Red's A-10C Pit

[brydling]

Yes, I believe the transistor was the source of their problem. They generously kept me up to date with their progress. The update is a direct result of their feedback. The symptom was the inverter stage was inverting when it should not.

 

Ok! How could the circuit work for you and not for them? Is the new solution something that you got from them or did you make your own fix? I have designed a DTS-converter myself, but I hadn't looked at your design until today when I saw on the news page that there was an error with it. Finding errors in circuits is something that I find amusing, so it got me interested :-) Our solutions are very similar, I guess there's not alot of ways to do this :-)

 

Niclas

[Mike Powell]

Ok! How could the circuit work for you and not for them? Is the new solution something that you got from them or did you make your own fix? I have designed a DTS-converter myself, but I hadn't looked at your design until today when I saw on the news page that there was an error with it. Finding errors in circuits is something that I find amusing, so it got me interested :-) Our solutions are very similar, I guess there's not alot of ways to do this :-)

 

Niclas

 

I think the circuit worked for me due to my own dumb luck. The signal levels and the normal spread of component parameters just happened to fall together in my favor. Clearly, building more units and testing them under a variety of conditions would have been of benefit.

 

Their feedback called my attention to the transistor. I understood what they did in general, but not, I think, the specific details of their solution. In any case, all kudos to them. I was unaware of the problem until they found it. I think my answers to their questions were of limited (if any) use, and they developed a fix themselves.

 

The revision I posted is the result of me taking a fresh look at the application in general and FETs in particular.

[KLaFaille]

Might as well post this here too:

YUXa1ssfEIU

 

Niclas, thanks for pointing me in the right direction! Scratched my head a bit and then after a few experiments it all fell together. Never had anything go so painlessly either, super simple.

 

Do you have any thoughts appropriate stepper motors and interfacing options available to turn the synchros? They obviously don't need much torque, but definitely need precision as they are very touchy and the slightest movement sends the ball moving. I'm guessing you can get away with a slightly less precise stepper but up the precision overall with gearing?

 

I'm totally new to steppers, so even just a general shove in the right direction would be appreciated.

 

Thanks,

 

Keith

[brydling]

Niclas, thanks for pointing me in the right direction! Scratched my head a bit and then after a few experiments it all fell together. Never had anything go so painlessly either, super simple.

 

Do you have any thoughts appropriate stepper motors and interfacing options available to turn the synchros? They obviously don't need much torque, but definitely need precision as they are very touchy and the slightest movement sends the ball moving. I'm guessing you can get away with a slightly less precise stepper but up the precision overall with gearing?

 

I'm totally new to steppers, so even just a general shove in the right direction would be appreciated.

 

Thanks,

 

Keith

 

No problem! Quite a feeling to be able to control it, isn't it? ;-)

 

Stepper motors with good resolution is hard to find. Usually you get a step size of 1,8 degrees at best (0,9 degrees if the controller can handle half-stepping). You will probably need a gearbox. The problem then becomes speed. I don't think there is a perfect solution with stepper motors since it is a tradeoff between speed and resolution, but maybe it'll be enough to begin with. I would recommend any stepper motor controller that supports the use of an optical sensor to automatically find the zero-position at every startup, except for Opencockpits' USB stepper board. That one unfortunately looses steps and quickly drifts away from where it should be. If you can find a controller that supports microstepping, you may be able to get good resolution without a gearbox and get good speed at the same time.

 

Niclas

[brydling]

I think the circuit worked for me due to my own dumb luck. The signal levels and the normal spread of component parameters just happened to fall together in my favor. Clearly, building more units and testing them under a variety of conditions would have been of benefit.

 

Their feedback called my attention to the transistor. I understood what they did in general, but not, I think, the specific details of their solution. In any case, all kudos to them. I was unaware of the problem until they found it. I think my answers to their questions were of limited (if any) use, and they developed a fix themselves.

 

The revision I posted is the result of me taking a fresh look at the application in general and FETs in particular.

 

Ok, I'm glad you got it all worked out! :-) My solution uses a DAC-circuit that can multiply by -N to N instead of 0 to 2N, so I can invert the signal just by sending a number in the lower half to the DAC. The way you do it is a little bit trickier but you get one extra bit of resolution instead :-)

 

Niclas

[Mike Powell]

The first version of the DTS used a pair a MAX543 MDACs and no inverter stage. The price of the MAX543s jumped up and I redesigned to use the less expensive LTC1590 dual MDAC. I had some friends whose application required more precision, so the extra bit from using the inverter stage was attractive.

