Continue to Site

Welcome to our site!

Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

  • Welcome to our site! Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

LED Re-Drivers

Status
Not open for further replies.

crashsite

Banned
This is a pretty simple one but, I thought I'd run it by the Electrotech pool of experts to get some points-of-view.

I've been thinking about building some turnkey computer systems for basic multi-media functions (mostly aimed at video editing for specialty applications such as a home-town soccer team or car club, etc.).

The equipment selection would be driven by the application but, virtually all would involve LED (and sometimes incandescent or other) indicators that I would like to bring out to a common panel so the actual devices, that don't normally require direct access, can be mounted out of the way but still be monitored (modem, external hard drives, power control indicators, etc.).

I'm loathe to try to open the cases to tap directly into the LED leads and thus am thinking either fiber optics or a photocell sensors and re-drivers to light the indicators on the monitor panel.

Is this something any have experience with and thus have ideas, horror stories, comments? I'm especially interested in dealing with methods for attaching the sensors (or light fibers) to the equipment's LEDs with maximum reliability plus ease of removal for equipment replacement and would prefer methods that do NOT require any modifications to the equipment. Also, the best methods for handling bi-color LED indicators (which I assume would be done best with colored gel type filters for the electronic sensing option). I would also have no objection to a mix of methods so long as the displays can maintain at least a fair amount of consistency.
 
Sorry I don't know what you're asking and I'd guess that no one else does.

Please cut your post down to the questions you want to ask.
 
It seems that you understand all the issues.

How about attaching a sensor in front of each LED using a Velcro attachment. That would allow you to remove it for servicing.

Sensing of bi-color LEDs is another thing. I know they make chips that detect color but do not know what the cost would be. Maybe just stay away from equipment that uses them.

EDIT: Some of the signals like HDD indicators and power exist on the motherboard. The front panel indicators connect to these. You can unplug the OEM indicators and use your own remote ones.
 
Last edited:
Question

Hero999 said:
Sorry I don't know what you're asking and I'd guess that no one else does.

Please cut your post down to the questions you want to ask.

Sorry, I wasn't aware that a request for suggestions, rather than a "question" was illegal in this area.

But, in for a penny, in for a pouind so...

Primarily, my concern is the best way to interface either optical fibers or photo sensors to LEDs on equipment so that it has the right optical properties and a mechanical attachment that leaves the affected equipment unmodified (like glues, clamps, suction cups, etc.).

I figured that some of the hobbiests here may have been faced with this sort of thing and have come up with some novel and workable solutions they might like to share........or not.....

I have also found that attempting to pare down a "question" to the minimum of verbage generally has the effect of getting responses that are pretty much unrelated to the question actually asked. I could have posed it as:

"How to remotely light an LED" or "How to extend an LED to a remote panel".

It's pretty easy to see that the responses would be considerably "less focussed" and I hoped (apparently inaccurately) that a description of what I want to accomplish would get some practical and/or tried-and-true responses.
 
That's pretty much what I was thinking....

3v0 said:
How about attaching a sensor in front of each LED using a Velcro attachment. That would allow you to remove it for servicing.

Ha ha...as you were making your post, I was composing my last one.

Well, I understand what I want to accomplish and have some ideas. I must confess that I hadn't really thought of the velcro and the LEDs on the front of a computer do present some aesthetic concerns since the front of the comptuer will be visible to the user for access to things like front-panel slots and CD/DVD trays. Since I am thinking of custom systems, it probably wouldn't be out of line to get inside the box and maybe bring the LED signals out via a connector mounted on an unused blank slot panel on the rear of the computer (with a couple of extra components, the front LEDs could also continue to work).

I can see potential in the velcro. By punching a hole in a small, adhesive-backed piece it could surround the LED and the mating sensor (or fiber) could be attached with the mating velcro. Maybe even the 3M "Command adhesive" strips could be used for an even cleaner removal if needed.

Thanks, good idea.
 
Last edited:
crashsite said:
Sorry, I wasn't aware that a request for suggestions, rather than a "question" was illegal in this area.
There's nothing wrong with asking for suggestions. Please don't take this the wrong way. All I was saying that perhaps you should consider shortening your posts and cutting to the point more.


You can get some pretty small SMT LEDs and have you considered electroluminescent wire?

Optical fibre is good but it doesn't conduct that much light.
 
All aglow

Hero999 said:
You can get some pretty small SMT LEDs and have you considered electroluminescent wire?

Optical fibre is good but it doesn't conduct that much light.

Relocating the LEDs wont be a problem and I have tens of thousands of older but, perfectly suitable LEDs (I had planned to open an electronic surplus business but, damn computers and cheap, overseas mass assembly techniques sort of knocked a lot of the interest in "real" electronics out of most folks). Plus I live in....well, let's just say that, technically speaking, the Aussie aboriginals have nothing on the cowboys of Central Oregon.

