Best IR diode for remote control project?

Digsy 1971

Established Member
Joined
Apr 28, 2003
Messages
646
Reaction score
20
Points
149
Location
Norfolk, UK
Hi.

Not sure whether to put this in DIY or not, but I decided on here.

I am trying to make a "super IR eye" to go with my Dgigsender. Basically it takes the IR eye signal from the Digisender and uses it to illuminate a bank of high power IR LEDs. The idea eing that the eye can be placed anywhere in my lounge and operate any amount of equipment without having to place an individual eye on each component, or aim the unit at the HC stack.

In fact I am trying to emulate the IR performance of my OFA Mosaic, which can operate all my kit even when I am pointing it directly AWAY from the IR recievers.

Here comes the science:

At the moment I am using a bank of 6 SFH487 diodes at 100mA and 1.3V (according to the specs in the Maplins catalogue). I selected this diode as I had seen it used in another IR repeater DIY circuit.

I built the "super eye" today and it is not quite strong enough. It will operate half my kit via reflections off walls, etc, and will operate it all at 5m range if pointed directly at it. However the beam is much more focussed than on my Mosaic and more aiming is necessary.

Can anyone suggest a better diode to use? Ideally I would like to know what they use in the Mosaic as there only seems to be two in there. They must be REALLY powerful! ANother option is to drive the SFH487 at a higher current, which will hopefully make it brighter.

TIA,
 
Darren,

As you guessed at the end of your post, your "super eye" may simply be in need of a little more LED drive current. You indicated that you are using six SFH487 LED's at 100 mA and 1.3V per the spec sheet. Those are the DC maximum ratings for the LED at 25 degrees C, as if it were left turned on continuously. But the pulse ratings are much higher, and that's the secret to obtaining high output from a remote control. The SFH487 is actually rated for a pulse of 2.5 A, as long as the pulse width is only 10 uSec and does not occur more frequently than once every 2 seconds. The Siemens spec sheet provides a series of curves showing allowable combinations of pulse widths and currents at various duty cycles. The higher the current pulses the shorter they must be, and the more time allowed between pulses. It's actually a matter of allowing time for the LED die to cool off between pulses so it isn't overheated by the integrated effect of repeated pulses.

Most OFA remotes pulse their LED's at between 1/2A and 1A. The luminous output of an LED is approximately proportional to the current, so if you increase the current from 100 mA to 1/2A the light output increases nearly five times. This can be done safely in the remote because the optical signal consists of a series of pulses at low duty cycle. For example, a typical IR carrier frequency is 40 KHz, during which the LED is being repeatedly turned on for 10 uSec then off for 15 uSec. That's a duty cycle of 40 percent, but the carrier itself is also turned on and off in bursts to convey the message data, which further reduces the effective duty cycle.

Since your post doesn't give circuit details (power supply, etc.) and I'm not familiar with the "Digisender" I can't make specific suggestions. But here are some factors to consider. Although the SFH487 is rated 1.3V forward voltage drop at 100 mA, the voltage drop increases to about 2.5V at 1/2A and 3V at 1A, so use a value in that range when you calculate series resistance needed for a given supply voltage. To give you some idea of actual practice in OFA remotes, one 6V model with a single LED uses a 5.6 ohm series resistor. Another 6V model with two LED's in series uses a 1.6 ohm resistor. Still another 6V model uses two LED's in parallel, each with its own 2 ohm resistor in series. A popular 3V remote with two LED's in parallel uses a single resistor of 1 ohm. None of the resistors is ever greater than 1/4 watt, made possible by the low duty cycles involved.

Working with voltages as low as 3V is a challenge to remote control designers because every current-limiting factor must be taken into consideration, including the voltage drop of the saturated driver transistor. The higher the supply voltage available, the easier it is to stabilize LED current. Paralleling LED's and sharing a common current-limiting resistor can cause the current to divide unevenly as a result of different forward characteristics of the LED's, unless identical types are used, preferably from the same manufacturing batch. Wiring LED's in series is always OK, and offers an advantage in the driver circuit. For example, suppose you planned to drive six LED's in parallel at 1/2A each, using a 6V power supply and separate current-limiting resistors for each LED. Allowing 2.5V for an LED drop leaves 3.5V across each series resistor. 3.5V divided by 1/2A is 7 ohms, or the nearest standard value, 6.8 ohms. For this configuration your driver must be a hefty power transistor rated for a collector (or drain) current of (6 x 1/2A) = 3A.

