Phase 2: 5.4 3v VCT controller construction. The chronical continues.

Discussion in 'SN95 4.6L Mustang Tech' started by billfisher, Jan 18, 2006.

  1. ok, the engine is fine. it runs good. time for the next step i have 2 intakes to try, i have made a number of spacers for plenum volume adjustments, if necessary. either the stock plenum with volume increases via spacers or short runner.

    that is the url for downloading the pdf file for the PWM chip i am going to use to construct my VCT controller.

    this chip is under $10. the high power mosfets are $20 each, and feedback dampening diodes $.99 each .misc parts less than $30 for isolation PC board/compartment and misc resistors/rheostadts.

    The DRV102 is a high-side power switch employing a
    pulse-width modulated (PWM) output. Its rugged design is
    optimized for driving electromechanical devices such as
    valves, solenoids, relays, actuators, and positioners. The
    DRV102 is also ideal for driving thermal devices such as
    heaters and lamps. PWM operation conserves power and
    reduces heat rise in the device, resulting in higher reliability.
    In addition, adjustable PWM allows fine control of the
    power delivered to the load. Time from dc output to PWM
    output is externally adjustable.
    On (TTL-Compatible)
    Thermal Shutdown
    Over/Under Current
    Duty Cycle
    2 3
    (+8V to +60V)
    (Gnd electrically
    connected to tab)
    Gnd(1) 4
    The DRV102 can be set to provide a strong
    automatically switching to a “soft” hold mode
    savings. Duty cycle can be controlled by a
    voltage, or digital-to-analog converter for versatility.
    output indicates thermal shutdown and over/
    limit. A wide supply range allows use with
    The DRV102 is available in 7-lead staggered package
    and a 7-lead surface-mount DDPAK
    package. It operates from –55°C to +125°C.

    for those that don't want to download the file.

    basically i am going to construct a high impedance branch fed by the crank position sensor 0-13v analog. the branch is a voltage splitter used for 0-5v ttl on/off control. the chip will be in standby until >3.5v
    i reahed and controller activates preset duty signal. output will be connected to two high power mosfet gates/switches that feed a non-feedback loop/anto feedback loop to the solenoids. switching speed < 25000hz.

    the controller can initialize the solenoid with 100% duty to set the position to full open then drop to preset duty. the controller is sensative to EMF and heat. aluminum isolation inside passenger compartment with feeds to the solenoids. 1 mosfet for each solenoid. duty is adjusted actively by rheostadt.
  2. My brain just exploded.:D

    Keep up the great work.
  3. Make sure to have a beer to steady your hands before soldering. :D
  4. Some of it I understand but the majority of it is like reading a book written in a foreign language to me.

    Is all this (its probably just alot to me, you seem to have an excellent grasp on this stuff) really necessary for the 5.4 swap, or is this VCT controller your making just an improvement?

  5. :Word:
  6. "basically i am going to construct a high impedance branch fed by the crank position sensor 0-13v analog. the branch is a voltage splitter used for 0-5v ttl on/off control. the chip will be in standby until >3.5v
    i reahed and controller activates preset duty signal. output will be connected to two high power mosfet gates/switches that feed a non-feedback loop/anto feedback loop to the solenoids. switching speed < 25000hz."

  7. Have you actually measured the frequency at which Ford PWMs that solenoid? 25kHz is pretty fast for that much inductance. A frequency more like 125Hz might be more appropriate.
  8. I get it, good idea!
  9. frequency is simply a measure of precision. 2200hz is typical for precise wave (oscillation) control. frequenct has no effect on ampacity of solenoid. number of coulumbs stays the same. duty determines total amperes passing through circuit. higher frequency has the appearance of higher current, but it actually has nothing to do with it. the chip maker sets a frequency for all possible applications. further steps are the end circuit engineer's responsibilty. stepping it down is easy. a double JK flip flop counter can halve the freq. but i am going to use it as is.

    there is no need to do this. a machinist can alter the drive sprocket for any timing adjustment. i want 80% of 365 lb-ft 2 1000 rpm. and 350+hp 2 5500-5700. the mainfold changes were done for 5700 rpm max hp. retarding the cams 4-5 degrees at 3800 is the goal. i will let the track decide that for sure.

    these solenoids are run of the mill PWM hydraulic servo emulators. 12v 2 amps max, any freq above 2200hz. low frequenct lowers precision too much.

    remember solenoids travel 1 full stroke in 1/2 of 60hz. 120th sec. low frequency allows the solenoids to travle back to start. the end duty is the same, but precision control is impossible. i need the solenoid to stand in one spot. cam oscillations are probably catastrophic.
  10. There are limits to how fast a solenoid can be run due to inductive reactance of the windings. Page 6 of the datasheet shows the effects of inductance on the current through the solenoid...

    I have a sneaking suspicion that, if driven with too high a frequency (i.e. on-time too low for duty cycles below, say, 90%), the solenoid may act nonlinearly, depending on the load and the inductance. If may be fine at high duty cycles where there's enough time for current to flow but as the duty falls and the ontime drops, the inductive effects on winding current may cause the solenoid to simply "fail" to respond and, instead of linearly operating, it will have a sharp drop off below some as-yet undetermined duty cycle threshold.

    Just saying you cannot ignore the effects of inductance when determining how the frequency and DC range with which to drive a solenoid. A solenoid of this nature may have relatively significant inductance in its windings. Indeed, even mechanical and load inertia need to be accounted for.

