Brett, great that you're working on this. I haven't had time to follow
closely and didn't realize the Arduino has a built-in PWM function. If it
lacks the resolution needed, how about using the Arduino to regulate a
dedicated PWM controller?

When the ISCU quit working on my Syncro I disconnected the ISV and drove
that way for a couple of years. The only time I missed the ISV was on cold
starts, say below 40-50F. Granted, I have no AC or PS. After a cold start
I'd have to hold the throttle open slightly to keep it idling, until it had
warmed for one to two minutes.

I added a 'manual' PWM controller so the engine would idle while I got out
to scrape the windshield. In stock form the PWM control was very sensitive,
but after swapping a potentiometer matched to the range I needed it works
quite well as a form of manual idle speed adjustment. Once the engine is
warmed somewhat I shut it off.

As I'm sure you're aware, dedicated PWM boards are super cheap. A few years
back I think I paid $10 for mine. Today you could likely find a suitable
controller for $2.

Thanks again for undertaking this, a computer controlled (cheap) idle would
be a nice thing to have again.

Angus

On Tue, Nov 26, 2013 at 8:10 AM, Brett Ne <brettn777@gmail.com> wrote:

> We have some data to work with!  One of our members has generously put in
> the time and effort to gather most of the readings we need.  Here is what
> we have so far:
>
>     *Duty Cycle Under Various Engine Conditions*
>
>
>
>
>
>
>  *Condition* *Duty Cycle* *Peak Voltage*
>   Engine Cold, Key On, Engine Off 34.0% 9
>  Engine Warm, Key On, Engine Off 29.4% 9 @ 157 mA  Engine Cold & Starter
> cranking (45 F) 55.0% 6
>  Engine Cold & Idling 25.0% 10.5
>  Engine Warm & Idling 20.5% 10.5
>  Engine Warm & Running 2000 RPM
>
>
>  Engine Warm, WOT Signal
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>  *Duty Cycle Under Various Load Conditions*
>
>
>
>
>
>
>  *Load* *Engine Cold* *Engine Warm*
>  None 23.0% 20.5%
>  AT in Gear 25.5% 23.0%
>  PS 27.0% 25.0%
>  AT in Gear & PS 30.0% 27.0%
>
>
>
>
>
>
>
>
>  *Miscellaneous:*
>
>
>  PWM Frequency constant at 148.2 Hz
>
>
>
>
>  Yellow Wire (11/ST1) PWM Power to Idle Stabilizer Valve
>
>
>
>
>  White Wire (4/ST2) Ground Connection? (Needs to be confirmed)
>
>
>
>
>
>
>  From this we can conclude:
> **The idle air valve is very sensitive.  A 7% increase in duty cycle is
> able to overcome the combined loads of AT in gear and PS at full pressure.
> The built-in PWM output on the Arduino only has 256 settings, which would
> give use increments of 0.4% in the duty cycle output.  This would work, but
> I think that there would be a noticeable unevenness in idle speed.  Not a
> big deal, it just means that we will need to use a timer interrupt to gain
> fine control over the duty cycle(the percentage of "on" time of the pulse)
> of the PWM.  It's just not as fun and easy as using the built-in PWM
> functions.  I'll explain interrupts in more detail when we get to the
> programming stage, but it basically is just a way to grab the computer's
> attention and make it suspend its current operations while it attends to a
> time-critical event, like a new hall sensor pulse coming in.
>
> **Most of the time, the duty cycle is pretty low, around 25%, but that
> changes during starting.  While the engine is cranking, the duty cycle
> jumps to around 55%.  When I first saw this figure I thought, "Wow, it's
> really increasing the airflow a lot while the engine is cranking."  But I
> don't think that is true because the current draw from the starter motor
> drops the available battery voltage considerably.  The stock ICU output
> voltage drops from 9v with engine not running to 6v with starter cranking.
> I think the dramatic increase in duty cycle is not for increasing airflow,
> but rather to compensate for the expected voltage drop during cranking.
>
> **The ICU apparently makes no attempt to send a fixed voltage to the idle
> air valve.  With engine off, it sends 9v.  With engine running, it sends
> 10.5v, which matches the increase in voltage from the alternator.  With
> engine cranking, the output drops to 6v, reflecting the battery voltage
