Jump to content

Update suggestions from a real life CRJ-900 pilot


Recommended Posts

First of all, I want to congratulate Mathijs, Hans, and the rest of the Aerosoft team for developing such a great aircraft.  It blows me away that you are able to replicate a multi million-dollar aircraft and its systems so accurately for an $80 product. I love how involved you still are in making this aircraft even better.  I see you responding to questions and concerns on the forum within hours, and that’s a big reason why I am taking the time to write this.

 I have been a flight simmer for 20 years. I fly the CRJ-900/900NG for a regional airline in the real world, and I have had an opportunity the last few months to fly the real thing and then come home and fly the Aerosoft CRJ-900 version immediately after and compare the two.  This has given me a great perspective on where the differences lie between the sim and the real thing. I have compiled a list of small things I notice that could be tweaked in the next update. I am not a programmer, and I am sure a lot of the things I am going to mention here are things you would love to fix but can’t because of programming issues, development costs, sim-ism’s, differences in my system ect.  Maybe some of the things you haven’t thought of, or they weren’t caught by the original beta testers.  Anyways, this is not me complaining about the aircraft at all. I really love it and you guys did a great job. It is simply me trying to help you guys create the most life like and accurate aircraft possible with this next update.  For the record I am running P3D V4.4 with the latest CRJ update. I use Navigraph to update the FMS database and Pilotedge when I fly.

 I am going to break this down into two categories.  Aircraft handling/characteristics and FMS/System differences.

Aircraft handling and characteristics.  Only two big things:

I would say the biggest difference in how the real airplane behaves vs the sim is in decent. The real CRJ-900 does NOT want to go down and slow down.  It is legendary how slippery this airplane is and managing the decent is really one of the biggest challenges of flying it.  Descending from FL360 at 3000FPM with idle thrust you will still be accelerating all the way down and will have a hard time keeping it under 300kts.  Speed brakes are pretty much useless until below FL250 because of the thin air and how small they are.  I find the speed brakes in the sim WAY too effective. In the real aircraft I use a 12/12 rule to make 250kts at 10,000.  The rule is at 12,000ft select 1200fpm (idle thrust/no speed brake) that will get you from 300-290kts to 250kts barley in time at 10,000.  That gives you an idea how bad this thing doesn’t want to slow down.  I find in the sim if you chop the thrust at 3000fpm you immediately start lose airspeed quickly. Most of the time I find myself having to use 50-60% thrust to mimic the rate and speed the real aircraft descends.  In real life I am always at idle thrust while in decent and it still wants to haul ass.

Second, I find the Sim aircraft performs too well in cruise at high altitude.  I realize there are a lot of factors that determine this, but I usually have a similar fuel, pax load and altitude as I fly in the real world and the performance is much different. When you programmed the sim aircraft, I imagine you used Bombardier’s performance numbers which makes sense.  I imagine these are achieved on brand spanking new engines in ideal conditions.  A real world CRJ-900 fleet that is well maintained will have tails that average 8-10 years old.  Those engines can get tired over time. In the real jet, in cruise from FL340-FL390 you will need 87-89% N1 and a fuel flow from about 1800GPH-1950GPH to maintain .78.  It is very very sluggish at altitude, especially anything over FL360.  Any small power changes can take 30 seconds to even see much of a change in airspeed. Not responsive at all. And to barber pole it up high would probably take everything it’s got.

 

FMS/Systems

Altitude windows on RNAV stars:

when the FMS is loaded with a STAR, and that STAR has an altitude window at one of its waypoint altitude restrictions, its is displaying the top of the window altitude or above (A) as the target for VNAV. Example: a window between FL230 and FL200 the FMS is showing FL230A as default. As I’m sure you are aware, the real FMS shows both the top and bottom of the window like this FL230B/FL200A. I’m sure this was simplified to keep development time and cost down in the sim which is fine. While the real-world aircraft’s default VNAV target is still the top of the window (non NG) its not correct that it is an “above” altitude.  If you’re not going to display both top and bottom altitudes of the window on the legs page, it makes more sense to have it default display the bottom of the window altitude and Above (FL200A in our example).  Often, we will just hard code an altitude between the two windows to get more effective vnav info.

