• Hi Guest!
    You can help ensure that British Car Forum (BCF) continues to provide a great place to engage in the British car hobby! If you find BCF a beneficial community, please consider supporting our efforts with a subscription.
    There are some perks with a member upgrade!
    **Upgrade Now**
    (PS: Subscribers don't see this gawd-aweful banner
Tips
Tips

Megasquirt question

Share this page

I had a long online conversation with Paul Tegler. Now, based on his advice and experience, I know how to program a Megasquirt throttle body injection system using the stock A-series carbs as throttle bodies.

All I need to build the neccessary parts is a complete intake manifold with carbs. The manifold/carbs do not need to be in working or good shape. I just need them to get dimensions and test fit. You still interested Kevin?
 
I'd still be interested. I actually just picked up a throttle body to start on a custom single setup. Tell ya what, give me just a little time to decide what route to go for right now - my birthday is tomorrow, and I'll either be getting Vizard's book as a gift, or I'll be immediately purchasing it. I'd like to see his thoughts on intakes so I can choose the optimal size/setup for what I am planning. I have a spare pair of shot HS2's somewhere (bad pistons). Maybe I can find a cheap manifold on Ebay to ship off to you if I decide to go this route. I'm trying to keep my entire stock assembly as-is so if the project is taking a while to get tuned and whatnot, I can slap the stock setup back on to use it. We only get 4-5 months of LBC-friendly weather in MI, I want to take advantage!!!!

I'd be interested in how the setup would work. Sounded like a lost cause without some fancy wiring.
 
Mainly the issue was a problem with my understanding. In most TBI setups, the injector or injectors only fire once or twice per engine cycle. This is the case for the GM ECM that Rick Patton uses in his EFI conversions. With MS, the injectors can be programmed to fire 4 very short bursts per injection cycle. This, in conjunction with pressure pulse through the balance tube, is capable of dispersing the fuel evenly across all cylinders.
 
Makes sense Morris. I was thinking about that, because I had read that MS can do either a single fire, or a banked fire. I was thinking of a total of 4 injectors, run in banks, so in the rear you'd have one injector fire, and then the other immediately after.
 
A slightly theoretical approach.

Sequential injection - A-B-B-A will not work properly because of the fuel film. Fuel has these nasty habits of
1. readily sticks to the walls of the intake manifold / port.
2. remains in liquid drop form for longer time than one engine cycle.
3. As warm as the manifold / port is, the fuel on the manifold walls does not evaporate any where near completely or even enough to allow a major portion to follow the airstream into the cylinder.
4. Most fuel residing on the walls of the manifold / port flows in liquid form directly from the fuel film into the cylinder during the intake stroke.

In the past auto manufacturers have injected fuel either on an open inlet valve, or on a closed inlet valve or both depending on running conditions. (Open valves at low engine speed, closed valves at high engine speed)

Unfortunatelly for the A-series, injection on a closed valve is ruled out. It would do as Morris said and all the fuel would go into the first cylinder for each port. (2 & 3)

Injecting on an open valve is the only solution. However as stated in points 1 and 2 a large portion of this fuel will end up on the port walls. In this case fuel destined for cyl 3 would be delayed on the port wall and be sucked into cyl 4. However an equal amount of fuel destined for cyl 4 will end up on the port wall and be sucked into cyl 3.

In order to inject onto an open valve it would be best to inject directly on the valve. anythng else would only inject onto the manifold / port walls. The latter could be accomplished with single point injection (TB) and is undesireable. Injecting in the inlet manifold at the port opening simulates the TB but worse since there is less warm manifold distance to evaporate the fuel film, and a less even distribution of fuel. One could imagine that the TB system is the lesser of the two evils and is probably therefore better received by owners.

It would be interesting if the fuel could be injected directly on the inlet valves from individual injectors.
 
<div class="ubbcode-block"><div class="ubbcode-header">Quote:]It would be interesting if the fuel could be injected directly on the inlet valves from individual injectors.[/QUOTE]

Yes with a "but." It could be accomplished with an ECU that uses a cam sensor to determine when valves are open. Megasquirt does not support that feature at this time. Also, at high RPMs/loads, the valve open time gets shorter while the injector open time gets longer. Ultimately it is impossible to always squirt to an open valve.

