Wind simulator fans

@lbwski - Nice install! Only thing I would suggest is a finger guard over that fan air inlet especially if there could be children around. The TerraBlooms have sharp blade edges and could possibly even remove a finger or skin/meat. Rally cars use wire screens over headlights to prevent rock damage and those are probably available in a perfect size for that inlet.

Edit - if you are interested, just search for “headlight stone guard”. There are lots of sizes, styles, mounting lugs, and finishes. Some very cool stuff and probably a good idea for anyone running wind fans if others could get a body part into air inlets. Mine face the floor so less likely to get a hand up in the inlet but not impossible. I’m going to add them to my rig as well.

Good tip! Was planning on adding some kind of guard but those headlight guards look like they’d make for a really clean fit!

In case it’s helpful for someone doing a similar project, here’s a closeup of the board I’m using so that you can see the wiring

And just one level shifter gate needed! I found these since my fans are 150mm OD at the inlet. They are listed as 160mm and others are measuring at the ID of the ring so hopefully a clean fit.

I was looking at velocity stack screens but photos of these for motorcycles and cars came up. People use them as dress pieces so a number of sizes and styles available. Was thinking about versions with mounting ears but may go for just the circle and mesh (like the earlier post photo) and print my own mounting pieces.

Have you tried it out?

Another thought while reviewing this thread - we might be able to overcome some of the lag in the TerraBlooms. There is an unused tachometer line (yellow) in the TerraBloom fans (forgot about that) that we could feed back through the level shifter and provide as an input to the Arduino. That closes the feedback loop if we monitor it.

The way the software is now it just sets a PWM value proportional to the GameDash command. For big jumps in speed like fast acceleration especially from a stop, those PWM values ramp up as the car or whatever accelerates. This somewhat slow increase in speed gets convoluted with the fan electronics time constant and slows down the overall fan response.

What might be possible is for the Arduino to monitor the fan tachometer and when the difference in desired and current fan speeds exceed some appropriately-chosen difference, send full scale PWM values (either 0% or 100%) to make the lag wholly dependent on the inherent lag in the fans and remove our own contribution from smaller changes over time in the PWM value. As the fan catches up and the difference between commanded and actual drops below that appropriately-chosen difference, the actual calculated PWM value can again be sent for precise control of fan speeds. The only time full off and on would be sent is when the fan isn’t at the desired speed by that chosen difference.

There would be some code customization required since your 8” fan is rated at 3800 RPM while my 6” units are rated at 5000 RPM. But just as the code includes a section to choose the PWM frequency, it would just have a line where you would define your fan’s rated max RPM.

I checked and this might be easy with a caveat. Arduino includes an interrupt driven pulseIn command that returns pulse durations. (https://www.arduino.cc/en/Reference.PulseIn) A problem might come if it blocks, which I bet it does. In that case, at the low end, like 30 RPM for example, that command would stop everything else (I think - though PWM outputs should keep cycling) for 2 seconds (30 RPM is 2 seconds per revolution and one pulse per revolution). The pulseIn command also includes a timeout parameter so at low speeds the command could just time out and deactivate in some maximum time period.

Anyone interested in this? I’d need to hook up my tach lines to even try it but that’s trivial. I think it might give some extra responsiveness but no idea how much.

@Zed I ended up going with the same 6" as you, mostly because it was easy to find a 6-to-4" Y splitter, 4" ducting and the 4" tapered output nozzle on amazon. Figured if it wasn't powerful enough I'd upgrade, but I do think there's plenty of wind coming out of it as is.

I like your idea; if I understand it correctly: when the target speed is more or less than (say) 20% away from the current speed, output either 100% or 0% PWM until the current speed gets to within 5% of the target. Then, set the PWM to the actual target to arrive at and maintain speed.

Based on some playing with the included control, it does seem like the fan ramps up and slows down a lot more quickly at 100% or 0% compared to something in between. You may be on to something; would be up for experimenting.

I haven't driven a lot with it yet, but my initial impressions are good. Like I said before, I always need some air flow while driving and having a source that can vary with the vehicle speed adds a nice bit of immersion.