  • The RTU volume knobs work backwards. Pushed out is muted, pushed in is on.
  • All altitude and speed changes on the legs page should require an EXEC to take effect.
  • Legs page Missing “cancel mod” option to discard any changes you don’t want to make anymore
  • The last page of the legs page should be for the alternate flight plan
  • The speed trend indicator is much too jumpy. In real life it is silky smooth
  • FO’s side FMS Direct Intercept page doesn’t display the modified vertical path programmed on the legs page of the captain’s side.  We often have these two pages up at the same time so we can see feet per minute requirements of waypoints further down the STAR. I don’t think this page is syncing properly with the other side.
  • Direct intercept FPM rate is jumpy. Its changing at much to fast a rate and its hard to read.  This is a very important page we use for decent planning. It should only show changes in hundreds of feet. It changes slowly in real life and is easy to read.
  • When hitting NAV mode on the FCU the FMS1 goes green right away without it arming at all.  In real life, even if you are right on the lateral path the FMS takes a second or two to turn from white (armed) to green (captured) this triggers very important standard calls we make, and if it doesn’t arm at all first its not realistic.
  • I find the aircraft ALTS CAP a bit too early (I’m being super picky) I know it depends on the rate of accent or decent but the real thing usually doesn’t ALTS CAPs until 100ish feet below and 300-500ft in decent.  Once and awhile it will stop climbing before I reach altitude as well and I have to VS it to complete the climb.
  • The VNAV snowflake is much bigger on the real display, and I find it hit or miss if I get it in the sim. Same with a TOD point.
  • I’ve had a problem flying the SSTIK 4 RNAV SID off Runway 1L/1R in KSFO.  The FMS makes the left turn at SSTIK but instead of heading to PORTE it starts doing 360deg turns after SSTIK.  This might be a similar problem some people have reported about the FMS getting lost sometimes and making high rate and high bank angle turns.  I have had this happen to me.  The real plane does not bank like this at all with autopilot on and with ½ bank selected 15 degrees of bank is the maximum it will do.
  • After an approach is loaded and executed, the hold should be automatically be loaded in the hold page.
  • The Fuel pumps should operate fully independent.  A lot of times we shut one off for single engine taxi
  • HCP values should clear when you start entering a new elevation or distance.  You don’t need to delete what’s in there first.
  • Might just be my system but the sound cuts out momentarily when I enlarge the FMS.

 

That’s it for now. Sorry for the length, but I am very passionate about this airplane. If you make these changes you will have an extremely life like simulation of the real thing.  If you have any questions, comments or would like to add me to the beta testing group I would be happy to send you my credentials to prove I am who I say I am. please contact me and I will be happy to help out!

 

Mike

Link to comment
Share on other sites

Mike, thank you so much for taking the time to explain some differences to us sim pilots.

I will try the 12/12 rule right now 🙂

 

Do you fly the descent always with "fixed" FPM, in your example with 3000ft/min?

 

The Sim-CRJ is very sensitive to speed during SID and STARS take-off. The CRJ can do a turn with 190KIAS, for example, but with 220KIAS it loses focus and turns off course with all the problems we've seen before. 


Two Questions:
How would the real CRJ fly off a SID as seen in the picture?
Airport:LEBB, PPN2E (SID)

Please login to display this image.

 

Does the real CRJ try to get back to the original course as fast as possible and does it sometimes cut this course very extremely?
(The magenta HDG Bug points to the next waypoint)

The Sim-CRJ runs here in danger of losing course at too high a speed.

Please login to display this image.

 

Regards, Wolfgang

 

Translated with www.DeepL.com/Translator
 

Link to comment
Share on other sites

I agree with how good this crj simulation  is and unbelievable of how much you get for the price.   It is definitely worth continuing to refine and will remain relevant given the number of pilots that will continue to operate this RW aircraft as they go through their RW careers and the fact that airlines are taking new deliveries.

 

   A few comments,  WHEN proper descent planning is possible typical descent rates are less than 2000 feet/min unless a big tail wind..    Using the 3 to 1 descent rule (such as being at 10,000 feet by 30 miles if sea level airport), you can also use the ground speed/2 and add a 0 as the descent rate.     For example 300 kts ground speed /2 =. 150 add a zero  = 1500 feet/min.   

