What's new
Van's Air Force

Don't miss anything! Register now for full access to the definitive RV support community.

SDS CPI Tricks and Tips

I?m following along on this post because I am near first engine start with my SDS system. Thanks to all that have posted so far. I?m a bit unclear of the pointer and the mark part. A picture would help me a ton. Does anyone have a photo of that part of the process?

Thanks.
 
I?m following along on this post because I am near first engine start with my SDS system. Thanks to all that have posted so far. I?m a bit unclear of the pointer and the mark part. A picture would help me a ton. Does anyone have a photo of that part of the process?

Thanks.

Mike S has a good write up. Look here: http://www.vansairforce.com/community/showpost.php?p=1101538&postcount=3

It's important not to over think the pointer thing. All you have to do is establish that the engine is in the desired position for firing using the factory timing marks or other means, and step #2 is to add a pointer "anywhere convenient" on the ring gear. Wherever that pointer falls, simply mark that point. That is now your new "desired position".
 
Timing Pointer Ideas

Ralph Inkster came up with this simple way to project a pointer to use the rear flywheel timing marks.



He used a long drill bit and some duct tape to extend the case split centerline.





He put some additional marks on the flywheel as well.

Hope this idea helps out a few folks in this area.
 
That is similar to what I set up with piano wire from a seat hinge. Now building a stable platform to stand on while aiming the timing light, trying to avoid being macerated by the prop.
 
I don't think there is any need to be within 2 feet of the prop but you could also tape the timing gun in place and a video camera or phone, zoomed in with the monitor screen facing you in the cockpit.

I hope to show a couple ways to do it in the upcoming videos on the RV6 and -10.
 
I need to verify the config for my G3X RPM2 input that utilizes the green tach wire from the control box. How many pulses per rev does it output? I think it's two, but I need to verify.
 
I did a test start on the engine today. Ran smoothly at 760 rpm on the EI. I didn't have a helper, and the mag was not running well after being overhauled, so I did not do the timing. My question is what rpm is best to do the timing at? Can it be accomplished at idle, or should another range be used?
 
I like to use 1000 rpm. Set that range to a value you have marked on your flywheel 10/ 25 etc. Make sure rpm reads 1000 in gauge mode.

Shoot it with the gun and set Magnet Position so mark coincides with your timing entered.

Default is 90 on a Lycoming. With single pickups we usually see 89-91, dual picks around 86 on one and 93 on the other.
 
Any suggestions on a flight test card for verifying proper operations in flight? Any procedures on tweaking performance variables? Typically I fly 2-4500', and sometimes a bit higher on longer flights, so far never higher than 8,500 because I rarely travel far enough to justify the climb. Plus I like looking at the ground.
 
I can send you a performance worksheet that allows you to keep track of your advance/performance relationship. This is used for refining a custom curve though. I thought you were running the curve Ross did for you- if so, then treat it like any other ignition. Do the EOR checks, monitor temps, fly.
 
I can send you a performance worksheet that allows you to keep track of your advance/performance relationship. This is used for refining a custom curve though. I thought you were running the curve Ross did for you- if so, then treat it like any other ignition. Do the EOR checks, monitor temps, fly.

Please send what you have. I will be running Ross' curve at first, but the point of this ignition is to tweak :)

Do you by any chance have a procedure to test the installation? Once I get the recalcitrant mag behaving and the compression test completed, it will be ready to button up and test fly.
 
Aside from the normal runup checks, it's hard to think of a formal test card for the ignition. Assuming you have the correct circuit protection and won't pop a breaker on climbout, all you really have to do is monitor the displayed advance in the gage window (to make sure it doesnt do something unexpected). The rest is really looking at CHT/EGT to see how the engine responds. Ross set you up with a very benign curve though - I wouldn't expect to see anything even remotely different.
 
A heads up to those interested in CPIs, we are sold out at the moment but hope to have more PCBs from the soldering facility sometime next week. They've been in process for around a month now as we saw supplies dwindling quickly in June and put another batch in for loading and soldering then.

Thanks to all those folks who've ordered and given us lots of positive feedback.
 
Got the engine timed today after the rains let up. We attempted to create as much shade as we could with a hanger door and afternoon sun position, but it was still not easy to see the marks. I would suggest timing at dusk, or not being a cheapskate like me and invest in a daylight readable timing light.

Ended up with 92˚ magnet position. We timed while only the EI was operating; the mag was grounded. Hope this was the correct procedure.
 
