Return-Path: Received: from mail15.syd.optusnet.com.au ([211.29.132.196] verified) by logan.com (CommuniGate Pro SMTP 4.3c1) with ESMTP-TLS id 726613 for flyrotary@lancaironline.net; Sat, 12 Feb 2005 17:05:30 -0500 Received-SPF: none receiver=logan.com; client-ip=211.29.132.196; envelope-from=lendich@optusnet.com.au Received: from george (d220-236-11-223.dsl.nsw.optusnet.com.au [220.236.11.223]) by mail15.syd.optusnet.com.au (8.12.11/8.12.11) with SMTP id j1CM4bkv032422 for ; Sun, 13 Feb 2005 09:04:39 +1100 Message-ID: <004d01c5114f$27681bb0$df0becdc@george> From: "George Lendich" To: "Rotary motors in aircraft" References: Subject: Re: [FlyRotary] Re: Vapor lock in sump/header tank. Date: Sun, 13 Feb 2005 08:06:50 +1000 MIME-Version: 1.0 Content-Type: multipart/alternative; boundary="----=_NextPart_000_004A_01C511A2.F8A934E0" X-Priority: 3 X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook Express 6.00.2800.1106 X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2800.1106 This is a multi-part message in MIME format. ------=_NextPart_000_004A_01C511A2.F8A934E0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Paul and Todd, Todd I believe I support your slant on things nearly 100% i.e.=20 I don't like the complexity of header tanks and their problems. I figure if your going to have one - vent return to Tank!! This is indeed what I would do with a high wing gravity feed to the = header tank. The only problem with this that liquid finds it own level ( see water = level), the line soon full with fuel - without a valve to release the = air/ fuel vapour, but keeping the fuel in the header tank. George ( down under) ----- Original Message -----=20 From: Todd Bartrim=20 To: Rotary motors in aircraft=20 Sent: Saturday, February 12, 2005 9:18 PM Subject: [FlyRotary] Re: Vapor lock in sump/header tank. Hi Paul; Caution this is a long story relating my previous vapour lock = issues and how they may relate to yours. Recent comments have differed = about the need to vent or not to vent a sump/header tank. As I = understand it you have a sump tank in which it is gravity fed from your = mains, then pumped up to your engine. However IIRC a previous post in = which you said you had copied Ed's system which is a header tank that = has boost pumps feeding it. (I may have that incorrect as I couldn't = find the post in question) Now I went through a couple different = versions of a header tank before scrapping the idea altogether. My first = one was always vented, but this required level control, via the boost = pumps as it required fuel to be pumped up to the tank, then it gravity = fed my FI pumps. If not filled by pumps, it would just drain the tank = due to the vent. If pumps were left on it would over fill and fuel would = go out the vent lines which were tied via a common header to the main = tanks so negative draft would draw most of the fuel back to the feed = tank, but not all. I'd planned to use an electronic level control system = to help manage this but supplier problems caused me to look closely at = what was really an unnecessary failure mode.=20 At this time I decided to use a non-vented system much like Ed's, = however I was certain that I could improve upon it by increasing the = size of the header tank and having it located on the cool side of the = firewall. But I found that it required a vent as there was no other way = to purge out the air from the tank which would accumulate there when one = would run a tank dry before switching tanks (I have 6 tanks so it is = important to completely empty the aux tanks). So I installed a vent with = a pilot accessible valve. This allowed the venting of all air out of the = tank at which time it would be closed. Then the fuel would draw up from = the main tanks to replace the fuel that had been drawn out and not = returned through the return line from the fuel reg, without the need for = pumps. I had a clear sight tube to see the level of the header tank and = another short clear section of hose on the vent line, so that I could = see whether I was venting air or fuel.