Signal dampening

[tvrgeek]

First, I am totally new here. Just planning my first build, likely an MS2. My plan is to attack the sensors, then move on to the ignition side eventually moving to the fuel. A complete system is a lot to learn in one stroke! So, I have been spending quite some time understanding the mechanical system in my MGBs which will be my target. One is fitted with the Weber DGV downdraft carb. In my testing to convert it to a manifold vacuum signal for advance, I found it was necessary to provide a very small orifice in the line. This slowed the rate of change of the signal to the mechanical distributor can solving several nasty transition issues. I have not stumbled onto anything in the ECU mapping description suggesting a limit on rate. It does seem intuitive for shifting or just coming off the throttle quickly we would not want to change the advance that fast. The signal could be far faster than the RPM change. Can anyone shed any light on this?

Yes, I overthink things. You'll all get used to it.

[tvrgeek]

I found some more information on this. Turns out it was quite common as emissions and economy got important. I have not found any discussion of determining the correct rates. It would seem logical this could be a petty simple calculation off a base map.

[trakkies]

With a mechanical dizzy, the problem is that high vacuum advances the ignition. And you don't want that at idle, so the vacuum takeoff for a dizzy (usually above the throttle blade, so you don't get true engine vacuum near closed), isn't ideal for MS which wants 'pure' vacuum. As you can map the MS to give any timing you want referred to vacuum and RPM.

[tvrgeek]

Why would I not want more advance at idle? It seems to run much more efficiently, which means cooler and better mileage, lower emissions etc.

What about the concept of a dip at the preferred idle speed. is the engine slows, it gets more efficient so it speeds up. We would still need a start retard, but that is easy with a computer. From what I gather, the lack of fuel above the idle speed limits it on the top side. So the trick is getting from the efficient high advance to what is optimal under load when coming off idle. The dip could be made to migrate for cold start. 1200 cold, backing off to 850 when hot.

Now, If I start with less advance during cranking, then jump it up for the 400 to 1000, ( maybe with the dip as described) then let it drop to the start of acceleration where I want to see how much MAP to show how hard I am loading it. I do all this with just the RPM vs advance mapping?

I am very new to this, so please explain where I am going wrong. I get the advantage of limiting the MAP signal to an RPM range, and I guess even to TPS if desired.

[trakkies]

Point I was trying to make is the vacuum take off used for a dizzy may not be ideal for MS. MS (may) use engine vacuum for more than just ignition timing.

[tvrgeek]

Where I was going is in a mechanical car, the restricted port from the carb to the dizzy slows the maximum rate of change. I was wondering if the vacuum signal, manifold of course as there is no ported carb signal, has programmed in a limit on rate of change, and if so what kinds of time constants does it have?

From what I think I understand:
I want a little retarded for easy cranking. 5 advance is a good number.
Jump up to quite a bit for idle, 20 or so for cool efficient idle, but with a dip to provide idle stability,
Then dropping as the load increases pulling away, 14 or so
Climbing +/- load through about 3000 ( for the MG anyway) with a total of 34-ish under WOT, more under light load.
Then slowly tapering off 4500 or 5K to 6K redline a couple of degrees. This part I actually do not understand at all.
All this with the addition of a knock sensor, so the maps can be a tad aggressive above 1500 or so.

I'm over thinking this aren't I? I like to understand things. I would not try and do all that first shot. Testing various dampening methods in the mechanical version on my cars, I see the problem in matching rate the manifold signal falls off to the rate the mechanical advance comes on, causing a big dip in advance right as you try to pull away. The MS does not have the delay the springs and weights do, but it also does not have the delay of the restricted orifice in the vacuum signal either unless this is programmed in. If we are setting ignition timing from MAP and RPM ( temp & TPS maybe) then how do we make the transition graceful?

[tvrgeek]

I found some patents on this subject. Mechanical blocks that managed rate of change of the vacuum signal. I don't know if they were ever used on any car, but interesting anyway. Based on the dates, I expect computers were taking over so it became moot.

[Matt Cramer]

These don't need mechanical damping - cranking advance and the advance curve are set electronically.

[tvrgeek]

Of course! My original question was if the software had any features to limit rate of change to mimic some of the features done in mechanical systems. ( dashpots, accumulators etc). It seems not. I am still interested in the subject and found some patents addressing it in the mechanical realm.

This came about as a question when I was trying to get some of the advantages of a electronic system in my mechanical one before the conversion I am just starting. I wanted to use manifold vacuum to give me more advance in idle that I could use as a normal curve, but not fall off so fast when I opened the throttle as to bog down. It got me thinking about controlling rates of change and how that would apply to the MS.

With a film can and a pinhole, I was able to get it pretty good, but not as good of a transition as not using vac at idle. With the MS, I can dial in whatever idle advance I want, ( 16 or so makes it very happy) ignoring vacuum until above 1200 or so then figure it out. But, if 1200 under full load needs less advance than I have at idle, how fast do I let it drop? If it is just following the map, it will behave like the carb and bog down when it sees low vac, low RPM asking for less advance suddenlly?