If the "system being timed" is also the "system running the engine" (think about that for a second), there is no issue as long as the engine is stable. That said, if you have a split system (a magneto in your case), its best to let that other system run the engine while you fool with the timing (assuming the magneto is functioning and timed properly). The reason for this is that modifying magnet position in situ is also changing effective timing and this will change RPM some. With the engine running on the magneto at data plate timing this is no big deal because the mag is so far advanced from the EI that the latter is just sparking on smoke... Fooling with magnet timing will have no effect on RPM.

Glad you got it sorted. Time to fly and enjoy!
 
What I did notice today during trials is when the magneto is engaged, I'll get a 100 rpm increase at around idle speed (680-700rpm). If I kill (depower) the EI coil and run on the mag, I get a 30rpm drop. If I ground the mag, I'll get ~100 rpm drop. Is this to be expected when running a split system? Is it a result of the timing differences at low rpm?
 
Last edited:
I was reading the SDS CDI installation manual.

Be sure to remove any parts which are not positively held in place once the mags are removed. If these come loose inside the engine, severe damage or complete engine failure could result! If you’re installing a single coil pack, remove the right magneto and any gears or bearings which rely on the magneto to hold them in place

&

When replacing both mags, you need to again remove any gears, drive couplings or bearings which are normally held in place by the mag. Lycomings may be fitted with various different mag drives and parts so you need to understand what to remove.


I have a O-360-A1A (180HP). Besides removing magneto and gear/coupler (attached to Mag), installing cover, what else do I need to do? Can anyone explain.

Accessory drive has crank shaft gear, cam gear, two idlers. These drive two mags, one prop gov and vacuum pump. Oil pump is direct drive off of crank. I assume the accessory drive gears don't rely on the magnetos to be present for the Lycoming engine's. Just checking... the instructions scared me...:eek:
 
Last edited:
Nothing extra to remove on 4 cyl Lycs. On 6 cylinder Lycs, we have a diagram showing which parts need to be removed but they all come out through the mag holes.
 
Flywheel

I have an O-235 which uses the smaller SAE-1 prophub. I am having trouble finding the larger flywheel with that bolt pattern. Would it be possible to use the SDS automotive sensor with the smaller flywheel?

Bill Newkirk
 
We had a customer with a 320 and small flywheel make his own mount on the left side of the engine and install the magnets in the fore/aft plane. You'd have to find your own positioning by scribing a circle on the flywheel face and machine a new mount. I could send you photos of what he did.
 
Small flywheel pics

Ross, a picture or two would be great. Once I get the flywheel figured out, the next thing will be to decide on whether to get a single or double system.
 
CPI/Mag combo's work just fine. Slight difference in LOP timing to compensate for the retarded timing of the magneto, but thats about it.

I will emphasize again that you buy the dual hall sensor on the first go, even if you dont think you are going to add a second system down the line. Its only a few extra bucks and makes it SIGNIFICANTLY easier to add a second system if you change your mind.

From post #19
 
Small flywheel pix

Here are some photos of what the other guy did:




8-32 Allen set screws go into threaded holes here, magnets go in from aft side to face the red Hall sensor

 
CPI-2 remote LOP switch ?

Hi - we are looking at a single CPI-2 install for our IO540 and would prefer to have the contoller mounted out of (normal) sight, either integrated into the iPad mount (behind the iPad) or in a flip down mount under the panel. Reason being don't expect to be accessing it during normal flight, only during set up and fault finding.

Anybody know if the CPI-2 supports the use of a remote LOP switch ?

Many thanks

Andy
 
I'm looking at the wiring diagram for the CPI-2 and I don't understand why the coil and ECU power wires need to be switched, but not the backup battery.

The docs say that the backup battery draw is minimal, but how does the system know it's off?

In the documentation it says:

It is important to switch both purple and red wire for proper operation and to also keep the backup battery from slowly draining current when the system if shut off. Also allows proper power-off operation.

So, if in normal operation, both the purple and red wire loose their power, it doesn't go to battery backup? It just switches off?

Also, why does this system need 3 fuses (6 for a dual)? Can't I use one fuse to power it (red and purple) and one fuse for backup battery? I don't understand why the ECU and coil are powered separately when loosing either would make the ignition fail.
 
The docs say that the backup battery draw is minimal, but how does the system know it's off?

. . .

So, if in normal operation, both the purple and red wire loose their power, it doesn't go to battery backup? It just switches off?

