General disclaimer, I am going to be talking about how we do a "descend via" arrival in the United States. This is an arrival with a sequence of waypoints often with airspeed and altitude restrictions. The complicated part of this is that they are not all just "Cross FIX at ALT", rather they are "Cross FIX between ALT1 and ALT2, at 290kts".
So how do we do that in the CRJ? Well, it is a bit more complicated than a Boeing or Airbus product. The CRJ series technically doesn't have VNAV (I am lying, but I will qualify that in a second though). We have advisory VNAV which is a very basic descent profile that the FMS generates based on entered crossing restrictions. I want to explain a bit of how it does this. To illustrate, lets go look at an arrival.
This is the VANZE1 RNAV arrival into Memphis, TN.
Imagine you are the FMS. You have a desired descent angle which is 3.0 degrees. This is of course editable in the VNAV page 3/3 of the FMS. 3.0 works fairly well though and unless there is a massive tailwind will allow you to descend at near idle thrust maintaining profiled speeds. The way that the FMS builds the descent path goes as follows. Start at the end of the arrival with the first hard altitude (which is to say a restriction that isn't "at or above/below/between). On the VANZE arrival, that is HEXIN. Now, the FMS will draw backwards up the arrival a path of 3.0 degrees. At each waypoint it logic checks to see if you meet whatever restriction is entered in the LEGS page (it is important to note, before you fly any arrival like this you must manually verify each altitude at each fix is correct by cross referencing the chart). Now, if it fails this logic check and the 3.0 degree angle doesn't meet the proceeding restriction, then it will adjust the angle such that it does. So the result is you may see any angle of 2.8 or 3.2 on the LEGS page between waypoints to signify you will need an adjusted descent rate for that section of the arrival. The physical manifestation of all this is the white snowflake which appears on the PFD where the glideslope appears for the ILS. Following the snowflake will keep you within all restrictions of the arrival.
If you are still with me, good job. That isn't the hardest part though.
We need to now mention a few things about our friend, the snowflake.
1. It doesn't build deceleration stages like the 737 or A320. All it is concerned about it altitudes. So if the arrival requires you to decelerate to, say, 210KTS, then you must plan that into your descent on your own.
2. It doesn't account for your TAS changing as you descend. Hence, while one VS may have been working at a higher altitude, you will need a lower one as the descent progresses. Remember, we are attaching a VS to an angle. We are still doing a 3.0 degree descent, but much like the VS of a CRJ on a 3.0 ILS will be greater than that of a C172 doing the same approach because of the speed difference, the higher true airspeed for a given indicated airspeed at altitude will require a greater VS.
*aerodynamics side note (skip if you want): a swept wing high altitude jet like the CRJ makes a lot of drag at higher altitudes going fast (drag = speed^2), so the higher vertical speed required for a 3.0 degree descent out of FL340 actually compliments this aerodynamic principal quite well. In my experience, the CRJ can easily do 3000FPM out of the high flight levels without accelerating, but below 20,000 often times it is hard to descend at more than 2,000FPM.
3. You cannot couple the autopilot to our friend the snowflake. The snowflake is a function of an angle. No where can we select a 3.0 flight path angle on FCP. 10/10 engineering. Even worse, the CRJ doesn't share with you what VS its currently using to compute the snowflakes rate of descent. I have to imagine this number totally exists somewhere within the Rockwell Collins avionics, but they decided to not tell us. So this can lead to you just chasing it with the VS wheel in hopes you are ballparking it correctly. There is however, a better way.
*side note number 2: there are a few CRJs flying that actually have a VNAV button on the FCP. Guess what it does. It couples the snowflake to the autopilot. They're lucky. None of my companies airplanes have this button though, and I don't think that the AS CRJ7/9 has it either. So as near as I can tell, it doesn't exist.
If you've been reading this up till now, I need to issue an apology. Everything we just talked about it kind of irrelevant because we don't actually use the snowflake in day to day operations. At least, most of us don't. I would consider it secondary descent profile information. Enough theory then, how do we actually do this?
Good old airmenship and the DIR INTC page. The DIR INTC page is extremely useful for a singular reason. It will tell you the VS required to cross FIX at ALT, based on what you put on the LEGS page. A simple example to start, shall we.
