[Supras] Boost controller

Wed Jul 22 19:46:04 CDT 2009

Way out of my pay grade to answer but the damned things DO work ;-)

And they work very well.

----- Original Message ----- 
From: <berniek at technicaldevelop.com>
To: <cthommes at adelphia.net>
Cc: "Khalid Almufti" <kalmufti at yahoo.com>; <supras at supras.com>
Sent: Wednesday, July 22, 2009 7:42 PM
Subject: Re: [Supras] Boost controller

> Craig:
>
>
> As always, thanks for the reply.  Guess I must have a bit of
> misunderstating about one point:  Please understand that what follows
> here is not a challenge.  Rather, it is an expression of the apparent
> way in which electronic controllers must operate in order to prevent
> spiking of consequence.
>
>
> We know that static coefficient of friction is higher than dynamic.
> Hence the "stiction" in the actuator will prevent its movement until
> enough diaphragm force is reached to break it loose to move.  Then,
> lower force dynamic friction coefficient takes over (probably at the
> grommet where the actuator rod penetrates the housing).  Once that
> happens, the hysteresis loop is entered (same as in magnetic B-H curves)
> and the wastegate opens suddenly.  If you set a B&S controller for
> sustained desired boost, it will spike before the actuator rod starts to
> move due to static friction coefficient.  If you stop a couple of tenths
> of a PSI less than the "hysteresis pressure" (electronic or mechanical),
> the waste gate will not open when it should, again resulting in a spike.
>
>
> It seems that the electronic controller needs to let pressure off the
> diaphragm VERY quickly in order to avoid spiking once the boost reaches
> a spike INFLECTION slope or point, which with good enough frequency
> response can probably be pretty quick.  Use of PWM is a tool  to change
> the effective flow coefficient (Cv) of the valve.  It would appear that
> the real advantage of electronic controllers must lie in the bandwidth
> and tuning.  Rate of change of PWM allowed or provided by the EBC would
> be important here.  In PID process control, the derivative term does
> exactly this.  It is often looked upon as an anticipatory term in the
> algorithm.
>
>
> One way of combating "stiction" is dither, usually used in final control
> elements like proportional electrohydraulic servovalves.  Dither is used
> at a high enough frequency that it does not affect performance within
> the closed loop bandwidth.  The amount of spool valve movement due to
> dither in a servovalve is very small, but that is all it takes to avoid
> effects of both static and dynamic friction characteristics.  This does
> not appear practical in wastegate applications because of a couple of
> factors:  The fluid is air (compressible, a cushion), the actuator
> connection is very small, and the actuator rod takes a long time to
> settle to a given  position.  The actuator appears to be the lowest
> frequency pole if we were to consider a properly operating control loop
> Bode plot.
>
>
> I am beginning to think that my actuator is the problem with 5 PSI of
> hysteresis.  Conversely I can understand how to design an EBC to fight
> this and limit spiking to perhaps 1 PSI or less, but it would take two
> PWM'ed valves (one to pressurize, one to vent), a pressure transducer,
> and fast acting circuitry to vent in a programmed way once the desired
> level of boost is reached.  If this is the way the HKS or other
> controllers actually work, then I'm sold.  So much the better with uP or
> uC based adaptive learning.  Any comments would be appreciated.  Craig?
> Khalid?  Nick?  Jon?  Anyone else in the list with control theory and/or
> electronics experience?
>
>
> I'll apologize for a lot of material here that comes from specialized
> knowledge in my field.  However, others in the group do share this
> involvement.
>
>
> BernieK
>
>
> cthommes at adelphia.net wrote:
>> Hi Bernie,
>> I refrained from this discussion until now out of respect for your 
>> original wish to not be convinced of the worth of an EBC.  I had nothing 
>> else to contribute.  :)
>>
>> You've just described the beauty of the electronic control.  While 
>> monitoring boost pressure, it will begin modulating the boost control 
>> solenoid *near* the desired boost level to avoid spiking.  Nearly all 
>> EBC's worth their weight in plastic have a "learn" function that allows 
>> the controller to adapt to the boost characteristics of the car it's 
