Standard disclaimer for these discussions: DOT-approved lighting units experience this same exact issue, this has nothing to do with street legality or lack thereof

With that out of the way, I have an old GM car that has a double function of the front corner lights (194 bulbs). One side goes to the turn signal circuit, one side to the parking light circuit. I've also seen some cars that have illuminated side lights that are grounded to the turn signal circuit to make a side blinker mod. The problem is usually the front parking light bulbs, which are dual filament (1157 in my case). With all the bulbs incandescent, everything works as expected. The problem comes in when you upgrade everything to LED's. Reason being, the front corner 194's ground through the parking light 1157's. With incandescent bulbs, the resistance is low enough where it doesn't matter much. With LED's in the front parking light 1157's, the resistance becomes so high that the circuits get really screwy with resistance and backfeed etc. to the point that depending on what you pick, you can even wind up having the bulbs try and ground through the instrument cluster turn signal indicators and disabling your turn signals entirely whenever the parking lights are on. Not good.

I decided to try some multimeter measurements with the LED's installed in the front parking light 1157's, but both front corner bulbs removed.

The problem?

With both circuits off, the parking light circuit is connected to GND. The turn signal circuit however is left floating. While I could get a turn signal (and Hazard for that matter) relay that has an actual ground, I want to rewire my car to use the 8-pin Subaru modules so I can throw an old TapTurn relay I have on there for the approach lighting feature, and it too leaves the outputs floating instead of grounded, so a more proper solution is needed.

Here is, effectively, what I am dealing with.



[Link to Falsad]

Note that C1-* and C2-* will effectively be an extension cable with some circuitry in the middle to connect any floating grounds. C1 goes to the bulb itself, C2 comes from the bulb feed wires. S1 and S4 are purely to test and see if floating GND vs connected GND works as intended. S2 and S3 flip the turn and parking circuits on and off.

The goal is to connect any floating grounds when a circuit is otherwise off. So in the picture above, since S2 is OFF but S1 is OPEN, C2-T should ground through C1-GND. Current passes from C2-P to C2-T to illuminate L1 as a parking light. For full clarity, let me show other screenshots of the same circuit in different states.



This one is working as intended. S2 is ON, so the turn signal circuit's floating ground should not be connected to C1-GND, and S3 is ON but S4 is CLOSED, so there is no floating ground on the parking light circuit. Current passes from C2-T to C2-P to illuminate L1 as a turn signal.

With both circuits energized, there is no ground through either side. This behavior needs to be preserved:



This one is also working as intended. S2 and S3 are both ON, so no ground to C1-GND should be added for either circuit.

My first thought was a couple N-channel MOSFET's in a criss-cross pattern, which fixes the floating grounds!



[Link to Falsad]

As intended, S2 is OFF, and S1 is OPEN, so C1-P opens the Gate of T2, passes through the Source of T1, goes across L1 from C2-P to C2-T, and grounds through the Source to Drain of T2, out to C1-GND. Perfect!
...until you try and have both S2 and S3 ON, that is



Now both circuits manage to turn on T1 and T2 and ground themselves, bypassing L1 entirely, creating a short circuit. Icky. Obviously an XOR or NAND is needed. So I gave that a go too, but kept getting a "Convergence Failed!" error, so I gave up on that.



[Link to Falsad]

Thinking maybe this is a simulator quirk, I looked through the NAND examples and tried to use the RTL one:



[Link to Falsad]

Unfortunately this doesn't work. Because current is flowing from C2-P to C2-T on L1, both sides of the NAND (DP, RP, and TP for Parking, DT, RT, and TT for Turn) are receiving power, turning on T3 through R1 and C1-GND, which makes the NAND effectively an OR, and bypasses L1 entirely

I've been at this for days throughout Google and such, I'm not an electronics engineer, just a hobbyist who still mostly doesn't understand what I'm looking at (but am trying from time to time). Ultimately my goal is to make a PCB that I can throw on GitHub for anyone else who wants to solve it with a circuitboard mounted by their own quirky blinking 194 bulbs wired the same way and is okay with some soldering and running a ground wire, but until I can figure out what's going on, I can't really do a whole lot of good with it.

