Suzuki V6 "Mini Merlin" for RV-9(A)?

[clarkefarm]

Hi Dan

I have about 90 hrs on my NSI EJ25 package and seem to have sorted it out OK but the following may be of interest.

The engine dynos 162 hp at the prop flange at 6000 RPM so with maybe 10 or more horsepower being consumed in the gearbox, I am never going to match a Lyc for fuel burn.

The performance is less than an O 320 with a max of 155 KIAS at a DA of 2000' although it will exceed 180 KIAS (more than VNE in the 9A) very easily in a shallow descent such as a high speed demo pass down the runway (it sounds fantastic with its modified cams and tuned internal exhaust) but of course this is not smart.

Cooling has been an issue, especially during downwind taxiing but has been pretty much overcome with the addition of electric radiator fans and header wrap for the muffler and upstream exhaust pipe. In flight cooling is good and there is probably room to reduce cooling drag with further work on the cowl outlet and possibly an adjustable cowl flap. I suspect that turbo installations will require a lot more work to reconcile cooling/cooling drag.

The aircraft is heavy at 1,155 lbs but it has tip tanks, a BMA Efis with 2 axis a/p, redundant engine computers, sensors and coils, a backup battery and an extended storage compartment running about 6' back from the baggage area.

In a funny way I have enjoyed dealing with the problems over the last few years which were probably responsible for NSI's demise but it has been expensive. Short of copying or buying a proven package, prospective auto conversion enthusiasts should probably be aware of the liklihood of design issues leading to substantial delays and expense.

Having said that I really like what I have ended up with, especially the ease of starting, smoothness, virtually no engine management, power, very quiet cruise and the fexibility of being able to use the prop as an air brake if required.

Regards.

Rupert Clarke
RV-9A Melbourne Australia

[djvdb63]

Thanks for checking in here Mr. Clarke. That's some of the first solid feedback on NSI that I've encountered in years of watching this area of alternative engines. I've always thought the NSI setups appeared to be rather elegantly designed. They were doomed, as I understand it, mostly by questionable business dealings, combined with the reality that even an elegantly designed package like this probably needed years of testing and tuning before it could be honestly promoted as a reliable package.

Could I ask a few follow up questions based on your post?

As I understand it your first NSI engine needed to be replaced, correct? Was it overheating that doomed the first one?

Might you briefly summarize for us the "bugs" (besides cooling issues) you've solved on the installation?

At what RPM's is the engine turning at 155 KIAS? Am I correct that the true airspeed would be somewhat higher? Are you comfortable cruising at this speed...or do you basically feel like you are beating the snot out of the engine to get 155 KIAS?

What is the redrive (PSRU) ratio on your package?

Do you have the NSI CAP propeller? Do you like it? Do you trust it?

Where is the radiator located on your installation...horizontally behind the engine?

Finally...off topic here...are you satisfied with the BMA EFIS? Did you have alot of trouble before getting it right?

Thanks

Dan
Chicago
RV-9A Tail kit

[clarkefarm]

Hi Dan

Please call me Rupert.

The first engine failed with two seized big end bearings being the piston rod/crankshaft bearings for cylinders 1 and 3 (luckily on short final) after about 5 hours of flight time but substantially more engine time.

The reason for the failure was probably due to overheating while ground running, although dirt contamination during assembly or the consequences of the severe backfires experienced while trying to start the engine with the supplied engine computers (which would not handle the variable frequency ignition trigger signal coming from the 58 tooth sender during cranking with a light flywheel) could have been factors. The extensive ground running, mostly without cowls fitted and therefor radiator effectiveness was due to efforts to identify and remedy the starting problem.

The second rebuild (NSI replaced the first engine) was precautionary because of concerns that in the dying days of the company assembly may have been compromised by use of substandard parts or poor workmanship. It was also an opportunity which became available when I decided to dyno tune the engine to properly fuel map the Motec M4 ECUs which replaced the original units.

By far and away the biggest "bug" was the innapropriate ECUs. I never adopted the NSI wiring setup so complications which arose with that were avoided. Other items which needed attention were the used valve springs which failed to meet Subaru specifications and were certainly too weak for an engine expected to run at 6,000+ RPM on takeoff. I also had a problem with frayed wires in one fuel injector connector. Normally the loss of one cylinder on an EJ25 would be hardly noticeable but my system is now running a closed loop O2 sensor so the remaining 3 cylinders were over-enriched by the 8% maximum trim when the sensor recorded the unburnt oxygen from the failed cylinder. It definitely caught my attention on climb out.

NSI used a lot of second-hand parts to instal its SOHC 2.2 heads on new 2.5 blocks. The fuel injector connectors including terminals and a short length of wiring were spliced onto the new loom. The engine/prop packages offered by NSI's successors Maxwell Propulsion Systems (complete unit should be on display at NW EAA Arlington next month ?) use a new unmodified SOHC 2.5 engine so this stuff is really ancient history.

