New Motor Mount for Polini

We’ve received in some engine mounts for the Polini Thor 250.  These mounts have the advantage of allowing direct attachment to the firewall — with a couple of caveats, which I’ll get to.  It’s all amazing light and strong.

The Mounting plate is CNC machined and comes with rubber bushings to attach to the motor and to the firewall.

Polini Engine Mount

View of the Polini Thor 250 mounted to a Belite ProCub.

Polini Engine Mount-2

View behind the engine, showing the plate and one of the angle aluminum firewall reinforcements.

Polini Engine Mount-3

View of the angle reinforcement on the other side of the engine.

Polini Engine Mount-4

View from the top of the engine looking down at the firewall.

Polini Engine Mount-5

View *inside* the cabin, looking forward at the firewall and the large aluminum angle reinforcement. The middle three bolts are attached to the Polini motor mount plate.

Polini Engine Mount-6

A view from below the engine, looking up.

Polini Engine Mount-7

The motor mount plate is virtually invisible behind the motor.

Here’s the caveats:

1)  This motor is so light, you’ll need to be wary of Center of Gravity issues.

2)  The installation is so small and tight, you may need to mount the motor 4 or 5 inches away from the firewall in order to have the correct placement of the propeller relative to the cowl.  This may be done by using aluminum rectangular tubing to ‘extend’ the firewall forward.    (Not described here, but eventually coming.)

Not shown is how to mount the radiator or the water overflow tank.  (We’ve previously posted many pictures of the radiator on the ProCub / Polini installation.)

Rear Fuselage Construction, continued

Let’s continue our work on the rear fuselage.

You will have two top rear fuselage plates; they have a notch in them which goes around the rudder post.  They are laminated together as we’ve done for other pieces.

After laminating the top rear fuselage plates together using a high quality wood glue, the resultant piece is sanded and then glued to the foam using gorilla glue or 2216 glue.  The edge is also routed round; this is easily done either before or after gluing to the foam.

Rudder Test Failure-9

Top rudder reinforcement plate. Note that the edges have been routed round.

The rear rudder post receives a 4″ strip of 2 oz glass on each side.  After sanding, we apply some lightweight spackle over the glass to smooth the joint.

Carbon fiber cloth is cut to fit over the top and the bottom plate.  Each piece is, in turn, epoxied in place.  The overlap area should be at least 4″.  Cut around the exit point for the rudder cable.

There are two pieces of vertical stabilizer foam:  the main and the front piece.  They are glued to the rear rudder post and to the front rudder post using gorilla glue.  The front piece is also ‘pinned’ into the rear fuselage foam using scrap pieces of 1/4″ aluminum tubing:  a 12″ piece and a 6″ piece.  Drill a quarter inch hole, clean the scrap pieces of tubing; and insert into the holes using water mist and gorilla glue.

Rudder Test Failure-10

Orientation of 1/4″ aluminum pins in front rudder piece.

The above photo shows the carbon fiber cloth reinforcement which was epoxied over the wood and fuselage.  A bottom piece is also epoxied over the bottom wood piece; the overlap area should be at least four inches.

Carbon Fiber over bottom wood

Carbon fiber cloth epoxied over bottom wood. Note cloth overlaps with carbon fiber cloth which was epoxied on the top wood. Make cut in cloth for rudder control cable tube as necessary. Overlap area is at least 4″.

After the glue sets, we route a round edge to the top and front edges of the vertical stabilizer, and then apply a 4″ fiberglass reinforcement over all edges, and also over all tube / foam glue joint areas.  The front foam piece is also glassed to the fuselage using 4″ pieces on each side and around the front.

There are three more pieces of wood to be installed in the rear fuselage.  The first is the lower front fuselage reinforcement.  The pieces are laminated together as shown:

Lower front fuse wood reinforcement

Lower front fuselage wood reinforcement laminated together.