I saw someone mention SimTools as an option for adding effects to flight sims. Will check that out.

SimTools is great for flight sims! And there are more plugins that support wind now too. It wouldn’t help you for driving much I think since apparent wind is mostly from directly ahead, but for flying, the code I use already accepts commands to control right and left fans separately and could hopefully mimic slips, skids, and turns on the ground.

And that’s exactly what I was saying on how the routine could help remove some lag. The speed control circuit has a simple R/C time constant in it and when the difference between desired and actual gets bigger than some threshold, we just bang that R/C network as hard as we can instead of nudging it the right direction. Your test results sure indicate it could work to reduce that lag to the minimum in rapid speed rise and fall scenarios - pretty much just what is inherent in the TerraBloom circuitry depending on how big or small the difference threshold is. Thanks for that test!

When I was talking with TerraBloom tech support, they said that the time constant could be shortened up some but it would stress the electronics more in dynamic applications like this. At some point it could break things so I never tried that. They had no guidance at that time on how small it could be made without hurting things. Plus it would be warranty-voiding surgery (not that big a deal) that could cost me for new fans.

I am interested in building a wind simulator now that most of my rig is almost done. I am pretty sure it will be either the Sea Flos or some powerful alternative. The TerraBlooms are interesting but the shipping to EU is not a good thing in case I want to return them if the lag is too much.

I found an alternative in Ruck fans:
www.ruck.eu their EM..EC line looks promising. I think they are quite similar to the TerraBlooms. The pwm signal they recquire is 5-10v and 1-5khz

I have to do some more digging in the wiring and components you did and see if I can do it with these fans.

Maybe it's too big, isn't it?

I want power

its these im looking at:


i cant post direct links yet as im new to this forum, but theres plenty of i nformation on these on their website.

www.ruck.eu/ventilatoren/rohrventilatoren1/rohrventilatoren-ec-motor/em-ec?filterFrequency=50

Sorry something broke on my previous post.

@Erik van Uijtert - the only thing the TerraBloom fans really bring to this are ball bearings which quietens them down and makes them last longer. Eliminating the MonsterMoto and being able to control them with the same Arduino and code was a bonus. Other than that they are pretty comparable to SeaFlo fans with the exception of the lag and a possibly valuable tach wire that may let us minimize some of that lag.

Those fans listing a 5-10 volt PWM signal is very promising since Arduino Uno uses 5 volts on its I/O lines. Ruck tech support can probably supply the rest of the control details if it’s not in their literature. Even if it can use the same voltage, current or other requirements could make it where you can’t just directly connect them, though. The TerraBlooms basically filter and average a 0 to 10V analog voltage or a 10V PWM to hit full speed. That 10V part is important for the TerraBlooms. Unfortunately there’s some quirks hidden in the TerraBloom control specification that may or may not be a factor with these. That rotor looks a lot like the TerraBloom rotors so it could be similar electronics if they ultimately source from the same overseas manufacturer. They both have identical flow vs. back pressure curves though I didn’t compare values. Hard to say.

At first blush, it looks like those fans could be an EU alternative. Kind of slick looking too. But you may need to sort the control signal if it’s oddball in some way.

Exactly the SeaFlos are too noisy in my experience. The Ruck fans are more quiet i believe 70db under heavy load? And eliminating the MonsterMoto is a good thing. Ive read they can become quite hot and risk failure because of this.
As soon as my triples are mounted im going to look more into this and probably order 1 of these fans, see what it does.

At least using an iPad sound level measuring app, the numbers aren’t too different between the TerraBloom and SeaFlo fans. I think the sound itself is somewhat different and less whine, so the TerraBlooms feel quieter but YMMV.

Avenga76 did a really good job with stick on heatsinks and a fan for cooling the MonsterMoto that had no problems at all as I burned through SeaFlo fans. MMs can work great. There’s photos of his setup design here and in other threads. There were a number of people having problems with knock off MonsterMoto units failing but that seems to be an issue with counterfeits in the market.

I may just have had bad luck with the SeaFlos. I don’t know that others have had them fail the same. The bearings were the big driver for me because of that. The TerraBlooms have a 70,000 hour (8 years continuous) MTBF rating and after a year, mine sound like they did out of their boxes while the SeaFlos were squealing and dying. Mine are/were mounted with the rotational axis vertical and that could also be a factor.

