X-Virus-Scanned: clean according to Sophos on Logan.com Return-Path: Received: from exchange.carey.wa.edu.au ([118.82.44.212] verified) by logan.com (CommuniGate Pro SMTP 5.4.1) with ESMTPS id 5108035 for flyrotary@lancaironline.net; Tue, 30 Aug 2011 04:39:04 -0400 Received-SPF: none receiver=logan.com; client-ip=118.82.44.212; envelope-from=stevei@carey.asn.au Received: from exchange.carey.local ([10.10.0.5]) by exchange.carey.local ([10.10.0.5]) with mapi; Tue, 30 Aug 2011 16:38:21 +0800 From: To: Date: Tue, 30 Aug 2011 16:38:19 +0800 Subject: Re: [FlyRotary] Freq analysis of Knock sensor? Thread-Topic: [FlyRotary] Freq analysis of Knock sensor? Thread-Index: Acxm8CwPU+P4lSkBR4io6RGP4qi0qA== Message-ID: <04384081-9494-478D-AAA6-D8F4D33573E0@carey.asn.au> References: In-Reply-To: Accept-Language: en-US Content-Language: en-US X-MS-Has-Attach: X-MS-TNEF-Correlator: acceptlanguage: en-US Content-Type: multipart/alternative; boundary="_000_043840819494478DAAA6D8F4D33573E0careyasnau_" MIME-Version: 1.0 --_000_043840819494478DAAA6D8F4D33573E0careyasnau_ Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: quoted-printable Hi Ed My Renesis came with the knock sensor, but as you say the bandwidth of the = sensor is probably to wide. Having the sensor I built up a kit knock sensor amp and filters. But as we = recognise the knock of a rotor may sound vastly different from a piston. I plan to listen in when I get to full power ground runs. Problem being I don't want to punish my engine to hear what it sounds like = when its complaining! Cheers Steve Izett On 29/08/2011, at 10:36 PM, Ed Anderson wrote: A thought occurred to me. If anyone has a spectrum analyzer (Steve??) and = could obtain an Rotary Knock sensor - that perhaps determining the frequenc= y response of the knock sensor MIGHT indicate the frequency band pass requi= red. Although it is likely to be too broad to be much help if ( as likely)= they do the filtering through a bandpass filter within the engine CPU. Ed From: Ed Anderson Sent: Monday, August 29, 2011 9:54 AM To: Rotary motors in aircraft Subject: [FlyRotary] Not so Simple was Knock sensor was tune with ADI: [Fly= Rotary] Re: How to tune with ADI? As Tracy also indicated - in may cases, what initially appears simple - fr= equently turns out not to be the case, when you dive into the details. Another seemingly simple project I thought about, was to add a wide band O2= (WBO2) sensor option for the EFISM. This in order to provide more accurat= e info on the air/fuel ratio (although one could argue just how useful that= might be for our applications). While providing an algorithm for the different voltage input curve of the W= BO2 sensor (it is much more linear than the narrow band O2 sensor curve) is= straightforward, when you dive deeper you discover that: 1. For the wide band O2 sensor to provide accurate information the sensor = temperature must be held within a critical narrow range. 2. To do this properly, the WBO2 manufacturer either programs into the au= tomobile's CPU or provides a separate "control" chip that uses a PID (Prop= ortional, Integral, derivative) the control algorithm necessary to hold the= required critical temperature range. 3. Failure to hold this temp range renders the sensor output of questionab= le accuracy and will lead to degradation/failure of the sensor. 4. The PID temp control is much more complex than the algorithm required t= o read the output voltage So while I could add the WBO2 sensor - without the separate algorithm (or c= ontrol chip), the validity of the data is uncertain and likely misleading. = So no WBO2 option any time soon. Yet this project is much more straightforward than a knock sensor. Most au= tomobile engine manufactures can afford to do the research and they can spr= ead the cost over millions of engines sold. As you might imagine, the resu= lts of their R&D is considered proprietary information and seldom available= to the public. Ed From: Tracy Sent: Monday, August 29, 2011 9:24 AM To: Rotary motors in aircraft Subject: [FlyRotary] Re: How to tune with ADI? What Ed said. There is a knock sensor input on the EC3 but after finding = the same facts as Ed mentioned, it takes a lot more R&D to implement it so = that's why it's currently unused. Tracy Sent from my iPad On Aug 29, 2011, at 8:33 AM, "Ed Anderson" > wrote: The problem with Knock sensors is that they need to be able to identify th= e "knock" sound characteristic of the engine you have it on. The circuit i= s basically simple - the challenge is design the parameters of a signal fil= ter that permits the circuit to properly identify (filter) the "Knock" char= actertistic of your particular engine. So I would suggest that a knock system for a rotary might have different ch= aracteristics than for a piston engine. So what might work fine for a pist= on engine might not work so well on a rotary. But, on the other hand it mi= ght work fine. Ed From: stevei@carey.asn.au Sent: Monday, August 29, 2011 3:45 AM To: Rotary motors in aircraft Subject: [FlyRotary] Re: How to tune with ADI? Hi Dave If you are chasing a circuit for knock sensing try: http://siliconchip.com= .au/cms/A_108910/article.html Cheers Steve Izett Perth WA On 29/08/2011, at 12:24 PM, David Leonard wrote: Ok, So I have the new ADI system installed and working.... I think? Some of you smart racing guys probably have experience with ADI. Most of the instructions that come with tuning it are directed toward car g= uys that have computers and knock sensors. All they do is turn the system = on and the computer allows more advance because I is not sensing any knock.= The only real advise I have so for for me is that I should be running jus= t less than the amount of ADI that it takes to bog down the engine. My sys= tem is a fixed flow so I don't have much choice but it seems to be about ri= ght. If I use straight water (which I did for initial testing), the engine= bogs after about 10-15 seconds of ruining the ADI system. But when I uses= a 50/50 mix of water and denatured alcohol (too hard to find methanol arou= nd here, but close enough) the engine does not bog. So assuming my amount of ADI injected is about right, now I am supposed to = be able to lean to better power and advance the timing a little. But how m= uch? For the last testing flight I was using EGT for lack of a better guid= e. When the ADI comes on, my EGT drops by about 30 deg C, and there is a h= int of roughness to the engine. I can then lean a little and the roughness= clears up. For my test flight I leaned until my EFT came back up to where= I normally limit it (890 deg C). But if I then add 5 more inches of MAP, = there is still a hint of roughness even at 890 deg and I get the sense that= it wants to be leaned a little more. But I am afraid to do that... We are= talking 45" MAP and 7300 RPM here (185 KTS indicated (Vne) and climbing at= 500 fpm). So here are my questions: Is EGT a useful tool here, and does leaning to get the EGT back to 890 make= any sense? Since Methanol will do such a good job at preventing detonation, is it safe= to lean further toward best power (peak EGT)? (note, that will be hard on = the turbo... I need to remember to bring the spare one to Reno) How much advance do you think i can add? I was previously ruining 5 deg ah= ead of Tracy's recommended initial set up. Can't give a number because the= EC2 retards the timing with higher MAP's. How much, I have not heard. I = can easily add one deg at a time using Ed's EFI monitor. But without a kno= ck sensor am really in the blind. Thoughts? -- David Leonard Turbo Rotary RV-6 N4VY http://N4VY.RotaryRoster.net http://RotaryRoster.net ________________________________ The contents of this email are confidential and intended only for the named= recipients of this e-mail. If you have received this e-mail in error, you = are hereby notified that any use, reproduction, disclosure or distribution = or the information contained in this e-mail is prohibited. Please notify th= e sender immediately and then delete/destroy the e-mail and any printed cop= ies. All liability for viruses is excluded to the fullest extent of the law= . --_000_043840819494478DAAA6D8F4D33573E0careyasnau_ Content-Type: text/html; charset="us-ascii" Content-Transfer-Encoding: quoted-printable Hi Ed

