Guaging Interest in a ACC project - Page 2 - PeachParts Mercedes-Benz Forum

bd1308 · #16 10-26-2012, 10:36 PM

Well I'm quite sure that would require something with a bit more pep than a 16Mhz arduino. That would be AWESOME plan for a 68000 or a 40Mhz Z80 though.

Scratch that, there are 72Mhz 32-bit arduino microcontrollers out now, I forgot

NC-Diesel · #17 10-26-2012, 11:41 PM

You could build it out of one of those lego mindstorms kits. lol. Just being silly here. I think they built a thing out of legos that solves rubiks cubes...

my123ca · #18 10-27-2012, 12:00 AM

Quote:

Originally Posted by bd1308

Right now there are tiny solenoids that control the flow of vacuum between the pods and the rest of the system (its right near the fiber optic spider web looking thing for lighting).

Replacing the pods with big solenoids is trivial, however a solenoid is either on or off...A servo can be rotated in any degree (up to 360) in software using PWM. This way we can control things with a fine-grained level of detail all in software, and enable things like ECO mode and AC + recirc in the summer (something that I'm not convinced happens with the current system although i'm not a ACC II genius).

The tiny solenoids are called switchover valves. They just open and close. They dont control the flow of vacuum.

KarTek · #19 10-27-2012, 06:29 AM

The newer cars have vacuum transducers that receive a PCM signal (I assume) from the ECU and actually vary the amount of vacuum on a line.

I don't know if these would be useful in this situation. Switchover is all that's probably required to simply open and close flaps to control air flow.

Doktor Bert · #20 10-27-2012, 09:18 AM

Quote:

Originally Posted by KarTek

If you're going to the trouble of automating the ALDA, you might as well use an EDC pump and let the processor drive it completely. That way you have absolute control over all fueling.

So we can all walk everywhere we go, like what has happened to heavy-duty trucks....

bd1308 · #21 10-28-2012, 11:00 AM

Quote:

Originally Posted by my123ca

The tiny solenoids are called switchover valves. They just open and close. They dont control the flow of vacuum.

I'm not sure I understand how they dont control vacuum....

Mercedes Benz 300SD ACC Problem

my123ca · #22 10-28-2012, 01:12 PM

If you want it 10% open or 90% open, it wont happen in the ACCII.

cook · #23 11-01-2012, 08:43 PM

I think this is a great idea, if I understand it correctly.

Using an 84 300D as the basis for my question...Are you going to replace the pushbutton unit in the dash, and the two controller boxes located behind the glove box, and the vacuum pods?

For my car, the vacuum pods are working fine, but the electronics are the problem. I think that is rather typical, so maybe start with a system to replace the electronics first. Then as a second phase add on the ability to replace the vacuum pods/flaps. You could sell them as 2 seperate products.

I think it is a great idea.

Silber Adler · #24 11-01-2012, 09:00 PM

Great that you are getting your AC working again.

One of the best attractions of my Mercedes for me was to get another car that works without electronics.

EMP proof.

cullennewsom · #25 11-01-2012, 09:24 PM

Quote:

Originally Posted by bd1308

First: I know this is just on the edge of insane and complex, but this is definitely something that I am interested in doing.

Okay, as we all know (or will know) the ACC system works great when it does work. As thing age in our cars, expensive parts (like the vacuum pods) break/leak air and cause problems with other vacuum systems in our cars. With the advancements in modern technology, things like controlling a seemingly complex ACC system is something that can be implemented in something like $40-50 worth of parts.

What I am attempting to do (as soon as I find a worthy parts car anywhere near me) is augmenting/shoehorning/replacing the ACC system and its vaccum-related solenoids with electronic servo motors and PWM to control the flaps for each component, control the monovalve, and finally put some safety circuits in between the ACC and the aux coolant pump.

The only reason this makes financial sense is I could retrofit servos and a good AVR/Arduino microcontroller for the cost of 1-2 vacuum pods and not worry about this part of the car for a while. Eventually things like IP timing can be monitored (with the right retrofitting of sensors, etc).