=20 This seemed like a fine system, until it began seeing the heat of = flight operations. Ground runs seemed mostly successful and running a = tank dry until the header tank was empty resulted in approximately a 20 = second delay in getting fresh fuel to restart the engine and completely = purging the air within 2-3 minutes. But under flight conditions it was = much different. Heat would cause vapour problems that were very = difficult to deal with. The pressure would build within the header tank = and the sight tube would not always give a true reading as it should. = Twice I ran a tank dry while at +10000' above the airport and had a very = difficult time in getting fuel back resulting in approx 6000' of glider = alt loss each time. Another time I was doing circuits and was on my = downwind leg when the pressure in the tank caused a vapour lock even = though I had plenty of fuel in the tank(s) feeding the header tank. The = pressure had just built up to the point where even the 7 psi = boost/transfer pumps couldn't overcome it to keep fuel in the tank. In = this case I was able to make an emergency deadstick landing on an = intersecting runway and fortunately I even kept my speed up enough to = coast all the way to my tie-down spot. That's where the firetrucks met = me... kinda embarrassing!=20 In each of these cases, I found that when opening the vent valve = to release the pressure, I could see through the clear vent tube section = that I was venting a boiling fuel (air/fuel) mixture. And it took an = uncomfortably long period of time to relieve that pressure and refill = the tank with fresh fuel. I then installed optical fuel level sensors = (hi/low) on the header tank, a vacuum/pressure gauge and a temp sensor. = I found that it would begin by having a vacuum in the tank as the system = would draw in new fuel to replace that which was being consumed, but it = would soon begin to build as a pressure as the fuel heated and would = then begin to push the level down. Opening the vent relived this = pressure but it couldn't be left open, or it would just empty the header = tank. Keeping the pump(s) running would over flow the tank.=20 Can you see where all this was going? I surmise that the greater = capacity of the header tank (as compared to Ed's) allowed some heat = absorption as it took longer to become affected, but once heated it was = far less manageable. My pilot workload was too great and became centred = around fuel management. I had so much time, effort and $$$ invested in = making that header tank work that I didn't want to let go of the idea, = but one day I just had enough and tore it out of there. I've now = converted to a returnless system with no header tank and I couldn't be = happier. What I'm trying to show is that there is more going on with heat = in the fuel system than expected. I'm not familiar with your system, but = the way I understand it, you have your main tanks draining by gravity = into your sump tank. Your FI return line feeds into this tank bringing = heat from the engine with it. I surmise that it is bringing enough heat = to begin a vapour build-up in your sump tank overcoming the gravity = feed. How long after your emergency landing did you restart your engine? = If it was immediately and it ran fine, then this theory may not be = valid, however if there was sometime elapse (while you kissed the ground = & changed your shorts :-), then it could be that the tank had time = enough to cool and/or relive the pressure allowing more fuel to enter = the tank. One way that I could see overcoming this would be..... A vent line from the top of your sump tank up to the top of both of = your main tanks, but not tied into the main tank vent system. This would = allow any fuel vapours to immediately flow out of the sump tank, = eliminating any chance of vapour build-up, while allowing the cool fuel = to continue to be gravity fed to the sump tank. These fuel vapours would = then flow into the main tanks where they should immediately condense, = preventing the loss of any fuel through a direct atmosphere vent system. = This would (may) only work if you have your main tanks located above the = sump tank. Having your FI pumps located at or below the sump pump would = surely be a help as well. But I would not expect the tank to work = without a vent or with a vent to atmosphere. I hope some of this is relevant to you and helps. Todd Bartrim (top posted all the way to the bottom) ------=_NextPart_000_004A_01C511A2.F8A934E0 Content-Type: text/html; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable
Paul and Todd,
Todd I believe I support your slant on = things=20 nearly 100% i.e.
I don't like the complexity of header = tanks and=20 their problems.
 