Here's a fun Idea: set up an EDIS system for the car thus eleminating the need for the distributor, then use the distributor to time your injections. On the MS set up, injectors have constant power running to them and MS just grounds them for the duration of the injection event. With this mechanically timed set up you would just use the dizzy to send juice to the injector you want to fire. With a slight modification of code and hardware, you might be able to use an optical sensor in the dizzy to trigger the injecion event a few degrees before the valve opens.
 
Oops. I jus re-read your thread and realized my response is not really germain to what you where suggesting.

If I understand correctly, what you suggest would require a custom or highly modified head. The racing head I mentioned earlier in this thread has four injectors that sit right abovet the intake valve and shoot fuel straight down into the cylinder.

Also, if I fully understand your comments, the problems you mention are problems for any car injected or carburated with the possible exception of the new direct injection engines we are starting to see.
 
Injectors big enough to accept maximum fuel mass flow in the time allowed for open intake valve at maximum engine torque speed.

Maximum flow is not at maximum speed or maximum power but at maximum torque.

If one assumes 4000 rpm = 0.015 sec/rev and intake valve is usably open for 270 degrees, all the fuel and air enter in 0.01125 seconds. At 90% VE at that speed the cylinder of a 1275 pulls in about 287cc = 0.344g of air and needs 0.0234g fuel = 0.0335 cc making a flow rate of 2.97 cc/s = 179 cc/min. With a duty cycle of 80%, this would be 224cc/min. This would be doubled up if one injector is used for two cylinders which is not possible yet. Of course these numbers are just a guess. But injectors of this size are easily obtainable.
 
If it runs, and runs decently, I am happy. I'm not out to run any races. I just want to improve over what I have now ;-)


I'll be either using my 2.8L TB in a downdraft setup on a custom manifold, or Morris's concoction. I'm all about keeping it as simple as possible, because in my mind, it'll be much better than what I currently have.
 
Morris said:
<div class="ubbcode-block"><div class="ubbcode-header">Quote:]It would be interesting if the fuel could be injected directly on the inlet valves from individual injectors.
Click to expand...

Yes with a "but." It could be accomplished with an ECU that uses a cam sensor to determine when valves are open. Megasquirt does not support that feature at this time. Also, at high RPMs/loads, the valve open time gets shorter while the injector open time gets longer. Ultimately it is impossible to always squirt to an open valve.

Here's a fun Idea: set up an EDIS system for the car thus eleminating the need for the distributor, then use the distributor to time your injections. On the MS set up, injectors have constant power running to them and MS just grounds them for the duration of the injection event. With this mechanically timed set up you would just use the dizzy to send juice to the injector you want to fire. With a slight modification of code and hardware, you might be able to use an optical sensor in the dizzy to trigger the injecion event a few degrees before the valve opens. [/QUOTE]

First off, sorry to drag up a somewhat old post here. I've been of course reading more of David Vizards book, this time focusing on carbs. From what he's saying in there, there really is no perfect carb for a boosted 'A' series engine. On top of that, a carb which would provide the proper amount of flow seems to be pricey.

This is why I, for fun, rekindled my interest into the mega quirt option, this time much more informed than the last. Obviously sequential injection is the answer to the overall problem and I was thinking something along the lines of a electrical way to turn one of the 4 different injectors on at a certain time.

The idea I had (which I haven't put a lot of thought into yet) is to use a pertonix optical ignition. From how I understand this device, it hooks up mechanically to the dizzy, which connects to the cam.... Then inside of the pertonix optical device is a cylinder with a rotating arm. This arm has some sort of LED on it. The cylinder has 4 optical sensors which then set the timing. Am I right so far? (this eliminates the need for a cam position sensor so far)

My idea is to place 4 additional optical sensors a calculated distance before the current ones (say if the current ones are located at Pi/2, then place the new ones at , I dunno, say 7Pi/8, this I would need to know more about the ignition system to figure out....

Then setup the mega squirt to inject as if it were controlling only one injector in a throttle body, but split the signal to 4. Use the mega squirt in conjunction with the pertonix optical signal to electrically only connect to one injector at once.

If the signal from either isn't strong enough I can think of a couple high speed relays or depletion mode transistors which, although will leak current, can take care of that problem.

What do you all think?

Ben

PS. I'm an EE and although I focus on digital logic, this is somewhat up my alley.
 
I am pretty sure the Pertronix uses Hall effect instead of optics. Other than that, I can't really comment as I am not an EE.

I do know that the Megasquirt designers are working on a router board that will make sequential injection with Megasquirt much more practical. As you are an EE, maybe you should contact them and see if you can be on the development team. Megasquirt is very much a community effort.
 
Back
Top