Please keep us posted. Those look like nice fans.

So I tried the Ruck fan, hooked it up to an Arduino Uno with SimHub. Worked pretty much straight forward. Unfortunately the lag for it to reach top speed is unacceptable. I don't know about the TerraBlooms lag, but this fan takes about 20 seconds to reach full speed.

It is a shame though, the fan itself is very nice & easy. It is easily as powerful as the SeaFlo's.

I sent an e-mail to the manufacturer. Ask if there is something that can be done to that soft start. I have very little experience with this kinda stuff, but maybe its something simple as removing a chip or bypassing it.

that spool up time does seem quite long. It's about 10-15sec to get from zero to max on the TerraBloom. Would like to see if there's a way to be more aggressive with the speed up, but it's not *bad* by any means.

I suspect the Ruck and TerraBloom fans are built by the same manufacturer and just branded. The rotors and stators look the same. There seem to be differences like the Ruck apparently able to use a 5V PWM where the TerraBloom needs 10V, but those could be regional preferences/requirements.

What the tech support engineer told me about the TerraBlooms is the power supply and circuitry doesn’t have lots of excess power or power handling capacity. That’s needed for fast acceleration. The TerraBloom circuit fits completely in the fan hub itself. No outside boxen. But to get more power handling for faster acceleration would mean bigger and more expensive power supplies and switching circuits, cost more money, etc. For their intended application, that ramp is sufficient. For ours it’s kind of weak.

The input circuitry in the TerraBloom just has a simple R/C circuit as the front end which forces a minimum time constant on any input voltage step changes. That limits current draw. The engineer said it was possible to change the resistor or capacitor (or both) to reduce the time constant. IIRC, he said we could probably cut the time constant in half and still be ok. The fans were adequate for my needs so I never tried it. If you get a reply, I’d bet it’s similar.

If you cut the time constant in half, that 20 seconds could turn into 10 provided it’s within the limit of power supply and power handling circuity but I’d be surprised if they tell you that you could get any more than that. Would 10s be acceptable, though? Obviously we have to defer to them on if and how much improvement you could get. It could be you can’t even do this on the Ruck fan. At least on the TerraBlooms, the resistor and capacitor are easy to spot but components are crowded around them. It would take some soldering skill to swap one out or piggyback another. Not too hard, though.

As an aside, I had little adhesive seals on my electronics covers that void warranty if they are opened.

I’m working on a code change for the Arduino that might reduce that to a minimum. I bet @Erik van Uijtert turned the controller knob fast from 0 to full which would simulate what the software change does so 20 seconds might be as good as it gets with the Ruck without a circuit mod (provided that it’s possible on the Ruck fans).

Another way, if it does have an R/C input filter, could be to use a higher voltage on the PWM during periods where the fan speed is significantly different from the target speed, at least on acceleration. (I doubt these actively brake.) That kind of force feeds the R/C filter but would be complex to implement. Better would be to modify the time constant to eek out a little better acceleration provided the hardware can tolerate that.

Another way would be to use an out of spec (high) voltage on the PWM and rely on the PWM at a certain percentage to create the equivalent of 100% after the R/C filter. That would reduce the PWM resolution while in steady state, but allow the PWM to go 100% (for the R/C force feed) when the fan is much slower than commanded.

But all of that depends on how much headroom there is in the circuitry. If the margins are tight, there might not be much wiggle room to get better acceleration.

I havent opened the Ruck fan to see the electronics. It is glued shut. No easy way to open it. I am going to return it. As you made it clear there is no easy way around this lag.
I am going with either the SeaFlos and accept the stupid noise at low speeds or I might try some pc fans with straighteners.

If you look closely in the clip i made you see me putting the fan from 10 to 100% pretty much instantly. So ye its the fan lagging. It does stop rather quick, but its not actively braking.

I may play around with that input circuit and see if I can improve the acceleration. Were you able to identify which resistors/capacitor are involved? I’m assuming lowering the resistance would decrease the delay, correct?