I want to use this as a learning opportunity and avoid the overused gimmick of just throwing a relay on it Hopefully someone is able to help me figure out what I'm doing wrong. Or was I right with the last approach and a simulator quirk is misleading me? (Wishful thinking though I'm sure). Thank you!

 

That is indeed quite a challenge. The simplest trick would be to go back to incandescent bulbs for both positions.
The other would be to chage the scheme so that each light had the other side tied to body ground thru an added wire.

 

The other would be to chage the scheme so that each light had the other side tied to body ground thru an added wire.

I'm not sure I understand what you mean. C1-T or C1-P to GND is a short circuit.

 

OK, I am suggesting a rather total change in that scheme. The idea will be to provide two diodes to isolate the two power sources for the replacement LED light. That will be much less complex, and much easier to install.
It will require two diodes rated to handle the current drawn by the LED light, and some heat shrink tubing.
The negative side of the replacement LED must connect to the vehicle power negative, which is the body (ground).
In series with each of the wires, C2P and C2T, will be a diode, with the arrow pointing towards the LED connection. From the remaining ends of the diodes a wire to the POSITIVE LED connection. That is all that is needed.
When either switch is closed, that diode will be "forward biased" and the LED will light. The other diode will be reverse biased and have no effect. That is as simple as it can be without a major re-design of the system.

 

In series with each of the wires, C2P and C2T, will be a diode, with the arrow pointing towards the LED connection. From the remaining ends of the diodes a wire to the POSITIVE LED connection. That is all that is needed.
When either switch is closed, that diode will be "forward biased" and the LED will light. The other diode will be reverse biased and have no effect. That is as simple as it can be without a major re-design of the system.

Actually this breaks things in a way that makes it no longer street legal. When parking and turn signal light circuits are on, the bulb should be off. The way you're describing it, the bulb would always stay illuminated when the parking light circuit is on. Basically, it lights up as an XOR logic gate output, not an OR.

Now, I could make it an XOR gate, then connect L1 between the XOR output and GND, however I'm looking into a setup that will offer "switchback" functionality, where the parking light circuit would keep them lit white, and the turn signal circuit would light them up amber, which means that the direction of current flow matters. The LED bulb I am using for now though has a built-in diode bridge, so it doesn't mind much which way current flows through it and will illuminate no matter which way you plug them in.

So basically, the problem becomes how to ground C1-T or C1-P only if they are currently not connected to 12V. Floating ground or connected ground

Maybe a better question to ask: how can I detect that an input isn't connected to anything? A half diode bridge might not be a bad start. Is there a way to hook up transistors and pull-up/pull-down resistors in a way that'll get what I'm after?

 

Rather then spend time with an XOR effort, just add a decent turn signal light, that dinky 194 bulb mad a lousy turn signal when the sun was bright. Solve the problem by doing it right.
OR, don't use the parking lights, which are actually intended to warn that a parked car is there during darkness. Excpt that my dumb 2017 car switches the parking lights OFF after just a few minutes! The parking lights should stay illuminated until intentionally switched off. OR until the battery goes flat dead! That would be my rule.

 

Rather then spend time with an XOR effort, just add a decent turn signal light, that dinky 194 bulb mad a lousy turn signal when the sun was bright. Solve the problem by doing it right.

