What's new
CK5
Register an account today to become a member! Once signed in, you'll be able to participate on this site by adding your own topics and posts, as well as connect with other members.

Electronics Gurus? (resistors, diodes, etc)

dyeager535

dyeager535

1 ton status
 Premium
Joined
Dec 13, 2000
Posts
31,569
Solutions
1
Reaction score
2,908
Location
Roy, Washington
Looking for the simplest, cheapest way to reliably get a check engine light (CEL) to work properly, without being dimly illuminated at all times.

Tried a bunch of different resistors, but there is no "happy medium". If the resistor value is enough to keep the light from being dimly on (when it's not supposed to be) it's also enough to make the CEL be dim enough to be harder to see than it should be when it's actually on.

Problem is that the LED doesn't offer enough resistance to keep the ECM from "leaking" as I've seen it called. ECM grounds the circuit when the CEL should be on, but with an LED, the light is always on dimly because the LED has little resistance.

I can live with it dimly lit, bulb life isn't an issue, but I am sure there has to be a specific component that can solve the issue, and I'd like to if I can.
 
Last edited:
The easiest way, although its seriously low tech, is a small low power relay.

There is no ambiguity with a relay. Its either on or off.
Run 12 volts through the relay contacts to the LED with the proper resistor, and let the ECM turn on the relay.

If you need high tech, there are voltage sense circuits that will block the current until the voltage reaches a certain point.

What we need to know is what is getting to the LED when it should be off.
I suspect that when its supposed to be off, there may be almost full voltage leaking to it at very low current.

If so, then you need to determine the correct resistor for the LED at max brightness at 14.5 volts to keep from burning out the LED.
Then you need to put a drain resistor to ground before the LED/resistor pair. This would bleed off the trickle current so it does not light the LED.

Probably about the same value as the load resistor for the LED, but I would start with more to minimize the load on the switching transistor in the ECM.
 
On the + side the LED is seeing voltage around 13.5V at all times (when key is in run).

The ECM supplies the ground. The LED's work fine at the vehicle voltage as-is, they are designed as replacements for the 194/168 series bulbs. Obviously need to account for like you said 14.5V on the upper end, down to about 10V for those wild possibilities where the dash voltage goes weird, failed alternator, poor idle, etc.
 
Use a basic transistor circuit like the one here (about 1/3 of the way down the page under "Testing a transistor". Just connect the ECM supply the ground for the circuit.
http://electronicsclub.info/transistors.htm

Or you can use this one. It just uses two resistors as a voltage divider to control the 'on' voltage to the base of the transistor. Obviously you'd only be using one LED instead of the 5 in the circuit.
http://www.fritz-hut.com/transistor-switching-leds/
 
Yeah, darn it. ECM outputs are always sinks, never sources. I don't why I have so much trouble remembering that.
The drain resistor to ground would make the problem worse.
However, hooking it to +12 instead might fix it.

Either of those transistor circuits have the same problem. They are designed to turn on with a positive voltage. Here we need one that will turn on when grounded to a negative voltage.
Its simple enough to fix, but for what you are doing, I would go with a relay. I suggested that because its really easy if you do not have the breadboard stuff to do transistors.

If I were doing it, I would probably go with transistors, since I have lots and the equipment to fab a board and all the extra parts needed.

A small cheap fairly high sensitivity relay with the coil hooked between +12 and the ECM's CEL line would get the job done.
If you are using an LED that is designed to replace the bulb already, then it should have the correct resistor built in for normal auto voltage.

Just hook one relay contact to +12, the other to the LED, and ground the other LED lead.
I'll scrounge around and see if I can come up with a relay model that closely duplicates the load of the original bulb.
 
OK, I looked up the specs on the original bulbs, and they equal to about 50 ohms in one case, and 40 in another.
So any coil above 50 ohms is going to be less load on the switching transistor in the ECM and therefore not much chance of blowing it.