 

 

Link to comment
Share on other sites

  • Developer

Thanks everybody for the comments. I will look into all of this and decide what can be added and what can't. Especially the information that Mike posted in the first post, contains some things that I honestly have never heard about. Others are the complete opposite of what I have been told (VNAV alt/speed entries require EXEC, the "on" position of the volume knobs on the audio panels - I even reversed those once already based on pilot information....). And finally, extending the flightplan structure by an alternate flightplan won't be easy, but I will try.

 

Link to comment
Share on other sites

I confirm  as well the speed and alt changes in legs page require an exec and volume button pushed in is “on”.     I also agree on the speed trend being to jumpy.  I made sure to disable any wx that produces turbulence.    In the rw aircraft even in turbulence this speed trend does not jump around.      Good stuff and thanks Hans for sticking around with us CRJ pilots!

Link to comment
Share on other sites

  • Developer

I already fixed the speed trend a few weeks ago, so consider that done - it was indeed too jumpy, not just for the looks but especially for the CLB/DES and IAS modes. It's now a lot smoother. I'll fix the LEGS page entries and audio panels next.

Link to comment
Share on other sites

wwittkoff, The reasoning behind staying up high as long as possible and then making a steeper decent further along the route is for fuel savings.  You burn a lot less fuel at high altitudes and the longer you are able to stay up there the less you burn.  The ideal decent is a steeper and constant decent with power idle the whole time and no speed breaks down to the start of the approach. This isn't always possible due to the altitude and speed restrictions on different sids and of course ATC instructions for spacing ect but that's the goal.  Just got back from a 4 day pairing and I watched the bank angle as the autopilot was making turns.  It will turn a maximum of 30 degrees without 1/2 bank selected and 15 degree's with it selected for those who are wondering about the aggressive turns in the sim.

 

Hoffie3000, in your screenshots the FMS is displaying what we call a devils tail.  It is a result of the waypoint being an "overfly" waypoint denoted on north american charts as a star with a circle around it.  This essentially means the procedure calls for the airplane to actually fly over the waypoint rather than smart turn before it as it transitions to the next fix in the legs page.. the autopilot should fly the aircraft over the overfly point and start a turn in the direction of the devils tail and then attempt to get back on course to the next waypoint using 30 degree bank turns. Maybe this is what's causing the strange turning behavior in the sim. The waypoint "STTIK" on STTIK 4 SID in ksfo is also an overfly waypoint and the FMS doesn't seem to know how to handle it properly.

 

Mike

Link to comment
Share on other sites

Great write up and as a new CRJ2/7/9 driver I agree with all the above, especially flight characteristics up high. Long time simmer and have tried to help with other projects from my IRL experience, unfortunately the numbers just do not "plug n play" into the sim nor does the "feel" of a particular subject aircraft, at least not FSX/P3D. The multi-million dollar Level-D sims are not exact, but pretty close. Coming from the Proline 21 set up on the King Air, I have yet to see a fully functional FMS, but that's understandable given the complexity of the "box" and trying to mirror all those calculations in a PC sim. I'm also glad you guys did not build a CRJ-200!

Link to comment
Share on other sites

Hum... yes...  procedure or technique differences ...or...  .fuel savings or reasonable angle for beverage carts and flight attendants in the back... the great thing about simulation is the that the simulator pilot can set  rules that he/her chooses to create for the experience.    Thanks for the high end developers who make this possible.     

 

A little virtual trick I use for the developers that create a great product that shows great investment to model an aircraft  (8 to 9 digits price points) for only 2 to 3 digits is to use a “custom MEL” while the issue is resolved (if possible within the scope of this hobby’s platform)

 

The way this works is to create a work around for a known simulator issue and treat as an MEL .   For example the APU power/fuel switch that does not behave properly if managed early in the simm load up (at least for me).     My solution to this is to keep the aircraft on ground power while I do other stuff and start the gpu later which for me produces the right switch light logic.     

 

My recondition for virtual simmers is to support developers with feedback on improvements while at the same time not getting stalled on specific issue.  

Link to comment
Share on other sites

  • 8 months later...

Hmmmm... I'm a captain on the CRJ RW and have been waiting for this to come out for a good many years now. I'm still on the fence though.  Thanks for the review. I agree that half the challenge/fun of flying the CRJ is in proper descent planning. 