I'm not sure how I would use, electroluminescent wire (I assume that's the wire that glows). I have some of that from a halloween display that used it to outline a jack-o-lantern's eyes, nose and mouth in different colors. The colors I have are not a good match for most LED displays (orange, blue and purple) but, I must admit that, with the exception of having to buy lots of halloween pumpkins for the power sources, using the wire is somewhat intriguing for the display.

Yes, I share your trepidation about the optical fibers. They also have tighter entrance angle restrictions and something like velcro might give less than consistent light transfer due to it's non-rigid nature.
 
Last edited:
Yes, yo could use a CDs photocell (cadmium sulphide- they're cheapest) to monitor the light, then use a simple light/dark indicator circuit to activate external LEDs. for incandecents, you'll need to use a higher current transistor.

These circuits are so simple anyone can make 'em - blindfolded. all it is is a transistor-as a-switch circuit. google "light dark sensor circuit"

Because you intend to have many LEDs you may want to consider an 'analog switch' which is a chip with switches that are activated by current. The advantage of that, is that they dont need an extra resistor where you'd have one at the base of a transistor. Plus they're a whole lot cheaper and smaller.

In fact, they're free. Intersil has samples of up to five chips - just follow this link to the sample page
The datasheet is here
and the part number is DG412. Of course, you can choose your own part from the intersil(.com) website, but this one I reccommend. I use these chips.

Hope this helps somewhat
-michael
 
mike11298 said:
These circuits are so simple anyone can make 'em - blindfolded. all it is is a transistor-as a-switch circuit.

Hope this helps somewhat

It all helps.

I have lots of transistors and transistor arrays on DIPs. I do anticipate the probability of needing to use an IC comparator if filtering bi-colors, especially when the LED illumination is on the lower side.

Since this is not a "production item", I'll have some extra latitude in being able to use what I have on hand but, don't want to have to re-invent the redrivers with each system built. I'll probably put on my blindfold and create some "general purpose" circuit boards that I can put some simple sense (as you say, "light/dark" switches) and current drivers on. As such, I'll check out your suggested chips but, will tend to use my stock and reserve getting samples for when I more desparately need them (vendors seem to prefer to sell the parts after just one batch of samples....unfair, I know).

I was thinking probably CDs cells and have some but, my supply is a bit spotty and inconsistent. I'm interested in a cheap, reliable source of small ones. I've seen some that looked ideal in cheap motion sensor lights (used for sensing ambient light to prevent the lights from activating in daylight) but, not sure of a good source and don't want to buy a $10 light to extract a $0.10 photocell.
 
Hero999 said:
There's nothing wrong with asking for suggestions. Please don't take this the wrong way. All I was saying that perhaps you should consider shortening your posts and cutting to the point more.


You can get some pretty small SMT LEDs and have you considered electroluminescent wire?

Optical fibre is good but it doesn't conduct that much light.
Hero, he's looking for ways to attach sensors, not light emitters. He has equipment that will be buried in an enclosure, and he wants to be able to monitor the indicators on that equipment without having to open them and attach wires.
 
mike11298 said:
Because you intend to have many LEDs you may want to consider an 'analog switch' which is a chip with switches that are activated by current. The advantage of that, is that they dont need an extra resistor where you'd have one at the base of a transistor. Plus they're a whole lot cheaper and smaller.
Michael, DG412 is activated by voltage, not current. And they are very expensive if you have to buy them.
 
Yes

Roff said:
Hero, he's looking for ways to attach sensors, not light emitters. He has equipment that will be buried in an enclosure, and he wants to be able to monitor the indicators on that equipment without having to open them and attach wires.

Yes.

In an ideal world where the manufacturers mount their LED indicators such that they can be plucked off the equipment and relocated, without heroic effort, there wouldn't have even been a question. The LEDs would have simply been extracted and re-mounted on the remote panel and wires run back to the equipment.

Pretty simple in the '70s...virtually impossible in the new millenium. Ah, the joys of progress.
 
crashsite said:
Yes.

In an ideal world where the manufacturers mount their LED indicators such that they can be plucked off the equipment and relocated, without heroic effort, there wouldn't have even been a question. The LEDs would have simply been extracted and re-mounted on the remote panel and wires run back to the equipment.

Pretty simple in the '70s...virtually impossible in the new millenium. Ah, the joys of progress.

You might be able to use some of the thousands of spare LEDs you have as the sensors instead of the slower CdS cells. Wire them up as photodiodes feeding transistors to control the remote LEDs. Might have trouble with coloured LEDs not giving a strong enough response though.