Another problem with my NSI package was the prop pitch gauge or lack of it and to a lesser extent the safety pins within the prop hub which were intended to prevent a runaway prop (someone leaving their finger on the pitch switch on the stick grip) taking you into negative pitch while in flight. The pitch gauge is really important to set the takeoff/go around pitch although in flight the tacho is the indicator of choice.

I never received a gauge but many who did experienced failures. I have been running with a small laser bouncing a beam off a mirror on a blade collar with the reflected signal being collected by a row of optical fibres and displayed in the cockpit.

The safety pin within the prop hub can be sheared off by the much stonger motor which was supplied with later packages and will then cause binding within the hub. I have not yet really addressed this problem but will probably remove the pin and rely on a much more "in your face" pitch indicator.

I mentioned the remedial heating mods in my last post but omitted a change which I made to reduce the "free" airflow over the engine. As you are aware the NSI installation has a large and deep horizontal radiator behind the engine which is fed by ducts on the inside of the cowl carrying air to a plenum chamber above the radiator. Negative pressure from the cowl outlet would normally assist with airflow over the radiator but this was (in my opinion) being lost to the excessive amount of air flowing straight through the balance of the cowl inlets, over the engine and out through the same outlet as the radiator air. I have reduced and diffused this free air with ss mesh screens but probably have a long way to go to obtain optimum cooling and minimum drag.

At 155 KIAS the engine is running at max continuous power, WOT and 5,800 RPM. The gearbox ratio is 2.12:1 so the prop is spinning at 2,736 RPM. At this power setting the engine feels busy and a bit coarse (it is much smoother at WOT and 6,000+ RPM which is what I use for takeoff) but the airframe is also beyond its sweet spot and starting to experience a fair bit of buffet. TAS was a bit higher but less than 160. Of course there is no guarantee that my plane is perfectly rigged or that I was flying dead level so these numbers should be taken with a grain of salt.

I do have the NSI CAP propellor which is a 72" two bladed (Whirlwind manufactured composite blades) +32? to -22? electrically adjustable unit. Apart from the comments above I am really happy with mine and would not swap it for anything.

My BMA G3 Efis One has been a bit of a learning curve and although I think that I will soon receive a terrain database for Australia which will enable a VFR map display I am probably not the person to give you a balanced opinion of its functionality. Hopefully when I become more familiar with it and perhaps also some software/database tweaks it will become the feature that I was looking forward to.

Sorry about the long reply - maybe a single picture would have been better.

Rupert

[rv6ejguy]

Fascinating. Thank you very much for posting this information! Best real info I have ever read on on a flying NSI package.

[djvdb63]

Originally Posted by clarkefarm

Hi Dan

Please call me Rupert.

Sorry about the long reply...

Rupert

Rupert...

Thanks for this excellent information. Nothing "too long" about it. This is rare and solid information. Thanks.

The second rebuild (NSI replaced the first engine) was precautionary because of concerns that in the dying days of the company assembly may have been compromised by use of substandard parts or poor workmanship.

Other items which needed attention were the used valve springs which failed to meet Subaru specifications and were certainly too weak for an engine expected to run at 6,000+ RPM on takeoff.

NSI used a lot of second-hand parts to instal its SOHC 2.2 heads on new 2.5 blocks. The engine/prop packages offered by NSI's successors Maxwell Propulsion Systems (complete unit should be on display at NW EAA Arlington next month ?) use a new unmodified SOHC 2.5 engine so this stuff is really ancient history.

I've always felt that Eggenfellner's philosophy of not opening the engines is a really good one. Obviously there are compromises involved, but leaving the basic engine unopened/unmodified instills a certain level of confidence. Your experience bears out the reality that opening the engine in order to optimize cams or valve springs or whatever comes with a certain level of risk.

It occurs to me that if the successor company...MPS...is well funded, honest, and meticulous in preserving the best NSI features in a new package...that something pretty good could come out of the whole mess. Right now there is very little substance on their website other than a promise not to take customer's money until they can deliver engines. They'll have a big reputation deficit to climb out of, however.

Dan
Chicago
RV-9A Tail kit

[rv6ejguy]

My view on hot cams in aircraft conversions has always been- don't- unless the mods are really well researched by someone who really knows what they are doing. The high lift, high duration cams I see advertised by many just plain scare me with stock rockers and unknown valve springs. I seriously doubt most claims of increased power at the same or lower revs than stock. Never saw much of that on my dyno.

I've seen so many cases of coil bind, retainer to guide interference, valve to piston clearance problems, improper surface treatment of reground cams, base circle reductions and resulting improper rocker arm geometry that reliability is a real question.