 

Now I am considering lower precision, lower cost approaches. Many MDACs are only available in difficult to use SMD packages. A pair of discrete R-2R DACs coupled with firmware to synchronize to and track the 400Hz supply is a possibility. Most synchro engine gauges do not need 12 or 13 bits. ...perhaps a single board that provides the synchro signals and enough 400Hz power for a single gauge...?

[brydling]

The first version of the DTS used a pair a MAX543 MDACs and no inverter stage. The price of the MAX543s jumped up and I redesigned to use the less expensive LTC1590 dual MDAC. I had some friends whose application required more precision, so the extra bit from using the inverter stage was attractive.

 

Now I am considering lower precision, lower cost approaches. Many MDACs are only available in difficult to use SMD packages. A pair of discrete R-2R DACs coupled with firmware to synchronize to and track the 400Hz supply is a possibility. Most synchro engine gauges do not need 12 or 13 bits. ...perhaps a single board that provides the synchro signals and enough 400Hz power for a single gauge...?

 

Ok, interesting! I use MDAC's in MSOP package, it's not the easiest thing to solder. I have all my passive components in SMD packages too. Don't know why I did it, it takes more time to produce by hand :) I guess the idea from the beginning was to use SMD because of their better performance, but you can always use software calibration to correct errors.

 

I know of another group of sim-builders in Sweden that are developing a DTS with the approach that don't need a separate 400 Hz power supply. The DTS itself generates 400 Hz for the synchro rotor. It's not finished, but I am very interested in seeing the performance once it's done since the 400 Hz is generated by a PIC and a low-bit resistor ladder.

 

I didn't think there were a lot of synchro-only gauges out there? The ones I have seen are simulator gauges, and I believe these are even more hard to get than the real thing?

 

Niclas

Edited September 9, 2011 by brydling

[Mike Powell]

Ok, interesting! I use MDAC's in MSOP package, it's not the easiest thing to solder. I have all my passive components in SMD packages too. Don't know why I did it, it takes more time to produce by hand :) I guess the idea from the beginning was to use SMD because of their better performance, but you can always use software calibration to correct errors.

 

SOIC I can do, but I haven't tried MSOP.

 

I know of another group of sim-builders in Sweden that are developing a DTS with the approach that don't need a separate 400 Hz power supply. The DTS itself generates 400 Hz for the synchro rotor. It's not finished, but I am very interested in seeing the performance once it's done since the 400 Hz is generated by a PIC and a low-bit resistor ladder.

 

I'm looking at several approaches. There are interesting trade offs between sampling rates and numbers of bits. With the low cost of PICs, it may even be best to use multiple PICs and go PWM. One of the largest costs is printed circuit board area.

 

I didn't think there were a lot of synchro-only gauges out there? The ones I have seen are simulator gauges, and I believe these are even more hard to get than the real thing?

 

Niclas

 

I think basic, single pointer engine gauges in military and commercial A/C are simply synchro receivers with a needle on the shaft. More complex gauges with mechanical rolling-digit displays under the needle are servoed with a synchro transmitter used as a position sensor. The general aviation engine gauges I've pulled apart often have a moving coil mechanism.

[brydling]

I'm looking at several approaches. There are interesting trade offs between sampling rates and numbers of bits. With the low cost of PICs, it may even be best to use multiple PICs and go PWM. One of the largest costs is printed circuit board area.

 

I think basic, single pointer engine gauges in military and commercial A/C are simply synchro receivers with a needle on the shaft. More complex gauges with mechanical rolling-digit displays under the needle are servoed with a synchro transmitter used as a position sensor. The general aviation engine gauges I've pulled apart often have a moving coil mechanism.

 

I think I would just build a simple analog 400 Hz oscillator instead of doing it with a PIC, but that's just because I'd rather do things in hardware to keep software simple :)

 

I haven't seen real military single pointer gauges made of only a synchro receiver (except for one, but that may be a simulator instrument), but I have mostly seen swedish stuff. May be different in US aircraft. I have seen a lot of single pointer gauges for simulator use that are made with a single synchro receiver though. Aren't real engine gauges usually made with a spring-loaded coil?

 

/Niclas

[KLaFaille]

The one gauge I've taken apart, a Percent RPM gauge, seemed to be a simple rotor & stator configuration, the faster the rotor turned the greater the deflection on the two needles.

 

Niclas, you wouldn't happen to know these guys or have more information on their DTS would you?

 

NovelAir AB Simulatorteknik Digital To Synchro (DTS) usb-interface

 

Thanks

[Mike Powell]

I haven't seen real military single pointer gauges made of only a synchro receiver (except for one, but that may be a simulator instrument), but I have mostly seen swedish stuff. May be different in US aircraft. I have seen a lot of single pointer gauges for simulator use that are made with a single synchro receiver though. Aren't real engine gauges usually made with a spring-loaded coil?

