|
The procedure for tuning that I outlined was not based on staging MAP particularly. The staging MAP can be set to any reasonable value that avoids the primary injectors going into saturation. 15.5" Hg would be fine.
The attached image shows injector and CAS signals from one of my EC2 controllers connected to a 13B CAS. I have two EC2's and they behave the same. The captured signals show operation with primary injectors saturated, followed by primary injectors at
close to the maximum pulse width the EC2 allows before saturation, followed by the stage transition to using both the primary and the secondary injectors. These changes all occurred in 5 revolutions of the engine at about 6200 RPM. The EC2 modes were at
default values which assumes that the primary and secondary fuel injectors are all the same size.
It is useful to note the % fuel flow relative that produced at 9 ms pulse width which is near the maximum and is indicated by the labels at the top of the image. When the primary injectors have been in saturation for 2 revolutions, the fuel flow is 114%
of that at 9 ms PW. If the primary injectors had been continuously saturated, the fuel flow would have been 124% of that at 9 ms PW. The result is that the mixture will go richer when the primary injectors saturate until increasing fuel demand exceeds the
saturated fuel flow rate at which time the mixture will become leaner.
Note also that the fuel flow from both the primary and secondary injectors is 87% of the fuel flow from just the primary injectors below the staging MAP. At staging the EC2 simply divides the PW in half and does not take into account the 1.2 ms dead
time or lag of these injectors. The PW and the actual injector open times are indicated in the labels. The result is the shorter PW above the staging MAP makes the injectors behave as if they became smaller since the dead time is now a larger fraction of
the PW.
This demonstrates the difficulty encountered when trying to tune the system by the method in the EC2 manual. The sudden mixture change from one table location to the adjacent one when staging takes place is to be accomplished by manually making changes
to the table values in these locations. Successfully adjusting the table values from the control panel requires a stable condition with the MAP at the exact value corresponding to the table location to be changed. This stable condition is unlikely to be
achieved when the engine has just undergone a significant mixture change.
Setting mode 3 just below the staging MAP and setting mode 6 just above the staging MAP is a work-around for the EC2 not compensating for the injector dead time. Mode 6 was intended to aid tuning when primary and secondary injectors of different flow
rates were used. The effective flow rate of the injectors changes as a result of staging as explained above. Mode 6 now takes care of this regardless of the relative size of the primary and secondary injectors. The mode 3 or 6 adjustment does not have to
take place with the MAP corresponding to adjacent staging table locations. After the initial adjustment, one can even change the staging MAP within reason without having to make adjustments to the table. The EC2 handles that change fine.
I do not know if later EC2 or EC3 controllers incorporate dead time compensation. For those with access to the code, this can be verified and if not, the dead time of the injectors in use (usually in the range of 0.8 to 1.2 ms) can simply be added in
a line of code to the PW determined by a controller which doesn't have dead time compensation. Proof of concept of this approach has been demonstrated by adding 1.2 ms to the PW of my controllers with hardware between the controller and the injector. The
result was much easier tuning equivalent to the alternative procedure outlined above.
As a point of possible interest, my take on the history of the EC2 development follows: The early EC2's used diodes to clamp the voltage spikes to one diode voltage drop above the power supply (~14V) when the injector is turned off. This gave a path
for current to flow from the collapsing injector field which delayed the closing of the injector. This delay was similar to the injector dead time and largely avoided the mixture going lean (engine bog) upon staging. There was a difficulty in achieving a
good idle, however, since very small injector open times could not be attained if the delay on closing was too long. This was addressed by replacing the diodes with a device which still clamped the voltage but at a higher value. This minimized the injector
closing delay and improved the idle tuning characteristics, but the leaning (engine bog) upon staging then became an issue due to the uncompensated injector dead time. Two methods for dealing with this have been presented above. Personally, I observed arcing
in the relays when switching between the A and B controllers with the higher voltage clamping setup. I didn't like the RF noise from the arcing in close proximity to the controller processor possibly corrupting the controller operation or memory. So I installed
classic RC snubbers in my controllers which appears to have eliminated both the arcing and the injector closing delay. This discussion is in no way meant to be critical of Tracy's work, but rather as simply indicative of a work in progress.
Steve Boese
From: Rotary motors in aircraft <flyrotary@lancaironline.net> on behalf of Stephen Izett stephen.izett@gmail.com <flyrotary@lancaironline.net>
Sent: Monday, June 15, 2020 5:26 PM
To: Rotary motors in aircraft <flyrotary@lancaironline.net>
Subject: [FlyRotary] Re: EC Staging Point
◆ This message was sent from a non-UWYO address. Please exercise caution when clicking links or opening attachments from external sources.
So a couple of things off note:
1. Steve Boese' procedure for Mode #3 Inj Floe Rate is below staging whereas Tracy’s is above.
2. Staging point was set at near 20” and with the Renesis 4 ports smaller Primaries this does seem a bit high.
This makes sense of her going lean prior to the secondaries kicking in.
Thanks
Steve Izett
> On 16 Jun 2020, at 5:57 am, Stephen Izett stephen.izett@gmail.com <flyrotary@lancaironline.net> wrote:
>
> Thanks Jeff for finding this. Much appreciated.
> Steve
|