train of pulses.

Quite likely. Their normal voltage drop is about 1.4 volts, so a two cell supply will drive two of them with little current regulation.

The failed LEDs should show either a low resistance in both directions, or some leakage in the reverse direction and normal drop (about 1.2 to volts with1 mA forward bias) with a diode check function.

package).

Yes, I intend to take a peek when I get home. The g/f says it was broken again so I need to confirm and investigate further. From a light visual inspection the first time, there was nothing I could see wrong, and I know the solder joints were good both times. I'm going to poke around when I get home, though.

I'm not sure it is an *intermittent* connection because I beat it up pretty good looking for a bad connection and it didn't work (I know, bad idea) even for a split second during the beating it up phase. But gently replacing the LEDs made it work. Last night, with the radio shack replacements, it was working great. My g/f turned on a DVD, put the remote on the table and fell asleep, and first thing in the morning, it wasn't working.

I never did notice a red light from the remote. I had thought the reason that they have you use a digital camera is because the sensor can pick up a light range that your eye cannot. Are you sure the digital camera can't pick up 940nm? You might try it if you never have. It was pretty cool :) Aim a remote at your camera and watch in the viewfinder for a purple light.

That's what I was afraid of. If it were the driver, would you imagine it would "die" instantly, never to work agian, or would you suspect that the driver could have an intermittent problem that may kill diodes daily or weekly as it "dies" slowly? If it were more a "dies" slowly scenario, then replacing LEDs may not do the trick. If perhaps the LEDs were slightly overloaded from the driver from the factory (by a poor design), then I suppose it is possible that they did not last as long as anticipated because they were run at or just over their maximum ratings for a long period of time. In this case, I would suspect a higher rated diode may correct the problem.

Yes, another reason I suspected them to be in series was because I replaced one of the diodes at first to see if it would fix the problem before I opened the second package, and it did not work. I then tried replacing the second diode, and it worked. This led me to believe it was in series, or there was something else fishy going on, but I was so happy it worked again, I sorta forgot about it :P

Yeah, I'll have to inspect the diodes to see how they failed. I'll also poke around a little more while I'm in there to see what else I can see.

Thanks for your help everyone!

Most digital cameras will respond at least a little to either wavelength, but more strongly to the 880 nm.

Honestly, I would expect it to never work again, if the driver failed. The fact that you restored operation, for a little while, with new LEDs, sounds like you just need beefier ones.

I expect wavelength anywhere from 860 to 950 nm to be OK.

I think better would be LEDs with higher current capability, both continuous and peak.

Then again, I see enough inconsistincies between LED datasheets and actual performance... If the unit is indeed out of warranty, one thing I would do is add a resistor in series with each LED. Go ahead, splice resistors in, do a bit of surgery/hacking if the unit is out of warranty! I think that a 10 ohm resistor in series with each LED should make a difference - by dropping half a volt at 50 mA. 1/8 watt resistors should be adequate and are very compact.

The above Lite-On LED does have a datasheet making a claim for higher current and power. I give some fair chance this is true and likely to help. But beware, it has a wider beam width (35 degrees), and that can detract from the beam intensity and reduce your range - especially if you need resistors to keep eve these from blowing. Should you need resistors, go for higher mW/sr intensity LEDs (they tend to have narrower beams) and be prepared to use 15 possibly 22 ohm resistors to keep the LEDs from blowing.

==============================================

By any chance, are you using lithium batteries instead of alkalines? The extra .1 or so volt per cell could make a difference if the product design is not tolerant of that extra voltage.

==============================================

You may do well by using rechargeable batteries (NiMH or NiCd) - the voltage is a little less than that of fresh alkalines - that may save your LEDs.

=====================================

- Don Klipstein ( snipped-for-privacy@misty.com)

It appears to me that in general, digital cameras have some ability to "see" the "glow" of 940-950 nm LEDs. They do indeed, as mentioned, usually appear purplish, often a shade close to "lavendar". However, 880 nm ones show up with a similar, maybe slightly different color (still usually purplish/"lavendar") to digital cameras.

- Don ( snipped-for-privacy@misty.com)

snip

I just built a dual beam IR Tx/Rcv using the TSAL6100 and Panasonic detector. If you don't have a scope you will be limited in determining any problems.