 

If you guys are still trying to tweak idle-descent, the following works as a pretty good approximate in the airplane to get STAR speeds at idle: 

Thrust Idle:

2700fpm =  280KIAS

2000fpm = 250KIAS

1500fpm = 230KIAS

1200fpm = 210KIAS

 

There is obviously variation based on the conditions, but these get you in the ballpark.  Example would be to cross "ABCDE" fix at 280KIAS, you can wait until your required descent rate is 2700fpm and then set V/S to that number, pull the thrust all the way back, and magically watch the airplane eventually slow to the desired speed of 280-ish... I say ish because it's like everything else with the CRJ in that you always have to be flying it or it will start flying you AKA it'll get away from you pretty quickly if you're not careful.

 

Anyway, I'll probably buy it but I was curious about a couple things that I haven't seen answered.  Does it properly display white/blue/green/yellow needles?  Does the wx radar work with reasonable accuracy?  Does the APU only break when it's extremely hot outside and while parked at an outstation on go home day?  How angry are the FA's when that happens? 

On 6/16/2019 at 7:20 AM, Hans Hartmann said:

VNAV alt/speed entries require EXEC

 

  VNAV alt/speed entries do not, LEGS page do.

Link to comment
Share on other sites

  • Developer
On 18.3.2020 at 14:57, saguaro5 sagte:

VNAV alt/speed entries do not, LEGS page do.

I have different information on this one. All from real CRJ pilots. Some say any LEGS page change (incl. ALT/SPD) requires EXEC, others say, all changes except ALT/SPD require EXEC. The first who sends me a video, that clearly shows one way or the other, wins 🙂

Link to comment
Share on other sites

I have ~1100 hours in the CRJ, mostly the 200 with some in the 700.  Changes made on the legs page required EXEC. If you do a search for the Rockwell Collins 4200, there are a few PDF copies available. Collins Manual Page 4-38 deals specifically with vnav and says to push the EXEC to execute. There was a call out to verify any changes made to waypoints, speeds, or altitudes before the EXEC key was pressed. 

Please login to display this image.

Link to comment
Share on other sites

  • 2 weeks later...

For what it's worth for my 2 cents:

 

In the HGS, the acceleration indicator (little chevron beside "corsair"/flight path symbol) could have its sensitivity boosted an easy 50 times. Teaching guys to use the HGS always surprises ppl with how sensitive the acceleration cue is. It allows for amazing "walking" of the thrust levers, to the point where an FO not looking through it actually can't read the minute changes a captain can make through the HGS. The cue should also be able to travel as far as the speed error queue that hangs/pops up from the Corsair "wingtip."

 

The speed error cue is also a bit off - a cue rising to match the top of the flight path symbol indicates a +5 knot deviation from the bugged speed, and it's a linear cue so double the height of the flight path symbol = +10kt error. The same is true in reverse.

 

Also in the HGS, non-conformal symbology wasn't simulated - the flight path indicator can't actually travel off the sides of the screen - it just runs to the edge, but stops and become dashed to indicate the true flight path is outside the limits of a conformal display - this allows the pilot to continue to use the flight path cue during more aggresive maneuvering.

 

Next, I can't recall ever landing a -900 with a nose-low attitude, once flap 45 is selected and you slow to Vref the plane settles approx 1 deg nose up and holds there with an appropriate Vref. Even the largest gust factor additives don't end up flying us nose-low on the approach in landing config.

 

Finally, to add to giltender's comments, I agree that the ALTS CAP happens too early, and should temporarily actually read "ALTS CAP" in the Flight Mode Annunciator until the actual attitude is fully captured, then the FMS will swap over to ALTS. 

 

giltender's comments are spot on, my experience with the plane are the same as his, plus these added pieces.

 

That being said, I'm intentionally being nit-picky for the sake of chasing perfection, because excellence is already accomplished. Beautiful work.