 

/Niclas

 

I can't make an acurate assessment of that. I've been interested in aviation all my life, but mostly on the general aviation side as a private pilot until having children sucked my wallet dry. My time in the Air Force Reserve sadly (or maybe luckily) kept me on the ground.

[Mike Powell]

The one gauge I've taken apart, a Percent RPM gauge, seemed to be a simple rotor & stator configuration, the faster the rotor turned the greater the deflection on the two needles.

 

 

Thanks

 

Sounds like an old style tachometer with a permanent magnet three pahse synchronous motor coupled to a magnetic drag plate or cup. These tachs were feed by a tach generator which was really a three phase alternator.

[KLaFaille]

Sounds like an old style tachometer with a permanent magnet three pahse synchronous motor coupled to a magnetic drag plate or cup. These tachs were feed by a tach generator which was really a three phase alternator.

 

Could be, I don't know enough about them to be sure but you are likely correct. It's a GE 8DJ81 series Percent RPM Gauge, FWIW, and I do believe it was feed off a tach gen.

 

Edited to add that I opened it up again and I think I see what you mean by the drag plate. There is a rotating assembly driven by the coil in the rear which has on it two sets of magnets with a round copper perforated plate in between. As the assembly turns clockwise, looking from the face of the gauge, the plate turns counterclockwise and the rotation of the plate moves the needles.

 

Very neat setup!

 

K

Edited September 10, 2011 by KLaFaille

[brydling]

Niclas, you wouldn't happen to know these guys or have more information on their DTS would you?

 

NovelAir AB Simulatorteknik Digital To Synchro (DTS) usb-interface

 

Thanks

 

I know them very well and I know very much about their DTS interface. I'm one of them and I developed the DTS interface :smilewink:

 

Niclas

[KLaFaille]

LOL.. Well I guess that would make you familiar then... I watched some of the vids of the sims you have, very very impressive stuff! :beer:

[KLaFaille]

Here's a doodad I put together to hold the synchros that control the roll and pitch of the ADI. Parts are salvaged from the donor ADI and re-purposed to fit my needs. The little term board has all the connections in and out soldered to it and wired into little connectors. The three on the rear are for the signals going to the ADI, and the one on the bottom is the 26V in to feed the two synchros. The front should allow for easy gearing the interface between synchro and stepper motor.

 

Synchro Assembly (1 of 4).jpg by KLaFaille, on Flickr

 

Synchro Assembly (2 of 4).jpg by KLaFaille, on Flickr

 

Synchro Assembly (3 of 4).jpg by KLaFaille, on Flickr

 

Synchro Assembly (4 of 4).jpg by KLaFaille, on Flickr

 

Red

[brydling]

LOL.. Well I guess that would make you familiar then... I watched some of the vids of the sims you have, very very impressive stuff! :beer:

 

The DTS in the video was the first prototype but worked very well. A nice thing about it is that it shows up as an Opencockpits' USB Servo board, so that it can be controlled directly from SIOC. Since we are still using SIOC and will be doing for some more time (until we have time to develop our own scripting software), it's nice to have the DTS controlled from there too.

 

Thanks for the kind words! The Draken truly is an amazing simulator. The Viggen is in a bus so that we can travel to various flight events with it. Due to the space constraints in the bus, there has not been any development made for it since it was put there (many years back). Next summer we will make the simulator room twice as big and put the Viggen in there too. Then it will get a huge overhaul. Just an upgrade to FSX will be a huge improvement, and that will also give us much more possibilities to develop simulations of various systems in the aircraft like we have done in the Draken :) Due to the nice interfacing possibilities in FSX we have been able to make accurate simulations of both the autopilot, the radar, weapons and the various navigation systems in the Draken. The Viggen is missing both radar, weapons and autopilot.

 

Here's a doodad I put together to hold the synchros that control the roll and pitch of the ADI. Parts are salvaged from the donor ADI and re-purposed to fit my needs. The little term board has all the connections in and out soldered to it and wired into little connectors. The three on the rear are for the signals going to the ADI, and the one on the bottom is the 26V in to feed the two synchros. The front should allow for easy gearing the interface between synchro and stepper motor.

 

Red

 

Remember that you need to have an optical sensor on the shaft of the synchro so that you don't need to manually calibrate the stepper motors at every start-up.

 

Niclas

Edited September 12, 2011 by brydling

[KLaFaille]

Hey Niclas,

 

When you mention an optical sensor to calibrate the steppers, I am guessing that you mean that there will be a bit of code that spins the stepper until an optical sensor detects that it is in the proper position to zero out the pitch and roll axises on the ADI?

