Dave's World – A WWII pit

Directional Gyro

Here's how I got the directional gyro working. As you may recall from an earlier post, I bought original (non-functioning) gauges from ebay and will modify each one for my simulator needs. I'm finding it interesting to see how the original gauges worked by taking them apart. (Before hacking them up to put in place stepper motors!) Here's the directional gyro:

In the airplane, you would turn the knob to set the correct heading while on the ground. From then on, the gauge will show your heading as you fly around. The internal gyro is the key. The gyro is essentially a heavy spinning thing mounted on an axis. When spinning, the gyro wants to hold its position. Inside the gauge, the gyro has a strip mounted on it that shows the heading degrees. Once the gyro is up to speed it doesn't want to move so when you change the heading of the plane by initiating a turn, the case and the entire airplane move around the gyro. Since you are in the plane and your reference moves with the plane, it looks to you like the heading strip is the part that's moving.

Here's the gyro, removed from the assembly. It spins by a stream of air fed through the gauge and blows across the gyro fins.

Here you can see the frame that holds the gyro with the heading strip on it. The gear on the bottom is how it turns when you manually set the heading. The heading knob has a gear that when pushed in mates with the gear on the frame and allows you to manually rotate the heading/gyro assembly.

My design is to have a stepper motor drive the gear and spin the heading gauge based on the heading of the simulated plane. I need the manual setting function to work so the design includes mounting a switch that will be activated when I push or pull the knob to rotate the heading manually. The switch will trigger a relay which will disengage the stepper motor allowing the heading strip to spin freely. (Much as the mechanical setup used to disengage (cage) the gyro when you wanted to manually spin the heading strip.) The picture below shows the completed assembly with the top half of the new back shell removed. If you look closely, you will see the motor mounted in the center geared to the gyro/heading assembly. To the right of that, the blue and black wire run to the cage switch on the front. To the left of the motor is a reed switch and you can just make out the magnet that triggers it glued to the bottom of the frame where the heading strip is mounted.

This is the inside of the case. I've mounted a switch to the original knob/gear mechanism.

Here is the knob mechanism removed so you get a clearer view of the switch and my fancy 3D printer work. It's a little difficult to see but when you pull the switch, the gear moves back towards theframe and causes the black plunger to push against the lever of the micro switch.

And finally, the back of the gauge all closed up with the new backshell and circuit board for the relay, fly back diodes, connectors,etc.

Projectors Move

As I expected, using the projectors and screen over the longer term is revealing issues to be dealt with. One thing I noticed is that fairly often I have to realign the image in Immersive DisplayPro. It seems the projectors move a little bit. I'm not sure what's the culprit but I have a few ideas. I decided to just started fixing them one at a time until I get it to stop moving over time. The first thing I tried is to stabilize the mount itself. The projector mounts I use have a safety mechanism that makes them easy to mount. There are two tabs so when you push the projector up into the mount, the tabs will hold it in place while you put two screws in from either side. The screws keep the projector from moving in that axis, but the axis with the tabs allows for some movement. In the picture below you can see where I've drilled a hole next to the tab and put a self tapping screw in. With this screw in place, it stops the movement in that axis. I'll use it this way for a while to see if it stops the movement. Otherwise, my next thought is that the wood frame may be flexing and need some re-enforcement.

In an earlier post I mentioned that the velcro I used to attach the screen material came loose from the backer board. So I used super glue to reattach it. That didn't last very long. Now I'm considering to either use a staple gun to hold the velcro to the backer board, or just give up on it altogether and permanently attach the screen material to the backer board. More on that after I decide how to approach it.