 

Cheers

Link to comment
Share on other sites

Now almost 3 years of flying both the Aerosoft crj and the real one, I was worried about the risk of picking up any bad habits that could transfer over.    This also in respect to final approach phase, landing, flare phase.    I have been pleasantly surprised that has not been the case an I would argue just the opposite.   I would agree that final flare on the P3d crj does not feel quite as “under control”as rw (yes, less nose up flare than the real thing), however the level D simm does not do that well either or p3d simm in general.      I am amazed how energy management (flying the appropriate ref speeds) in both is one of the key ingredients for a great landing.   As we fly with different pilots often it evident that exact technique with successful outcome varies greatly and there is no one right answer.      My own technique adopted from an IOE instructor that has not failed me is: 


For the 900

-  ref at 50 feet (very important) 

- begin a very very slight pitch up and slight power reduction.  

- at 30 feet begin to reduce thust further 
- at O to 10 feet have power out and pitch slowly further and be patient  until the mains touch.   
 

I had my wife fly in the back with me once when she filmed the landing.   The sunrise was at a perfect angle to cast a neat landing shadow. I will share If able.  

Link to comment
Share on other sites

2 hours ago, wwittkoff said:

Now almost 3 years of flying both the Aerosoft crj and the real one, I was worried about the risk of picking up any bad habits that could transfer over.    This also in respect to final approach phase, landing, flare phase.    I have been pleasantly surprised that has not been the case an I would argue just the opposite.   I would agree that final flare on the P3d crj does not feel quite as “under control”as rw (yes, less nose up flare than the real thing), however the level D simm does not do that well either or p3d simm in general.      I am amazed how energy management (flying the appropriate ref speeds) in both is one of the key ingredients for a great landing.   As we fly with different pilots often it evident that exact technique with successful outcome varies greatly and there is no one right answer.      My own technique adopted from an IOE instructor that has not failed me is: 


For the 900

-  ref at 50 feet (very important) 

- begin a very very slight pitch up and slight power reduction.  

- at 30 feet begin to reduce thust further 
- at O to 10 feet have power out and pitch slowly further and be patient  until the mains touch.   
 

I had my wife fly in the back with me once when she filmed the landing.   The sunrise was at a perfect angle to cast a neat landing shadow. I will share If able. 

 

I've always been a fan of sims that make you work harder than the real plane (for normal ops).

 

Even our level D sims have some quirks. Simulations will inherently have differences. Ours climb way better on takeoff than the real planes do, but it's good training for the crews because it forces them to really know their stuff for that busy phase of flight. 

Link to comment
Share on other sites

5 hours ago, A_Pilot said:

 

I've always been a fan of sims that make you work harder than the real plane (for normal ops).

 

Even our level D sims have some quirks. Simulations will inherently have differences. Ours climb way better on takeoff than the real planes do, but it's good training for the crews because it forces them to really know their stuff for that busy phase of flight. 

I’m a mechanic on the -200. As far as I know, the only Level D sims for that airframe are in Toronto and St. Louis. (There is also a non-Level D procedural sim in Peterborough). I did engine run and taxi training on the one in Toronto, and that one really shows its age - (at least cosmetically). It seemed to function OK, but the switches and cockpit furnishings are really worn out!

Link to comment
Share on other sites

41 minutes ago, JRBarrett said:

I did engine run and taxi training on the one in Toronto, and that one really shows its age - (at least cosmetically). It seemed to function OK, but the switches and cockpit furnishings are really worn out!

 

It a little worn-in but the flight dynamics on that sim are top-notch, especially for high altitude aerodynamics, and especially for an older sim. Spins, Mach tuck, and stalls are super fun in that sim.

Link to comment
Share on other sites

39 minutes ago, A_Pilot said:

 

It a little worn-in but the flight dynamics on that sim are top-notch, especially for high altitude aerodynamics, and especially for an older sim. Spins, Mach tuck, and stalls are super fun in that sim.

We were pressed for time, or the instructor would have let me fly it, even though that wasn’t on the syllabus for maintenance training.


That sim sits right next to the Dornier 328-100 sim. Our company used to operate that model, then we switched to the Challenger 850. The pilots now train in St Louis, but our mechanics go to FlightPath in Toronto. I always enjoy going up there.

Link to comment
Share on other sites

Archived

This topic is now archived and is closed to further replies.

×
×
  • Create New...

Important Information

We have placed cookies on your device to help make this website better. You can adjust your cookie settings, otherwise we'll assume you're okay to continue. Privacy Policy & Terms of Use