MickC's snail-pace build

[mickc]

So I thought I should post some bits and pieces I have been working on.

I am currently rebuilding and converting real items for use in my Sim.  I shared my wet compass build in another thread, and will continue here.


1) RMI rebuild and  conversion.

I obtained an RMI from a United Airlines 737-300, a Sperry/Honeywell RH-300C.   This is one of the predecessors to the unit fitted to the NG series.

I stripped it down and removed all un-needed parts, and went about fitting 3 separate stepper motor to handle the heading card, and the 2 VOR/ADF bearing needles.
The unit has synchro-resolver unit for these, and it was fairly simple to attach gears to the resolvers, which would operate the needles at a 1:1 ratio.

The heading card was a bit more difficult, as it is geared at 5:1.   With the parts from a second donor RMI, I was able to drive the card at 5:1, and have a small torque motor geared to the card to supply a reciprocal  1:5 speed signal to a photo-sensor so I could have a single gear train control the whole thing.

Once the 3 steppers were set up and geared properly, the 2 ADF/VOR switches were cut down and re-used. These switches use long shafts from the front panel, that clamp to gears which turn the switches.  The switch shafts are illuminated by bulbs, which send light through light tubes into the knobs.

As part of the upgrade to NG specs, I fitted 3 micro servos to operate orange flags for Heading failure and needle 1/2 bearing failure.  the servos are NewPower XL-2 srevos, which weigh only 2 grams, and only 16 x 8mm in size.

These servos turn small brass frames that the orange plastic flags are glued to.  The artwork for the flags is done in Corel Draw as a clear label.
the HDG flag is the opposite to the real NG flag, as one of the bearings for the heading card is located in the position needed for the flag, so i inverted the location.

Photo-sensors for each stepper motor were added using a combination of Tetris and wizardry

New knobs were fitted to NG specs.

The unit is run via OC Stepper and Servo cards.

The Heading flag is controlled via Prosim's IRS variable, so if IRS L is not aligned, the HDG flag appears and heading card motion stops.  FSUIPC variables for VOR and ADF signal strength control the flags for the 2 bearing needles depending on the switch positions.

Here is a video of a test of the RMI during construction. Flag graphics are not installed at this stage.

https://www.youtube.com/watch?v=Fm1HaRsru5E

Since this video, i have modified the script to make the card and needles move every 0.1 degree. in this video they all move at 1 degree steps, which makes it quite jerky.

After testing, the original cover was cut short to provide a mounting method for the device, so the short section of case it attached to the RMI with screws into tapped holes, and the whole unit it attached to the MIP through 4 brackets.

Original illumination is retained for the RMI  as well as the knobs. 

[Trevor Hale]

Speachless!  Like really, I can't believe how awesome that is.  Everyone here has some of the most incredible engineering talent that I have ever seen. 

Very Jealous!

Trev

[727737Nut]

Awesome JOB!!  Do you have any close ups of how you mounted an interrupter on the gears to pass thru the photo sensor? 

I have the exact same conversion to do myself.

Thanks
Rob

[mickc]

Thanks Trevor!

Awesome JOB!!  Do you have any close ups of how you mounted an interrupter on the gears to pass thru the photo sensor? 

I have the exact same conversion to do myself.

Rob,

Its not a very elegant solution, but the gears are epoxied to steel collars which clamp to the resolvers and motor shafts. 

I got the collars from a local R/C hobby store, and they come in various hole sizes.

http://shop.dubro.com/p/dura-collars-1-8-qty-pkg-4

For the interrupters, i used the off-cuts from LED legs as they only need to be small.  Basically I used superglue/cyanoacrylate to get the interrupter in the correct position, then epoxied over it to lock it in place. 

Photosensors are Sharp 1A51-HR units, which only need 4mm of interrupter travel into the body to trigger.

http://sharp-world.com/products/device/lineup/data/pdf/datasheet/gp1a51hr_e.pdf

I mounted them on long threaded bolts so the height could be adjusted easily with 2 locking nuts.  Once adjusted all the nuts and screws were hit with Loctite 290 to keep them in place.

