X-Virus-Scanned: clean according to Sophos on Logan.com Return-Path: Received: from rv-out-0506.google.com ([209.85.198.235] verified) by logan.com (CommuniGate Pro SMTP 5.2.10) with ESMTP id 3305944 for flyrotary@lancaironline.net; Tue, 18 Nov 2008 11:35:25 -0500 Received-SPF: pass receiver=logan.com; client-ip=209.85.198.235; envelope-from=rwstracy@gmail.com Received: by rv-out-0506.google.com with SMTP id f6so3398297rvb.7 for ; Tue, 18 Nov 2008 08:34:49 -0800 (PST) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=gamma; h=domainkey-signature:received:received:message-id:date:from:sender :to:subject:mime-version:content-type:x-google-sender-auth; bh=Y4KdRAECvBUEwxcJZWkerdacEzzUgAhYIeXgajH3Q/k=; b=yH/894Vu6NxtziglF6+mDZcthDAY3Uq+Daphp0y0tCKR+032smtShTm8Ni5xc3ybDE l+tCKS/0mVfAInJnGQF5+CU+yAiykAmHIsVX8/VPhda9XnuAYaIli8wh1iYwhr36nnw8 b7wK2BJrq6Gi2EzbyBaCrEPFdyghQKIQI3rPA= DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=message-id:date:from:sender:to:subject:mime-version:content-type :x-google-sender-auth; b=uJj/D4ZPJmfVq0TYcEfx2zOZCcGJ84qGG8SJTNwj5hca803ke+SFLkTwE+R8MRMsKE Ro5cuQ7Y0t9tMUwuRNdLV7phK4elGnupIXI5ia06E+IukrIuP6ZiQgvaJOLLUa8W1CR0 1oNMcPKUglhQSkyoy/pgjtNNcbRY2QmItQiL4= Received: by 10.140.172.20 with SMTP id u20mr9724rve.244.1227026089641; Tue, 18 Nov 2008 08:34:49 -0800 (PST) Received: by 10.141.29.15 with HTTP; Tue, 18 Nov 2008 08:34:49 -0800 (PST) Message-ID: <1b4b137c0811180834g1698d26bkd46040d2a6f642b0@mail.gmail.com> Date: Tue, 18 Nov 2008 11:34:49 -0500 From: "Tracy Crook" Sender: rwstracy@gmail.com To: "Rotary motors in aircraft" Subject: Re: [FlyRotary] Re: Trip Report / Remote Troubleshooting MIME-Version: 1.0 Content-Type: multipart/alternative; boundary="----=_Part_52548_20510357.1227026089588" X-Google-Sender-Auth: 91b78f744239b57f ------=_Part_52548_20510357.1227026089588 Content-Type: text/plain; charset=WINDOWS-1252 Content-Transfer-Encoding: quoted-printable Content-Disposition: inline Here's my latest trip report and an illustration of the difficulty of troubleshooting at a distance. A couple of years ago I agreed to develop a GM LS-1 - LS-7 V8 engine controller based on the EC2. Part of the agreement was that the developer would do extensive testing, report all test results and get all problems resolved prior to any customer deliveries. The developer reported that the first unit worked perfectly and began customer deliveries. "Damn, I'm Good!" I (prematurely) said to myself. Soon, I began receiving calls from the developer's customers. They were having various difficulties with no identifiable pattern based on their descriptions. When I had them return the units for testing, no problem was found with the controller in all but rare cases. In talking with the developer, I asked if they had seen any of these problems in their testing. Over the course of many of these conversations it became clear that no real testing had been done other than to start the engine and idle it on a test stand with no load. I'd been "had" and would now be forced to deal with dozens of individual builders with unknown engine skills to resolve their problems or watch my reputation go down in flames. Worst of all, I now had no confidence that the basic software design was any good since it was essentially untested other than on my engine simulator on the bench (I have no GM V8 to test on). After more than a year of email exchanges with many of the developers customers, only a few were successful in getting reasonable results but all of them assured me that the installation instructions had been followed to the letter. I had no choice but to look at one first hand. With the RV-4 packed with everything I thought I'd need, I departed for Savanna Georgia where the closest installation was located. Beautiful CAVU day with the Renesis humming sweetly. Set fuel burn at 6.5 GPH, tune XM radio to desired music (40's music on this day), follow the course line to SAV on the Garmin with no weather in sight. Only a few Restricted areas to dodge south of the airport. Marvellous flight, the best possible use of 8.= 5 gallons of 87 octane regular (with 10% Ethanol) on the 200 mile flight. As expected, I saw how rusty I was in dealing with Class C airports. Ge= t ATIS, contact approach control, get squawk code, carefully hold instructed altitude (never easy in an RV when tuning radios, etc), follow vectors for approach , switch to tower & contact, get clearance to land. Had a rare perfect cross wind landing on concrete then plead ignorance to tower and be= g for progressive taxi to the desired FBO, contact ground control find a parking spot then finally shut down. Really complicated compared to my normal straight in approach to the grass at Shady Bend. I guess you'd eventually get used to it. Checked out the builders plane which was a beautiful RV-10. It was equippe= d with virtually everything (including airconditioning) you can put in an airplane, not to mention the LS-2 V8 with supercharger. All this added ove= r 300 pounds above the heaviest IO - 540 powered RV-10. If flown solo, a minimum of 100 pounds of ballast was to be carried in the back seats to maintain CG at forward limit. Anyway, below are my notes from the troubleshooting session that followed. I have to point out that in over a year of email correspondence I never got any report that remotely resembled the symptoms I observed except for the high throttle miss. OK, I feel better now after bitching about it. Such are the perils of long distance troubleshooting. Tracy Crook On 11-16-08 made trip to Savanna GA (KSAV) to check builder EC2 installation and found the following. 