M= y Renesis came with the knock sensor, but as you say the bandwidth of the s= ensor is probably to wide.
Having the sensor I built up a kit kno= ck sensor amp and filters. But as we recognise the knock of a rotor may sou= nd vastly different from a piston.
I plan to listen in when I get= to full power ground runs.
Problem being I don't want to punish = my engine to hear what it sounds like when its complaining!

<= /div>
Cheers

Steve Izett
On 29/08= /2011, at 10:36 PM, Ed Anderson wrote:

A thought occurred to me.  If anyone has a s= pectrum=20 analyzer (Steve??) and could obtain an Rotary Knock sensor - that perhaps=20 determining the frequency response of the knock sensor MIGHT indicate the=20 frequency band pass required.  Although it is likely to be too broad t= o be=20 much help if ( as likely) they do the filtering through a bandpass filter w= ithin=20 the engine CPU.
 
Ed

Sent: Monday, August 29, 2011 9:54 AM
Subject: [FlyRotary] Not so Simple was Knock sensor was tune wi= th=20 ADI: [FlyRotary] Re: How to tune with ADI?

As Tracy also indicated - in may cases, &nbs= p;what=20 initially appears simple - frequently turns out not to be the case,=20  when you dive into the details. 
 
Another seemingly simple project I thought about,= was to=20 add a wide band O2 (WBO2) sensor option for the EFISM.  This in o= rder=20 to provide more accurate info on the air/fuel ratio (although one could arg= ue=20 just how useful that might be for our applications). 
 
While providing an algorithm for the different vo= ltage=20 input curve of the WBO2 sensor (it is much more linear than the narrow= band=20 O2 sensor curve) is straightforward, when you dive deeper you discover=20 that:
 
1.  For the wide band O2 sensor to provide a= ccurate=20 information the sensor temperature must be held within a critical narr= ow=20  range.
2.  To do this properly, the WBO2  manu= facturer=20 either programs into the automobile's  CPU or provides a separate "con= trol"=20 chip that uses a PID (Proportional, Integral, derivative) the control algor= ithm=20 necessary to hold the required critical temperature range.
3.  Failure to hold this temp range renders = the=20 sensor output of questionable accuracy and will lead to degradation/failure= of=20 the sensor.
4.  The PID temp control is much more comple= x than=20 the algorithm required to read the output voltage
 
So while I could add the WBO2 sensor - without th= e=20 separate algorithm (or control chip), the validity of the data is uncertain= and=20 likely misleading.  So no WBO2 option any time soon.
 