This piece fits on the lower bottom of the fuselage, where it butts to the cabin.  The notch allows proper operation of the elevator bellcrank.  In order to install it, the foam is cut out exactly 1/2″ (the thickness of the piece), and the part is Gorilla glued in place.

lower front wood with gorilla glue

Lower front wood with generous Gorilla glue.  Note use of peel ply below to ease cleanup.

In order to make it easy to glue, and in order to ensure that the Gorilla glue didn’t get over the fuselage, the fuselage is lifted into place.

lower front wood gluing with weights

The fuselage is carefully squared to the wood piece, and held in position with same added weight.

The view on the floor:

lower front wood gluing with weights 2

The view of the gluing process. Note that the Gorilla glue is expanding outward and down, making cleanup easy.

Next we laminate the lower front load spanner.

Lower front load spanner being laminated using high quality wood glue.

Lower front load spanner being laminated using high quality wood glue.

And we place it over the bottom of fuselage, using a sharpie pen to mark areas to be cutout.

lower front wood load spanner_-2

Cutout areas marked.

lower front wood load spanner_-3

1/2″ depth waste areas routed out.

lower front wood load spanner_-4

Test fit. This example wasn’t cut as exactly as it could have been. You can do better.

lower front wood load spanner_-5

Use only 2216 glue to attach this piece.

24 hours later, after the 2216 glue has set, sand the entire surface smooth.  Large gaps must be filled with additional 2216 glue.  Small gaps (less than 3/16″ depth) may be filled with spackle.  There was a slight warpage in the wood piece, it was also sanded out.  Note that the front piece has also been sanded to uniform fit perfection.


After sanding.  Small gaps may be filled with spackle.  Large gaps must be filled with additional 2216 glue.

The next piece of wood to be installed is the trim ring around the cargo area.

The two pieces are laminated together.  Here they are, before lamination:

Cargo trim rings, before lamination.

Cargo trim rings, before lamination.

The location of the foam cutout is marked with a Sharpie.

cargo trim rings-2

Cargo trim ring shown in place with sharpie marking the route line.

cargo trim rings-3

A router bit set to a depth of 1/2″ (the depth of the trim ring) helps make a clean cutout.

cargo trim rings-4

The trim ring is glued in place with Gorilla glue.

After the glue is set, excess glue is cutout and removed.

Breaking An Airplane Fuselage While Testing…

If you want to look at some really nice photos of a failed rear fuselage on our ProCub, you need to read the following PREAMBLE first.  Here it is:


I’ve received two types of comments concerning the ProCub and the foam / carbon fiber construction that cause me angst.  They boil down to these two statements:

1)  “Push on something until it breaks.”

2)  “If you do push on the fuselage, the glue joints in the fuselage will snap.”

The first statement makes no sense; as we are doing our testing to some fairly exacting standards.  In particular, we are testing our ProCub design up to 100% of an upgraded gross weight (albeit with some modifications — to be explained elsewhere), where 100% is generally defined as either 2x (negative G) or 4x (positive G) of the force of gravity; or the sum of all calculated flight loads (which is also an interesting calculation — saved for explanation at another time).  Then, for some of the tests, we increase the test load to 150% of load.  WOW, if nothing breaks, we are delighted.  If something breaks (or bends) we have to fix / redesign / retest.  If it passes these tests, I have absolutely no interest in pushing things further.  When the tests are passed, we have real joy in our workshop, and at that point, the structure of our little bird is a Thing Of Beauty.  Ultimately, the sum of the design ends up in what we call a “Bill Of Materials (BOM)”, and that (along with our evolving manual) forms the core of a future quality control system.  We’re not there yet, but we think and talk and prepare for it.

The second statement rankles me as well.  First of all, it’s just not true; and second of all; it reflects a mental conclusion making process which puts supposition ahead of facts.  That’s not a good way to go about life.

I’ve observed Gorilla glue in operation now for several years, and I’ve learned that the glue (when properly applied) is stronger than the Foamular foam we use in our ProCub, and will tenaciously adhere to foam, aluminum, and wood.  I’ve taken a butt joined foam glue joint and had it stressed sideways; the foam snaps before the glue joint.