Manifold Pressure Gauge

Each gauge I convert seems to yield an interesting nugget of how things were engineered back in the day. :smartass: With the manifold pressure gauge, it's essentially a mechanical amplifier. Here's the gauge:

There is a tube connection on the back of the gauge that brings in the pressure. The pressure causes a shaft to rotate against a spring. The more pressure, the further the shaft turns. The mechanism then uses a gear ratio to work as a mechanical amplifier. The movement due to the pressure is quite small but needs to cause a much larger swing of the gauge indicator needle. By attaching a large gear (or a segment of a large gear) to the pressure driven shaft and a much smaller gear to the indicator needle shaft, a very small movement of the pressure driving shaft causes a much larger rotation of the indicator needle. Here you see the parts removed from the gauge to make room for a stepper motor:

And here after the stepper motor and related wiring installed:

And the new 3-D printed end cap complete with connector:

Tachometer

The single pointer gauges are getting a bit easier to convert now that I have finished a couple and have some items to re-use, such as connector boards, stepper boards, printed brackets, etc. The main difference now is coming in how I attach the stepper to the original gauge movement. That varies depending on how the original gauge is designed. For example in the manifold pressure gauge, it was easier to use a small mini stepper motor connected through a gear set to drive the pointer. In the tachometer, I found a nice fat shaft attached to a beefy bearing and decided to run it with a regular stepper motor. Here's the tachometer:

Removing the back half of the gauge revealed there is actually a small electric motor in there. You can see the coil and brushes in the half on the left. On the right, you can see magnets mounted on the shaft that turns.

In the front half of the gauge, there is a mechanism that drives the pointer. I haven't figured out exactly how it works but the faster the motor turns, the more pointer deflection. (Makes sense! :doh:)

Here's a look at the shaft with the nice bearing. This is why I chose to use a regular stepper motor and drive it directly. It's a bit too heavy for the mini stepper.

This is the small stepper motor I used. I've attached it to a bracket I printed that will attach to the new gauge housing I printed. The coupler on the shaft will attach to the shaft in the gauge movement and it houses the magnet I use with the read switch to mark the home position.

And finally, the completed gauge with new housing on the back half, mounted in the panel.

Next up....Airspeed Indicator. :joystick:

Like usually really nice work:thumbup:

Like usually really nice work:thumbup:

Thanks!

I've been watching your pit build for a while. You are doing some fantastic work. I think mine will not be as realistic as yours, but hopefully every bit as much fun! :joystick:

Airspeed Indicator

The airspeed indicator looks a lot inside like the manifold pressure. Makes sense because both gauges use pressure to drive the indicator. (The airspeed indicator uses the pressure from a pitot tube.)

This airspeed indicator doesn't come apart at the back. This is the first one where I've had to remove the pointer and face plate. A bit nerve racking since I didn't know if the pointer was meant to be pulled off and replaced or pulled off and re-used. I only have the one so if it's a one time use I'd be stuck. Being a resourceful guy, I found an old sale on ehay of a technical order for this instrument. By zooming in on the sample photos I was able to see the instructions indicating to remove the pointer. I took the plunge and fortunately the pointer can be pressed back onto the shaft. Here you see the mechanism with face plate and pointer removed. There is a large copper colored diaphragm at the back of the gauge which expands based on the pressure coming into the gauge. The other movement hardware in between converts the diaphragm movement into a rotational movement, amplified by the large gear ratio. The gear ratio is so large they don't need the entire large gear, just a small arc as you can see in the photo.

Because everything is mounted to the diaphragm and the back of the enclosure, about the only thing I could reuse was the face plate, pointer and the pointer shaft. Then from the outside, I was able to reuse the front half of the bezel and the glass. The rest of the internal movement and back half of the enclosure had to be designed and printed. So this one didn't present any new technical challenges to solve, but did involve a lot more designing of mounting brackets, housing, etc.

All cleaned up and reassembled

Vertical Speed Indicator

This one is also known as variometer, climb and dive, rate of climb and probably a few other names. From my days working on the S-3 simulators for the Navy, it was known as the VSI and that's what stuck with me so that's what I call it.

The VSI looks pretty much like the other pressure driven instruments except it has a gear attached to a shaft that can be turned to provide an offset to zero the pointer. For the simulated instrument that isn't needed so for my purposes it is ignored.

In converting this to a simulated, stepper driven gauge, the only new challenge was the design of the housing to fit this particular movement frame. I had some lexan left over from other work and used the CNC to cut out a new lens to replace the broken glass from the original gauge.