One thing you can check with a DVM is the DC voltage that is applied to the drive circuit. If you can determine the regulator then verify with a DVM it's regulating properly. If it's a low dropout type and it's failed this might explain why it works for a while after you change the LED's. All components have probably been derated for at least an ambient temp of 50 deg C. So if an extra 0.3V to 1V is applied to the circuit it won't immediately be destructive. Depending on the duty-cycle it may even be able to take substantially more voltage (standard three terminal regulators) 2-5V for a time. I designed mine for max range and 60 deg C ambient and they could survive an extra three volts for a minute or so (this is continuous use 10% overall duty cycle).The extra voltage drops across the current limiting resistor, in my case almost doubling the current, and the power dissipation in the LED's.

The TSAL6100 has a Radiant Intensity = 80mW/sr min - 130 typ mW/sr @

0.1A,and 650mW/sr min - 1000 typ mW/sr at 1A.

Versus the 18 to 37 mW/sr at 0.1A for the LTE-5228A.

And yes a digital camera can see IR at 940nm. I used it to align my TX/Rcv.

If you can get the model of detector for your remote you want to match the emitter wavelength to the detector for max sensitivity.

Andrew, I have the same problem with my 880 Harmony, it happened Tuesday afternoon (Sept 25th). Everything on the LCD screen looks normal, but my devices don't respond to the remote. I took it apart and measured about 1.6v on the diode with a digital multi meter. I also concluded that the diodes were bad. One thing I have to add to the disscussion is that a couple of days ago I noticed that the IR shield was very warm, the only source of the heat was the IR diodes. The remote was sitting flat and nothing was hitting any buttons, but I was a little freaked out, so imediately I pulled out the battery. I rebooted the remote, and everything worked fine for a couple of days. Untill it stoped working...boo. I sent a letter to Logitech customer service, but it seems like they might not be too helpful, I asked for the diode specs, but that's really just wishfull thinking. Let me know how some beefier diodes hold-up and I'll keep you advised of anything I find.

[and another Harmony user has a remote that almost works, but low/no output]

As you suspect, the peak current rating for the remote may be higher than the Radio Shack LEDs; the replacement pair could have failed due to overcurrent.

I've seen a weak battery kill a remote that functioned fine (used under

1 mA to run the smarts) until a command was sent (using 100 mA to run the LED). Always try replacing the batteries first.

Because you have the phototransistors, you can make a simple integrating detector to calibrate the remote output. Use a shoebox, with a hole in the leftside to shine your remote in, a hole in the right side with the phototransistor, and a separator in the middle of the box (tape in a card, with a small hole in the center, to leak light from the illuminator side to the detector side). You'll need to use a pullup/pulldown resistor pair to bias the phototransistor from (for instance) a 5V supply. I'd use a couple of 1k ohm resistors. If the signal (into an AC voltmeter or oscilloscope) from a known-good remote control is 0.5V, you're sensitive enough. Change the hole in the separator card to change the sensitivity.

It's easier to find unneeded remotes than to buy the LEDs new. Scavenging is your first source of supply!

Logitech support said that their warrenty department would contact me in 5-6 days. So, even if my remote is covered, which it's not, it would take probably 4 weeks to get it repaired (1 week for contact, 1 week to ship it there, 1 week to repair, and 1 week to ship it back to me). This is not really acceptable, so I'll continue trying to fix it myself.

I shorted out one of the harmony diodes with a diode from an old remote, which got the other harmony diode working. Only one diode was blown. Also, the replacement diode did not light up; I'm assuming this is because the current is too low to run the replacement diode.

I made some more measurements on the remote, and here are my results:

Angle between the two diodes: 25 deg Current through one diode: 54mA Diameter: 5mm

Note, I measured the current by connecting my DMM to the leads of the bad diode, and pressing the volume button which seemed to give the fastest flashes, but the current measurement may be low?

With these measurements in mind I think a diode with a 30 deg emission angle and a rating of 100mA continuous forward current will work. I'm going to order a couple of diode sets from Digikey, and I'll let everyone know if I find one that works well.