Let's say ATC instructs us to cross SHONN at 11,000ft. We will enter that restriction into the LEGS page (because this is an EXPECT altitude, it will not have loaded when we load the arrival). Now, execute the change and switch over to the DIR INTC page. You will see the SHONN, an arrow pointing down, and a vertical speed. 11,000 should be in small font right above the VS. That is the VS required to cross SHONN at 11,000 based on your current altitude. There will also be your current angle from present position directly to SHONN. To cross SHONN at 11,000, dial in 11,000 into the FCP, wait for the angle to read 3.0 (or less, I like to descend a 2.5 sometimes just to make the ride a bit smoother and give me more options with a bit of power in on the descent), and start descending at the VS shown on the SHONN line of the DIR INTC page. I should note, if you do what I do and start down at an angle <3.0, you will never see the snowflake because you have never intercepted the 3.0 degree path created by the FMS which the snowflake obviously represents.
Same situation, same arrival, same crossing restriction at SHONN, but now ATC asks you to cross SHONN at 250kts. Now you must slow down and go down. This is not one of the strong suits of the CRJ. In level flight, the thing will decelerate quite rapidly. It won't while descending though and in all honesty, the flight spoilers aren't too stellar slowing the bird down. Remember how I said that the snowflake doesn't build in deceleration legs, thats on you, the pilot! The only difference in our descent planning we will need to make is that we are going to need a level segment in order to decelerate. As we approach a 3.0 degree descent towards SHONN using the DIR INTC page, start down a bit early. Maybe at 2.8, however, add approximately 100 or 200FPM to the required rate of descent. This will ensure that we get to 11,000 a bit before SHONN and have enough time to decelerate to 250kts.
The reason I wanted us to go through that example is so that we can now go to a more complicated one. Lets go back to the VANZE arrival into MEM that I linked to earlier, back at the top of this increasingly long essay.
Pretend we are cleared for this arrival and we are starting it at the TALLO transition, landing South (18L/C/R). The restrictions are all in the LEGS page and ATC clears us to descend via the arrival. Great. The first crossing restriction is FASON at or above FL240. That is pretty easy. The next one is CRAMM at or below FL230, and VANZE has the same restriction. So what should we do?
The honest answer is make it up as you go. You need a mental picture of what the entire arrival looks like in terms of vertical profile, but in reality you can't keep track of every set of restrictions in your head. There are way too many! So just worry about the next four, or so.
Hypothetically, we cross both CRAMM and VANZE at FL230, are we going to be able to go from VANZE to MASHH and descend to 16,000ft to make the top of the gate? Probably not. Here it what I would do. Go look at your DIR INTC page. Look at the VS required for each waypoint. I am not sure exactly which altitude it will show for MASHH on the DIR INTC page, but it'll either be 16 or 14,000. It will sometimes change based on the arrival and profile and that is wayyy beyond what we are talking about here today. We have done enough pseudo coding of the airplane's avionics today. We need to now pick a vertical speed that will hit all the restrictions. Ideally you want something less than the VS to cross FASON, and greater than the VS to cross CRAMM. This will ensure you can cross FASON above 240, and CRAMM below 230. More than likely, that VS required to cross MASHH at 16,000. So that is probably the one you want to use. Wait until you are 3.0 degrees from MASHH and start down at that VS, assuming that VS is less than the VS for FASON and greater than that for CRAMM. If it doesn't meet those requirements, then you'll need to adjust it until it does.
Lets say we cross FASON above 240 successfully. We can now start thinking about the waypoint after MASHH.
dist 22 (skip w/ turn at HLI, no crossing restriction there)
Run the same algorithm in your head. You need to select a VS that will comply with all of the above. Rinse, repeat, until you hit the bottom of the arrival. And make sure you are complying with speed restrictions too. If that example was a little hard to follow I understand.
The summary to all this: if you are flying an arrival with multiple crossing restrictions to hit, use the DIR INTC and your brain to select a descent rate that will comply with as much restrictions as you can. Look forward and ensure to not box yourself in a corner where you have to do an impossible rate of descent to make a restriction. Constantly be monitoring and adjusting.
The snowflake is a good piece of supplementary information and if you're doing the above procedure correctly, you'll end up following it all the way down usually. It is important to understand the limitation of the system, such as no accounting for speed restrictions. For this, there is no substitute for human interaction and mental processing during these procedures. There is no simple way either, rather just practicing and engaging your brain.
This is a huge pain. The CRJ avionics suite was built in the early 90s, well before any of these complicated RNAV procedures. We are using 25 year old tools to solve a modern day problem. This is genuinely the hardest thing to do in the CRJ and will make you feel like you've done some mental gymnastics. And once you get good at it, its very satisfying. Jumping back into the NGX or Airbus after this will make life seem pathetically easy.
If any of that needs clarification, I will be happy to help.