The hardest part is finding the room for the sensors and interrupters without them hitting each other!  It took me a few attempts to get it right 
On some I glued the interrupters into one of the set screw holes in the collar, and on one I had to put a clearance bend into and glue directly to the face of the gear as shown below.

After these pics they got another coat of epoxy and the ends of the interrupters were painted matt black.

[Caflyt]

Good Gracious!
And I thought I was doing well to fashion a stinkin' window frame in my sim, I have a long way to go before I visit the lofty perch you occupy!
Simply amazing what our group of folks is capable of.
I continue to be inspired by work like yours Mick!!!! (And all of you other crafty fabricators)

I sometimes get a little disappointed and discouraged when I see posts from those who are "selling the sim" and seem to have lost their passion for our shared hobby, but I'm quickly reminded I still have guys like you to continue to motivate me to complete this damn thing and start enjoying the "flying" instead of just "building" all the time....(which I secretly really enjoy)!

Thanks Mick!!!!!

[727737Nut]

Very Very True, I have built and sold a few sims and now i'm building 2 at the same time!  LOL  I really like building now as much as flying them!  Scott W. was my motivation for having 2 sims to choose from. Helps fight the boredom from flying only one type all the time. 

Rob

[XOrionFE]

I am also in complete awe.  This is incredible work.  Inspirational.

Scott

[Nat Crea]

UNREAL Mick!

I bought a real RMI 4 years ago..actually...Axel Grau brought it down to the 2010 open house for me.
I opened when I got home, passed out, and then slipped the cover back on 
Hooked up the back lighting but that was it...hats off to you.

Nat

[wildcat02]

Mick,

Great job !!!

[mickc]

Thank you all for the comments, they mean a lot, especially from the caliber of people on here.   



_________________________________________________________________________________

Next item is part of the rebuild of my yokes, the trip reminders or "Memory Device" as it is referred to in Boeing parts catalogues.

As you can see, they were in pretty bad shape when I got them.
All the numbers had just about disappeared, and they were full of crud.

I stripped them down, not much to them really.  I cleaned everything up and re-primed and painted everything.

The numbers are printed on a clear acetate film which wraps around the thumb-wheel's clear centre section. this is then illuminated from behind.

I drew the number labels up in Corel Draw, and after about 43984 test prints to get the spacing right, i had them printed onto light-box vinyl and die-cut.

Inside the body is a thin spring steel plate  that is used to both provide the detent for the wheels and acts as a contact strip for the bulbs (a bit hard to see in the photos.)   These had lost their "spring" over time, so I used some stainless steel shim stock to add some pressure to them and tighten up the detent. 

I also sourced new bulbs as a couple had blown.

Reassembled with a light smear of silicone grease, and everything is cool

The test of the light in these pics are done with a 9v battery, so they look quite bright.  I did this to make sure the heat from the bulbs was not going to affect the vinyl numbers.

I ended up using 3 of these units to make 2 good ones, as some of the spokes on the thumbwheels were  broken off.

[Trevor Hale]

Holy CRAP!  Big difference there.  Love the look of them now though

[geeman47]

That was like restoring a 1968 Mustang or something.  Great job. 

Those light bulbs look to be the same in many ATC radio selector panels.  I've replaced a few of those when our maint. guys were not looking (hated to bug them with the little tasks that we could do and let them do the serious items)

Great job Mick!

Only 43984 test prints?  Would have taken me many more LOL

[mickc]

New victim has arrived

More details to follow

[mickc]

So with a bit of research and tinkering, it looks like this rudder trim indicator is going to be quite easy to convert.

The unit runs off 28VAC, and has its own internal circuit board and transformer, which supply a small DC motor depending on the signal provided from the trim actuator in the tail.  This motor turns a drum that has the indicator tape attached to it.  The drum is mounted on a potentiometer, which provides feedback for the position of the tape.

There are roll pins in the motor shaft that are user for full travel stops, and they contact adjustable plates attached to the drum. The motor uses a simple plastic tube to transmit drive, which acts as a slip clutch when full travel is reached.

Interestingly, the "OFF" flag that shows when there is no power or position signal is not a traditional flag, but is actually printed onto the indicator tape.  There is a huge capacitor on the circuit board that stores enough charge to move the tape to the far end with a loss of power, which will allow the flag to come into view.