I figure if your going to have one - = vent return to Tank!!
 
This is indeed what I would do with a = high wing=20 gravity feed to the header tank.
 
The only problem with this that liquid = finds it own=20 level ( see water level), the line soon full with fuel - without a valve = to=20 release the air/ fuel vapour, but keeping the fuel in the header=20 tank.
 
George ( down under)
----- Original Message -----
From:=20 Todd = Bartrim=20
Sent: Saturday, February 12, = 2005 9:18=20 PM
Subject: [FlyRotary] Re: Vapor = lock in=20 sump/header tank.

Hi=20 Paul;
    Caution this is a long story relating my = previous vapour=20 lock issues and how they may relate to yours. Recent comments have = differed=20 about the need to vent or not to vent a sump/header tank. As I = understand it=20 you have a sump tank in which it is gravity fed from your mains, then = pumped=20 up to your engine. However IIRC a previous post in which you said you = had=20 copied Ed's system which is a header tank that has boost pumps feeding = it. (I=20 may have that incorrect as I couldn't find the post in question) Now I = went=20 through a couple different versions of a header tank before scrapping = the idea=20 altogether. My first one was always vented, but this required level = control,=20 via the boost pumps as it required fuel to be pumped up to the tank, = then it=20 gravity fed my FI pumps. If not filled by pumps, it would just drain = the tank=20 due to the vent. If pumps were left on it would over fill and fuel = would go=20 out the vent lines which were tied via a common header to the main = tanks so=20 negative draft would draw most of the fuel back to the feed = tank, but=20 not all. I'd planned to use an electronic level control system to help = manage=20 this but supplier problems caused me to look closely at what was = really an=20 unnecessary failure mode.
    At this time I decided to use a non-vented = system=20 much like Ed's, however I was certain that I could improve upon it by=20 increasing the size of the header tank and having it located on the = cool side=20 of the firewall. But I found that it required a vent as there was no = other way=20 to purge out the air from the tank which would accumulate there when = one would=20 run a tank dry before switching tanks (I have 6 tanks so it is = important to=20 completely empty the aux tanks). So I installed a vent with a=20 pilot accessible valve. This allowed the venting of all air out = of the=20 tank at which time it would be closed. Then the fuel would draw up = from the=20 main tanks to replace the fuel that had been drawn out and not = returned=20 through the return line from the fuel reg, without the need for pumps. = I had a=20 clear sight tube to see the level of the header tank and another short = clear=20 section of hose on the vent line, so that I could see whether I was = venting=20 air or fuel.
    This seemed like a fine system, until it = began seeing the=20 heat of flight operations. Ground runs seemed mostly successful and = running a=20 tank dry until the header tank was empty resulted in approximately a = 20 second=20 delay in getting fresh fuel to restart the engine and completely = purging the=20 air within 2-3 minutes. But under flight conditions it was much = different.=20 Heat would cause vapour problems that were very difficult to deal = with. The=20 pressure would build within the header tank and the sight tube would = not=20 always give a true reading as it should. Twice I ran a tank dry while = at=20 +10000' above the airport and had a very difficult time in getting = fuel back=20 resulting in approx 6000' of glider alt loss each time. Another time I = was=20 doing circuits and was on my downwind leg when the pressure in the = tank caused=20 a vapour lock even though I had plenty of fuel in the tank(s) feeding = the=20 header tank. The pressure had just built up to the point where even = the 7 psi=20 boost/transfer pumps couldn't overcome it to keep fuel in the tank. In = this=20 case I was able to make an emergency deadstick landing on an = intersecting=20 runway and fortunately I even kept my speed up enough to coast all the = way to=20 my tie-down spot. That's where the firetrucks met me... kinda = embarrassing!=20
    In each of these cases, I found that when = opening the=20 vent valve to release the pressure, I could see through the clear vent = tube=20 section that I was venting a boiling fuel (air/fuel) mixture. And it = took an=20 uncomfortably long period of time to relieve that pressure and refill = the tank=20 with fresh fuel. I then installed optical fuel level sensors (hi/low) = on the=20 header tank, a vacuum/pressure gauge and a temp sensor. I found that = it would=20 begin by having a vacuum in the tank as the system would draw in new = fuel to=20 replace that which was being consumed, but it would soon begin to = build as a=20 pressure as the fuel heated and would then begin to push the level = down.=20 Opening the vent relived this pressure but it couldn't be left open, = or it=20 would just empty the header tank. Keeping the pump(s) running would = over flow=20 the tank.
    Can you see where all this was going? I = surmise that the=20 greater capacity of the header tank (as compared to Ed's) allowed some = heat=20 absorption as it took longer to become affected, but once heated it = was far=20 less manageable. My pilot workload was too great and became centred = around=20 fuel management. I had so much time, effort and $$$ invested in making = that=20 header tank work that I didn't want to let go of the idea, but one day = I just=20 had enough and tore it out of there. I've now converted to a = returnless system=20 with no header tank and I couldn't be happier.
    What I'm trying to show is that there is more = going on=20 with heat in the fuel system than expected. I'm not familiar with your = system,=20 but the way I understand it, you have your main tanks draining by = gravity into=20 your sump tank. Your FI return line feeds into this tank bringing heat = from=20 the engine with it. I surmise that it is bringing enough heat to begin = a=20 vapour build-up in your sump tank overcoming the gravity feed. How = long after=20 your emergency landing did you restart your engine? If it was = immediately and=20 it ran fine, then this theory may not be valid, however if there was = sometime=20 elapse (while you kissed the ground & changed your shorts :-), = then it=20 could be that the tank had time enough to cool and/or relive the = pressure=20 allowing more fuel to enter the tank.
    One way that I could see overcoming this = would=20 be.....
A=20 vent line from the top of your sump tank up to the top of both of = your=20 main tanks, but not tied into the main tank vent system. This would = allow any=20 fuel vapours to immediately flow out of the sump tank, eliminating any = chance=20 of vapour build-up, while allowing the cool fuel to continue to be = gravity fed=20 to the sump tank. These fuel vapours would then flow into the main = tanks where=20 they should immediately condense, preventing the loss of any fuel = through a=20 direct atmosphere vent system. This would (may) only work if you = have=20 your main tanks located above the sump tank. Having your FI pumps = located at=20 or below the sump pump would surely be a help as well. But I would not = expect=20 the tank to work without a vent or with a vent to=20 atmosphere.
 
    I hope some of this is relevant to you and = helps.
 
Todd=20 Bartrim  (top posted all the way to the=20 bottom)
------=_NextPart_000_004A_01C511A2.F8A934E0--