 

Red

[brydling]

Hey Niclas,

 

When you mention an optical sensor to calibrate the steppers, I am guessing that you mean that there will be a bit of code that spins the stepper until an optical sensor detects that it is in the proper position to zero out the pitch and roll axises on the ADI?

 

Red

 

Exactly. It will first spin until it finds the zero position. Then it will spin until it finds the sensor again, counting the number of steps in a full revolution. That way it doesn't matter what gearing you are using and what step size the motor has.

 

I know of at least one stepper motor interface board that has this functionality built-in, and that is the Opencockpits' USB Stepper board. Unfortunately that board loose steps when it passes the zero-position, some bug in the code that we have been telling them how to fix but it hasn't been done and probably will not. We have even sent them a piece of code and said "add this and it will work better", but nothing... So if you have a motor that will pass zero alot, it will have a significant error pretty soon. If you can find another stepper board with this functionality, use that one instead. For the pitch you can do with a geared RC-servo since that axis only has 180 degrees of travel.

 

Niclas

[KLaFaille]

Wait, you mean you haven't designed and built your own stepper control board too? You designed everything else you need. ;)

 

Seriously though, thanks for the info, I'll have to do some searching around and see what I can come up with.

 

As an aside, SIOC and opencockpits stuff is yet another area where I have zero experience, much like most everything related to this cockpit undertaking. Things are starting to click here and there and it's always fun learning new things, though it seems like the more you learn the less you know sometimes. :D

[brydling]

I have reverse-engineered the protocol to the USB Stepper board about a year ago, so I have begun ;) All I have to do is write a new firmware for a PIC and replace it, the hardware of Opencockpits' board is already fine. However that board is on our low priority-list, since we will replace the stepper motors in all critical instruments with DTS-interfaces or other solutions anyway.

 

The more you learn the less you know is very true ;)

 

Niclas

[KLaFaille]

Mike,

 

I was reading more into your circuit for the DTS, and was curious of your thoughts of using FET Optocouplers in lieu of an N-Channel JFET.

 

I have a few H11F3 couplers laying about that switch "on" at as low as 20uA (resistance is in the Megohm range here) and "off" are upwards of 300Megohm. However with 5 volts going across a 150 Ohm resistor into the emitter it gives a forward current of about 25mA, which equates to around 58 Ohms across the detector. Of course with more current, the more resistance across the detector drops, but as your design basically just dumps the voltage to ground when a switch is "on" and whatever resistance we have at this point is certainly less than the 10K Ohm in your schematic, I don't see an issue.

 

I know you stated you picked the JFETs you did because of their low resistance, but just how sensitive is your circuit and how low a resistance do we really need here? With the emitter being an IRLED there is certainly no need to produce a negative voltage like the JFET so the extra opamp circuitry is again removed. There may be other lower resistance OCs out there, I just had these laying around.

 

Thoughts?

 

Red

H11F3.pdf

[Mike Powell]

It should work. The higher minimum resistance can be compensated for by scaling up the resistances around the opamp. It's an issue of accuracy. Ideally you'd have the stage provide a gain of +1.000 or of -1.000. Too high an on resistance will reduce the inverting gain. The FET forms a voltage divider with the 10K resistor. The ratio of those two resistances is a key accuracy factor. However, given tolerances of other components, you'll likely not see a problem. The 13-bit precision of the circuitry was developed to produce a smoothly varying output rather than a strictly 13-bit accuracy.

[KLaFaille]

Thanks for the info Mike, that's good to know. I see what you mean regarding the accuracy and I agree that if this were built with all 1% tolerance parts there would be an issue. When I get around to building a couple of these, I'll try out the OCs and see how they work. I do like that using the OC simplifies the design, less to go wrong and all that.

 

Thanks!

[KLaFaille]

Finally got back to some 3D modeling of the pit and managed to finish the left console structure and get started on an emergency brake handle / seat height adjust housing. The main housing is done, it just needs the screw holes and the pull handle hole added. The quickly made backplates are to verify the dimensions will be correct for use in real panels and the length of DZUS rail that will be necessary. Now with the structure modeling done I have a concrete set of dimensions to work to when it comes time to build the console and I know for a fact that it will work with real panels and parts. I need to add in some small details such as nut plate locations and the such, but overall it's good to go.

 

On the part acquisition front, I found an AOA gauge, one of the Oil Pressure gauges, the Cabin Pressure gauge for the environmental panel, the proper switch for opening and closing the canopy, and a few other little odds and ends.

 

Left Console by KLaFaille, on Flickr

 

MIP and Console by KLaFaille, on Flickr