School Me on Relays

[the.ronin]

I have a CM Scout and plan on wiring a cold cathode to the LED on/off switch that comes equipped with the case. Unfortunately, the switch is built to deal with only two of the case fans it comes with.

I came across a post made by an enthusiast who managed to tap the switch using a micro relay. I’m having trouble deciphering his diagram and he is no longer active in the forums.

And the specific micro relay he used for this application:

Which terminals on the micro relay correspond to which terminals in his diagram? In particular, the NO (naturally open) side of the relay?

 

[Mohonri]

Typically with RadioShack items, there's a diagram on the packaging which will tell you which pin is where. Do you have this relay in hand?

 

[the.ronin]

The relay came in a fairly nondescript plastic bag bearing only the Radio Shack logo and some specs. I will take a closer inspection to see if anything is indicated on the bag or more likely on the relay itself.

I came across this diagram of an SPDT relay which is drawn to the actual physical configuration of the terminals ...

I suspect the middle pair of terminals may be the coil which I hook up to the LED and switch such that when the LEDs are turned on, the current is sent through these coils activating the electromagnet.

My guess is that the NO is the corner terminal in the same side of the single, unpaired terminal. I’d wire the cold cathode to these two terminals. This would provide a configuration where no current from the LEDs leaves the switch “open” or no power to the cathodes. Conversely, wiring to the corner terminal opposite the single, unpaired terminal, I would have a default “close” or power running to the cathode while the LEDs are off.

Does this sound about right?

Ok, so another noob question – I was planning on simply soldering my wires to the terminals and using heat shrinks. Is this the best way to ensure a secure connection? Or are there special wire clasps / connectors for this type of application?

 

[keenan]

Relays will typically have a diagram printed right on them, especially those intended for automotive use. Otherwise you will need to find a datasheet or deduce which pins are which by measuring or guessing. If you go to Radio Shack it should be printed on the packaging.

edit: Sounds like you've got it right, assuming that diagram is correct. Beware that it looks like this relay includes a built-in snubber diode, so make sure you energize the coil with the correct polarity. The ground side is the one on the bottom in the diagram. As far as connecting to it, these relays are designed to be PCB mounted and there's not really a better way to do it. Just make sure you don't melt the body of the relay too much and you'll be fine with your method, though using a relay designed for use with quickconnects might be worthwhile even if it's a bit larger. These are common automotive parts and should be available cheap, though finding a 5V one might be a little more difficult.

 

[the.ronin]

Thanks, keenan ...

Beware that it looks like this relay includes a built-in snubber diode, so make sure you energize the coil with the correct polarity. The ground side is the one on the bottom in the diagram.

To be clear, I have the black wire running from the switch to the LED to the middle terminal on the same side as the unpaired terminal. The middle terminal on the opposite side of the unparied terminal runs (red wire) to the LED to the power source to the switch.

On the cathode side, I have the black wire running to the corner terminal on the same side as the unpaired terminal. The unpaired terminal runs (red wire) to the power source which runs to the cathode.

I will draw up a diagram when I get home.

 

[keenan]

I'm not really clear on what you mean, so I'll wait for the diagram

One other thing - LEDs typically have a voltage drop of somewhere between 1.5 and 3V, which might not be sufficient to pull in the relay, and they need to be current limited which might also prevent the relay from operating off that pin. Unfortunately if you're working with a prebuilt device you don't have a lot of choice on this fact, though if you can avoid any current limiting resistor for the LED you should be able to get the full 5V to your relay. Just be aware of this and connect as close the controller as possible - there's not much you can change without modding the controller - since if there is a resistor integrated in the controller (which seems likely), the voltage and current allowed might not be sufficient to drive the relay and you may need a transistor buffer similar to the second diagram you posted.

 

[the.ronin]

Thanks again, Keenan. I’m not as considered with the power draw from the LEDs being sufficient to trigger the relay as this solution has already been proven by someone else for basically the exact same application. Only problem is that guy is no longer active in the forums where I came across his solution.

 

[the.ronin]

Ah I forgot you can render PowerPoint presentations as images haha ... anyway, here is what I was thinking ...

 

[keenan]

I would connect the relay directly to +5V, not through the LED, but otherwise this looks correct.

 

[the.ronin]

Thanks for bearing with me on this Keenan, I do have one additional question. I suspect I’ll be needing to extend the wiring than what’s available. What gauge would be a safe one to go without encumbering cable management with huge wires? I’m thinking 18-22?