    All you can do is try. I was wondering aloud whether you'd determined what Ford chose the drive the thing at versus some arbitrary frequency driven not by application specific requirements but by what's easily available in silicon.

    Should be no harm in trying.

    There's a balance between switching too slowly and inviting mechanical effects like you cite and trying to do the impossible in switching it too fast against the reality of inductive reactance. Sure, 100Hz might be too low just as 25kHz might be too fast. Perhaps some more research (like opening the hood of an F150 and scoping) might be beneficial :shrug:

    Or you can just try it and see. :)

    Regardless, it's good to see someone venturing out like this :hail2:
  11. the effects in self inductance is controlled dy the two diodes.

    in reading the patent papers files by ford, they failed to mention driving frequencies. the lowest freq typical PWM solenoids is 2200hz. i was going to build my own JK flip flop driver, but 1 percent precision requires too many chips.

    ok i see what you are saying. time periods for inductive resistance to current flow in dc applications. i forgot about that. i guess i should take one out and measure how many henries it is.

    ((2xPIxfxL) = inductive reactance is AC)

    but 5 time periods inductive resistance = 0 ohms. it is L dependent. but at time period #1 resistance is too high to let current pass. if switching freq is too high enough time periods for field collapse and current flow won't happen.

    i guess i had better look into that closer. you know anything about this, or you just read closely? good catch.
  12. The flyback diode in parallel with the solenoid will protect the output stage of the driver from the inductive "ignition coil" effect of the collapsing magnetic field impinging on the inductor coil. Reading the datasheet, it appears the series diode is there to help ensure the internal transistor is truly off during the flyback time.

    This would be a nifty project for a microcontroller like the M9S08 series from Motorola/Freescale. Its TPM modules have highly programmable PWM capability. Unfortunately, you'd also need a proper IDE and tools (e.g. BDM debugger) to do it.

    If you have an inductance meter you're half way there. If the inductance of the coil is comparable to the numbers in the datasheet, you'll have pretty much eliminated electrical reasons why the solenoid couldn't be run at that frequency.

    Yep. Like I said, the coil may function nonlinearly: predictable for high DC times when conduction is long enough to get reasonable current in the coil but as DC drops and conduction time falls, the solenoid may not act predictably.

    These are only "maybes" might just be fine.

    I'm a medical electronics designer and deal with high-speed switching circuits. My expertise is not with solenoids per se although the fundamentals are similar to the challenges I face: high speed PWM switching of power MOSFETs where gate charge and capacitance dominate the ability of the FET to actually turn fully on and fully off.

    For one of our circuits we want to PWM a MOSFET matrix with DCs low enough that we see 4uS (0.000004 second) pulse on the FET gates. Gate charge and capacitance dominate with times on this scale, moreso because we're using power FETs and asking them to conduct 3A or more for this duration. Although we're dealing with capacitance and you're working with inductance, the potential problems are similar in nature.

    It just struck me when reading this that a high frequency drive of a comparatively large solenoid might prove tricky. A 50% duty cycle at 25kHz gives a conduction time of just 20uS...not alot of time to get current moving in a large inductance coil. Of course I might be completely wrong and my guesstimates of the solenoid inductance, inertia and linearity and of load effects might be out of the ballpark. Just thought it was worth noting.
  13. Either you guys are geeks or electronic engineers.

    Good job.
  14. rock out dude. it is good to talk to an educated member(electronics). no offense.

    disclaimer: i mean no offense to anyone past ot present concerning their education. you are all intelligent fellas. you drive fords so you have a doctorate in "common sense". i just mean electronics guy.

    now with the possible insults aside.

    i used the characterisation "i forgot that" earlier. what i mean to say is the focus of this exercise is precision. one element is all it takes for failure. if the inductance is high enough to hold DC current near zero for a time constant too high for the frequency, i am screwed.

    one element is all it takes. so with that said i will have to search TI for a lower frequency PWM controller. thanks again.
  15. I didnt understand much of anything you two said :shrug:. But im not an electrocs guy...lots of snap/common sence, but not that in depth. I manufacture silicon wafers @ freesacle semiconductor, so i should know a little about electronics...but i dont know much haha. Maybe one day I will come back to this thread and understand it all LOL.
  16. Ditto! :hail2:

    I don't think there's any harm in prototyping what you've got and trying it. You could scope the crank and cam sensor signals and vary the PWM at the frequency you've got to see if the solenoid is responding as expected in situ. If it's not, you could modify just the frequency & driver section until you find a frequency that works. You could take that knowledge and build a final "hardened" version that would live under the hood and do the job.
  17. that's exactly what im going to do. we have a portable scope at work.

    if it doesn't operate properly, or doesn't have the required precision, then i might have to go with the original plan. add a clock,maybe 6x 4 jk fl/fl chips and an 'or' gate and of course a driver and see how that works. if the cam phases between -3 to -6 degrees, that can really be good enough. the problems arise with oil flow. diluted oil from gasoline blowby or sludge build up WILL change phasing. i am not willing at this point to feedback correct it. i do have in my possesion a 7mhz 8086 contoller with a comparator and PPi, but i have to admit, i am just trying to get out of this as easily as possible.
  18. "I need a nuclear reaction to create the 1.21 Jigawatts to power the flux capacitor!!"


    Maybe I should have gone to college.. :p

  19. What are youi missing out on right now by not using this system ?

    How does you car run ?