So, hack apart a classic vehicle to install some cheap failure-prone light assembly, drilling holes, cutting metal, nevermind screws that'll invariably break off or the rust and corrosion, having to fish wires through everything, take apart half the car to connect it all to the right circuits, spend days installing something that looks absolutely ridiculous and leave in a non-functional lens assembly... I'll pass on that suggestion. Besides, there's very literally thousands of guys I can find on Google that have this same problem and wind up settling for load resistors, incandescent bulbs, or just losing either the parking light or the turn signal of their front corner lights. I figure if I can get this working, I can throw the KiCAD up on GitHub with a BSD license for anyone to make, use, even sell if they want Great learning opportunity too! I know it's possible, just have to work through it. Someone at some point probably made their own PCB fix -- probably a few somebodies -- but AFAIK no one's made it publicly available for others to use. At least that I could find.

Well, other than a turn signal relay that adds a third pin to switch between 12V and GND, but the TapTurn relay I want to mod the car to use leaves the turn signals at a floating ground, so I'd rather get this working first before I get too carried away with that whole endeavor.

OR, don't use the parking lights, which are actually intended to[...]

What? Parking lights are always on when headlights are on. Are you saying to remove the parking light bulbs entirely? Or the entire housings? Or drive without headlights at all? I truly don't know what you're trying to suggest here, but I probably won't be doing that anyway. Kinda scratching my head at that one too.

I'm gonna go out on a limb and assume you're unfamiliar with the setup I'm talking about, so I pulled up a video of the same era of Ford (in this case it's a Taurus) to show you the intended function when headlights (and thus, parking lights) are on. First it shows with only turn signals (no headlight or parking light), then it shows with turn, parking, and headlight (though turn and parking without headlight operates the same way).

Dedicated and quite-visible amber parking lights separate from the headlights are very much a staple of that era of automotive design, and it was everywhere in the late 80's and early 90's, you really couldn't escape it. Dual filament front-only bulbs, single-filament side or corner bulbs, both amber-only (sometimes the rears did this too, but in red). They operate in unison unless the parking lights or headlights are on, in which case they alternate.

It did this in a rather jank way that unfortunately makes the LED path trickier than it should be. The 1157's obviously had two different inputs with their own dedicated grounds, but the 194's tapped into both turn and parking filaments of the 1157's and grounded through the 1157 filaments themselves. No dedicated ground to speak of in a single-filament dual-purpose 194. Here's the service manual for my particular car I'm doing this on if you want to see how the factory wired everything up. Note how the FRONT MARKER LIGHT (the side/corner ones) basically just wire up to either side of the FRONT PARK/TURN LIGHT, meaning the marker lights ground through the park/turn lights.



This is also something a ton of guys intentionally modify their cars to do, usually called the "side blinker mod" -- basically, you take the side parking lights and change the ground side so it "grounds" to the turn signal wire of that corner's turn signal bulb using a T-Tap. I know a ton of Subaru guys do this for example.

Now, for me in particular, the biggest reason I'm putting LED's in the front is for daytime running light usage. Switchback LED's illuminate white for parking lights, but amber for turn signals. They're only about as bright as modern DRL's, so a lot of guys run parking lights only to get the same effect of DRL's with the added bonus of the taillights being lit up too for a bit of extra visibility without needing full headlights. Here's an example of that modification on a Camaro this time:

While there is technically always a ground on each corner light / side marker circuit, it's such an extraordinarily high amount of resistance thanks to the LED units the front that they ground through that it's far better to design it as if the ground is floating entirely. I multimetered it in the range of megaohms, whereas most of the grounds on the car are far below 25 ohms. The instrument panel light ground for example is only 7 ohms. (I was having issues with it where it wouldn't light up at all -- connector worked its way loose and I didn't realize it).

That all being said, your mention of diodes did give me a lead, and pointed me on the right track, I think. I tried taking a page from the Arduino playbook and using pull-up and pull-down resistors and transistors and such to detect floating grounds, actual grounds, and VCC originally, but couldn't get it to work as a whole bypass circuit. I've revisited that idea, starting with the diodes instead of adding them later, and so far I have the detection logic working beautifully! I also cleaned up the schematic quite a bit while I was at it, since the ground lines were quickly becoming an absolute mess (though most of the labels are missing -- I'll work on that). Here's the progress:

[Link to Falsad]

Top logic section is for the turn signal circuit, bottom section for parking light circuit.