By the way, as you probably guessed, the colors on the wires are also meant to correspond with the actualy wire colors themselves.

 

[pissboy]

I would connect the relay directly to +5V, not through the LED, but otherwise this looks correct.

It should be fine, they're in parallel.

 

[keenan]

What gauge would be a safe one to go without encumbering cable management with huge wires? I’m thinking 18-22?

For the coil wires you can use basically anything. It's only 90mA. However for the load wires I would use something a bit heavier, but still not really worth concerning yourself over it. The typical CCFL will draw < 1A. If you've just got one lamp I'd say anything at least 24AWG is fine.

 

[the.ronin]

Guys, thanks so much for taking the time to provide your feedback and insight.

I&#8217;m hoping to get to this once the cathode arrives from Newegg. I&#8217;ll report back as to whether I managed to blow my whole system up haha.

 

[the.ronin]

I updated the schematic ...

So ultimately, I&#8217;m going to have two molex plugs with a single red wire running through it? If so, I&#8217;plan on using a splitter to connect both plugs to a single molex power.

What&#8217;s throwing me for a loop is that the actual wiring that comes with the Logisys cathodes I have on order incorporate a switch and molex extender in a series circuit (I think it&#8217;s called) ...

It think the power in is the 2-pin yellow & black molex with the extender as the full 4-pin molex but the switch has the red and yellow voltage wires. Not sure how to go about hacking this to accomplish my goal.

 

[keenan]

The switch needs to be actually switching something. As you've shown it's connecting +5V to ground, which is bad news bears.

Here's what you're looking for. (Go MS Paint! This looks way better than I could have done in XP in the same 3 minutes!).

Edit: It's also wrong, assuming your inverter runs on 12V, but I think you can figure that out...

 

[the.ronin]

Thanks for the diagram, Keenen ... I'm lovin the cathode light beams haha.

Yes, the switch is running the fan LEDs. Can you elaborate on how this is bad news bears?

Well, I warned I was an electrical dolt ... yes, there are schematics on the back of the plastic packaging of the relay ... and I was all wrong haha. I will post back up after this other thing I wanted to do - I want to try switching out the red LEDs from these "proprietary" CM fans with LED on/off wiring to blue ones.

 

[keenan]

In your latest diagram the switch connects a red wire (which also connects to the molex) and a black wire (which also connects to the molex) when it's turned on. You're shorting power to ground, and this should cause your PSU to shut down before disaster occurs. Otherwise something is going to start burning pretty darn quickly.

The switch needs to be breaking the circuit either on the power->LED leg or the ground->LED leg, not connecting the two.

Note also you don't really need to use two separate molexes, just pull the lines for the cathode off before the switch.

 

[the.ronin]

Ah I think I understand, Keenan. For what it&#8217;s worth, here is the diagram from CM on their LED switch ...

I take it then that although both wires are running into the switch, only the ground wire is actually being switched? The red wire is passed right on through to the molex?

Let me stew over how I run everything over the same molex. Question though, 4-pin molex always have Yellow/12V, Ground, Ground, Red/5V right? How do you tell which is pin 1 (yellow)?

 

[the.ronin]

After reading some craziness about inverters melting and blowing out, I decided to just go with an LED strip instead. It's 12V so I still need to go with the inverter. So basically same concept here ...

I'm just going to splice into the LED leads and bring those wires to the relay.

The relay diagram is verbatim to what's printed on the plastic packaging it came in - but I'm not sure if that relay position is with or without power on the 5V line (no indication of NO or NC). I'll just have to test.

I'm just going to share a molex for the LED strip and the switch. There seems to be 4 wires coming from the switch to the molex. A set of red and black wires go into each of pins 3 and 4 - but I assume the pair going into 3 are both ground. The LED strip's ground is going straight into pin 2 and the 12V wire is going to the relay and from the relay to pin 1.

Hope this works.

 

[keenan]

Looks good to me. You could just use a transistor, but you've already got the relay, so

The diagram shows the relay's rest state, so you've got it right.

 

[the.ronin]

It works!!! Thanks so much Keenan for schooling this electrical noob.

Right now I just have test wires flying out from all over the place so I can test the wiring before installing. Will be installing later (have to spend "quality time" with She Who Must Be Obeyed) and take some photos.

Thanks again - it's really cool when people like you are willing to help complete strangers. Makes me feel all warm and fuzzy.

[edit] Turns out the polarity matters on the relay and when I had it wired up based on the diagram above, the relay would click but the LED strip stayed on. I switched the polarity and I'm hearing an even louder click and everything works as planned.