This one should do fine.
http://www.alliedelec.com/search/productdetail.aspx?SKU=70132373#tab=specs

The only two problems with it, are the mounting and the fact it really should have a suppressor diode.
But, it fits a standard 16 pin DIP socket, such as this.
http://www.alliedelec.com/search/productdetail.aspx?SKU=70206323

Also I doubt its needed, but you might want to pick up a diode or two just in case.

http://www.alliedelec.com/search/productdetail.aspx?SKU=70214892#tab=specs

Given the price of all these, you might want to get a couple of each, maybe more of the sockets.
 
So it's not truly grounding the output but more like the output from a buffer/driver? I see in the schematic the CEL is coming off of a 'driver', but it doesn't say what exactly that is.
 
The simplest thing I can think of is to put a zener diode in series with the LED. Something like an 8V zener would keep the LED off until the ECM sink gets 11V below the battery voltage. Then you resize the current limiting resistor to about 1/3 what it is now. Can you measure the voltage when it is lit dimly? It would also be nice to know the voltage when it is "on" to do the math.

Another idea is to put a resistor in parallel with the LED/resistor combo. If you bring the load up similar to a bulb, it might make that circuit in the ECM behave better. This defeats the efficiency gain of LED indicators, but for the amount of time a CEL is on, who cares?
 
Why not hook up a stock CEL bulb between the computer and the LED bulb? Use it like a redneck resistor.
 
I had thought of the Zeners, will one work just as I need it to? I vaguely understand them, just wasn't sure what the "rating" meant on them. You mention current limiting resistor, where would one of those be? As I understand the components of the LEDs I'm working with, the only component outside of the PCB they are installed on, is the diode that won't let you blow them by hooking up backwards.

I forgot to measure the voltage with key on/light off on it last night when I was working with the cluster.

I have looked at those type relays before, been playing around with some real low voltage stuff. If the Zener would work, that would seem the least bulky/complex arrangement, which would be good.

I did get the PWM potentiometer replaced with resistors though (yay, something that worked as expected, finally!) so at least I made some progress.
 
Oh, the resistor is built in. Well that's harder. The zener could still work, but the "on" brightness will be somewhat reduced from where it is now.

"Zener" diodes (actually avalanche) work like a normal diode in the normal direction (forward biased), but they have a nicely controlled reverse breakdown voltage in the opposite direction (reverse biased). So it stays off until the voltage reaches the rated threshold and then it turns on with that voltage, across a wide range of current. So it's a voltage regulator, but also like an on/off switch.

diode11.gif


So the question is how much of a voltage threshold is required to keep the LED off until it's supposed to be on. If it's just a couple/few volts, this idea could work. If it's like 6/8/10 volts, then the LED will end up pretty dim.

I think the parallel resistor is worth a try, though. An LED normally has about 3V on it when on (forward biased). If you can get the ECU line within 3V of the supply (check engine light off), the LED should be "all-the-way" off. Let's say your dim LED is about 10mA. That would make the sink path in the ECU about 1.2kOhm. A parallel resistor across the LED assembly sources current to that ECU path and the I*R brings the voltage up. To get that line within 3V of the supply voltage will take about 300 Ohms. That's over 1/2W when the light is on, so it should be a 1W resistor or larger. The ECU will be fine with this, as it's a smaller load than the stock bulb.

Not knowing any actual voltage or current, though, these are guesses. You could try larger R values, like 1K, and work down to see what works. If a larger value works, it could end up being a little 1/4W thing that you can stuff anywhere.

Do you have DMM with current?
 
Here is what we really really need to know. When the CEL is supposed to be off, but the LED is glowing faintly, we need to replace the LED with a resistor or two.

First about a 50 ohm, and then a 400 or so. Measuring the voltage across the resistor each time.

Here is what is happening. The hot lead is hooked up all the time, so you will see full battery or alt. voltage to ground at all times measured with a high impedance voltmeter.
Also, you will probably measure most if not full voltage from the hot lead to the ECM.
Under no load.

In theory, there should be infinite resistance to ground at the ECM's CEL line.

But whatever switches ground on that lead is leaky. It is letting some small amount of current through. With no load, the voltage will rise to full voltage. But, with a load, it will drop.
By measuring the voltage at the two loads, that will tell us the equivalent resistance of the leaky switch.

