Turbo GpZ550 Ignition misfires starting above 10 lbs boost

loudhvx wrote: An inverter to drive LS1 coils can be easily made with a transistor and maybe a resistor or two.

However, when you go to a current-limiting feature, you have to start considering heat. Current-limiting means the current gets held at a steady level. This is done by a transistor acting as a variable resistor. When the coil reaches the desired hold-current, the transistor jumps up in resistance. This works fine when the dwell time is just a little bit longer than desired. But the transistor is acting as a load and is dissipating a large amount of heat to do this. It is the excess heat the coil would have been destroyed by if the transistor wasn't there. So if the transistor goes into current-limit mode for too long, it can also eventually be destroyed by heat.

( On the HEI ignitions, if you use a high-resistance coil such that the current-limit is never reached, and thus you never go into current-limit mode, you don't really even need to heat-sink the module. )

So current-limit mode should only be used sparingly unless the system is designed specifically for heavy current-limit use. I would expect the LS1 to be driven by a computer which would be programmed to only use the current-limit sparingly.

I believe the LS1 coil's internal driver handles the current limiting independent from an ECU. The bike's 18 amp charging system wouldn't be able to keep up with 4 of them all maxed at 8 amps:) . The idle speed dwell from the Kawi ignitor is probably too long.

NOTE . The LS1 coil internal "intelligence" limits dwell to 8 mSec even though the ECU, or in my example the Kawi ignitor, is providing more. Not sure how it does that. It can't simply be cutting off the coil early. That would cause early spark. So it must be operating 1 or more cycles behind the ECU. That way it knows what it needs to do BEFORE it has to do it.

I checked Megasquirt's site. The dwell time limit on the LS1 igniter is simply a safety limit so the driver doesn't melt down. It's really more of a time-out function. After 8 msec the coil shuts down. If there hasn't been a fire signal after 8 msec, it assumes an ECU malfunction and overrides the dwell signal of the ECU.

They recommend sticking to a designed dwell of about 5.6 to 5.8 msec. It does take longer for the coils to reach current-saturation than 5.8 msec, but they must be reaching magnetic-saturation sooner since they report no increased spark energy when going above 6 msec.

So it is likely a simple current-limit circuit much like the old HEI modules. This works well as long as the ECU knows to use short dwell in the first place.

The problem in the case of using a ZX igniter to drive the LS1 is that the ZX igniter doesn't use a short dwell. It uses a steady dwell angle with varying dwell time, or at the very least uses a steady long dwell time for the much slower ZX factory coils. Simply using an inverter to drive the LS1 units will result in the LS1 units running much hotter than designed because it will be using the current limiter much longer than designed and/or worse, it will simply time out and possibly create a spark way too early.

At 1000 RPM, a ZX will be sending very roughly a 17 msec dwell signal. This will steadily drop to about 2 msec by 10,000 RPM.

www.megamanual.com/seq/coils.htm

loudhvx wrote: I checked Megasquirt's site. The dwell time limit on the LS1 igniter is simply a safety limit so the driver doesn't melt down. It's really more of a time-out function. After 8 msec the coil shuts down. If there hasn't been a fire signal after 8 msec, it assumes an ECU malfunction and overrides the dwell signal of the ECU.

They recommend sticking to a designed dwell of about 5.6 to 5.8 msec. It does take longer for the coils to reach current-saturation than 5.8 msec, but they must be reaching magnetic-saturation sooner since they report no increased spark energy when going above 6 msec.

So it is likely a simple current-limit circuit much like the old HEI modules. This works well as long as the ECU knows to use short dwell in the first place.

The problem in the case of using a ZX igniter to drive the LS1 is that the ZX igniter doesn't use a short dwell. It uses a steady dwell angle with varying dwell time, or at the very least uses a steady long dwell time for the much slower ZX factory coils. Simply using an inverter to drive the LS1 units will result in the LS1 units running much hotter than designed because it will be using the current limiter much longer than designed and/or worse, it will simply time out and possibly create a spark way too early.

At 1000 RPM, a ZX will be sending very roughly a 17 msec dwell signal. This will steadily drop to about 2 msec by 10,000 RPM.

www.megamanual.com/seq/coils.htm

Great info
thanks

17 msec dwell at idle kills that idea.

Question for Lou

Have you had a chance to scope a full electronic ignition to confirm the 17 sec dwell at idle?

120 deg of dwell would give 2 msec on-time at 10000 rpm and 17 msec at idle but it's hard to imagine the ignitor would use a fixed angle.

It should have been simple to adjust dwell based on rpm. I believe spark is delayed based on rpm.

The ZX reluctor has the single long totally flat ramp with leading edge timed at max advance and the trailing edge at min advance.

When the leading edge aligns with the trigger coil, the ignitor would need to calculate an appropriate spark delay based on rpm.

Zero delay if rpm is high enough and max delay if rpm is low enough. That would adjust timing based on rpm.

If it can do that, it should be able to delay turning the coil back on.

Zero delay would theoretically approach 360 deg which would maximize dwell time for high rpm.

Max delay would limit dwell time at idle. Only enough on time for coils to reach max energy.

