I'd like to build some LED versions of the 194 wedge base bulbs used in older automotive dash lighting. My goals are:
1) powered by directly from the dash dimming signal, 9.3-11.0 VDC
2) fit in the stock 194 wedge base bulb holder and within the confines of the factory green tinted lens (about 0.600"x0.750" board area)
3) provide up to 4X the brightness of the original incandescent bulbs
4) dim from zero brightness to 100% using the same the 9.3-11.0 VDC range
5) be reasonable cheap
6) model in LTSPICE

I've got a few simple circuit designs modeled, but am concerned about brightness and directivity of the LEDs. With brighter LEDs having about 3.2V forward voltage I can only get two in series before running low on voltage, and with 120 degree viewing angle I'm thinking they may be too directional. These go inside an instrument cluster and there's a bunch of reflection going on in there to spread the light around though, so it might be ok. But if I could find a way to add more LEDs at different orientations it would be better, though that would take a boost converter. Here's the circuit I was starting with: https://link.springer.com/content/pdf/10.3758/BF03203658.pdf

Do any IC's come to mind that might work for something like this? It would be nice to find an IC with DC dimming and built in LED driver transistor, but most that I find that come close aren't by major manufacturers so there's no spice model for them, and I'm not confident building anything without pencil whipping it to death first.

Thanks!

 

1) powered by directly from the dash dimming signal, 9.3-11.0 VDC

Is that an analog (not PWM) voltage? Are you sure the range is that narrow?

2) fit in the stock 194 wedge base bulb holder and within the confines of the factory green tinted lens (about 0.600"x0.750" board area)

Hmmm... That's going to mandate surface mount. Unless you move all the control to a single LED controller unit that has nothing to do with the individual LEDs.

3) provide up to 4X the brightness of the original incandescent bulbs

That may be a challenge. Bright LEDs still make heat that needs to be dissipated. I suggest you start talking in lumens, and then see if dissipating the power of 4X lumens LED is practical. Of course adding more LEDs gives you brightness and more area to shed heat.

4) dim from zero brightness to 100% using the same the 9.3-11.0 VDC range

Doable. You need a PWM controller for the brightness adjustment, and a buffer that converts the 9.3-11.0 to, for instance, 0-10V.

5) be reasonable cheap

You need to define this.

6) model in LTSPICE

Not gonna happen. I could be wrong but I think this will be a custom assembly and you'll need to build your own model from the components.

 

Thanks for taking the time to reply.

The dimmer voltage hops from 0V to 9.3V when you when dialed past the "off" detent, and then changes pretty linearly up to 11V over the rest of the rotation of the knob. That was with 12V at the battery at the time, so it may vary a little when the car is running.

I love SMD stuff, and this is going to be the smallest thing I've made yet in order to get it to fit in the stock location with no mods to the green colored lens.

The typical 194 wedge bulb is 3.8 watts at 14V, 25 lumens, and 2800k. I haven't spent the time yet to learn how lumens and mcd relate, but that's going to be important I'm sure.

PWM is one way to get the dimming to happen, but I don't want to rule out current control as a possibility if it might make for a simpler design. The circuit doesn't really need to be extremely accurate - the car is 50 years old and stuff is relatively unrefined. My first design used PWM, but I was able to do it with current control with fewer components. Both were limited to two LEDs by the limited voltage available though.

As cost goes I'd like to keep the driver IC under $5 and the LEDs under $1 each, USD.

I'm pretty decent at modeling in LTspice and can put together just about any typical circuit, but there's a specialized driver IC needed for the circuit I'd want it to have a model available so I can integrate it into the circuit model. So manufacturer's like Taiwan Semiconductor would be off the list, in favor of Linear/Analog, TI, Maxim, and similar.

Thanks again for your help.

 

PWM is one way to get the dimming to happen, but I don't want to rule out current control as a possibility if it might make for a simpler design. The circuit doesn't really need to be extremely accurate - the car is 50 years old and stuff is relatively unrefined. My first design used PWM, but I was able to do it with current control with fewer components. Both were limited to two LEDs by the limited voltage available though.

Well if we don't care about efficiency, and why should we?, then a simple scheme would be a single voltage controller for the dash and a current-limiting resistor for each LED. You choose the resistor to give the maximum safe brightness (current) at the max system voltage, ie. 15V. The resistor could easily be built into the LED module and in fact you can buy "12V" LED bulbs ready made. A big advantage of placing all the LEDs in parallel is that the failure of any one won't bring down any others. But you could certainly put 2 or 3 in series with a single resistor if you prefer.

 

That would definately be a more straight forward path, but I'm trying to make a "bulb" if you will that can directly swap for the originals with no other mods to the dash wiring. It's more challenging than I expected!

Improving efficiency would be really nice. The insulation on the 50 year old wiring in these cars is pretty dried out and the connections all have a bit of oxidation of course, so anything that can reduce the electrical load is probably a good thing.

One thing that I will need to include is a P-type mosfet as a reverse voltage protection, as the sockets aren't polarized so you can plug in the bulbs either way.

 

... I'm trying to make a "bulb" if you will that can directly swap for the originals with no other mods to the dash wiring.

Yup, that's well beyond my (albeit meager) expertise. How exhaustively have you searched for a commercial solution?

 

How did you measure the input voltage?

Measuring the voltage on the existing lamps while they function?
If you remove the lamps and try again, you may see dramatically different results.
Designing around voltage alone might be a failing strategy.

The LED lamps might need to have resistors installed to simulate the current draw of the incandescent lamps they replace.