How to add interval function to Formacar wipers

[Rogier]

If you want to add interval wiper function to the forma car wipers you will need
1 N-channel power MOSFET (rated at least for 15V)
1 P-channel power MOSFET
1 resistor 320omh.

See schematic how to wire up the components


120436

[Hojo]

Okay. I can solder up a circuit as well as anybody who can barely read a schematic and has no idea what’s going on. You have peaked the interest of all of us who like gadgets and don’t want to buy 10 of each mosfet and don’t know where to put them. Please, more details so I can figure out how to make this into a 10 day project.

[Rogier]

When you look at the wiring diagram off the Subaru wiper system you will see that the wiper motor has a switch which connect the green yellow wire to ground when in park and when in motion it connects the green yellow wire to the yellow black wire. So for the interval system of the Subaru to work we have to recreate the switch which is on the original wiper motor. We could do this by adding a mechanical switch to the Forma car wiper motor or we could do this with electronic components if we would have a signal which indicates when the wiper is in park position.
120474

The forma car wiper has a switch which when in motion connects the wire red wire to the white green wire. In there park position its isolated. So when we don't connect the white red wire.we get a 12V signal when the wiper is moving and 0V when the wiper is in park.
We use this to drive the two mosfets. The N-channel is not conducting between source and drain when the gate is at 0v where the p channel is conducting when the gate is at 0v. With !2 V on the gate the N-channel will be conducting and the P-channel will not. The resistor is to pull down the voltage to 0V on the gates of the mosfets when the wiper is in park.

Thetypical pinout for a to220 MOSFET is 120475
G= Gate
D= Drain
S = Source

I sodded up the components like this120476

Connected both gates and one side of the resistor to the white red wire. The drain of the p channel and the other side of the resistor to the ground wire. the drain of the N-channel to the yellow black wire. The source of the two transistors to the green yellow wire.

[lsfourwheeler]

How did you mount/waterproof all this? Or is that still tbd?

[Rogier]

well I printed a little housing for it an was thinking to put some epoxy resin on the top. only the little circuit stoped working as soon as I ended foddering it in place. :(
So I replaced it with a relay like this120624

[aquillen]

A relay is the easy solution as you note. I did not study how the rest of the circuit works but at first glance seems right - so if it works with the relay then cudos- simple and good fix with no challenge to building a more complex circuit.

Why the MOSFET design fails:

12V in this discussion is "nominal" car battery voltage...

One or both mosfet transistors will be damaged in short order in your design because for an instant during each change of the 12V control signal applied to them, from the wiper motor (wire WR_WIPER_M in your drawing), they are both on at the same time. When both are on, they suffer a short circuit path from ground to the battery supply, through the transistors. This is a very brief condition and likely results in delayed failure which might vary from a moment to even an extended run time, but eventual failure is predictable.

In general you designed this circuit based on:
The 12V gate voltage will turn on the N-MOSFET, turn off the p-MOSFET. Likewise 0V gate voltage will turn on the P-MOSFET and turn off the N-MOSFET. While true in steady state conditions there is a third condition that occurs during the "transition" of the gate voltages. When the signal from the wiper module (WR_WIPER_M) switches between 0v and 12V it is not instantaneous, but instead takes a few microseconds to even some milliseconds (millionths to thousands of 1 second) to swing from one voltage to the other.

How long this voltage change actually takes depends on the contact/switch "bounce" inside the wiper unit switch (often mechanical switches open and close more than once as they "wipe" from connected to not-connected and vs.) A second, even bigger factor in how long this takes is capacitance inherent in the FET transistors. It takes time to charge or discharge the gate structure in the transistors, which is essentially a capacitance that slows down how fast each transistor can transition between on and off.

Next, the P and N mosfet devices have a turn on/off voltage "band" if you will, such that the N channel turns on around +4 volts coming up from 0V. The P channel turns on somewhere around 4V below 12V, i.e. about 8V. I'm not as well versed on the P-channel devices and may be just a bit off on their voltage rules, but in general I think I'm summarizing this correctly. So as the gates on both (they are tied together in your circuit) swing between about +4 to +8V they are both pretty much fully turned on. As you can see this is creates a short circuit across the power supply... wire B_GROUND to wire YB_SUBARU_+12V. This is a very short duration event, but happens each time the switch opens or closes at WR_WIPER_M. The duration can be short enough that the fuse cannot react as it doesn't heat up fast enough to melt/blow open. Each transistor however suffers an over-current pulse at that moment. They won't take that abuse for long, and the only prevention to instant failure is the resistance in the wiring circuits and any other resistance in the overall circuit paths, and the ability of the transistors to withstand high current pulses intermittenly. End result is one or both transistors fail either quickly or at some point "down the road".

In practice, using pairs of P and N channel mosfets in switching circuits must always be done with other additional components or at least specific design considerations to prevent the above described momentary short circuits, usually referred to as "crowbar" conditions, where the power supply is "dead-shorted" by the transistor pair during the dynamic moment of switching. Often you can find pictures of integrated circuit "internal designs" that show no components other than the N and P channel parts across a power supply. Those drawings can be misleading because there is more going on in there to prevent the "crowbar" situation than meets the eye.

I've had my share of blown parts in similar circuits (more or less) while designing a couple switching power supplies some years back using these "totem pole" (stacked N and P channel FET) style circuits. Some of the results were quite spectacular, although not necessarily successful. :)

[Hojo]

Okay, what I needed more info to be able to build won’t work , but a relay will. I can’t quite read the spec on the relay and certainly can’t see how to wire it. Please take a more detailed shot at the working relay solution for us Luddites.

[aquillen]

I'll let Rogier cover his relay, since I'm not needing to do that on my car. Just wanted to explain the behavior of the FET based setup. In general you will use the switch inside the wiper unit to connect a relay coil on/off to 12V. The relay then switches on and off. Use a relay with both on and off contact (has a common pole and one open, one closed contact).Then wire those contacts to switch the input to the GY_SUBARU_INT between 0 and 12V. You would want the relay coil to be off when the wiper is in the OFF condition so it doesn't draw power when the wiper is not in use.

[Rogier]

well Art explained way better then I can why my MOSFET switch failed, and in hindsight it makes perfectly sense to me know. For the relay part he basically explained how to wire it up. :-)
I used a general automotive 12V 40amp rated switching relay the pins are standardized. To makes things simpler to wire up the relay.
Ground wire (black) to pin 86 and 87a
GY_SUBARU_INT to pin 30
YB_SUBARU_12V to pin 87
WR_WIPER_M to pin 85.