Thats great to see inside them, the amount of work to design and built such accurate mechanical instruments is fantastic to see.

Thank you and great work converting them to electrical, can you imagine the designers and manufacturers would never comprehend how their gauges would end up!!

Thats great to see inside them, the amount of work to design and built such accurate mechanical instruments is fantastic to see.

Yes, I find it quite interesting and a bit humbling how much they did without the use of stepper motors, integrated circuits or even transistors. They had to account for such details as the exact movement of things based on pressure and material properties and account for spring strength (which typically changes as a spring coils or compresses). And they did it without the use of PCs and CAD programs.

Accelerometer (G-Meter)

The accelerometer was interesting to see how it worked originally. Converting it to a simulated stepper driven gauge was mostly a challenge in designing a new shell that also housed all the little parts needed to make the reset function work.

The accelerometer has 3 pointers on it. One shows the currrent g load while the other two show the maximum positive and negative g's that have been experienced since the pointer was last reset.

The mechanism to handle the max pointers is self contained and I re-used it entirely as is. This way I only have to drive the center pointer with a stepper and the other pointers and reset function work the same as before. Each of the max pointers is driven when the main pointer moves in that direction. Then it is held in place by a ratchet mechanism. When the reset knob is turned, both the positive and negative pointer ratchet mechanisms are released and the two pointers are driven back by springs to the neutral position.

The main pointer (current g load) is driven by a combination of weights acting against a set of springs. As the aircraft makes a maneuver, pulling g's, the weights swing around, moving the pointer. As the aircraft g load decreases, the springs pull the weights back in to neutral. The two weights can be seen at the bottom of the next picture. The return springs are about in the center, running across the gauge. The ratcheting mechanism is at the top, just under where the face plate attaches.

The reset mechanism is relatively simple in design, but a bit of a pain to put into my own housing. It is essentially a pin on a spring loaded shaft. As you rotate the shaft, the pin turns and pulls on the reset arm of the ratchet mechanism. A couple of pins hold the spring in place and act as movement limits for the shaft. The next photo shows it all mounted in my new housing.

Next I mounted the stepper motor using a 3D printed bracket mounted to some existing holes in the frame (where I removed the weights and springs).

...and all cleaned up and back together.

Repeater Compass

The repeater compass was pretty easy. The course adjust is self contained in the front part of the gauge inside the bezel. I replaced the entire inner movement with some printed brackets and Bob's your uncle.

The original movement was driven by an Autosyn motor. This is essentially a synchro system. This indicator is just a repeater or remote compass, driven by the source housed somewhere else in the aircraft. The rotors in an Autosyn system are two-pole electromagnets and the stators are delta connected three-phase, distributed-pole windings. The rotors in the transmitter and indicator are connected in parallel and are excited with 26v, 400Hz AC. The rotor in the indicator follows the movement of the rotor in the transmitter.

Since I don't have any intention of generating 26v, 400Hz, I just removed the autosyn motor and mounted my stepper motor.

All cleaned up and put back together...

Your work is awesome!!!

Thanks a lot for all the pictures showing the inside of instruments and how you interfaced it.

Your work is awesome!!!

Thanks a lot for all the pictures showing the inside of instruments and how you interfaced it.

Thanks for the kind words.

I took a look at your pit build page. The work you've done is impressive. I like the ingenuity where you combined the metal work with wood pieces and plastic (looks like 3d printed) to get the best of all worlds.

Thanks a lot fbfan64!!!

I hope to see more real instruments conversion from you soon.

Thanks a lot fbfan64!!!

I hope to see more real instruments conversion from you soon.

I have some more coming. The temperature gauge is next. I'm temporarily side-tracked with work so it might be another week or two before I get back to it.

Great!!! I stay tuned.

Two things one you should try and sell all those removed parts from those guages I ex’s-ectopic them to beneeded to repair broken units. Second you should write a book with the step by step procedures you are doing while building your simpit since you’re doing so much detail. I’ve seen other books on sims sell to niche markets you probably want get rich but it might help recoup some of the monies spent building the pit.

