[noisymime]

So support has been in place for a while now for a single PWM output for idle control. This works well for virtually all the 2 wire PWM idle valves out there and OK for the 3-wire valves if a high current resistor is used to tie the other line to ground. One of the things on my checklist though had been getting these working the way they're supposed to with both lines being driven by PWM signals. I'd never really looked into these though and wasn't sure what the opposing PWM signals were meant to look like.

It struck me that potentially the 2nd signal is simply the opposing duty cycle of the first signal. Eg if one signal is 70% duty cycle, the other is 30%.

Does anyone have any experience with driving these valves that they can confirm whether the above is correct? If this is the case (And I've yet to bench test it, but hopefully soon), then it is trivial to code in.

[edc_atl]

We use them all of the time since they are very cheap ($8.00) at salvage yards. It is a simple reversible motor activated valve. The center pin on the valve receives battery voltage. If you connect your positive lead to the center pin and tap your outer terminals individually to ground, your valve should move one way and then the other. Valve coils should measure about 12-15 ohms from terminal 2 to each of the outer terminals.

On some of them there is thumb wheel used to set the base initial travel or idle. In any event, you will need to add twoTIP122 transistors to effectively switch terminals 1 and 3 to ground via ECU control. The second way is to tie pin 3 to ground, via a 35 Ohm/50W resistor which acts as a spring to shut the valve off, when the ECU is not sending the advance command to pin 1. In any event, you can control this vale effectively with the FIDLE output only if you tie pin 3 to ground as described.

Also, this valve will require tuning since the advance pulse width is a little weird.

[androidcho]

So here is some info from Autodata. This is BMW M50B20 engine with Bosch Motronic 3.1, it uses 3-pin IAC valve. Hope this helps.

[noisymime]

Juan - Things are working reasonably well already with the resistor method, but it has some drawbacks (Slow response speed, potential for the vlve to not fully open/close). What I'm trying to do now is figure out is if I use all 3 wires (ie 2 PWM signals), how should those 2 signals relate to one another.

Eg if I want the valve 70% open, should I send a 70% duty cycle on one wire and 30% on the other? Or should there be more 'pull in each direction with something like 80% on one wire and 40% on the other or will this lead to damage, buzzing etc?

I think I probably just need to do some bench testing to figure it out, but thought I'd see if anyone had played with them before.

[noisymime]

So here is some info from Autodata. This is BMW M50B20 engine with Bosch Motronic 3.1, it uses 3-pin IAC valve. Hope this helps.

This is interesting and kind of confirms what I was thinking. The 'typical' idle values give are 70% and 26% for the 2 lines, so the theory that they need to be equally opposed (ie adding to 100% duty cycle) seems like it might be right (Roughly anyway).

Thanks for posting this!

[PSIG]

FWIW, some years ago, a 'squirter used a 555 timer and transistor (probably a TIP12x type) to set a PWM counterforce instead of cooking a big resistor with constant power. It apparently worked well as the 'closing spring' with better action as well. No other helpful info. I have used the 2-wire Bosch type, that default to a closed-position fixed opening (limp home), then with PWM control move to fully closed and further to a gradually increasing opening. Being spring-loaded, they have a safe 'homing' action, and 1-wire control. Would the counterforce method on a 3-wire type permit a safe auto-closure of some type to prevent a runaway idle if signal was lost? Those Bosch valves can pass a lot of air.

David

[edc_atl]

Juan - Things are working reasonably well already with the resistor method, but it has some drawbacks (Slow response speed, potential for the vlve to not fully open/close). What I'm trying to do now is figure out is if I use all 3 wires (ie 2 PWM signals), how should those 2 signals relate to one another.

Eg if I want the valve 70% open, should I send a 70% duty cycle on one wire and 30% on the other? Or should there be more 'pull in each direction with something like 80% on one wire and 40% on the other or will this lead to damage, buzzing etc?

I think I probably just need to do some bench testing to figure it out, but thought I'd see if anyone had played with them before.

Problem is that you have to overdrive the valve when you use the resistor method, otherwise it will not fully close because it will never reach 0V. If I remember correctly, you should measure about 8Vdc across the resistor or a 67% constant close signal. But you are right on the money with the duty cycle numbers if you drive both pins directly.

A better method that can be looked at is using a derivative of the schematic below, but changing the MOSFETS to something like an IRF640. We are in the process of developing a plug in board for a customer, that covers all of the fast idle valve types available. Should be done about the first few weeks of the year.

[androidcho]

This is interesting and kind of confirms what I was thinking. The 'typical' idle values give are 70% and 26% for the 2 lines, so the theory that they need to be equally opposed (ie adding to 100% duty cycle) seems like it might be right (Roughly anyway).

Thanks for posting this!

Yes, it seems to work like you said. Sorry, I don't have oscilloscope so I can't measure what signals are going to the valve. I have access to engine with this type of valves.

[edc_atl]

Since these IAC valves are very common and are becoming mostly a standard for a few people, we created a small board that replaces the DRV8825. This new unit plugs into the DRV8825 socket and provides the needed three wire support. One pin will require re-mapping or a selector jumper added to the board, but otherwise it is just a plug and play device.

I will have these added to the store and available for shipping next week.

[androidcho]

Good news, but I think we don't have a software support right now. Another question is what is the current consumption of this valves?