Scott-S6

3D model fine. Exported display in 3D model aircraft fine. Popup display in other aircraft - immediately dark when summoning popup. Exported display in other aircraft - immediately dark when summoning popup.

Also happens when just exporting the NS430 display.

Yep, that thread was one of the ones that I went through for research before building my ejection handle.

Can confirm that it's RAM related. Stuttered badly when moving at speed with 32GB RAM, dedicated NVMe drive for DCS, dedicated NVMe drive for swap. 64GB highly mitigated the problem (stutter is now very occasional).

The trick with three buttons is that you have to ensure that the actuated switch is reset before the next is actuated. The timing of that is important. Or, you could have a saw tooth arrangement that presses the same switch three times during the ejection handle stroke. Both a bit messy imo. The electronic/software/programmable solutions all leave the handle mechanism extremely simple which I think is desirable considering that it needs to be pretty robust.

They're virtualfly ruddos. The pedals run on linear bearing skates and the toe switches are force transducers. Fancy stuff but all off-the-shelf components so very easy to repair should I ever manage to wear them out. (The whole springing mechanism and all the stuff behind the metal enclosure is DIY) A damper is fine for fixed wing flight as well (desirable even, it can stop oscillation around the sprung center), as long as it's adjustable. I plan to have the damper set very low, just to stop drift. At the moment the pedals run so smoothly in copter mode that if you tap a pedal it'll glide all the way to the stops.

Here's a link to that particular 555 unit but there are lots that are suitable. AliExpress link If you aren't in the UK the relay link won't be very useful to you. eBay link Virtually any 5V relay will do, 10A switched current or less (the lower the better). Get whatever relay and if you're not sure about the pin layout post a pic and I'll explain how to read the diagram. Given the very low voltage & current at play together with the low frequency, back emf isn't really a concern here but it certainly wouldn't hurt to stick a diode on if you wanted. Leo's boards use 3.3v for sense voltage. It'd probably work if you skipped the relay, they're pretty resilient boards, and you could throw a resistor in there to drop the voltage but it's a potentially messy scenario that can be completely avoided with the relay.

Thanks, but I don't want to pull this thread off topic. I'll start a thread with my rig at some point... I think your idea of a throttle that slides left to right is a good one for people that want to keep the throttle in use. Pretty easy to do, just needs a pin or thumbscrew to locate it. I might do that myself, just because more buttons is never a bad thing. One thing that is relevant to this topic though. Part of the switching between jet and copter is pedals. I wanted to be able to disengage the sprung center. This is a prototype that's working pretty well - removing the two thumbscrews disengages the springing (the bar pivots around the center on roller bearings and all four eye bolts have spherical bearings). Still to do is making some adjustable travel stops for copter mode and brackets for a damper. Those aren't virpil pedals, obviously, but the idea is highly transferable.

That would have required compromising on the ergonomics (pushing the throttle out to the left) which, for me, wasn't a good tradeoff. I have sufficient panel functionality that I rarely use the buttons and switches on the throttle base anyway and I have the rear mounted throttle quadrant for helicopter engine management.

My solution was to mount both the throttle and collective on locking hinges. Very quick and easy to switch between them. The extension on the collective is because I am hoping for an Apache grip... (Of course, I am using a custom restrictor plate)