1. EC2 was not grounded to airframe close to EC2 per instructions. Had 2 ground wires attached but they were longer than acceptable and went to a noisy destination on electrical system. The case was electrically isolated from ground by foam tape between it and the mounting bracket, contrary to instructions. Could not sync up EM3 due to electrical noise induced communications errors. Even without EM2 adapter in line, unit would not consistently enter test modes (1 or 8) also due to noise. Unit also showed a couple of instances where map table was programmed spontaneously due to noise. Installing a short wire to airframe eliminated these symptoms. 2. At one point in testing, found that cylinder 5 and 8 were not firing. = This turned out to be due to Blue Mountain tach connection to the 5 & 8 coil driver which is not recommended. This shunted the signal to ground when the Bluemountain was not on. 3. The crank signal was measured with an oscilloscope and found to be of low amplitude and was modulated by off center trigger wheel. The signal amplitude varied by as much as 30% (average) and 100% after the sync gap initial pulse. Even with the engine at highest RPM tested, the CAS signal amplitude was only 2.4 V P-P at the highest point. Note, tach on plane was way out of calibration and measured prop rpm instead of engine rpm, (no= t recommended). I estimate that the engine rpm was ~ 2400 - 3000 RPM. Thi= s is much less than the sample Electromotive sensor evaluated here (Builder'= s spare) which gave at least 12 V P =96 P at only 1200 RPM. I think this i= s due to the mounting of the sensor offset axially by approximately .22". T= his must be corrected. 4. The EC2 flow rate was found set to maximum and even this was not enough to get enough fuel flow at ~ 25% of available manifold pressure. When hi= gher manifold pressures were eventually tested, the mixture went rich and was well within range of the available flow rate. I determined that the dynamic range setting of the EC2 was way too high for the LS-2 injectors an= d this particular engine =96 prop combination. This is why the customer was always setting the flow rate to the inappropriate max setting because they were setting Mode 3 at too low a manifold pressure setting, probably due to the rough running at higher throttle. The dynamic range was adjusted usin= g Mode 2 and the mixture could then be well adjusted at all tested throttle settings when using the default inj. flow rate. Must change the default dynamic range setting for LS-1/2 type engines and clarify the tuning procedure to caution for this situation. 5. The one remaining problem that remained in testing this installation wa= s an occasional miss at higher throttle settings. I think this is due to th= e low amplitude and modulation of the crank angle signal causing the EC2 to loose synchronization. If the amplitude were higher, the modulation due to non concentricity would not be a problem but the combination of the two is unacceptable. It is expected that moving the sensors axially so that they are centered on the trigger wheel and making sure that the sensor gap is close to .020" will eliminate this problem. Action Plan: Configure an EC2 closer to the required injector flow rate and dynamic range for this engine so that the builder will have a much simpler job setting the EC2 up. See if I can make the EC2 crank sensor less susceptible to crank sensor signal modulation at even low signal levels. Even if possible, this does NOT eliminate the need to remedy the crank sensor mounting misalignment. Send this configured controller to builder to swap out his current EC2 controller. Builder Changes Required: 1. At minimum, install a short (< 12") wire (16 ga or heavier) to airframe ground near the EC2 to the EC2 ground stud. If possible, the mounting method for the EC2 should directly ground the EC2 case to the airframe. Th= is is the preferred mounting method for the EC2 but the short ground wire should still be installed. 2. The crank sensor mount must be modified in order to have the crank sensor centered axially on the trigger wheel. Crank sensor gap should be set at .020" and should not vary by more than .003" around the circumferenc= e of the trigger wheel. If gap varies more than this, I recommend that the wheel be removed and machined in a lathe in order to take the high points off the teeth (equalize the gap around the wheel). 3. I strongly suggest that the connection to the cylinder 5/8 coil trigger from the Bluemountain EFIS be removed. The injector driver on Pin 17 of the EC2 is a suggested replacement. Check with Bluemountain if this is a workable solution. 4. When the replacement EC2 is received and after doing the above changes= , carefully follow the suggested tuning procedure in the EC2 installation guide. Adjust both the Mode 3 and Mode 2 setting before doing any adjustments in Mode 1. Report results to me at RWS. ------=_Part_52548_20510357.1227026089588 Content-Type: text/html; charset=WINDOWS-1252 Content-Transfer-Encoding: quoted-printable Content-Disposition: inline Here's my latest trip report and an illustration of the difficulty of t= roubleshooting at a distance.