What do you all think?

PS=I know i'm crazy.

Count me in.

cullennewsom · #26 11-01-2012, 09:33 PM

Quote:

Originally Posted by NC-Diesel

You could build it out of one of those lego mindstorms kits. lol. Just being silly here. I think they built a thing out of legos that solves rubiks cubes...

CubeStormer II - YouTube

KAdams4458 · #27 11-02-2012, 04:07 PM

Actually, I do wonder which ACC system you've been subjected to. For me, it's the old kind with the evil servo. Honestly, the control units in the dash on these seem pretty stout, but those servos are a weak link. The spare core I have is from a monovalve equipped car. (Oddly enough, the monovalve is fine, but the control panel is falling apart.)

I'm not opposed to putting it on a bus, but I'm still guessing that a used one sourced locally might still cost less than that bus ticket. LOL. They're actually not too hard to remove.

gear-head · #28 11-02-2012, 05:30 PM

I'm in! Look forward to hearing when the development is done and you have some parts to sell!

Doktor Bert · #29 11-02-2012, 06:51 PM

Quote:

Originally Posted by KAdams4458

Actually, I do wonder which ACC system you've been subjected to. For me, it's the old kind with the evil servo...

You must tame the evil servo....

franklynb · #30 11-03-2012, 12:01 PM

bd1308 wrote:

Quote:

Originally Posted by bd1308

I'm not sure I understand how they dont control vacuum....

Mercedes Benz 300SD ACC Problem

You need a factory shop manual for ACC2/3. Its around 300 pages and covers ALL the
details. And since you'll need to read/refer five pages simultaneously, prepare to spend
the ~.$07/page to print it/buy it on ebay.

I've torn apart around (10) ACCs of various vintages, for cleaning, repair, and
just to understand 'what's going on in there', and starting with a similar
"how hard can it be?" attitude, and a couple of ancient degrees in curriculae
now called 'mechatronic engineering'.

A few observations/things to consider:

1) The pods actuate differing lever travels, depending upon door application;
the pod design has enough compliance in the actuator rod to allow severe
ploar mis-alignment over a relatively long stroke, as the door lever swings
through its polar movement.

Rubber dashpots are also very forgiving of "bump stop" homing -- unlike
a rigid electric/hydraulic/air servo.

Hence, "actuators" may not be as simple/cheap as you may suspect. Stepper motor
actuators, which I don't consider to be "a real servo" would need to be
joined through some sort of spherical axis, meaning the current/torque profile varies over the travel.

it "a real servo", agreed? How will you encode the existing lever travel?>

2) There is a PID "loop" of sorts that handshakes between the 5 speeds of the blower
<3 in ACC2 and manual heaters>, the monovalve which tries to modulate
the amount of heated water to the exchanger,
and the cycling of doors -- implementing partial defrost, upper/lower/fresh air blending,
etc. and interior temp sensing through a "profiled" air passageway insulated in foam,
trying to protect it from conductive swings.

The programming built into the typical ACC3 controller isn't as simple as one might hope.

3) Any of the ACC2/3's can be retroverted to the manual control two-cable setup from a 240D,
eliminating all the pain, for about the target $50 and about the same install
labor. its the same/very similar box and doors with a set of cables/vacuum dash switch to actuate the
doors. The monovalve can be replaced by a two zone water valve, also cabled.

So where's the advantage of a 'process controller'? I'd argue "not simpler" given the
blending requirements of the basis ACC3 setup; 5 blower speeds, hot water/A/C
"bi-level" defrost, and somewhat sophisticated air door switching.

Some aggressive boneyard hunting will deliver the control parts from a 240D for the
$50 .... but yeah, an arduino based PID controller could be fun! But then I'd
consider it just as much fun to re-interface the bowden cables with campagnolo/
shimano/magura levers and hoods ... or deliver a set of simple "lever position"
readouts on my phone to at least know when the darned footwell pod had stuck
closed -- again!

Ping me off-list for more discussion. I have several of every ACC2/3 part ... and several o'scopes.