> drop under the heavy load of the starter.  For running conditions, we don't
> need to worry about the output voltage level because it will be
> automatically compensated for by the engine speed feedback.  But for
> starting, we could have the Arduino measure the voltage coming from the
> battery and calculate the appropriate duty cycle change to have better
> control over the airflow.
>
> **The idle air valve draws 157 mA when supplied with a 9v 29% duty cycle.
> This translates to a 535 mA draw @ 100% & 9v or 4.8 W of power.  The final
> drive transistor in our circuit should be chosen to handle 1A of current to
> assure durability.  Lots of options here.
>
>
>
> What we're still looking for:
>
> What happens above idle speed?  Does the idle air flow shut down, or stay
> at some predetermined level?
> What happens at Wide Open Throttle?
> What does the Hall sensor output signal look like?  I haven't found
> anything definitive on the internet.  There are hints that it is a square
> wave and that the peak pulses are up near battery voltage.  I would like to
> know peak and base voltages and confirm that it's a clean square wave.
> Also, has anyone taken apart an idle air valve?  Is it just a motor
> operating against a spring?
>
>
> Brett
>
>
>
>
> On Wed, Nov 20, 2013 at 1:41 PM, Brett Ne <brettn777@gmail.com> wrote:
>
> > Before we can go any further with this project, we need those
> > measurements.  I'm not interested in investing time, materials, and money
> > into a project if it's based on assumptions and guesswork.  My Vanagon
> > can't be used because it doesn't have the Digifant injection system.
> >
> > So, we need either:
> >
> > Someone with an oscilloscope or a digital multimeter w/pwm functions who
> > has access to a Vanagon with Digifant FI that has a healthy ICU as well
> as
> > power steering and AC.  Bonus points if it's an automatic.
> >
> > Or,
> >
> > Someone in the Portland Metro area with a van that fits the above
> > description who would be willing to work with me for a couple of hours.
>  I
> > do have an oscilloscope and could take the measurements.
> >
> >
> >
> >
> >
> >
> > On Wed, Nov 20, 2013 at 1:23 PM, Brett Ne <brettn777@gmail.com> wrote:
> >
> >>
> >>
> >>
> >>
> >>
> >>> Assuming a properly functioning engine and IS system, if one measured
> >>> current draw at the ISV, loads on and off, at idle, cruising speed and
> >>> WOT, could they get a really good idea of when the ISV is used? (valve
> >>> changes position). Maybe a graphing meter would be useful in this
> >>> regard?
> >>
> >>
> >> Current draw or voltage measurements would be kinda okay.  They should
> be
> >> proportional to the duty cycle, but again, knowing the duty cycle tells
> you
> >> exactly what the control unit is doing and is much preferred over
> indirect
> >> measurements.
> >>
> >>
> >>
> >>> My guess is that at idle, no loads (PS, AC) the valve is
> >>> closed. Or.....
> >>>
> >>
> >> ...or partially open.  We can all guess until we're blue in the face,
> but
> >> if we want to create a re-engineered control unit that solidly replaces
> the
> >> old one, we have to establish the behaviors of the old one first
> >>
> >>
> >>>
> >>> Isn't the ICU basically taken out of the picture any time the throttle
> >>> is just off idle or at WOT?  (throttle enrichment switch activated or
> >>> not).
> >>>
> >>
> >> Maybe, maybe not.  Maybe the idle throttle opens fully under full
> >> throttle to help the engine breathe better.  I just don't know.
> >>
> >>
> >>>
> >>> Did you find the correct "Vanagon" Digifant pro training manual
> >>> online? If not, can send more images. There may be some useful in
> >>> there for your needs.
> >>>
> >>
> >> No, I haven't, though to be honest, I haven't been looking.  Typically
> >> training or repair manuals may give enough info to identify a faulty
> part
> >> but don't give all the details on how that part interacts with all the
> >> other parts of the system.
> >>
> >>
> >>
> >> --
> >> Brett in Portland, OR
> >> "Albert" '82 VanaFox I4 Riviera
> >>
> >
> >
> >
> > --
> > Brett in Portland, OR
> > "Albert" '82 VanaFox I4 Riviera
> >
>
>
>
> --
> Brett in Portland, OR
> "Albert" '82 VanaFox I4 Riviera
>