[Sinatra]

I seriously cannot wrap my head around what you're attempting with a high advance number at idle. It's the exact opposite of what every manufacturer has done for more than half a century. Just to clarify: with the MegaSquirt running the ignition, the only effect manifold vacuum or pressure has on timing is what you program into it. There is absolutely no "electronic mechanism" inside the MS ECM that makes the timing do anything other than exactly what you program it to do. You have to tell it what to do at every single one of your (chosen) RPM/load points for optimum performance. It's a whole map. Everything is programmed:



The plan is to not run MBT (maximum brake torque) timing at idle. The idea is that you actually retard it a few degrees, which leaves you with a reserve of torque. At engine speeds slightly above and below your idle speed, you run the MBT timing, so that any change in engine load (such as putting an automatic transmission in gear, turning on the AC, or a less-than-clean clutch engagement) can be accounted for with the additional timing. The result is similar to throttle-kicker solenoids used in the old days. Those would increase the idle speed just enough that the additional drag would pull it back down near the intended idle speed. That's how they got the "electronic" advantages back in the mechanical days: Band-Aids. Know them and learn them if you want to get computer-controlled advantages in the mechanical world--but you'll never, ever match what can be done with the timing higher up the RPM scale mechanically, at least not automatically. Dash-mounted advance lever, anyone?

With the computer-controlled ignition configured as above, when you accelerate from idle you're immediately at a higher timing point due to the RPM and MAP change, so the result is a smooth, torquey pull from idle rather than a stumble or bog. That's why vacuum advance canisters were run to ported vacuum rather than at full manifold vacuum--the timing increase cleanly pulls you from a stop. Not trying to run a stoich or lean mixture is another tip to achieveing smooth tip-in; engines like to be just a tad rich both at idle and immediately leaving it.

Remember, the whole idea at idle is to have ZERO net torque. The engine should make just enough torque to overcome frictional, pumping, and accessory losses, and nothing more. That's how it stays at a low speed without hunting around.

Of course, the only truly right way to find MBT timing at any combination of load and RPM is with a dynamometer, but you can certainly get very close via tinkering.

[tvrgeek]

Yes, I'm trying to understand the "old" way first to put the new way into perspective. Physics does not change, just how easy it is to adjust.
Let me put this project into perspective. I am doing an ignition only setup on this car. ( 65 MGB) I want to see how far I can get by actually providing a decent spark at the most opportune time. Something one can not do with a distributor. I'll do the complete system on my other B. ( 74 GT) It will be a MS2 if I run direct coil pack, or Micro if I run LS2's.

This first instance will still have a Weber DGV carb. So if anything I know, the AFR will be wrong almost all the time. Running too little advance in a little British car is why they are known for overheating. OEM settings of 10 degrees requires way too much throttle and can't stand sitting in DC traffic jams. On my setup, it is actually above the carb ported takeoff so it sees vacuum in cold start idle. I don;t like idleing at 1200, though many resort to this. 16 degrees increases the efficiency so much, you can back off the butterflies and pump a lot less gas through it keeping cool. These cars do not have modern high turbulence combustion chambers, they have solid lifters and pushrods. They have a pretty aggressive cam. We do not have adjustable vac cans in the dizzies either.

It is my understanding the retarded timing on modern cars is to increase the exhaust temp to lower emissions. Not to make the car idle smoother. On a 65, I am free from those restrictions. To reach the "barely stable" point you point out, can be done with more precise AF/ control if you have EFI . Something I will not have on this car.

I do get your point that coming off idle I want to see an immediate advance so it will "want" to rev. But not for just a little, as I don't want idle to creep up as it warms up too much. This is the root of the problem with trying to configure manifold vacuum at idle to overcome a distributor with too much mechanical advance. My experiments in this was to see if slowing down the vac signal would allow the RPMs and weights to catch up before the vac came off too much. Same kind of thing coming off throttle when you go to high vac suddenly. Slower is good. With the maps, one can even raise the below idle advance to give some amount of idle stabilization with a carb. It just occurred to me some modifier to the maps based on the engines mechanical transition times would be beneficial. Hence my inquiry.

[trakkies]

You seem to be talking about a different MGB to the one I had many years ago. It would idle happily all day long at 800 rpm on the hottest day with no signs of overheating. And was completely standard. Even when I did some tuning - flowed head, different cam and bigger SU carbs, it still showed no signs of overheating with the standard cooling system.
And although things were different way back then, I used it to commute through London every day, so saw plenty heavy traffic.

[tvrgeek]

Only a little different. DC beltway traffic ambient temps on the road were in the 120F range when I was pushing too hot. Just pushing, not puking yet. (220 or so) At 10 MPH, that was enough to bring it down. Going from book 10 degrees to 14 and backing off the carb, (my 65 has a DGV on it) I can now sit there all day long at 190, get the idle to 850 and the throttle plate is below the vac port. Advancing 4 degrees let me back off by over 500 RPM.

We got more aggressive emissions than you did sooner. Gas has changed a bit over the years and of course, we get seasonal oxygenated blends along with our 10% ethanol. With our gas, we seem to top out at about 32 to 33 degrees around 3K by un-instrumented troglodyte style tuning. Dynos around here only want to do Harleys and kiddy racer Hondas. They would not know a carb from their, well... let's say they won't have the knowledge to tune them anyway. I MAY be able to talk my way onto one that does WRX/STI turbos once I do the full EFI. Other than that, the MS dashboard will be all I have.

Now, If I could just decide, coil pack/MS2 or LS2's/Micro. Leaning toward the LS2's Knock sensor in feedback or just as a meter tuning aid? What I have learned, maybe just an aid as it seems I don't want to push it that close anyway. Several around the table at MegaMeet suggested going about 3 to 4 degrees below detectable knock, not the 2 I had previously been recommended. That reconciles with what I am being told here.