When I see the profiles selected by some with high ramp rates and crazy overlap. I shake my head. This stuff has bitten many. Taking engines apart is fine if you know what you are doing, a disaster if you don't. The more OE parts in an engine, usually the more reliable it is.

[djvdb63]

Originally Posted by rv6ejguy

I know Dan Hawken quite well (just visited him a couple weeks ago) and we supply the ECUs for the Mini Merlin V6. Dan is a very smart fellow with decades of practical mechanical experience. The drive was designed and built by him. It is a belt unit, robust and well flight proven.

Dan has offloaded the Mini Merlin design and drawings to Titan last year who I think now has subcontracted machining and assembly to someone else.

For the RV9, the latest atmo EJ25 rated at 173hp could do a good job at close to Lycoming weights if the redrive ratio was selected to allow the engine to spin up to 6000 for takeoff (about 2.2 to 1). I'd expect cruise fuel economy with this setup to be at least 10-15% less than an O-320 but auto fuel could offset this in $ perhaps. 4400 would give you about the same cruise power as an O-320 at 75%.

Ross...

Just half joking here, but what is Dan Hawken doing now that he's offloaded the Mini Merlin project? Is there any way you could talk him into developing a FWF engine package for RV's? If not Dan...someone else?

I think it would be a labor of love, but the potential is there for a good package to sell in the hundreds. Taking nothing away from Jan (who has survived and even prospered in a historically brutal marketplace)...but I think Egg has cornered the market at this point more by achieving basic competency and by default than by the overwhelming excellence of the package. It would be clearly a tough, tough endeavor, but...basics of a new package might be...

1. Goal of net performance equivalent to a 180 Lyc. Notice I used "net performance" and not "horsepower" equivalency. I suppose a new 180 hp package would need to generate 190 hp at the prop flange (maybe 200 hp from the engine) to overcome the inherent cooling drag issues with a liquid cooling engine and the power cost of a redrive.

2. Goal of a "bolt on" package for RV's that requires no airframe mod's for installation.

3. Choose one of Subaru's latest 4cyl 2.5L engines....with no internal mod's...unopened.

4. Mildly turbocharged...to allow for more power at lower rpms and decent performance at altitude. But turbo'd gently enough to avoid the need for an intercooler. Maybe this is impossible?

5. The turbo could alleviate the need for a muffler, correct?

6. Design an engine mount in a style more like NSI's than Egg's.

7. Careful cooling design. Incorporate massive cooling capacity. Carefully consider how air flows through the engine compartment. This is the hardest and most important part, I think. Perhaps use a rear mounted radiator something like Belted Air Power's. As I have surveyed this (and having no personal experience, of course) it appears theirs is the best designed liquid cooling system out there. They have a large single radiator for simplicity. According to their website, as well as the various magazine articles on BAP that I have encountered, they have had zero engine cooling issues (some oil cooling issues, however) while producing 180 to 200 genuine hp. I have no way of confirming the claims here, of course.

8. Figure out some way to get prop blast air to flow through the radiator while operating on the ground. Avoid the complexity/weight of cooling fans.

9. SDS/Racetech EFI & ECU's...fully redundant, etc., etc.

10. Maybe design a header tank to alleviate the need for dual high flow fuel lines in and around the cockpit and back and forth to/from the wing tanks? Maybe a flush NACA scoop to cool the header tank and diffuse the vapor lock issue inherent with having a fuel tank in a hot cowling.

11. New Hawken-designed single belt redrive. 2.2:1. Or get Belted Air to develop one with this kind of ratio. From what I gather recently, the cooling issues of Egg's geared redrives are significant and annoying...and and have been possibly under-appreciated in the past. I'm guessing that a single-belt redrive could save weight too.

12. Create a distinctive but tidy cowling to enclose all of this.

13. Fly it around & tweak it for a hundred hours before actively promoting it.

Of course...there is always the good old American liability problem to contend with.

Well...I'm daydreaming too much. I'm going out for a run along Lake Mich before dark...

Dan
Chicago
RV-9A Tail kit

[N131RV]

Originally Posted by djvdb63

Listers...


4. Crossflow seems to have burned their bridges with the RV community. http://www.crossflow.com/.

Boy, you got that right. I gave up ever getting support from Crossflow. Since I was not the original purchaser of the package, they did not even want to give me the time of day.

Too bad, because the Subaru boxer engine is just about one of the strongest and most reliable designs there is. You have to really, really work to break one.

I have friends who race the EJ20 (2.0 liter) in sand cars, running 30+ lbs of boost and putting out 450+ horsepower, and they run the same engine for a whole season.

Just about the only thing keeping me from using one on an RV is the lack of a good, reliable, reasonable cost, redrive.

The other thing I like is the cheap availability of the engines. Low mileage used ones can be had for as little as $400 plus shipping.