How would delay be calculated? A timer circuit?

Maybe someone at 750turbo.com have a solution for your ignition Problem.

www.750turbo.com/forum/index.php

Garylbishop, do you hjave a thread on your bike build? Seems pretty awesome. We want more pics :silly:
What turbo are you running? CV32s are from a ZR7S or Zephyr? Jetting?

May be this ignition can help:

www.c5ignitions.com/

Scirocco wrote: Maybe someone at 750turbo.com have a solution for your ignition Problem.

www.750turbo.com/forum/index.php

Thanks for the link.

I'm searching the site and can't fine any reference to the very real problem of misfire at high boost.

Many who raise boost above 15 lbs but no mention of ignition misfire. Some raise it higher than 15.

There appears to be a "race" mode.

The ignitor may be controlling dwell which would be great.

I have a used one coming

undiablo wrote: Garylbishop, do you hjave a thread on your bike build? Seems pretty awesome. We want more pics :silly:
What turbo are you running? CV32s are from a ZR7S or Zephyr? Jetting?

I don't have a build thread but I do have many pics that I can put together.
Here's a link to most of the pics. photos.app.goo.gl/6FgkjTfdVXuzb6T68
Some quick info:

T25 from late 90's Saab 900
CV32 carbs were from a ZX600 C model. Spacing will match IF you use the Zephyr 550 offset rubber carb spigots on Cylinders 1 & 4.
OR you move the outer carbs to match the ZX/KZ. Several mods needed for that.

Jetting for the CV's on NA motor with pod filters ; 12:1 pistons, 4-1 exhaust, long duration cams
1 size up pilot jets /screw at 1-3/4 turns out
raise needles with 3 shims
130 main jets.

For Turbo ;
1 size up pilot jet/ same screw setting
needles at stock height (no shims)
116 main jets. Size depends on the pitot tube signal /float bowls see boost pressure through pitot tube

High pressure electric pump with boost referenced fuel pressure regulator

Turbo oil outlet is below engine crankcase so it needs a scavenge pump. I copied the 750 turbo using a mechanical pump driven off the secondary shaft.

17" front wheel w/300mm brake rotors and Zephyr 550 calipers ( wheel from FZ600)
17" rear wheel w/ stock rear caliper and smaller Busa rotor. (wheel from Bandit 650)

Turbo placement is smart. nice work

I do have scope traces somewhere. I actually use an airmotor to run the pickups then use an audio recorder to record the traces. You can actually play them back on a computer and see and hear them. It's kinda like listening to the actual motor. I will try to find a way to post them when I get a chance.... just for the curious... it doesn't really show anything than the actual signal.

But to answer the question... I've only mapped out the dwell for the KZ type ignition, and only recall thinking the ZX was similar when I checked it briefly on a scope.
The KZ does not use a fixed dwell angle, but the angle does not vary widely. In the world of old-timey ignitions you can try to reduce the effects of wide ranging dwell time by either actually controlling the dwell time, or build in a safety margin to handle the excess heat created at low RPM. The KZ and I assume the ZX do a little bit of both. They make the coil big (considering it only fires two cylinders) to handle a lot of heat, and they designed a system that varies the dwell angle slightly.

As I mentioned, there are some design changes and/or manufacturing tolerances that effect the final dwell values, but for a rough average on KZ's, the idle dwell angle is about 100 deg (crank), and above 2000 RPM or so it stays around 120 deg (crank). The difference seems slight.

However, a slight change in dwell has a significant difference in heat. This is because in addition to reducing the amount of time heat is being generated, you increase time for cooling. So a small bit of dwell-angle variation is probably all that is needed... this is evidenced by the fact that these ignition are 40 years old and the majority are still working fine.

I think I still have a ZX igniter and pickups here at work. If I can rig up the air motor and a scope I'll post the hard numbers.

Ok so I was able to do a few runs at work. My buddy at work helped me get things setup, but we were still rushed so this is really only preliminary.

The ZX dwell is quite a bit more efficient than that of the KZ. At 1000 RPM the dwell is about 6 msec (quite a bit better than 17 msec). At 10,000 the dwell is the expected 2 msec.

However, below 1000 RPM, especially during startup type RPMs, the dwell is longer than 8 msec. I measured roughly 10 msec at 450 RPM. So that makes startup using an LS1 possibly problematic. Then again, maybe it won't matter. The spark will be coming about 2 msec early, but that's out of 120 msec for a rotation. Still, I'd rather not have early sparks during startup.

A couple side notes I noticed... unlike the KZ, the ZX ignition is not two independent ignition circuits in one box. Both pickups must be connected and functional in order for the dwell to work properly. With the other set of pickups not connected, the dwell drops to being very short. At times it drops well below 1 msec.

Also, unlike the KZ, the ground side of the pickups appear that they may be able to be swapped. That is, it appears the negative terminal of the pickups are simply hard-wired to ground.

I'll have to look back at the video (I'll post it after I edit it down hopefully with some screen shots) , but I think the igniter box alters the shape of the pickup signal. So how the pickup signal looks will depend on whether it's connected to the igniter or not.