Idle PWM - rattling sound?

[mattias]

Usualy only acceptable value is 0. signal isn't true pwming, but some kind of pseudo pwm - irregular and with interrupts.

Any diode should do the job.

May be some capacitor can help to mask rattling?

Settings for IAC speed  are between 0 and 15.

0 = ~ 250 Hz
15 = ~50 Hz

A high frequency makes the solenoid lazy, reacts slowly to small changes. Too small frequency makes it oscillate and might hit the end positions (rattle noise).

I use 6 which is about 100 Hz and works very well with dual coil ICV on my BMW E30 (1980s 3-series). It is a bit quieter than the stock frequency which is lower, easily heard with ignition on, engine off. It is a 3 pin deal, with  two solenoids working against each other and uses two outputs on the ECU (one inverted) and is probably designed this way to prevent dependancy on battery voltage.

Without diodes to control solenoid flyback voltage you won't get anywhere in adjusting the idle.  I use the diodes that come with the rescue kits, 1N4007 which can handle 1A. Not sure if it's the perfect one for the job but a Schottky  diode  like SR806 that can handle 8A and is faster than the 1N4007.

• Anode to switched point of the solenoid (channel output, EC18-pin..)
• Cathode (marked with a line) to the other pad of the solenoid: the same where you connect the positive +12V (switched and fused) power

[GintsK]

Interesting.
On my table I see: it is not PWMing. And 250Hz is not 250Hz.

I will try to explain what I see on the scope.
250Hz is 250Hz only at 50% duty cycle.
Any other average duty cycle is achieved with combinations of positive and negative pulses of exactly same duration as 50% pulse.
Lets say 80% is one positive and four negative pulses (in fact one 4 times longer negative). But frequency is 2.5 times lower than at 50/50, because polarity changes only every 5th pulse.

80% is simple. But something like 82% means time to time more than  4 pulses at once will be together (every 5th period IMO). And this is a cause of rattling. such pulse can be long enough for full IACV opening and clack.

10% and 90% gives 5times lower frequency.
5% and 95% gives even 10 times lower frequency. (only 25Hz at 0-setting)

Closer to 0% or 100% duty cycle means more rattling and lower frequency.

Can some hardware specialist clarify why this pseudo-PWMing strategy is used?

Gints

[SamSpade]

Basically it's an incorrect implementation of PWM in the VEMS firmware.  "IAC actuator speed/pwm freq" (config iac_speed) is correct only when the solenoid is at 50%/50% (iac_max_steps = 128).

As you've observed, the frequency changes as the solenoid moves away from the 50/50 duty cycle since the pulse length is fixed.  It's clearly not an appropriate approach for an electromechanical device like the ICV which operates reliably only within a specific PWM frequency range (probably 80-300 Hz).

Another problem I've observed is the intermittent spurious pulse lengths on the first channel that causes rattling even at iac_speed=6 (100 Hz) and 50/50 duty cycle.


Related thread here: http://195.159.109.134/vemsuk/forum/index.php/topic,917

No hardware can fix the issue.  The firmware IAC routines have to be re-implemented.