Two things one you should try and sell all those removed parts from those guages I ex’s-ectopic them to beneeded to repair broken units. Second you should write a book with the step by step procedures you are doing while building your simpit since you’re doing so much detail. I’ve seen other books on sims sell to niche markets you probably want get rich but it might help recoup some of the monies spent building the pit.

Good ideas. I'll have to see if I get the time to follow up. Right now I'm so crazy with work I haven't even worked on the pit for I think a month.:cry:

Suction Gauge

Got a few days off over Thanksgiving so working out a few more gauges. I started work on the altimeter, but I had to order some parts and decided to get the suction gauge working while waiting.

Inside I find the familiar diaphragm driving a tiny spring loaded gear movement. The gear amplifies the small movement of the diaphragm enough to make the pointer rotate almost a full revolution (about 300 degrees).

Turns out there was very little I could re-use. I ended up only using the original face plate, pointer shaft, and pointer.

Oxygen Cylinder Pressure Gauge

Just by looking at how small this one is, I'm gonna predict I don't re-use much, including the housing.

Inside is a twist on the familiar diaphragm mechanism. In this gauge, the pressure from the cylinder enters the back of the gauge and goes into the flat ring that goes around the outside of the mechanism. The higher the pressure, the more the ring expands, driving the pointer. Kind of like a sideways diaphragm. Then it's back to the tried and true gearing to amplify the movement for the pointer.

Sure enough, there wasn't much I could reuse except the pointer, face plate and face plate mounting bracket.

fbfan64, did you complete the project? I am part of a team looking at building a cockpit for a museum and your information is extremely helpful.

fbfan64, did you complete the project? I am part of a team looking at building a cockpit for a museum and your information is extremely helpful.

Glad you found my posts helpful. Sounds like you have a very interesting project yourself. I am still working on my pit:joystick:, although it has slowed down a bit due to life getting in the way. Also, I fell out of the habit of making posts for my updates.:( I will try to get back in the habit of that. I've made some progress that's worth at least a post or two. I'll try to get those up within the next few days.

John, that's some really good work in your photos. :thumbup: I'm quite impressed. Your instrument face plates look great. (The dial part with the numbers and tick marks on it.) Are those engraved?

 

I'm using 200 step motors along with drivers from Pololu. The drivers can run up to 32x microstepping. I added a 4:1 gear box to smooth the motion even further. It looks really good for the most part. The only part I'm not completely happy with is very slow movement or very small movement. (Like in a fuel gauge that moves slowly over time.) If you stare at the gauge while it is at that slow movement, you can still see some stepping. Not sure if it's just my focusing on it or if it really should be smoother. (Also during flying you shouldn't be staring at the gauges :pilotfly: so I might be obsessing over this to no real immersion value.) At 200 x 32 x 4 = 25,600 steps per revolution I though I might not be able to detect any stepping but that didn't turn out to be true. So I'm still looking to see if I'm not doing something right or if I can smooth it with firmware somehow.

 

I recently got a new smaller motor that looks a lot like the ones in your pictures. They are used in lots of cars for driving the speedo and other gauges. I think they are X25 or something like that. Is that what you are using? I found that when switching to those motors (without gear box because they aren't strong enough to drive it) that the stepping is noticeably worse. Do you have any visible stepping in your motion? The X25 motors are way smaller, use a lot less power and cost a lot less so it might be worth the trade-off to use them. I watched the gauges in my car quite closely and can't see any stepping at any rate of movement so I wonder a bit how they achieved that. (No, I wasn't driving at the same time...)

 

One last note. I've only 'live' tested my instruments using COD to drive them. I found that COD doesn't update the data it outputs as often as it updates what goes on the screen. I think at best it's about 30 Hz. I'm not sure how much stepping this is introducing but I'm pretty sure it is contributing some amount. I am looking forward to when I update my software to interface with DCS to see how much that impacts the stepping. I've done a lot of electrical/mechanical work on the instruments since the last time I drove them offline with a test program so I may do that again soon too. That will be another indication of how much of the stepping is being introduced by COD.