Yet this project is much more straightforward tha= n a knock=20 sensor.  Most automobile engine manufactures can afford to do the rese= arch=20 and they can spread the cost over millions of engines sold.  As you mi= ght=20 imagine, the results of their R&D is considered proprietary information= and=20 seldom available to the public.
 
Ed
 
 

From: Tracy
Sent: Monday, August 29, 2011 9:24 AM
Subject: [FlyRotary] Re: How to tune with ADI?

What Ed said.   There is a knock sensor input on the EC3 but afte= r=20 finding the same facts as Ed mentioned, it takes a lot more R&D to impl= ement=20 it so that's why it's currently unused.

Tracy

Sent from my iPad

On Aug 29, 2011, at 8:33 AM, "Ed Anderson" <eanderson@carolina.rr.com>=20 wrote:

The problem with Knock sensors is that they nee= d to be=20 able to identify  the "knock" sound characteristic of the engine you= have=20 it on.  The circuit is basically simple - the challenge is design th= e=20 parameters of a signal filter that permits the circuit to properly identi= fy=20 (filter) the "Knock" charactertistic of your particular=20 engine. 
 
So I would suggest that a knock system for a ro= tary=20 might have different characteristics than for a piston engine.  So w= hat=20 might work fine for a piston engine might not work so well on a rotary.&n= bsp;=20 But, on the other hand it might work fine.
 
Ed

Sent: Monday, August 29, 2011 3:45 AM
Subject: [FlyRotary] Re: How to tune with ADI?

Hi Dave=20
If you are chasing a circuit for knock sensing try:

Cheers

Steve Izett
Perth WA
On 29/08/2011, at 12:24 PM, David Leonard wrote:

Ok,  So I have the new ADI system installed and working....&n= bsp;=20 I think?
 
Some of you smart racing guys probably have experience with ADI.
 
Most of the instructions that come with tuning it are directed tow= ard=20 car guys that have computers and knock sensors.  All they do is tu= rn=20 the system on and the computer allows more advance because I is not sen= sing=20 any knock.  The only real advise I have so for for me is that I sh= ould=20 be running just less than the amount of ADI that it takes to bog down t= he=20 engine.  My system is a fixed flow so I don't have much choice but= it=20 seems to be about right.  If I use straight water (which I did for= =20 initial testing), the engine bogs after about 10-15 seconds of ruining = the=20 ADI system.  But when I uses a 50/50 mix of water and denatured al= cohol=20 (too hard to find methanol around here, but close enough) the engine do= es=20 not bog.
 
So assuming my amount of ADI injected is about right, now I am sup= posed=20 to be able to lean to better power and advance the timing a little.&nbs= p;=20 But how much?  For the last testing flight I was using EGT for lac= k of=20 a better guide.  When the ADI comes on, my EGT drops by about 30 d= eg C,=20 and there is a hint of roughness to the engine.  I can then lean a= =20 little and the roughness clears up.  For my test flight I leaned u= ntil=20 my EFT came back up to where I normally limit it (890 deg C).  But= if I=20 then add 5 more inches of MAP, there is still a hint of roughness even = at=20 890 deg and I get the sense that it wants to be leaned a little more. B= ut I=20 am afraid to do that...  We are talking 45" MAP and 7300 RPM here = (185=20 KTS indicated (Vne) and climbing at 500 fpm). 
 
So here are my questions:
Is EGT a useful tool here, and does leaning to get the EGT back to= 890=20 make any sense?
Since Methanol will do such a good job at preventing detonation, i= s it=20 safe to lean further toward best power (peak EGT)? (note, that will be = hard=20 on the turbo...  I need to remember to bring the spare one to= =20 Reno)
How much advance do you think i can add?  I was=20 previously ruining 5 deg ahead of Tracy's recommended initial set= =20 up.  Can't give a number because the EC2 retards the timing w= ith=20 higher MAP's.  How much, I have not heard.  I can easily add = one=20 deg at a time using Ed's EFI monitor.  But without a knock sensor = am=20 really in the blind.
 
Thoughts? 

--
David Leonard

= Turbo=20 Rotary RV-6 N4VY
http://N4VY.RotaryRoster.net
http://RotaryRoster.net


The contents of this email= are confidential=20 and intended only for the named recipients of this e-mail. If you have=20 received this e-mail in error, you are hereby notified that any use,=20 reproduction, disclosure or distribution or the information contained in = this=20 e-mail is prohibited. Please notify the sender immediately and then=20 delete/destroy the e-mail and any printed copies. All liability for virus= es is=20 excluded to the fullest extent of the=20 law.

= --_000_043840819494478DAAA6D8F4D33573E0careyasnau_--