I was confident that if we severely stressed the rear fuselage on a ProCub, something might break, but it wasn’t likely to be the glue joints.  And I was right.  Sometimes things break — but it’s usually not what you expect!  When it does break, you learn, you apply, and you change it so that it won’t break in that manner again.

OK, I’m off my high horse now and on to the story of the rudder test failure.


We needed to perform a side pull test on the rudder in the ProCub.  I’d done this in the past on the UltraCub, but had not taken the test to a very high value of stress.  I think I posted that test somewhere on the blog a couple of years ago.  For the ProCub, we’d analyzed the maximum flight side pull on the rudder (thanks to *Peter, an amazing German engineer); and we’d determined that the side pull test needed (at standard load) to be about 210 pounds. In other words, bolt the fuselage to the concrete floor at the base of the fuselage / cabin; then pull sideways on the tail until the test is passed.  The standard form of the test (equivalent to maximum flight load: 100%) was calculated at around 210 pounds.

To do this test, we rigged up a side pull board and clamped it to the rudder.  Then, we ran an adjustable clamp strap to a scale, and the scale was attached to the wall of our shop.  I should have taken a photo before the failure, but I missed that.  Here’s a photo after the failure, and you can see the side pull board.  You can also see the break in the front of the rudder; and the odd side angle of the entire rear tail feather section.  You can’t see the cable and scale as they are nearly perfectly hidden behind the rudder — it runs from the rear of the rudder over to the wall.

Rudder Test Failure

A good day of testing produces a broken rudder assembly. (Did you think I’d call it a bad day of testing?!)

When it broke, it definitely made some noise!  Around 170 pounds of tension was released as the foam snapped in two.

Some more photos, showing the damage:

Rudder Test Failure-4

A failed rudder assembly in a ProCub from Belite. The cable which applies tension was connected to the “U” and bolt assembly on the left side of the rudder clamp bar.

Rudder Test Failure-2

Another view of the failure. The foam sheared from top to bottom. Note that the shear line occurred on the edge of a fiberglass reinforcement line. Also, it’s not visible, but the shear start point occurred right where the carbon fiber longerons terminated on the top rear fuselage structure.

We got to work on rebuilding the rear fuselage.

Rudder Test Failure-5

Cleaning out the damage. New parts await installation.

Rudder Test Failure-6

Another view of the cleanout.  The hole beside the post is where the elevator torque tube passes through the assembly.

Rudder Test Failure-7

New parts easily glued in place.

Of course, I had to make a modification to the design to prevent this from happening again.  The solution was to add a plywood reinforcement to the top side of the rear fuselage, then reinforce that with a carbon fiber wrap.  Here’s the repair process and re-engineering.

Rudder Test Failure-8

Preparing to install a new rear rudder post. A small AN3 bolt is screwed into the bottom wood rear spring brace. This area is flooded with about a tablespoon of 2216 structural epoxy before the tube is placed. (The glue is poured into the tube, then the tube is placed over the bolt. Some masking tape outside prevents the 2216 from running out. The 2216 flows around the bolt and locks the tube in place.

Rudder Test Failure-9

The rear rudder tube was simultaneously gorilla glued to the foam; then a 4″ fiberglass reinforcement was epoxied on each side of the tube from top to bottom. Then after some light sanding, we added some ultra light weight spackle to help smooth the final appearance. The top board assembly is glued to the foam with gorilla glue or 2216 glue. The sides of the board are routed so that it’s easy to fold and epoxy carbon fiber cloth over the whole thing. Sanding is then done and it looks nice.

Rudder Test Failure-10

After the carbon fiber cloth is applied, the new rudder is installed. The front rudder piece is ‘pinned’ into the foam using two 1/4″ aluminum spikes; one is 12″ long and the other is 6″ long. Drill out first with a 1/4″ long bit and then use a lot of gorilla glue before inserting each spike. After all glue is dry; round the edges and apply 4″ of 2 oz fiberglass cloth over all joint areas and all edges. Not shown: carbon fiber is also applied to the bottom board and overlaps the top carbon fiber by at least 4″.  The base of the front rudder foam is also fiberglassed to the fuselage foam.

