Is it just because the 2554 spools so fast? it seems odd to me that when I look at my log it shows the duty cycle going to 100 right around the time I hit boost, and then not even attempting to drop off as boost climbs.

Post your datalog

Few Questions
 

I've read this through several times and I've a few questions I can't answer from the thread:

Why would MS or the EBC completely stopping working risk over boost? Doesn't it then just go back to mechanical wastgate?
I can understand the danger if the electronics aren't working correctly but that would be a problem whichever way things are set up wouldn't it ?

Having it set in closed loop, is that better? Does it allow you to say ok my max KPa is 190 and whatever boost that reaches it then holds it there?

I've been running open loop for a year now and with the 100% in my boost table it's running wastgate only. It's taken a while to work out the table slowly dropping the numbers but it's getting there. Biggest challenge I found was how each gear gives a diff PSI and therefore KPa at the same revs. I was aiming for 12psi in 2nd max 185 KPa on a WOT pull and it took a while to realise that those figures just hit over boost pretty quick in 3rd gear.
Seems there's a percentage climb from gear to gear to me.

Then you need to tune your P value. Read the thread.

Scott, when you say "depowered" do you mean the ign key is in the off position?
I wrote down notes saying you need to plum your waste gate to Normally Open (NO) and Common. Further saying if you can't blow through it, it's wrong. Thanks, I just needed some clarification on this basic step.

By the way, can closed loop bring boost on slowly or change it with throttle position like open loop can? Open loop is super super nice for getting the throttle pedal and torque output to feel just right.

Dann

it's tps/rpm based and then ms3x has gearing/vss options.

if you wanted you turbo to spool up like a centrifugal supercharger you could make that happen. Or like smart people do, you can limit the boost at part throttle.

I've had solenoids fail. two of them actually. The first was the GM solenoid that is notorious for failing. The second lasted YEARRRRRRRRRRRs but I bolted it into a position that sat way too close to the turbo and it failed; I moved the replacement behind my intake cold air box instead of outside it.

Now, let's assume I had the EBC plumbed in a way that it was normally closed.

That would mean that you'd need to ground the solenoid in order for it to open up and run wastegate (this should make sense to anyone that has the most basic understanding of wastegate plumb).

So what happens when the solenoid fails, and the MS can't open it because it's fried? You'll always be running full boost in this scenario.

Sure, you'll have overboost protection on, but that sucks. It's just not a safe way of doing things. Plumb it correctly, and if it fails then you'll fail to wategate and you can safely determine something is wrong.

It just doesn't seem smart to me to power an electronic device up and run it backwards. Cruising around it should typically be off, it's not until you need it will the MS fully clamp to ground to spool the turbo, then rapidly cycle the valve to maintain the boost target. Not rapidly cycling it to maintain 0 boost in cruise...seems silly.

yes.


yes, gearing and weather changes OL significantly. If you tune in 2nd, you'll overboost in 4th. Or if it's a cold morning and you tune, you won't be anywhere close to your desired boost level in the afternoon when it warms up.

CL for layman's:

You: Hey MS, I want to run 15psi.

MS: okay; here you go.

That suggests it's trying to work, but something's not working correctly for the PID algorithm. That's why I suggested MORE P.

The most basically understanding of PID will help you wrap your head around it.

But if I suggest that P0 would overboost to infinity and P200 would only run wastegate, that should give you a clue. If P-100-200 doesn't bring you to wastegate then you have a plumbing or setting issue.

P is the rate at which the code will reach the target. It's the first part of tuning PID.

Once you get it close, you tune I, which maintains the target over time.

Assume you have tuned P to hit 15psi--that will ONLY be during spool--like 3000-4000RPM, at 6000RPM it will drop down back to wastegate since you have no I.

Setting P should take 10-20 pulls quick pulls in 3rd to spool the turbo up and get it close, so you'd be hitting like 16-17psi.

Once you add I, you'll do full pulls to redline to hold the target over time. This will effect P and you might have to add some back in to acheive the target with added I.

This is a dance, and you will have to go back and forth adding a few to each.

Finally you'll add D; which is basically damping. It will help remove any oscillation, but it will effect both P and I and you'll have to maybe lower P again to speed the spool back up again and have P just peak past your target again.

Once tuned, you should spool quickly, breaching your target by 1-2psi, hold it to redline, smoothly/solid.

If you ever need to change your boost target, or what to hit certain levels at certain TPS values or gears or vehicle speed--you only have to input those parameters--you'll never have to tune PID again.

Yeah same as open loop is, I want it to both limit boost at part throttle and spool up like a rotrex to keep torque on the safe limit from as low as possible to redline.

Dann

yes. but OL is dumb as shit and is only useful as an initial duty table for PID to reference.


but basically you're saying you want a slow shitty car? just run wastegate.

or, just remove the turbo.

Now after this thread Ill go sort the CL.

No this setup has a stock bottom end, so ramp it onto 230wtq as fast as it can and hold it there until redline by adding boost only as fast as it can while holding 230wtq. If it holds it to redline it would make 300whp and be within the 'known limits' of stock rods.

However It will end up limited to 250 odd rwhp.

The other thing I wanted to make sure of is controlling torque in a really flat manner at all throttle positions.

Dann

rods break by tq, the force pushing them down. not hp, a math equation.

but yes you can make it do that if you want. I would add boost after 5K so the torque stayed flatter towards redline.

Ill just add boost till it hits the target torque then keeps adding it enough to keep it there.

Dann

pretty much what i did within reason. added rpms + cylinder pressure + tq = bent rods as well.

I'm going to CL as soon as I have the chance to get out and do a few pulls to tune PID, OL is for the birds, bleah.

Hell, I just checked my TS that is linked with my actual car's firmware and I've got this whole initial values thing, so I guess this article I've been reading about the boost control doesn't apply at all.

Boost Control

And this is the actual thing I should be reading instead.

Megasquirt MSEXTRA / MS3EFI • Boost control change testing (MS2 and MS3) (View topic)

Also, why did I tune Open Loop if the initial values table thing basically does the same damned thing? Well, I guess I can pretty much just transpose the open loop values into the initial values table.

I honestly felt like I knew what I was going to be doing with the old setup after Brain explained it, now I have to think through the thing again as the above article seems low on detail to me.

Edit: Now that I read through the thread everyone thinks the new method is way better than the old one and easier to manage, so there is that.

tuning the OL table is a great step forward if you're going to be using the reference table; not a waste at all.

Well, my CL seems to be working better now. I changed the "Closed Duty Cycle" to 100 (because that's the duty cycle when it's closed) and the "Open duty cycle" to 0 (because that's the duty cycle when it's open). They were flip-flopped by default..

I also set my P to 200.

Did a pull and I hit my boost target on the head and then held there (although it took me a bit to get there). So something worked. I dropped by P to 150 and did another pull, hit boost target again.

I'll tinker with the P a little more (lol) and see if its just massively high P values that 'fixed' it, or if switching the open and closed DC fixed it.

On to tuning the rest of the PID! woo!

i dont believe the code will work when you do the open/closed like that. unless they changed that as well. But it used to be that it would break if closed was less than open, but that was also when closed/open referenced the wastegate, not the solenoid, so that might have been updated as well.