It does not switch off if the primary power is lost in flight. The CPI has a tach pickup, so knows that the engine is turning and that its job is to keep it turning. Therefore, it goes to the backup battery. When shutting down on the ground, there is no RPM, so it does not automatically switch to backup power when the main power is cut.
 
Thanks for that. I see now in the documentation how it works:

Power down of the CPi2:
Due to the ecu having its own backup battery, power down is unique on the CPi2. When you turn off Mainbus 12 volt power to the CPi-2 it automatically switches over to its backup battery. If there is no engine rpms, and no keypad
keys are pressed, it will begin a power off countdown within 5 seconds, and after counting down from 10 to 0 it willshut off unless Mainbus power is switched back on. So total power off time is approximately 15 seconds.

Now I would like to know why it needs 3 fuses (7 if you include the built in ones). That seems pretty excessive:

If the documentation shows a double pole switch to power the coil and ECU at the same time, why not use a single switch and fuse to power both:

Power<-> fuse <-> switch <-> ECU and COIL

Why the extra wires, fuses, and larger switch?
 

Attachments

  • cpi2switches.jpg
    cpi2switches.jpg
    44.7 KB · Views: 250
Last edited:
The 2 power wires have different draws, therefore they should be protected with separate fuses/breakers rated for each circuit. Do not tie the 2 feed wires together & try to protect a 20g wire with a 10 or 15A fuse!
You have a double poled switch (or 2 ganged switches) so you can switch both circuits off during power down. If one switch was left on by accident, your BU battery circuit would read it open & drain it down.
The mini fuses in the CPU housing on the other hand are protecting the internal components, not the wires.
 
20g can't carry 10 amps sustained over any real length, but it can carry 10amps for around 3 feet as the voltage drop is around 1/2 volt which is 6 watts. If the run between the red and purple wires to the switch was only 2 feet, there wouldn't be a problem from a protect the wire prospective.

Here is the actual answer:

https://vansairforce.net/community/showthread.php?p=1445172

Basically they are separating the ECU power from the coil because the coil pulls a bit of current and they are worried about noise on that line.

However, the counter point is that having two switches (even if they are in the same body), two fuses, two wires, and 10 connections that all need to work in order for the system to not fail over is not making it less likely to fail, rather more likely.

Another point is that this mitigation doesn't exist on the battery backup side. The coils and ECU will draw from the battery over the same 18 gauge wire when in fail over mode.

Now that I know what the actual issue is I can choose how I want to wire it, and I have a scope so I can see how noisy it actually is.
 
1 year/85 hrs since installing the CPi-2. Everything is wired, installed and set up per the manual. Everything works as it should. Great system with great support.
 
Common feed for ECU and COIL

Per Bob Nuckles today on aero electric digest, a common feed can cause dirty power and confusion in the CPU.
 
Per Bob Nuckles today on aero electric digest, a common feed can cause dirty power and confusion in the CPU.

I replied on that list. I think this is more complex than "shared bus bad discrete bus good." I suspect it has a lot to do with length and gauge of wire. What I don't have a good handle on is how much fuses/breakers/switches change this.

Terminating the ECU and coil at the end of 10" of 10gauge wire is no big deal, doing the same over 18gauge wire matters more, but probably still no big deal, but through 18gauge wire to a shared switch and breaker? I have no idea.

I do know that the backup battery shares the coil and ECU over 18G wire and a single fuse.

schu
 
It’s been a long time since I started this thread and I have learned a bunch. That, coupled with the many others who have become avid SDS users means that it’s time to update this thread with more refined information. I hope the other users will join in and add their experience.

Since starting this thread I am now a full SDS EFI user and some of that journey crosses over to the “ignition only” product. What follows is “ignition related” unless otherwise identified. Also note that it has become apparent that effective tuning of this system is dependent on many factors: 6 vs. 4 cylinders, PV vs AV, or C/S vs. FP play a role in the numbers you see. Bottom line, don’t use my methodology as a recipe unless you have the similar engine configuration to me. The methodology should provide some guidance however. That said, The following has been effective on my stock compression PV 540 with a “heavy” Hartzel metal C/S prop as installed on my Rocket.

Also noteworthy is the fact that the SDS product uses RPM and MAP as tuning variables. I depart from Ross’ normal guidance in the sense that I use the MAP signal as the primary logic for engine demand. The RPM tuning in my case is a fairly simple linear slope that provides a conservative baseline that can be further tweaked with MAP settings. This is because with a C/S prop and the resulting narrow RPM range I fly, the MAP is the most dynamic and responsive to my needs. Recent experience helping a new SDS customer set up his fixed pitch Cozy illustrated some of the differences in my technique compared to what ultimately worked for him. It was also a clear reinforcement that a one size fits all ignition product generally leaves a lot of performance on the table for most.