There's a link to an EDN article on it on this page: https://dvgadre.blogspot.com/2006/11/led-as-light-sensor-and-more.html

You could hot-glue the LED sensors over the existing LEDs; it should be pretty secure and yet possible to remove without damage in most cases.


Just some wandering thoughts,

Torben
 
LRDs?

Torben said:
You might be able to use some of the thousands of spare LEDs you have as the sensors instead of the slower CdS cells.

I did a fair amount of experimentation with this several years ago. I even went so far as to drill holes into two plastic LEDs (red), nearly to the chip, and epoxied in a fibre optic between them. One was as the emitter and the other as both a photo-resistive and photo-voltaic sensor (tried it both ways). While it did sort of work, even with the optimization, I got pretty poor sensitivity. Better in the photo-resistive mode but, still pretty weak.

I also have some older phototransistors but, not too many and, like my photoells, a pretty motley collection. I would rather find an alternate (or a good source of cheap phototransistors).

Have you done any testing on this and accumulated some data?
 
crashsite said:
Have you done any testing on this and accumulated some data?

Nope, but I plan to do some simple tests tonight after the baby's in bed. I'll report back if I find anything interesting, but it sounds like you pretty much already covered it.

I thought it might be too easy. :)


Torben
 
Paper?

Torben said:
Nope, but I plan to do some simple tests tonight after the baby's in bed. I'll report back if I find anything interesting, but it sounds like you pretty much already covered it.

I thought it might be too easy. :)

Be interested to hear what you discover. I did my testing after reading an article in some electronics magazine (back when they were made out of paper rather than bytes of data) and, got a bug in my ear to try it.
 
crashsite said:
Be interested to hear what you discover. I did my testing after reading an article in some electronics magazine (back when they were made out of paper rather than bytes of data) and, got a bug in my ear to try it.

Hey crashsite,

Well, a couple of evenings of tinkering and figuring later and I can say I've had no luck whatsoever with the LEDs as sensors. I know that most examples have used op-amps for the high input impedance and I haven't tried that yet--just various transistor amps since I figured you might not want to dedicate an op-amp to each little indicator light you want to show. No joy. I made several nice touch sensors though. :)

However, a CdS cell feeding a simple Darlington made of 2 2N2222s works nicely, with a pot to set the input bias as a sensitivity adjustment. It's very sensitive and while not super-fast it seems more than fast enough to give a good visual replication of the hard drive activity light on my laptop (I hot-glued the CdS cell over the LED on the top panel).

So I figure you could do this with an integrated Darlington, LED w/limit resistor, 2 resistors or a trimpot to set the sensitivity, and a CdS cell. That's the best part count I've come up with for it.


Torben
 
That's Pretty Much My Thinking

Torben said:
However, a CdS cell feeding a simple Darlington made of 2 2N2222s works nicely, with a pot to set the input bias as a sensitivity adjustment. It's very sensitive and while not super-fast it seems more than fast enough to give a good visual replication of the hard drive activity light on my laptop (I hot-glued the CdS cell over the LED on the top panel).

Yep, that's pretty much what I was thinking. Actually, I was assuming that, depending on the LED (size, brightness, speed and color) I'd probably end up using a mix of transistors and IC comparators. The comparator especially for trying to capture the light from the bi-color LEDs with filters.

If you've still got a bug in your ear, you might like to try some different colors with cellophane filters. It's been awhile since I saw the color response curve for a CdS cell but it's likely that there are some conbinations that would be tough.

Actually, a chart of the resistance change of the cell with different LEDs and LED/filter combos (either measured directly or via voltage or current changes) would be very useful.
 
crashsite said:
Yep, that's pretty much what I was thinking. Actually, I was assuming that, depending on the LED (size, brightness, speed and color) I'd probably end up using a mix of transistors and IC comparators. The comparator especially for trying to capture the light from the bi-color LEDs with filters.

If you've still got a bug in your ear, you might like to try some different colors with cellophane filters. It's been awhile since I saw the color response curve for a CdS cell but it's likely that there are some conbinations that would be tough.

Hmm...filters could be interesting. The sensitivity range on the Darlington "re-driver" I built is pretty good, though; I could easily set it to respond either to bright room light or, when on the HDD light on the laptop, to an LED which is mostly hidden as the HDD LED is behind a cutout mask of the outline of the flowchart symbol for disc storage.

I agree that an accurate colour-matching re-driver would be tougher.

Actually, a chart of the resistance change of the cell with different LEDs and LED/filter combos (either measured directly or via voltage or current changes) would be very useful.

I might try to look one of those charts up, but I'm house-sitting for the next couple of months so anything I do will probably either be on paper or in LTSpice. :)


Torben
 
Status
Not open for further replies.

Latest threads

New Articles From Microcontroller Tips

Back
Top