A couple of years ago I agreed to deve= lop a GM LS-1 - LS-7 V8 engine controller based on the EC2.  Part of t= he agreement was that the developer would do extensive testing, report all = test results and get all problems resolved prior to any customer deliveries= .  The developer reported that the first unit worked perfectly and beg= an customer deliveries.   "Damn, I'm Good!" I (premature= ly) said to myself.

Soon, I began receiving calls from the developer's customers. =   They were having various difficulties with no identifiable pattern b= ased on their descriptions.  When I had them return the units for test= ing, no problem was found with the controller in all but rare cases.  = In talking with the developer, I asked if they had seen any of these proble= ms in their testing.  Over the course of many of these conversations i= t became clear that no real testing had been done other than to start the e= ngine and idle it on a test stand with no load.  I'd been "ha= d" and would now be forced to deal with dozens of individual builders = with unknown engine skills to resolve their problems or watch my reputation= go down in flames.  Worst of all, I now had no confidence that the ba= sic software design was any good since it was essentially untested other th= an on my engine simulator on the bench (I have no GM V8 to test on).

  After more than a year of email exchanges with many of the devel= opers customers, only a few were successful in getting reasonable results b= ut all of them assured me that the installation instructions had been follo= wed to the letter.   I had no choice but to look at one first han= d. 

With the RV-4 packed with everything I thought I'd need, I departed= for Savanna Georgia where the closest installation was located.  Beau= tiful CAVU day with the Renesis humming sweetly. Set fuel burn at 6.5 GPH, = tune XM radio to desired music (40's music on this day), follow the cou= rse line to SAV on the Garmin with no weather in sight.  Only a few Re= stricted areas to dodge south of the airport.  Marvellous flight, the = best possible use of 8.5 gallons of 87 octane regular (with 10% Ethanol) on= the 200 mile flight.

   As expected, I saw how rusty I was in dealing with Class C= airports.  Get ATIS, contact approach control, get squawk code, caref= ully hold instructed altitude (never easy in an RV when tuning radios, etc)= , follow vectors for approach ,  switch to tower & contact, get cl= earance to land.  Had a rare perfect cross wind landing on concrete th= en plead ignorance to tower and beg for progressive taxi  to the desir= ed FBO, contact ground control find a parking spot then finally shut down.&= nbsp;  Really complicated compared to my normal straight in approach t= o the grass at Shady Bend.  I guess you'd eventually get used to i= t.

Checked out the builders plane which was a beautiful RV-10.  It wa= s equipped with virtually everything (including airconditioning) you can pu= t in an airplane, not to mention the LS-2 V8 with supercharger.  All t= his added over 300 pounds above the heaviest IO - 540 powered RV-10.  = If flown solo, a minimum of 100 pounds of ballast was to be carried in the = back seats to maintain CG at forward limit.

Anyway, below are my notes from the troubleshooting session that follow= ed.  I have to point out that in over a year of email correspondence I= never got any report that remotely resembled the symptoms I observed excep= t for the high throttle miss.  OK, I feel better now after bitching ab= out it.  Such are the perils of long distance troubleshooting.

Tracy Crook

On 11-16-08 made trip to  Savanna GA (KSAV) to check builder EC2 installation and found the following. 

