- Mark 9 Development -
The Whole, Painful Story
In January of 1996, Jack Kane, (CEO of EPI, Inc.) presented the design of EPI’s lightweight 500-HP aircraft powerplant (piston-engine) at the Advanced Engine Technology Conference (AETC). At that time, the design of the engine had been refined and the prototype was being prepared for the first series of dyno-testing runs. The PSRU and the accessory drives for that powerplant were, at that point, strictly a paper design.
In February 1996, one of the principals of the original EngineAir LLC (EA) heard a tape of EPI’s AETC presentation, and contacted EPI to talk about a similar powerplant they (EA) were developing. EA asked lots of technical questions about the powerplant in general, and they expressed serious concerns about the PSRU for their powerplant. They said that the vendor with whom they had originally contracted for a PSRU three years prior had been promising to deliver for well over a year, but as yet, had produced no hardware at all.
On June 1, 1996, EA’s original vendor (NSI) delivered the long-delayed first gearbox. After a brief initial inspection, EA personnel took it apart because, in their words, “it sounds and feels like a coffee grinder”. What they found was a disaster in the making, which all but demolished their plans for Oshkosh that year. They phoned EPI asking for help, and we agreed to help.
EA immediately shipped the disassembled gearbox to EPI. As soon as it arrived (June 4), we inspected it and found an alarming number of actual and incipient component failures, as well as significant design errors and assembly errors. It was our stated opinion, with which EA agreed, that the NSI gearbox could not have survived the first hour of flight.
EPI analyzed the design and the components, and on June 9, delivered to EA a 10-page preliminary report describing the problems. Photos of the problems and failures followed soon thereafter.
At EA’s request, EPI redesigned both the innards of the gearbox and it's (completely-wrong) engine coupling system. We manufactured or modified every major part in that gearbox (including the design and implementation of a lubrication system!), and manufactured a suitable engine coupling system.
On June 26, 1996, (22 days later) EPI shipped a flyable, but quite different, gearbox to EA. This (heavy) interim gearbox contained EPI parts inside the EPI-modified original housing. The interim gearbox made it possible for EA to take their V8-powered Lancair-4P to Oshkosh-'96. However, problems with other systems on the airplane ultimately delayed the first flight of the aircraft until after Oshkosh-'96.
In August, 1996, EA successfully flew their Lancair V8, and quickly accumulated over a hundred hours of flight time, using the interim PSRU (the EPI-redesign of their original NSI PSRU). In early September 1996, EA approached EPI with a plea that we produce a PSRU of our own design for EA. They were convinced that their original vendor could not deliver a usable product.
EPI agreed, and in began to morph the existing set of proprietary ideas, sketches and calculations into hardware. Using CAD technology, we generated the detailed component, assembly, casting, and machining drawings for over 70 unique parts, produced the tooling for casting and machining those parts, cast the housings, acquired all the purchased parts and raw materials, machined and constructed the prototype gearbox in house, with only a few specialized services (broaching, hobbing, heat-treating and grinding) done outside. In parallel, we designed and built testing machinery.
On October 27, 1997 (less than 14 months later) we began testing the prototype EPI Mark-9 gearbox with a 400 HP engine and a 92” turboprop propeller. (The test-stand dyno is pictured below.)
A portion of that 14-month gestation period was consumed by several EA-requested design changes to accommodate the particulars of their installation.
These EA-requested design changes included an ill-advised insistence on a gravity-oil-return system, in spite of our strong recommendations against that change. Initial testing of the prototype proved that the EA-mandated gravity-oil-return would not work.
It also revealed two small internal gearbox problems. EPI altered the designs to cure those problems, and designed and built an integrated oil-scavenge system into the gearbox. We tested the revised prototype Mark-9 PSRU system thoroughly, and on April 3, 1998, we delivered the tested prototype to EA, complete with an extensive installation instruction manual.
It is interesting to note that, contrary to rumors originated by one of the EA principals, EPI funded the entire Mark-9 R&D effort, unlike EA's previous vendor who reportedly had received over $100,000 for R&D, and who (verifiably) delivered nothing useful. Our R&D effort happened while, concurrently, EPI was carrying on income-producing projects for other clients.
During flight testing, we discovered that the performance maps we received from EA’s propeller supplier proved to be a bit optimistic. As a result, the reduction ratio in the prototype Mark-9 PSRU was less than optimal, considering the size of the prop and the speed of the airplane. In order for EA to compete more effectively, EPI built them a special gearbox with a much larger reduction ratio for the July-1998 Kittyhawk-to-Oshkosh race (which they nearly won).
Soon after EPI delivered the prototype Mark-9 to EA, they requested that we begin production of the Mark-9. We told EA on several occasions that it would be foolish to begin production of a brand new, unproven design containing several untried features, until:
- the prototype accumulated AT LEAST 100 hours of trouble-free flight,
- the prototype showed no signs of internal problems after several thorough inspections, and
- EA accumulated enough real test-flight data for us to optimize the reduction ratio to the propeller-airframe combination.
Throughout the project, EA dismissed our caution as “paranoia”, and refused to send the PSRU back for the agreed-upon 25-hour inspections. They also were never able to provide even the most rudimentary flight test data, even when EPI provided them with the actual test-flight cards defining the tests and data to be captured.
EPI’s reluctance to begin production of the Mark-9 was solidified by the failure of EA’s deposit money to arrive. (The deposit check had been "in the mail" for several months.)