Contrast that to a circuit using a 12V power supply that drives two parallel groups each consisting of three LED's in series with a resistor. Allowing 7.5V drop for three LED's leaves 4.5V for the resistor. For 1/2A the resistor value calculates to 9 ohms, or the nearest standard value, 9.1 ohms. But now the driver transistor needs to handle only (2 x 1/2A) = 1A. In either case, make sure your power supply is equiped to handle the peak current demands. This doesn't mean that you need a power supply rated 3A or 1A in the above examples. A power supply rated for 100mA is plenty sufficient if there is a storage capacitor on the output. Most remotes supplement the already-considerable peak current handling capability of their alkaline batteries with a 100 MFD electrolytic capacitor. The average current used by a remote while actually transmitting rarely exceeds 60 mA.

Sorry if the "tutorial" nature of my message tells you stuff you already knew, but I was guessing by the way your question was worded that this information might be of help. I think the LED you're using is excellent, and six of them should be able to transmit all the way across the pond to my kit here in the colonies.
 
Hey Tommy. This is totally bizzarre! Look what I was just reading (three posts down).

I've spent the intervening time doing a search for "brighter" IR diodes and had pretty much come to the conclusion that there really aren't any, so more current was the way to go. Finding your post on RemoteCentral confirmed this.

I now have the datasheet for the Siemens SFH487 and plotted the Vf versus If curve and used it to try to work out the maximum I can drive my LEDs at. I've attached the spreadsheet for information.

A bit of background about the power supply and the circuit:

I want to use the same PSU for the Digisender and the super eye. Rather bizzarrely, the writing on the Digisender PSU says 0.5A / 9V but the transformer actually outputs 14V. interestingly, the Digisender comes with two of these transformers and they are both the same (boggle?)

The Digisender outputs 10V pulses which in "out of the box" form are sent directly to two ultra-short range LED eyes. I am using the output pulse to drive the base of a BC337 transistor, via a resistor to limit the base current but still ensure saturation.

The 6 LEDs are connected in series between the 12V rail and the collector, with a 62 ohm current limiting resistor.

Using the spreadsheet figures I want to bump up the current from 100mA to 200mA to 300mA or maybe even higher. I would be interested to hear your thoughs on whether I would need to uprate the PSU? You touched on this in your post. The other thing I may need to consider is a different LED arrangement as 6 x Vf starts to exceed the PSU output when If > 500Ma

Power in the current limiting resistor peaks at about 0.8W for continuous operation, but I assume that the reduced duty cycle means that a 0.6W resistor would suffice?

Thanks again for the response!
 

Attachments

  • sfh487 characteristics.xls.zip
    3.8 KB · Views: 32
Darren,

>> the writing on the Digisender PSU says 0.5A / 9V but the transformer actually outputs 14V

If you loaded it to 0.5A the voltage would probably drop to 9V. But the way you're going to use it the no-load value is valid.

>> The Digisender outputs 10V pulses which in "out of the box" form are sent directly to two ultra-short range LED eyes.

If these "eyes" are for sticking in front of IR receivers on TV's etc. they're what we Yanks refer to as "IR blasters", a rather unimaginative name in my opinion. Calling them "eyes" isn't much better because they're the "lights" that shine into the "eyes" of the equipment in your kit.

>> The 6 LEDs are connected in series between the 12V rail and the collector, with a 62 ohm current limiting resistor.

That explains the "corrected" notation on the spreadsheet that changed it to 50 ohms. So you've really been driving them at about 80 mA.

If I were doing it I would split the load into two groups of three LED's, each group with a 13 or 15 ohm resistor, and keep the Digisender PSU.