>> installed on.
>>
>> I'm not sure what sort of valve/setup you're using with your MBC, but are 
>> you completely cutting off the vacuum/boost signal to the actuator until 
>> your desired boost level is reached?  If so, that would definitely 
>> produce a substantial lag in actuation, and thus a boost spike and 
>> significant hysteresis around your desired boost level.  EBC's do not 
>> work this way.  I.e. the solenoid duty doesn't switch from 0% to say 80% 
>> immediately when the desired boost level is reached, then back to 0 when 
>> boost falls below the setting, etc.  It's the PWM allows it much finer 
>> control over boost level.  Some of the fancier ones allow you to monitor 
>> and even manually adjust the solenoid duty, but that's just useless fluff 
>> as far as I'm concerned.
>>
>> I prefer an HKS EVC-IV myself, just because it's the one I'm used to.  I 
>> tried the Apex-i AVC-whateveritscalled but the gigantic Tokyo-at-Night 
>> display crap was a big turnoff for me.  The EVC-IV can usually be found 
>> for around $100 used on e-bay.  I second the previous caution of used 
>> units, however, as I have bought one that was DOA as well.
>>
>> The bottom line is, yes you can control boost with an MBC.  But the best 
>> MBC will not provide the same smoothness, repeatability, or control 
>> reliability as the worst EBC.  I know that my desired boost will be hit 
>> spike-free and maintained through redline every time, without exception. 
>> The added side-effect is I can easily adjust that level on the fly from 
>> the cockpit with the simple turn of a knob.
>>
>> -Craig
>>
>> ---- "berniek at technicaldevelop.com" <berniek at technicaldevelop.com> wrote:
>>
>>> Khalid:
>>>
>>>
>>> I'm curious as to how the PWM'ed solenoid valve affects boost. There
>>> must be two, to feed the wastegate, and then to very quickly partially
>>> vent it to keep spiking down.  In my case, the spike emanates from
>>> stickiness or hysteresis in the actuator, which starts to move 5 PSI
>>> later than the boost controller setpoint.  Any information would be
>>> appreciated.  Manufacurers tend to not go into a lot of detail.  But
>>> detail is the inclusion which will convince me that the controller is
>>> worth something.
>>>
>>>
>>> Any information, specific to a controller or general would be
>>> appreciated.  Thanks.
>>>
>>>
>>> Bernie
>>>
>>>
>>> Khalid Almufti wrote:
>>>
>>>> Bernie,
>>>>      I bought a used EBC to replace my MBC last year.  I will never go 
>>>> back to MBC.  The difference is huge.  However, I'd like to caution you 
>>>> in case you run into a similar situation as I did with buying from 
>>>> Ebay.  Basically the guy who sold the EBC (Greddy Profec B) claimed it 
>>>> was working "fine" and "almost like new".  When I installed it and did 
>>>> a first test it was obvious that it was doing nothing!  Everything in 
>>>> terms of bells and whistles was working, but the solenoid was not 
>>>> switching at all.  Since I have a good background in electronics, I was 
>>>> able to trouble shoot it and it turned out to be a blown Darlington 
>>>> power transistor.  Luckily I was able to order an exact replacement, 
>>>> plus I had access to surface mount soldering and desoldering equipment 
>>>> at work.  I was also fortunate that it was not a bad programmable 
>>>> component, like the uC, which would have required a much more 
>>>> complicated repair (most likely I would've ended up returning it to the
>>>>  seller and contesting the sale with eBay).
>>>>
>>>> Additionally, the solenoid had the wrong NPT fittings on it.  I think 
>>>> the seller couldn't find the right fitting and just improvised!
>>>>
>>>> I am confident that you shouldn't have a problem to deal with a similar 
>>>> situation, however, I felt compelled to share my experience with you 
>>>> just in case you end up with an "in good condition" EBC that may not 
>>>> work right out of the the box.
>>>>
>>>> Regards,
>>>> /Khalid, 90T
>>>>
>>>> (balance snipped)
>>>>
>
>
>
> _______________________________________________
> Supras mailing list
> Supras at supras.com
> http://supras.com/mailman/listinfo/supras_supras.com

--------------------------------------------------------------------------------

No virus found in this incoming message.
Checked by AVG - www.avg.com
Version: 8.5.392 / Virus Database: 270.13.23/2254 - Release Date: 07/22/09 
05:59:00