 

1.  EC2 was not grounde= d to airframe close to EC2 per instructions.  Had 2 ground wires attached but they were longer than acceptable and= went to a noisy destination on electrical system.  The case was electrically isolated from ground by foam tape between = it and the mounting bracket, contrary to instructions.  =  Could not sync up EM3 due to electrical noise induced communic= ations errors.  Even without EM2 adapter in line, unit would not consistently enter test modes (1 or 8) also due to noi= se.  Unit also showed a couple of instances wh= ere map table was programmed spontaneously due to noise. =  Installing a short wire to airframe eliminated these symptoms.

 

2.  At one point in testing, found that cylinder 5 and 8 were not firing. = ; This turned out to be due to Blue Mountain tach connection to the 5 & 8 coil driver which is not recommended.   This shunted the signal to ground when the Bluemountain was not on. 

 

3.  The crank signal was measured with an oscilloscope and found to be of low amplitude and was modulated by off center trigger wheel.   The signal amplitude varied by as much as  30% (average) and 100% after the sync gap initial pulse.  Even with the engine at highest RPM tested, the CAS signal amplitude was only  2.4 V P-P = at the highest point.  Note,  ta= ch on plane was way out of calibration and measured prop rpm instead of engine rpm, (not recommended).   I estimate that the engine rpm was ~ 2400 - 3000 RPM.   This is much less than the sample Electromotive sensor evaluated here  (Builder's spare)  which gave at least 12 V P =96 P at only 1200 RPM.   I think this is due to the mounting of the sensor offset axially by approxima= tely .22".   This must be corrected.

 

4.  The EC2 flow rate was found set to maximum and even this was not enough to get enough fuel fl= ow at ~ 25% of available manifold pressure.   When  higher manifold pressures were eventually tested, the mixture went rich and was well within range of = the available flow rate.   I determined that the dynamic range setting of the EC2 was way too high for the LS-2 injector= s and this particular engine =96 prop combination.   This is why the customer was always setting the flow rate to the inappropriate max setting because they were setting Mode 3= at too low a manifold pressure setting, probably due to the rough running at higher throttle.   The dynamic range was adjusted using Mode 2 and the mixture could then be well adjusted at all tested throttle settings when using the default inj. flow rate.    Must change the default dynamic range setting for LS-1/2 type engine= s and clarify the tuning procedure to caution for this situation. 

 

5.  The one remaining problem that remained in testing this installation was an occasional miss a= t higher throttle settings.   I think this is due to the low amplitude and modulation of the crank angle signal  causing the EC2 to loose synchronization.  If the amplitude were higher, the modulation due to non concentricity would not be a problem but the combination of the = two is unacceptable.  It is expected that moving the sensors axially so that they are centered on the trigger wheel  and making sure that the sensor gap is close to .020"  will eliminate this problem. 

 

Action Plan:   Configure an EC2 closer to the required injector flow rate and dynam= ic range for this engine so that the builder will have a much simpler job sett= ing the EC2 up.    See if I can make the= EC2 crank sensor less susceptible to crank sensor signal modulation at even low signal levels.  Even if possible, this does NOT eliminate the need to remedy the crank sensor mounting misalignmen= t.   Send this configured controller to builder to swap out his current EC2 controlle= r.

 

Builder Changes Required:

 

1.  At minimum, install a short (< 12") wire (16 ga or heavier) to airframe ground near = the EC2 to the EC2 ground stud.   If possible, the mounting method for the EC2 should directly ground the EC2 ca= se to the airframe.  This is the preferred mounti= ng method for the EC2 but the short ground wire should still be installed.

 

2.  The crank sensor mount must be modified in order to have the crank sensor centered axially o= n the trigger wheel.  Crank sensor gap should be set at .020" and should not vary by more than .003" around the circumference of the trigger wheel.  If gap varies more than this, I recommend that the wheel be removed and machin= ed in a lathe in order to take the high points off the teeth (equalize the gap around the wheel).

 

3.  I strongly suggest that the connection to the cylinder 5/8 coil trigger from the Bluemountain = EFIS be removed.   The injector driver on Pin 17 of the EC2 is a suggested replacement.  Check with Bluemountain if this is a workable solution.

 

4.  When the  replacement EC2 is received and after doing the above changes, carefully follow the suggested tuning procedure in the E= C2 installation guide.   Adjust both the Mode 3 and Mode 2 setting before doing any adjustments in Mode 1.   Report results to me at RWS.


------=_Part_52548_20510357.1227026089588--