For instance, the relay I suggested has a 400 ohm coil, and a guarantee turn on voltage of 8.4 volts.

But I doubt that there is more than a volt or two across that load. Ohm's law, most of the voltage is going to be dropped across the higher resistance.
 
Blue, the only thing that worries me about a Zener is that its a voltage device. Our problem here is mostly current I think.
I suspect a zener in series would stay off presenting a very high resistance to the ECU.
Then, the voltage across it would rise because of no load until the zener fired. In which case the LED would light, the current flow would drop the voltage back down below the zener turn on, and it would reset.

So we would have a blinking CEL.

If he has some resistors, which I think he has, he can use a couple to see what the apparent resistance of the ECM line is so we can make some intelligent suggestions.
Remember the input to the system is full battery voltage, we just need to see what the ground impedance is....

Don't know what the forward voltage and current for that LED is, but if its red, I want to say 1.5 volts@about20ma for full brightness. The 3 volt ones are usually white or blue if I remember right.

I will drag out my Allied or Digikey book out and check in a minute.
 
The "bulb" is eight, white SMD LED's if that helps any.

Not at home right now, can't measure anything unfortunately.

I should have time to play with the resistors on that line when I get home tonight. FWIW I measured the resistance of a few 194 bulbs with the meter in the "20" position and saw 4-12...adding a resistor with a value in that range to the circuit of course did nothing.
 
Interesting fact, the resistance you measured on the bulb, is not the actual resistance.
If the spec sheets are correct, the actual resistance is in the 40 to 50 ohm range.

What you measured, is the resistance of the bulb filament cold. When it gets incandescent, the resistance goes up that much.
 
So what I need to do is use a couple of resistors, and when the light is "dim", measure the voltage on either side of the resistor?
 
Yeah, just replace the LED with a resistor when the light is supposed to be off but is really dim, and measure the voltage across the resistor.
In other words, put the voltmeter probes on each side of the resistor.

By knowing the value of the resistor and the voltage, we can determine everything else.
 
Alright, done!

Meter setting at 2K
Resistor one measures .24
Resistor two measures 1.029

"input" voltage was 14.25
Voltage with Resistor one: .05
Voltage with Resistor two: .31

I hope I didn't reverse the voltage with Resistor numbers, was working in the dark (how nice, interior lights decided to quit since its been sitting!). Do those numbers make sense?
 
Essentially you've measured current at 2 load values. I'm reading those as 240 Ohms and 1.029kOhm. Just to verify - resistor 1 is red/yellow/brown (red/red/brown) and resistor 2 is brown/black/red (or brown/brown/red)? If so:

1st case = 0.2mA
2nd case = 0.3mA

What's interesting is how the current changes only about 50% when the load changes by 4x. Is this with the LED removed? Your LED pack should be close to this range and I'm surprised such low current gives noticeable glow, which is why I want to make sure we're not off by a factor of 10 or something.

You say you added a resistor like 12 Ohms to the circuit with no effect. This must have been in series with the LED, no? In parallel would be bad, possibly damaging the driver in the ECM. The specs I see for 194 bulbs are 4W or less, which is 1/3A. That 1k resistor would be a good candidate to try in parallel, though. You're getting down close to 1/8W when the bulb is on and it can source more current than the ECM is leaking.
 
This is what I was wanting to know. Unfortunately I am about to head to the doctor, so I don't have time to think about it.
Like Blue says, double check the colors on the resistors. But I am thinking that we may have some non-liner stuff going on here in the ECM.

Also, Blue, consider this: There is a very good chance that we do not have a simple load resistor in the LED. We may have an actual buck/boost power supply like the better LED flashlights have.
My flashlights will keep shining quite brightly even as the batteries get weaker and weaker.
The voltage regulator in the light will boost the voltage until the batteries are sucked flat.

In this case, the "bulb" may be boosting what power is available to try to light it even though its too low to normally light an LED.
 
You must log in or register to reply here.

Latest Posts

Top Bottom