Here’s my conclusions…

I want to have a transparent aircraft design.  I want you to learn, along with me, how the testing occurs and what changes in the design as a result.  I hope you learn how I accumulate test data and incorporate it into the design.  I also want to show the good and expected results, along with the failures.

There are several ProCub builders out there right now.  We will be supplying these new parts to them, of course.

How to install undetectable malware on a hard drive

Short and simple:  How to install undetectable malware on a computer.

In 2012, I started giving public presentations on how to hide data in hard drives in ways which are undetectable with state of the art forensic software. I did this at computer forensic conferences.  You may read a sample abstract from one of these conferences here:

and I tweeted about it here, for another event:

Nobody really cared, at least no one I knew about.

Basically, what I said in those presentations is that a variety of hooks and tools exist which allow a very technical team to install data (or programs) on computers in ways which will survive a reboot, a cold start, an OS installation, or a complete reformatting of the hard drive.

You can see my concluding slide here, where I predicted the underpinnings of the hack which is just now making news:

Service Area Hacking

Sophisticated Service Area Hacking

For the following to make sense, you have to have a rudimentary understanding of sectors, partitions, and so forth.  Assuming you pass that test, in a nutshell, here’s the explanation of how it is done:

1)  Hard drives are composed of areas which may be accessed by the computer, and another area which is inaccessible by the user.    Your data, programs, partitions, and slack area all reside in the accessible area.

2)  The drive manufacturer reserves a special area, which is often called the service area.  It is inaccessible except by the specially priviliged; and that area is used to maintain spare sectors (when some magnetic media flakes off, and the hard drive automatically repairs itself, by substituting a replacement sector from the service area).

3)  That service area also usually contains the firmware which the hard drive processor literally boots from.  In other words, before your computer can boot off the hard drive, the hard drive must boot itself and start working.  It boots off the service area.

4)  If you hack the boot code, you can make the hard drive behave in any manner you want.

5)  The boot code is just another computer program.  You never see it, because it is hidden in the service area, and proprietary to Seagate, Western Digital, Hitachi, et al… all the major hard drive manufacturers.

6)  But if you can dig into the service area, and figure out how the code works, you can maliciously change it.

Data Hiding 2

The Service Area Contains Microcode for the HDD processor.

Just about everything in my presentations was culled from existing web resources.  I made a list of them and included them in a list at the end of the presentation.  Here it is, kind of heavily redacted:


up to 3 year old list of web resources on service area hacking — not really news

The bottom line is that anyone who knows the deep basics of hard drives has already done stuff you don’t know about, you can’t detect, and you can’t fix.  Feeling small?

About James Wiebe:

James Wiebe is a digital forensic pioneer and an expert on hardware utilized in the forensic acquisition of hard drive data.  He started WiebeTech in the year 2000 and was soon designing forensic write-blocking hardware and highly portable, rugged storage systems.  He developed a list of federal, state and local customers who were among the forefront of digital investigations.  James (along with his wife, Kathy) sold the business in 2008 to CRU but has remained active within the field, and is a popular speaker and lecturer on digital forensics, especially on storage systems.  James is a 1979 graduate of Tabor College (Hillsboro, KS) and has a degree in mathematics with an emphasis on computer science.  James and Kathy have two children, and live in Wichita, KS.  James loves to fly, fly-fish, and camp in the wilderness.  In his spare time, he is developing an aerospace company, Belite Aircraft.

Polini Thor 250 Installation

We’ve provided a batch of pictures which showed our installation of a Polini Thor 250 on a Belite ultralight aircraft.  The entire picture file can be found on our flickr stream.  Here’s a sample.

Polini Thor 250 Installation in a Belite ultralight airplane

Polini Thor 250 Installation in a Belite ultralight airplane

Interested in seeing more pictures?  Go to our flickr stream and search on Polini.