The problem - DCS requires the ejection command (button or keystroke) to be triggered three times for ejection to fire. This sucks for people that want a proper ejection handle because you definitely don't want to be pulling on it three times to get an eject. The other solutions - there are definitely other solutions to the problem, the two most commonly suggested are: Software - there are apps that will make joystick buttons pulse. If you're using one of these apps for something else then great. If not, adding an additional app just for this isn't a good solution in my opinion. Minimizing the amount of crap running in the background is a good idea generally. Arduino - definitely a viable solution but total overkill on cost and complexity if you're only using it for the ejection handle. If you've already got an Arduino in your cockpit doing other things then you should definitely go with that (but you probably just did that rather than searching and finding this guide). This solution - we're going to use a 555timer circuit (bought as a pre-assembled package) as an oscillator to make a relay flap open/closed triggering repeated button presses while the ejection handle is pulled. I know that sounds kinda complicated but it really isn't. No knowledge of electronics required. The disclaimer - I'm assuming that you can strip wires and join them together (soldering, terminal blocks, crimped connectors, whatever - I don't judge). If you aren't sure that you can do that then this is not the guide for you. Learn how to do that first. I am also assuming basic electrical knowledge. Don't go hooking this up to the mains and then blaming me when your house catches fire. High-Level design This is how it's going to work: Power source (for this guide, USB power) to Switch in the ejection handle to 555 board to Relay to Button Board When we pull the handle the switch connects the 555 timer board to power. The output of the timer board pulses on and off which causes the relay to open and close. The relay's output is connected to the button board so registers as a button press in game. Things you'll need: USB Cable Switch 555 board Relay Button Board (note, I'm referencing prices below in USD because that's easy for most people to relate their local currency to.) USB Cable - There are lots of things you could use for power but I'm going to go with USB for this guide. That means you need a USB hub (if you're building a cockpit I'm sure you've got a pile of those) or you could use a USB charger, since we only want power, which I'm also sure you have lying around. You also need a USB cable with one end suitable for your hub/charger. Ideally, use a power-only cable (only two wires). You could just ignore the data wires but I dislike having data wires floating around not connected to anything just in case they get shorted on something. Switch - This is a low voltage, low current application so any switch will do. The design of your handle mechanism is going to determine the type of switch you need. I went with a magnetic switch, the type used by security systems to tell if a door is open or closed. They're very cheap, $2-3 for a pack of 10. 555 timer - You'll find these extremely cheap on AliExpress and similar. Expect to pay $1 for the unit and $2 for shipping or thereabouts. There are lots to choose from but there are a couple of things we're looking for. We want 5V to be within the range of the input voltage offered and we want one of the frequency ranges to be suitable. What's suitable? We want ejection to go off in a fraction of second but we don't want the button to be getting pressed thousands of times a second or it'll get rejected as erroneous data. 10-20 times a second is perfect so we want 10-20Hz. The unit shown in the pic has 1-50Hz as one of the frequency ranges so that's perfect. The 555 board looks complicated but it's really not. Key things to look at: The jumper with four positions, that's where we pick frequency range. Make sure it's in the 1-50 (or whatever your's has that is suitable) position. The frequency adjustment screw, we're going to turn that to make the on/off faster or slower. Indicator light, not essential but useful for troubleshooting. VCC - this is where the 5v from the USB goes GND - this is where the ground from the USB and from the relay coil go OUT - this is where the other side of the relay coil goes Relay - The only things we need to worry about here are that it's a 5V relay (5V and 12V are the common voltages) and that it's small. I paid $3 for one domestically but they're much cheaper bought import (if you want a few) We don't want a beefy relay because it will take more power to run the coil and our 555 board may not be up to it. My relay is a 10A max output and it's nominally too much for the 555 board but it works fine. Definitely don't go bigger than 10A Lots of pins on the relay but the diagram explains it for us. Pins1 and 2 are the electromagnet. We're going to connect them to GND and OUT on the 555 board. Doesn't matter which way round. Pins3 and 4 are connected when the electromagnet is powered so we'll connect them to our button board. Doesn't matter which way round. Pins3 and 5 are connected when the electromagnet is NOT powered. We don't need that here but I mention it in case it is useful to someone for another project. Button Board - there are lots of these and I'm going to assume that as a cockpit builder you already have one. I like the ones that Leo Bodnar makes but use whatever you prefer. Total cost for these parts (assuming that you have a button board and a USB port) = ~$10 This is how everything hooks together: Red = positive Black = ground Blue = doesn't matter This does not indicate what color your wires will be. If you cannot look up the colors for the USB cable or use a tester to verify them then, again, this is not the guide for you because you should have basic electrical knowledge. And that's it. All you have to do now is tune the frequency adjust screw to something which gives you an acceptable delay between handle pull and ejection. As is mandatory, here's what mine looks like. You can see that I've used a piece of plastic board to isolate the circuit board from the metal panel and to keep the 555 board, relay and cable anchors together. You could, of course, put this into an enclosure but I like having easy access. YMMV. And here's the handle itself. Before anyone asks, the wire loop is a real martin baker part (they aren't expensive because making the metal block that the ends plug into is a bit of a hassle. Filing square internal bores isn't fun. If you have a rotary broach then you'll find it rather easier). The high position is because I fly sidestick, I based the handle position on the Rafale. EDIT - pictures now attached rather than linked.