All that said, let me cover a few of the “optimized” points on the curve that I did not address early in the thread.

IDLE TIMING: Anybody remember twisting the distributor on an old car to set the base timing? The vacuum advance hose was disconnected and the engine was running without benefit of the MAP advance component. Remember how labored the engine sounded and how much it perked up as soon as you hooked up the vacuum hose? Well, with SDS you can find your engines “happy place” right in the chocks with a little thought and a few keystrokes. Using the LOP (or “flight test switch” as I like to call it), with the engine warmed up, note the MAP and RPM of your engine. Zero out the LOP value, select the LOP to active, and press the up arrow to add timing. In most cases, the engine will respond favorably to advance and continue to add RPM and smooth out until it peaks and begins to decline. The peak of RPM and/or smoothness is your engines “happy place”. Note that advance value and edit your curve to apply that timing to that RPM and MAP. While editing, Smoothly ramp down the MAP advance to accommodate advancing the throttle and make sure it is all removed with any MAP above that required to taxi. Congratulations, You have now optimized your curve to the idle region. Took all of 5 minutes. Note that it will likely run so well that you will have to go back and adjust the throttle blade stop screw to pull the RPM back to your normal level. Expect the timing adjustment to add 100+ RPM to your normal value.

TAKE OFF POWER: High MAP at takeoff power is one of the more challenging areas of engine performance testing. Even with 100LL and well performing subsystems, some of us struggle with detonation and high CHT’s. This is in no small part to the fine balance that the timing of legacy magneto ignition systems had to contend with. With a fixed system, one had to strike the balance between too much advance at takeoff power and not enough when high and lean. That’s a tall order when the timing is fixed. But is the “data plate“ value optimum for takeoff power? Not always, as it turns out. Not a lot of experimentation in this area so when I had my 540 on the dyno I decided to try get some data. I published a thread on this site describing the process but to save you reading it, it turns out the PV Lycoming is largely insensitive to timing when this rich. My aim in the dyno test was not to see where peak timing occurred, but rather, how much timing could be removed before power fell off. Turns out, that value is A WHOLE BUNCH! So in my case, since I am primarily interested in auto gas use and therefore detonation prevention as the main consideration, I knowingly edited my ignition to remove a bunch of ignition advance in those MAP values associated with TO power. As a reference, I pulled 7 degrees OUT of the data plate baseline of 25 degrees. Yes, at a sea level airport, I’m only showing 18 degrees advance on my takeoff roll and initial climb. Dyno testing shows this value is only costing me a couple of HP (out of 336) but if you think that hurts performance in my airplane, I’m happy to take you for a ride. And obviously, my CHT’s are no problem - ever.

Those of you testing at home, there’s a couple of options. Start pulling 1 degree out and measure TO distance until it becomes obvious that the performance is declining faster than the CHT/detonation margin benefit, or just remove a few degrees and forget about the raw numbers.

MAP SENSOR FAILURE PROTECTION: Thanks to a report from another member concerning the failure of another brand ignition system, we all are aware of a previously unknown failure mode. Thanks to the flexibility of SDS, we can mitigate that failure mode through discrete programming. SDS, like several other EI vendors uses a GM style MAP sensor. This is a commodity part and has billions of hours of history - but they can fail. And the failure mode of this device is that it “tells” the EI brain that the MAP is zero at failure. In response to this low MAP signal, the brain box applies full advance (let’s call it 15 degrees) to the top of whatever RPM timing exists (Let’s say 25 degrees). And if you are on TO roll and the RPM timing is at 25 and the failed MAP sensor calls for 15 more, then you are lifting off at 40 degrees of timing. That’s a bad day.

So the mitigation for this failure is simply to navigate to the lowest 4 or 5 “slots” on the MAP page and zero out the advance. Typically, these settings are well below the MAP you will see at any point with the engine running, so they are effectively hidden from the useful part of the map. However, IF the MAP sensor fails it will ask for those very positions and the brain will deliver ZERO ADDITIONAL advance to the existing RPM timing. In the takeoff scenario above, the failed MAP would result in a timing of 25 at takeoff. No problem, thanks to a few keystrokes.
 
Last edited:
Back
Top