After some spec checking and measurement, i found that i could control the motor with the guts out of a large RC servo, and using the existing potentiometer to provide the feedback.  the unit fitted to the indicator is 30k.

I quickly bodged it into the circuit and connected it to a Pololu servo controller and it works great

Here is a dodgy test video, just using the position slider in the Pololu software.

http://youtu.be/o9b06HpwTqM

I set it up in Prosim and it follows exactly as it should.

I will however, be setting this up in SIOC and bypassing Prosim for the following reasons:

Prosim only goes from 0-10 units each side, the real A/C goes up to around 16 units each way.
With SIOC I can make the script perform the "OFF" flag functions.

I have removed all the un-needed components and will mount the servo controller and driver board in the open space.

After it is all mounted up, I will just need to replace the bulbs in the light board, and repaint the front bezel, then i will post the completed unit.

[fsaviator]

That is pretty interesting.  I'd never seen the inside of one of these.  It is clear that it is actually easier to modify a real one in this case than to actually build one.  Especially when you consider the indicator is actually mounted on a tape instead of as a needle that needs to be moved.

[Karl737]

Great post Mick 

I grabbed one of these as well. You have got me motivated to go have a play to see if I can getting working now...

Cheers,
Karl

[mickc]

Next chapter!

So with my RMI conversion being so successful, I have been toying with the idea of converting more instruments.

I personally don't like the look of the Flight Illusion standby instruments, but there aren't many alternatives.

My initial plan was to have the real RMI that i converted, and an ISFD, but have now decided that i need more of a challenge

I am ditching the idea of the ISFD as it is quite a simple thing to reproduce, and i have managed to source the final 2 gems in the 737 standby cluster...

I present my next few months of headaches:







Interfacing of these will be quite a challenge, as the A/H has the ILS deviation bars and flags to convert, and the Altitude/ASI is all mechanical with bellows and pressure chambers.

On the positive side, i will get that lovely sound of the gyro spinning up with power applied (everyone who has been in a real cockpit will appreciate this sound )
Also the ALT/ASI will make that familiar ticking sound to add to the cockpit ambience.

[wpslider]

Very nice indeed!

[mickc]

So.......














Just about complete,  just need to attach the Yaw damper label, fit the proper OEM knobs to the RMI and use the proper set of screws for the instrument clamps.  ( they were drying after being painted, so the ones in the photo are just fill-ins

My camera is pretty shit at taking photos in the dark, so the backlight shot is pretty washed out.

[quid246]

Was that post a teaser Mick?

[mickc]

Was that post a teaser Mick?

Sort of, trying to get imgur to work!

[fsaviator]

Very nice work indeed, Mick.  You are raising the bar on builds.

[mickc]

I have had a few requests for how this is all interfaced, so i will post it here and copy it to the other places it is posted

Yaw Damper:
This unit was purchased from the legendary Ivar Hestes. It was originally a military version, finished in military black with white backlighting. (bulbs were made of blue glass, giving off a white backlight)
It was stripped down and repainted to the correct Boeing grey and the movement was disassembled to replace the bulbs on the backlight PCB.
No modification was needed to the rest of the gauge, and it is driven by a Phidgets Analog 4 Output Card via limiting resistors (see below for details)

Artificial Horizon:
This was obtained through Swiss Sim Shop along with the Altimeter/Airspeed unit, after getting him to track them down for me.
The unit was stripped down and the gyroscope assembly was removed. Servos were geared up to control the pitch and bank of the ball.
Servos are driven by and OC Servo card and SIOC script. I know I could have used a Pololu/Phidgets servo card, but i have a script that simulates the gradual "tip over" and "self erection (lol)" of the gyro after powering off/ON, and
and the gyro caging mechanism has been removed and replaced with a microswitch, so when you pull the caging knob, the ball will bring itself back to the straight and level position.
The gyroscope was removed as it would be extremely difficult to try and drive the ball with the forces of the gyro acting against it.
I did retain the gyro itself, and the inverter board that runs it. It is mounted in a frame outside the gauge so you still get that wonderful sound of a real gyro spinning up when power is applied.
I retained the OEM logic board for the ILS & backcourse indicators and the flags, and these are all run by the GS and LOC signal valid FSUIPC offsets.  Gyro flag is driven by the inverter board that powers the gyro.
The unit needs 28vDC for gyro power (the inverter board converts this to 115vAC 3 Phase to run the gyro motor) also needs 28v for the Needle bias.  This powers all the flags and needles to hold them out of view when power is applied, and no ILS mode is selected on the front panel.  When an ILS mode is selected, the G/S and LOC flags will drop back into view, and once a valid signal is received on either channel, the flag will retract and the deviation bar will drop in.