It took several months for EA to agree to a reasonable (from EPI's perspective) financial arrangement for the production of 20 gearboxes. It wasn’t until November 13, 1998, that EA’s production deposit check arrived (less than 25% of what EA said they paid the original vendor for R&D alone). By that time, we were reasonably sure that the prototype had flown well over a hundred hours with no reported problems.
After the deposit check cleared, EPI began putting the Mark-9 into production. During that project, EPI regularly reported to EA (by text and by photos) on the progress and on the disposition of funds.
Of course, we encountered unexpected delays, including late deliveries and incorrectly made parts from outside vendors (which had to be scrapped and re-done).
Despite those delays, in May of 1999, EPI was ready to begin assembling and testing the run of 20 gearboxes EXCEPT for the fact that the gear-grind vendor had not yet delivered the 20 output-gears, which had been promised for early August, 1999. After seemingly endless additional promises, delays and phone calls, the finished output gears finally arrived on November 7, 1999.
While inspecting those gears, EPI discovered that the grind vendor had made a mistake, ignored our detailed drawings and process sheets, and ruined all 20 of the gears. (He later made good the entire cost of those ruined gears.)
EPI had no choice but to begin the process of making another run of propshaft gears. We immediately ordered the special material and received it in January, 2000. The remainder of the rerun (machine the blanks, hob the teeth, heat-treat the gears, finish-grind the gears, along with packaging, shipping, handling and transport of the parts to the various production vendors) took until April, 2000.
After the new run of (correct) output gears arrived, it took very little time to build and test the first production gearbox. In June, 2000, EPI delivered the first production Mark-9 gearbox, fully ground-tested, to EA.
As agreed, EPI then invoiced EA for the extant out-of-pocket expenses incurred during the production project. In response to that invoice, EA said they had no money. EA wanted EPI to continue to produce gearboxes and cover the costs, and take installment payments when the customers paid EA. We declined that offer, and refused to produce any further gearboxes for EA until the extant $20,000 invoice had been paid.
After considerably delay (with no payment having been received), and in the face of two significant EA breaches of their contract with us, we decided that it was no longer in EPI’s best interest to continue doing business with EngineAir.
In August, 2000, EngineAir declared itself insolvent. Several of EngineAir’s unsatisfied customers began the process of organizing to acquire EngineAir’s assets, including the the acquisition of the rights to the Mark-9 gearbox and the existing inventory of parts.
That process resulted in the formation of a new business entity which is owned by a group of previous customers and the two principals of the original EngineAir LLC. The new entity was named Engine Power Systems (EPSY).
In November, 2000, EPI and EPSY signed a contract which transferred to EPSY the following items:
- the right to manufacture a particular model of EPI gearbox,
- the tooling for that model,
- heat-treated housing castings and other finished inventory sufficient to build 15 more Mark-9's, and
- the drawings and process documentation which define every part in that that gearbox.
EPI retained ownership of the gearbox design and the proprietary technology embodied therein.
As part of the follow-on support which EPI agreed to provided to EPSY, we machined and built two more complete Mark-9 PSRU’s and a working cutaway (below) of the PSRU for trade show displays, consulted with various vendors EPSY solicited to produce the gearbox parts, and designed new parts to implement an EPSY-requested change to the reduction ratio.
Although EPSY acquired a significant inventory of finished parts from EPI, they couldn’t build gearboxes until they could get past the delays encountered by one of the vendors thay had chosen to CNC-finish the EPI housing castings they acquired (The same machining operations which EPI accomplished in-house on conventional machining equipment).
That CNC process encountered significant delays, and in the intervening time period, EPI produced several more PSRU's for EPSY, and in addition, inspected and overhauled a unit which had been on an engine that had lost it's oil (no damage to the PSRU).
EPI provided training courses to EPSY personnel on the assembly and testing of the Mark-9 PSRU, and in December of 2001, EPSY began testing Mark-9 gearboxes which they had assembled themselves. All the parts in those gearboxes were made by EPI, with the exception of the finish-machining of the housings.
After an extended effort to make the powerplant viable, and following the tragic loss of one of their main suppliers, EPSY went defunct.
The rights, tooling, and production facilities for the entire V8 powerplant, including the Mark-9 PSRU, were then acquired by a company called Silver Wings Aviation, a company wholly owned by Dr. Doug Pohl, one of the driving forces behind the survival of the original EngineAir powerplant and EPSY. Dr. Pohl had completed an EA-powered LancAir-4P (L4P) in 2001, and had been instrumental in the solution of many of the design issues which plagued the V8 engine and its ECU system. Dr. Pohl continued to develop the powerplant and to support other builders who were using it. He used his V8-powered L4P to commute between his home on the east coast of Florida and the hospital on the west coast of Florida where he was the Chief of Pathology.
On July 7, 2008, as Dr. Pohl began his return flight home from work, the engine in his L4P malfunctioned just after takeoff. The aircraft stalled and spun into the ground, killing a dear friend, skilled doctor, and dedicated family man.
When the NTSB investigated the crash, one potential problem they found was that the connecting shaft ("quill shaft") between the engine and the EPI Mark 9 PSRU was broken. They sent the shaft to a recognized metallurgical lab for analysis. The analysis confirmed that the shaft had been broken by the nose-down impact with the ground, which tore the PSRU off the engine and deposited it some considerable distance from the rest of the airframe.
The NTSB report listed 378 hours on the powerplant at the previous annual nearly one year before the accident. Based on the rate of usage and the fact that the inspection schedule for the aircraft was an Annual / 100 hour program (whichever occurred first), it is probable that the engine / PSRU had over 450 hours at the time of the accident.
A custom version of the Mark-9 is installed on one of the V-8-powered aircraft entered in the 2011 Reno Air Races.