 

[the.ronin]

Hey keenan, here's how the setup looks now. It may not look it in the photos but it's really a lot brighter than it was before with just front and rear fans providing light.

You can see where I tucked the strip into the top here ...

And for future reference this schematic is about as clear cut as I think it gets ...

The rear fan LEDs are definitely not as bright so I suspect that's because of it being tapped into the relay.

Thanks again keenan for all your advice!!

 

[keenan]

Very slick. Now all you need is a nice colour changing PWM controller and RGB LEDs!

 

[dandragonrage]

Looks good to me. You could just use a transistor

This would have been the right answer. Relay is the wrong way to do this. Relays are way overused.

 

[the.ronin]

Ok guys ... I looked into the transistor and also into resistors which to me, frankly, were way more easier to understand. I was wondering if I could use a terminal block and a resistor to power LED and CCFL lighting form a single switch this way (please excuse the newbie powerpoint diagram - I have no idea what proper electrical symbols should be) ...

The LED leads from the fans are hooked up to the terminal block in parallel as well as the power leads from the molex through the switch. But the 12V lead to the LED terminals will have a 470 ohm 1/2W resistor which (at least according to this calculator) bring the voltage down to 5V.

Does this look like it might work?

 

[keenan]

I can't remember if you're using bare LEDs here, or if the LEDs are in the fan and you give it exactly 5V and the resistors are built in.

If the LEDs are bare:

It will work and nothing will blow up, but you should use a separate resistor for each LED. Current will be half what you expect (though you could just half the resistor value), but the bigger issue is that unless the LEDs are very close on forward voltage they'll probably be slightly different brightness. If you're driving them close to their limit you might also be pushing a bit too much through one or the other. It'll work though, and probably look fine, but the right way is to use a separate resistor for each.

You could also wire them in series, add their voltages together for your calculator and then use a single resistor. This'll negate the issues I speak about above.

If the LEDs are built into the fan and designed to be run on 5V power directly:

Your resistor idea would work (and you wouldn't need the extra resistors I talk about above), but you need to know the actual current drawn by the bottom half of the divider to choose the value accurately. Measure it if you can, then use V=IR to determine the necessary resistor value to drop 7V. You should calculate the power dissipation as well, it's quite possible it'll be too much for a 1/2W. Probably not that important that it be exactly right, but if your estimates are off enough, you'll break things. You could probably also wire them in series without any resistor at all but there's a chance the extra 20% would break things.

 

[the.ronin]

Keenan, once again, thank you very much for your input. Yes, the LEDs I'm attempting to power at 5V are 4 blue LEDs in 2 case fans for a total of 8 blue LEDs. Since the point of this is both simplicity and modularity, I'll likely be tapping more blue LED fans to the terminal block.

... but you need to know the actual current drawn by the bottom half of the divider to choose the value accurately. Measure it if you can, then use V=IR to determine the necessary resistor value to drop 7V. You should calculate the power dissipation as well, it's quite possible it'll be too much for a 1/2W.

Hmmm ... ok this is a bit over my head. So use a voltmeter to see what the actual current is being pushed through the 12V lead? Power dissipation? What if I bring the ohms down to 390 which according to the nifty online calculator implies a pretty hefty 4.6V blue LED. Or can I just be lazy and maybe you have a recommendation for a safe resistor value?

You could probably also wire them in series without any resistor at all but there's a chance the extra 20% would break things.

I considered this but then I would need at least 3 fans in series to avoid overvolting and knowing me, I'll power everything up with just 1 fan connected. This is what brought about the resistor to keep connections simple and modular.

Thanks again.

 

[keenan]

If they're designed to be powered directly from 5V, you can do this:

1) Measure the actual current they pull all together when connected to 5V, it's probably around 120mA if you've got 8 LEDs on it.
2) You want to calculate a resistor that will drop 7V at that current so that the fan LEDs will see 5V. So plug the values into R=V/I, where V is 7V and I is the current you measured above (remember that it's in mA so use 0.12 or whatever). This will give you the resistor value to drop that voltage.
3) Calculate the power dissipation of the resistor at that current, the formula is P=VI, where again V is 7 and I is the measured current. The result is in watts, and you need a resistor rated at least as high as the value you get here. If you're close to the rated value the resistor will get VERY hot, so try and get one at least double the required rating. An alternative is to put two lower rating resistors in parallel, just remember if you do that you need to double the ohm value of the resistors to compensate. You may also do the calculations for a single fan connected (half the current) and then use one resistor to connect to each fan.