This is as far as I've gotten for now. I ran out of time today, but this was a major step in the right direction. I hope, anyway. All that's left is to see if I can successfully get the bulb hooked up to this whole thing and still have it work.

Thank you for the hat tip on the diode half-bridge idea! Seems to be the jump-start I needed. I do question the replacement unit thought but I won't bust your chops too hard on that. A lot of motorcycle guys would be super not thrilled if I gave anyone grief for installing extra lights Different strokes for different folks I suppose.

 

My comment is removed because evidently it horribly offended at least one individual.

 

I respectfully suggest that those lacking the resources, skills, and motivations to do the modifications properly, refrain from attempting the change. Certainly a change also requires patience, as well, to be done properly. And even more certainly there are many around who are not persistent enough nor patient enough to ever come close.

This isn't respectful, it is downright insulting and a disgusting choice to reply like this. Your version of "proper" makes you sound extremely young and immature, ESPECIALLY with a flippantly uppity reply like that.

You mounted a very random and unwarranted off-the-rails passionate attack on the notion of parking lights and decided that instead of doing a rather simple wiring mod, nah just install brand new lights and ignore the problem entirely, while calling that "doing it right" for some reason?! Which you're now doubling down on by saying "hey if you don't want to ignore the problem entirely, then you must be some lazy untalented slob so don't even bother." The arrogance here.

No, that wasn't respectful at all. Your version of "proper" and "doing it right" is -- again, ignoring the problem (I can't stress that enough, it's flying in the face of your post here), then getting a superiority complex when your own lazy hack of a non-answer is highlighted -- correctly -- as not an answer to anything.

I've tried to educate you on where you went wrong with your thinking there, with service manuals and videos, explaining the circuitry to you, hoping to get a little assistance along the way to come up with a good answer. Hell, I even went out of my way to thank you for the small brainstorm you did give me despite it being directly after an already arrogant and insulting reply, hoping I just misread things and a bit of a teaching moment on 80's and 90's tech could clear things up.

Nope. A very clear demonstration of the resources, skills, motivations, and dedication you accuse me of entirely lacking was lost on you, while you decide to go full-blown arrogance for no real reason.

Honestly, I very, very highly advise you to take a look at the advise you offer and internalize every bit of that. Then read through this thread again, top to bottom, everything I've outlined. Take your time, absorb it all in. Really just let it all sink in. Then ask yourself, is "screw everything, I don't care, let me just install some stupid light and call it a day" following any of what you said? Or -- hear me out -- is what I'm talking about, designing, and working towards, the actually proper answer?

All this especially given that I finished the circuit out, and the proof of concept now works fully as intended:



[Link to Falsad]

Play around with it, see for yourself.

- Resources? Check
- Skills? Check
- Motivations? Check
- Patience? Check
- Persistence? Check
- Condescending tone? Nope.

You?

I do hope this reply gives you a lot to think about. I don't like having to make replies like this, but frankly I don't know how DM's work here so I wasn't able to message a staff member to ask what the hell was the matter with you and how no one noticed your misguided narcissism that's off-putting and rude.

 

The tirade and criticisms are noted. And really,
Nor can everybody be a life-saving heart surgeon. Nor even a certified Master Mechanic.

That attack in post#7 was certainly rather critical, though.

I have seen modifications to stock vehicles that are better than OEM production! And probably others have seen them also. Consider that most folks can't create a prize winning show car. Far more extensive than revising a light housing.

The fact that some others have many greater resources than average is just how life goes, it is not being critical to point that out. The fact that "all are created equal" does not mean that all are gifted with equal abilities and resources.

Many of the participants here have experiences, abilities, and resources that vastly exceed those of the rest of us. That is just how life works.