Summary of Design Adaptations for

Addendum to Mazda 13B Engine Adaptation for Aircraft use

Edward L. Anderson

22 June 1998


Trials, Tribulations and Solutions of the Cooling System Problem

Some difficulties were encountered with the coolant system in December 1997. One of the radiator fittings had a heavy "T" fitting attached providing for a 16 stainless steel braided hose to the header tank as well as a return to the water pump. The radiator fitting was stressed by this weight and I had intended to brace it prior to flight to relieve the stress caused by weight of the "T" fitting. However, after approx. 16 hours of engine ground runs, the stress and vibration on the radiator fitting caused a seam to open in the radiator (Evaporator Core) right above the fitting.

I decided to redesign the fitting to remove the stress since I had to replace the radiator core in any case. I went through three welders who attempted to MIG/TIG well a new fitting to a new evaporator core. In each case, the heat generated in welding the fitting to the core caused a seam in the core to open up. Reason was that the temperature of the aluminum-brazing alloy that held the components of the evaporator core together had a lower melting temperature than that required to weld the fitting to the core. This process took three months with 5 trips to the junkyard for more evaporator cores and several welders before I found a radiator shop that successfully completed the work.

Once the new radiator cores were installed and new fiberglass inlet ducts built, I was again to the point I had been in December 1997. So, all was going well until I completely cowled the engine and then I found out that I still have the coolant and oil temp headed toward the redline temps at modest power settings on hot days. While I believe the cooling would have been adequate once in the air and moving more air thought the cowl, it was going to be a problem on hot days waiting in line for take off. Therefore, I was not satisfied and decided I had to identify and fix the problem.

The cause (as may be obvious) was the hot air exiting the forward mounted radiators was not exiting the cowl adequately and was dumping hot air back on the engine block - this certainly re-circulated the heat, but did little to cool the engine. I finally constructed two fiberglass ducts to direct the hot air exiting the radiator to the sides of the cowling. The hot air duct is attached completely covering the rear of each radiator and then ends in a 5X10 inch opening at the side of the cowl. This required cutting a 5X10 inch hole in the side of the cowling below the hinge line and the placement of louvers for the hot air to exit.

With the hot air exit ducts in place, I am happy to report that I can run the engine at a sustained 2/3 max power setting on a 90 Degree F day and temp limits remain within limits. Water temperature will climb to 210 Degrees F, however Oil temperature stays under 180 degrees, so I consider this adequate on a 90 Degree day running up on the ground. I believe the louvers, when installed to cover the 5X10 openings in the cowling, will also aid by creating a modest low-pressure area over the duct exit area.

Additional information on fuel subsystem. The HALTECH EFI CPU has a number of options, which, while providing lots of flexibility, also requires lots of decisions. The basic option is to decide on which of three fuel injection modes are best for the 13B aircraft application. The three modes to choose from are: Multipoint, Throttle Body, or Staged. Each has various advantages and drawbacks and with no baseline to begin from, I have tried each in turn. In addition, the creation from scratch of the fuel map for the engine takes a considerable amount of time.

The engine will run with all three injection modes, however, the Multipoint and Throttle Body modes delivering fuel through all 4 injectors tends to cause the engine to load up a low RPMs. The Staged mode is designed such that only two injectors (one in each throttle body throat) fire at low rpm/low manifold vacuum, which gives better idle/low RPM response. Above a selected manifold pressure, the second set of two injectors kick in and deliver fuel as well. The Staged mode appears to be the best suited for the 13B, however, getting the injector timing settings so that the engine does not stumble as it switches from two injectors to all four was somewhat time consuming.

So the number of decision points involve so far in programming my HALTECH EFI CPU are:

  1. The basic EFI Mode to select (Multipoint, Throttle Body, Staged)
  2. How often the injector is to be triggered (on every trigger pulse to CPU, every second pulse, every fourth pulse, etc.)
  3. Setting injector timing for each of 7 RPM ranges (0, 1000, 2000, 3000, 4000, 5000, 6000, 7000 rpm) and 64 manifold pressure points in each range to get the right air/fuel mixture at each operating point
  4. The cold starting settings for additional fuel on those freezing mornings
  5. The acceleration "pump" settings for those time you rapidly open the throttle
  6. Adjustment of the "leaning" control

These are some of the basic considerations for the HALTECH system, other EFIs on the market are undoubtedly suitable and may have different modes and options.

An important consideration, if using an EFI, is to know that EFIs require an external trigger signal of some sort to initiate the fuel injection sequence. Most will accept a variety of ignition signal sources such as off the distributor, from the negative side of an ignition coil, from the tachometer signal some ignition systems provide, and dedicated magnetic coil/infrared sensors. Some signals may require conditioning of the basic signal to shape it to meet the requirements a particular EFI trigger input circuit. Be aware of this and read on about some of the failure modes I have encountered.

EFI Trigger signal Conditioning and an Unanticipated Problem

As mentioned above, some trigger signals may require conditioning before using them to trigger the EIF unit (something you will want to check). The primary trigger source I use is off the negative lead of the leading ignition unit. However, I needed to condition the trigger signal which involved an electronic filter circuit that shaped the fuel injection trigger pulse. The pulse off the negative side the ignition unit has a 27Volt spike riding the leading edge of a 12 volt square wave. I designed a filter circuit to clip the 27 volt spike to prevent damage to the CPU.

After installing the filter circuit, I notice that when engine RPM rose above 4000 rpm I would get signal drop-out, the indication was the engine rpm (viewed on the laptop computer program) would vary between the nominal rpm and 1/2 the nominal rpm. At 4400 rpm the rpm would cycle between 4400 and 2200, at 4800 rpm it would cycle between4800 and 2400 rpm, etc. It turns out that a capacitor I had in series with the circuit did not have adequate time to discharge between trigger pulses at the higher rpms and therefore every other pulse would fail to get through to the injection CPU causing an indication of the rpm suddenly having dropping to the running value rpm. Taking this capacitor out of the circuit solved that problem.


Ignition Failure = EIF Injection Failure


Beware of Only One Source for your EIF Trigger Signal

On one engine test run, I abruptly discovered that you really need two independent sources for the EIF trigger signal. This realization dawned when the lead ignition coil (my trigger signal source) failed. While I still had the second coil providing spark, the engine immediately stopped because there was no longer a trigger pulse to the EFI from the lead coil and therefore no injection of fuel.

Fortunately, this failure occurred during ground engine run. While in hindsight, a single signal trigger source is clearly an obvious risk, it was not one I had thought of before the failure. Therefore, I strongly advise anyone using an EFI in an aircraft power plant to have a redundant source to provide a signal should the primary source fail. I now use a combination of the signal off the negative lead of the leading ignition unit and the tachometer output from the trailing (second) ignition unit as a backup.

This is a Critical Failure Mode so please take heed.