The deviation bars are driven by the same Phidgets Analog card that the Yaw damper uses, again with limiting resistors. 
The Phidgets card can put out +10/-10vDC but the deviation bars need only +/- 1vDC for full scale deflection, so adding the resistors gives a lot more flexibility in calibrating the gauge output.
These are simply calibrated in Prosim like any other gauge.
There is one issue I still have, where Prosim still outputs a signal to these bars even with battery power off.  I will ask to get this rectified, but I can always put a DPST relay in line with the outputs if this doesn't get fixed.
The flag / FSUIPC signals are all done through Prosim as well, using 2 of the "User In" switches set to the FSUIPC offsets, and the corresponding "User Outs" to a Phidgets 4 Relay Card.
This gives full ILS and backcourse functionality, with all the flags and bars working as per the real aircraft.

Altimeter / ASI:
This was the hardest to convert, as it is a purely mechanical device.
The only electrical parts in this unit were the lighting, and the electro-vibrator mounted on the frame.
The Altitude and airspeed were all bellows-driven with tiny gears and linkages, it was actually a shame to pull most of it out.  The altimeter drive was converted to use a small Pololu DC motor with a reduction gearbox to drive the needle shaft.
This shaft was geared at a 5:1 ratio with a Bourns 10-turn potentiometer for position feedback.  A Pololu JRK motor controller is used to drive the motor.
With the 12-bit resolution of the JRK controller, there are 4096 steps and the physical range of the drive is from -1000 to +48000 feet. THis gives an accuracy or around +/- 15 feet over a range of almost 50,000 feet.
The Barometric pressure adjustment has been uncoupled from the needle movement, but the gear train that operates the QNH number scale is retained and is coupled through to the back of the instrument using a complicated double universal joint linkage, and is connected to another 10-turn potentiometer
This is connected to the OC servo board above as an analog input and SIOC uses the position to adjust the standby QNH offset in FSUIPC. The standby altitude is read from FSUIPC, then calculated by SIOC back to a number from 0-4096. This is assigned to another "User In/Out" pair in Prosim, and sent to the JRK controller.
Airspeed is driven by a Micro servo, after 4 different attempts to get the gearing right, i ended up 3D printing some custom sized pulleys and used a Viton rubber O-ring between the pulleys to drive the ASI drum.
The frame vibrator and its oscillator circuit are still used, so the characteristic ticking noise is present when battery power is applied.

RMI:
The RMI is driven by 3 stepper motors, one for the compass card and one for each needle. Home position photo sensor are fitted to each of them Servos are used to operate the flags.

OC Boards are used for interfacing, and SIOC runs everything.  Prosim provides a signal that the Left IRS is aligned, which pulls the HDG flag, and enabled the compass card.

SIOC uses FSUIPC offsets for the NAV and ADF radios, and the VOR/ADF switches are connected the analog inputs of the OC servo card. 
The script checks the switch position for each needle, then pulls the flag if the signal is valid, and allows the needle to indicate the direction.  If the signal is lost, the flag drops in and the needle returns to the 90 degree position.


I am looking at replacing the SIOC functions with Arduino at some stage.

I will also post a step by step of the conversion for each of these in the near future, but for any specific questions please feel free to ask!

[Joe Lavery]

Talk about raising the bar, I think you shot it into space 

What fantastic attention to detail, I only wish I had the knowledge and skill to attempt something like this.

I look forward to the blow by blow images... Mick you really are a inspiration to us all.

Regards
Joe.

[mickc]

Next component up for conversion has arrived!