The BC337 is a good choice, as long as you do your homework. If you're using the -16 version (gain of 100 to 600) note that the "typical" current gain at a collector current of 500 mA is only 50. That would mean the Digisend signal would need to be at least 100 mA, which it probably isn't. If you're using the -25 or the -40 version the "typical" gain at 500mA is about 100. Again, if it were me I would use a separate BC337 for each group of three LED's rather than trying to get 1000 mA out of a single BC337 (even though it is rated for it) because the gain drops to 30 or 40 for that much collector current.

Assuming a target of 0.5A and using Vf of 2.3V from your spreadsheet, the resistor voltage = 14 - ((3 x 2.3) + 0.7) = 6.4V (the 0.7V is transistor saturation voltage). Using a 13 ohm resistor gives 492 mA, and the DC power dissipation is 3.14 watts. The duty cycle will almost certainly be 5 to 10 percent maximum, so a 1/4W or 1/2W resistor will suffice.

There are many other possibilities, and I'm sure you will wind up with something that works. After you get things going I'll tell you why I have a special interest in your project.

Tommy
 
OOps! Scratch all that about BC337 gain. I see I dropped a decimal place again. At 500 mA collector current and a gain of 50 the base current is only 10 mA, well within the Digisender output. Sorry about that.

Tommy
 
Hi again.

Yes, the standard attachments to the Digisender are "Blasters". You only get two and they have to be stuck on the front of the kit which (a) looks awful and (b) I have 5 pieces of AV kit to control anyway.

I'm going to push my luck at this stage and keep all 6 LEDS in series because I don't want to desolder them again. I've hed to bend the pins through 90 degrees to make them face forwards off the Veroboard but I had to bend them twice because (as I'm sure you know) the short lead on the SFH487 is the anode and I soldered everything the wrong way around initially. DOH!

Something else I might try is to drop down to 5 or 4 LEDs to free up some volts to go to higher currents and hope the tradeoff results in a net increase in brightness.

Anyway, initially I'll try the most I can get out of them with the 14V supply and all 6 in series, and I'll tweak the base resistor to give me saturation at the high collector current.

According to the catalogue, I am using the -16 version of the transistor, but it states the gain as 100 to 250. Didnt realise it dropped lower as the collector current increased. I'll have to get hold of the ful spec sheet...

Just one more thing: Could I further protect the PSU by wiring my own 100 MegaFarad capacitor across the 14V and 0V rails in the eye circuit itself?

Regards,
 
Darren,

You're right. The -16 has a gain of 100-250. I erroneously used the gain range of the entire BC337 series (all three grades). I'm starting to make too many mistakes to be of much more help to you.

And yes, by all means put a cap across the supply rails near where the LED's are, particularly if the PSU is a wall plug-in unit with a long cable to that point. I'm sure you meant to say microFarads instead of MegaFarads, but hey, what's a factor of a trillion between friends. (Glad I'm not the only one who ever makes a mistake.) Good luck.

Tommy
 
Cheers once again.

You're right, I did mean microfarads :)

I'll let you know when it's up and running (or melted, whichever happens first).

Best regards,
 
Bounce... :cool:

Well, it's done and it works. 350mA forward current and it will operate all my gear from reflections. It won't do my plasma screen, which is sited in a different location and presumably needs even more brightness from the LEDs to notice them. HOwever, the whole point of the "super eye" is so I can control my whole HC stack when watching on a second LCD screen in the bedroom, so I don't need to be able to turn the lounge screen on or off from up there. I could have bumped up the current to 500mA but I didn't want to risk the PSU once I had it working the rest of my kit.

Thanks to Tommy for the information :clap:
 
Darren,

Did you get your diagram off the web, or better yet is the any chance you can attach a circuit diagram of your supper eye as I have just brought a video sender from Maplin and found that the IR repeater is very directional and will only work at close range.
 

The latest video from AVForums

Is 4K Blu-ray Worth It?
Subscribe to our YouTube channel
Back
Top Bottom