You are having the same dissatisfaction as I have with the pointer movement on home made gauges - take a look at this link https://forums.eagle.ru/forum/english/dcs-world-topics/input-and-output/home-cockpits/278624-experiments-with-stepper-motor-drivers-for-gauges?t=276220 for some messing around I did. I am still trying to get the AX1201728SG stepper driver board to work as anecdotally it makes the movement smoother, and once I have any feedback (positive or negative) I will be putting it on here.

Can you clarify whether the steppers you are using are the X27-168 type?

Cheers

Les

 

You are having the same dissatisfaction as I have with the pointer movement on home made gauges - take a look at this link https://forums.eagle.ru/forum/english/dcs-world-topics/input-and-output/home-cockpits/278624-experiments-with-stepper-motor-drivers-for-gauges?t=276220 for some messing around I did. I am still trying to get the AX1201728SG stepper driver board to work as anecdotally it makes the movement smoother, and once I have any feedback (positive or negative) I will be putting it on here.

 

Can you clarify whether the steppers you are using are the X27-168 type?

 

Cheers

 

Les

Hey Les, I eventually got my stepper movement to be satisfactory. I use two different types of stepper motors depending on the gauge. For gauges with heavier movements (like rotating the artificial horizon internal assembly) I use what I call a regular stepper motor. These are larger 'standard' 200 step motors. I like the NEMA 8 sized one they offer at Pololu. For smaller lighter movements with just a small pointer (most gauges) I use X27-168 motors. I developed a small square PCB for mounting the motor that is relatively easy to mount in my gauge housings. This way I avoid soldering wires to those delicate legs. Then I can use either the holes in the PCB, or the holes in the motor to mount it in my housing, depending on the needs of the gauge. I prefer the x27 types over the NEMA 8 types because they are less expensive and if you don't need to rotate more than about 315 degrees, they are easier to use since they have stops built in and don't require a home switch. Because the x27 type motors are so small, they are a bit tricky. They have very little torque and the faster you move them the less torque still. There is a special run-up technique to find the initial home position. There are also acceleration and deceleration limits you have to comply with in order to avoid stalling (skipping steps). They don't step in the same 'standard' increments that the 'standard' steppers use. As I recall, they are designed around a 1/6th step rather than a 1/2 step. This means you won't get the best movement from them without using a driver that is designed for this 1/6th step. All of this is further dependent on the mass of the pointer you are driving. I was surprised how much smoother or bouncier I could make the movement just by adjusting the mass or length of the pointer. For example, lighter is not always better. Sometimes the very light pointers or very long pointers bounce making you think the movement is "steppy" But by shortening the pointer a bit and/or making it just a bit heavier you can dampen this bouncing and the movement becomes much smoother and more accurate. It's possible the bouncing can be so bad it actually forces the motor to lose a step or jump a step ahead. Of course if the pointer is too heavy, the motor starts skipping steps again...

The larger NEMA 8 motors have much more torque so while they suffer the same effects from pointer length and mass, they aren't as sensitive to it as the X27 types. They also aren't as sensitive to acceleration and deceleration rates. You can use micro-stepping to drive them in smaller increments for additional smoothing of the movement, but you may have to adjust the current to the coils appropriately (you have to overdrive them a bit) to avoid again running into skipping/losing steps.

I wasn't happy with any of the commercially available stepper drivers which is why I developed my own I/O modules for this. I have one module with drivers specifically for the 'standard' steppers and a different module designed for the X27 type motors.

Finally, no matter how smooth your gauge movement is, the actual smoothness of the indicator will be limited by the update rate of the data. For example, if the sim outputs engine RPM at 2 times per second your gauge will see big jumps in the commanded position when you rapidly increase the throttle. If the RPM data is output at 60 times per second, the incremental differences in the commanded position are smaller and less noticeable as individual steps. So you will be affected by the rate at which the data comes out of the sim and the rate at which it is then forwarded to the stepper driver.