[rv6ejguy]

Originally Posted by djvdb63

Ross...

Just half joking here, but what is Dan Hawken doing now that he's offloaded the Mini Merlin project? Is there any way you could talk him into developing a FWF engine package for RV's? If not Dan...someone else?

I think it would be a labor of love, but the potential is there for a good package to sell in the hundreds. Taking nothing away from Jan (who has survived and even prospered in a historically brutal marketplace)...but I think Egg has cornered the market at this point more by achieving basic competency and by default than by the overwhelming excellence of the package. It would be clearly a tough, tough endeavor, but...basics of a new package might be...

1. Goal of net performance equivalent to a 180 Lyc. Notice I used "net performance" and not "horsepower" equivalency. I suppose a new 180 hp package would need to generate 190 hp at the prop flange (maybe 200 hp from the engine) to overcome the inherent cooling drag issues with a liquid cooling engine and the power cost of a redrive.

2. Goal of a "bolt on" package for RV's that requires no airframe mod's for installation.

3. Choose one of Subaru's latest 4cyl 2.5L engines....with no internal mod's...unopened.

4. Mildly turbocharged...to allow for more power at lower rpms and decent performance at altitude. But turbo'd gently enough to avoid the need for an intercooler. Maybe this is impossible?

5. The turbo could alleviate the need for a muffler, correct?

6. Design an engine mount in a style more like NSI's than Egg's.

7. Careful cooling design. Incorporate massive cooling capacity. Carefully consider how air flows through the engine compartment. This is the hardest and most important part, I think. Perhaps use a rear mounted radiator something like Belted Air Power's. As I have surveyed this (and having no personal experience, of course) it appears theirs is the best designed liquid cooling system out there. They have a large single radiator for simplicity. According to their website, as well as the various magazine articles on BAP that I have encountered, they have had zero engine cooling issues (some oil cooling issues, however) while producing 180 to 200 genuine hp. I have no way of confirming the claims here, of course.

8. Figure out some way to get prop blast air to flow through the radiator while operating on the ground. Avoid the complexity/weight of cooling fans.

9. SDS/Racetech EFI & ECU's...fully redundant, etc., etc.

10. Maybe design a header tank to alleviate the need for dual high flow fuel lines in and around the cockpit and back and forth to/from the wing tanks? Maybe a flush NACA scoop to cool the header tank and diffuse the vapor lock issue inherent with having a fuel tank in a hot cowling.

11. New Hawken-designed single belt redrive. 2.2:1. Or get Belted Air to develop one with this kind of ratio. From what I gather recently, the cooling issues of Egg's geared redrives are significant and annoying...and and have been possibly under-appreciated in the past. I'm guessing that a single-belt redrive could save weight too.

12. Create a distinctive but tidy cowling to enclose all of this.

13. Fly it around & tweak it for a hundred hours before actively promoting it.

Of course...there is always the good old American liability problem to contend with.

Well...I'm daydreaming too much. I'm going out for a run along Lake Mich before dark...

Dan
Chicago
RV-9A Tail kit

Dan got out of it to pursue other activities. He was tired of customers wanting to change all sorts of things, then calling him up and telling him stuff didn't fit/ work etc. I couldn't believe the dumb things people we doing. No chance he'd be going back to it.

To get O-320 performance, the new 173hp EJ25 would work naturally aspirated. To have O-360 performance, a low boost EJ25 would do it at 35-38 inches and somewhat lower rpm. Leave the engines alone for sure. Lower profile intake manifolds would have to be used in most cases to use the factory cowling.

Intercooler is a must unfortunately. Muffler is not required with the turbo.

NSI's mount was pretty busy. With some good thought, a lighter and simpler mount could probably be made. The SOHC EJs have less suitable pickup points than some others.

A Cheyenne type inlet below the spinner could feed a single low mounted rad under the sump with cowl flap. A sealed duct with proper shape is a must. The Belted Air rad setup works but is high drag in my opinion and the plenum setup makes it hard to seal with the flat four shape. Low drag means no turns in the cooling airstream according to my tests and historical study. Use a Visteon type custom rad core with oval tubes. The inlet placed in this part of the prop arc should provide decent ground cooling. The aim would be to keep coolant temps under 110C at 90 knots and full power at ambient temps around 35C.

EFI is not a problem but we shun the redundant ECU idea. More failure modes actually due to isolation relays and extra wiring.

Unfortunately to avoid the possibility of vapor lock, the complicated full return fuel system is required IMO.

There are already several good drives on the market. A couple form NZ and I do like the Marcotte as well. Not a real fan of belt drives but if designed correctly, there are some really good ones out there too.

The Egg drives do seem to need a lot of cooling which may be due to the gearset layout. My Marcotte runs cool with no additional airflow ducted to it.

You've hit the nail on the head about what is out there and why so far.

Ok, get started...