This would be extremely useful.

Brilliant, thank you!

I am able to specify the location I want the NS430 display to appear - no alt-enter, no dragging and resizing. In the NS430\cockpit\scripts folder (note - not the aircraft specific folder) find the NS430_init.lua file. Add the usual dofile viewport handling and try_find_assigned_viewport lines you'd use for any other export (Lines 2 and 7 in my file). And, of course, add a viewport with matching name to your monitor config file with the co-ords you want, again just like any other export. For aircraft with cockpit integration the NS430 display will appear at your specified location automatically. For aircraft without, it'll appear at that location when you trigger the close look command instead of the default bottom right which is often off-screen on multi-monitor setups. (I'm running 5 monitors and have the usual un-utilized strip where the display appears normally) ETA - this only exports the display itself, not the display and bezel so it's not useful if you want to control the unit via mouse. Very useful however if you have some sort of button box to control the unit. NS430_init.lua

2x GTX 780

The problem does not seem to be related to the number of objects in view. If I create a mission out in the middle of nowhere (flat land, no trees, no bushes, no buildings, no vehicles) my single screen performance jumps up as you'd expect. In multimonitor mode it's still the exact same.

I've just been testing flying in the middle of nowhere vs. flying in the Black Sea free flight (air start) - lots of buildings, LOTS of trees. Flying in the middle of nowhere with one monitor gets me 140+ vs. 80-90 with lots of objects. In either location 50fps with two monitors, 30 fps with three monitors. Thanks for confirming that it's not just me! I guess that TH2Go is the fix for this - then DCS thinks that it's only one monitor?

I know this is a necro post but I'm seeing the exact same issue. I've been doing a bunch of work to resolve stutter and latency and am pretty happy with the results. However, I too see what appears to be framerate bottlenecking. My setup - i7-3770K. 2xGTX780, 16GB With a single monitor 1920x1080@144Hz My frame rate moves between 80-100fps (settings on high with HDR off, MSAA on 4x, TSAA off) depending on where I'm looking (slower looking at the horizon, faster looking straight down or at the sky, as you'd expect). MFD extraction (onto the screen) costs me 2-3fps. Changing tree draw distances, water, etc. affects frame rate as you'd expect. CPU core usage sits around 10/10/30/30% with core peaks at 16/19/68/71%. RAM usage is below 50%, VRAM usage is bang on 50%, GPU utilization is around 50% (on each) peaking at around 80%. Given that hardware does not appear to be getting maxed out (and drilling into more detail - VRAM controllers, etc.) supports this) I appear to be hitting the scaling limits of the game engine but there does not appear to be any throttling happening as the fps responds to the on-screen view and to the detail settings as expected. If I use two monitors then I get 48-50 fps regardless of where I look, regardless of MFD extraction, regardless of Helios running. CPU and GPU utilization drops slightly. Frame rate doesn't change if move trees from 1500 to 9000, if I put water on high, if I enable tree shadows. If I use three monitors then I get 29-30fps regardless of where I look, regardless of MFD extraction, regardless Helios running, etc. Frame rate again doesn't change if I change detail setting. The lack of effect of detail settings (same FPS with zero tree visible vs. hundreds (thousands?) of trees on screen!?!) looks very much like there is throttling or bottlenecking somewhere - either in DCS or in the Nvidia hardware/software.