— James

The Unexpected Virtue of Being a Computer Forensic Expert while building ultralight aircraft

I am a computer hard drive forensic expert.  I have had notable success with my former company,, and still enjoy helping that company (now owned by CRU) with a variety of product development and sales/marketing roles related to forensic and government applications of their products.

CRU has actively continued the development of some really esoteric forensic products.  One of them, namely, is a product called Ditto.  It does some really cool data extraction off of networks and hard drives.  But why would that be important to ultralight aircraft?  We’ll get to that in a moment.

Belite uses computers to drive our CNC machines.  The computers are in a really really dirty dusty cruddy shop environment, and I expect them to break.  And they do.  As there is no network in the shop (and precious little internet) we use a sneakernet thumb drive on a weekly basis to back up the computers.  Well… at least that’s what is supposed to happen.

Recently, I got a call over the weekend that the computer driving our Shopbot CNC had failed.  And then I discovered we had missed a few weeks of backups.  (Cough, cough.)  And then I realized that I had significant CAD work which I’d done in the last few weeks… and it was gone.  The stupid hard drive would no longer boot, but the computer itself worked fine.

This is what I did:

I took the hard drive home, and connected it to a product called “Ditto”.  Ditto reads data from hard drives which won’t boot.  (The reason being — the Windows boot code was clearly corrupted on the drive, but the data files which I needed were still OK.  And Ditto’s good with all that.)


Ditto Forensic Fieldstation from CRU / WiebeTech

Ditto saw the entire directory structure of the hard drive.

I told Ditto (via something called a “logical image” command) exactly which files I needed to recover.

I also told Ditto to save all of the recovered files in a special folder on another hard drive.

Ditto did what I asked it to.  It took maybe 20 seconds? once I’d given the command to start the logical image until I had my 240 megabytes of data safely tucked into a new folder on another drive.

I copied the saved folder back to my notebook computer.  Then I copied that saved folder from my notebook computer back to a portable drive.  Then I took the portable drive back to the shop, and *voila*, we had the lost files back on the production CNC shopbot computer.

And that is the unexpected virtue of being a computer forensic expert while building ultralight aircraft.

And now, we are really backing up our computers every week.

Interested in learning more about Ditto?  Read here:

Interested in computer forensics?  You might enjoy reading this:

Interested in getting unrecoverable data lifted off a dead hard drive?  Call DriveSavers:

(Warning — it can cost $3,000 to have data recovered off of a single drive.)

Rear tailspring assembly

The tailspring must take substantial abuse from bad landings.  It is a stout assembly, reinforced with carbon fiber.

Start by sanding each of the 6mm tailboards.  You will laminate them together in the next step.

tailspring boards

Tailspring boards, set of 2. Sand in preparation for lamination.

tailspring boards gluing

Using a high quality wood glue, clamp the tailspring boards together.

After allowing the glue to set, remove the clamps and remove excess glue.  Sand until perfect.

If you are using a composite tailspring, drill and bolt together the assembly as shown in the following two photographs:

tailspring assembly

Drilling and bolting the tailspring assembly together.

tailspring assembly 2

Another view of the same composite tailspring assembly, bolted together.

You will need to ensure that your rudder post is trimmed flush to the bottom fuselage line:

tailspring gluing preparation 2

Bottom of aluminum rudder post trimmed and sanded flush to fuselage lines.

Place the assembly into the correct flush position, and mark the location of the heads by pushing into the foam.

tailspring gluing preparation 4

Correct positioning of the tailspring assembly. Push down to mark the bolt head locations.

tailspring gluing preparation 3

Bolt head location for tailspring assembly indented into foam.

Using a flat bit, (or a sharp knife), open up the bolt head areas.  Then glue together using Gorilla glue.

tailspring gluing preparation 5

Gorilla glue applied to foam.

tailspring gluing preparation 6

Tailspring assembly held in position with weight, tape.

To be continued….