Diverging Diamond Interchanges

[jeffandnicole]

Many years ago, a woman was killed near my house when she took a diamond interchange ramp at 90 mph and hit a house across the street.

This scenario is probably the strongest argument against straight ramps, especially exiting straight ramps.

Well, good point.  Fatal crashes can occur at straight diamond ramps too.  I'll always remember that Atlanta bus crash on I-75 where the bus driver mistook an exit ramp for a freeway-lane.  The bus blew through the bridge deck and landed on it's side on the freeway killing 6. 

http://usatoday30.usatoday.com/news/nation/2007-03-02-atlanta-bus-wreck_x.htm

Ok J&N, I've always wondered why agencies seem to favor entering loop ramps onto the freeway as opposed to exiting loop ramps.  Agencies must have a reason they favor entering loop ramps.  What is it?

Well, first, that may or may not be true.  I've always thought agencies favored diamond interchanges.  I guess we would need to see a listing of all the interchanges that exist to see which ones are favored more than others.

If it is true, a few possible reasons:

It requires those on a highway to make fewer decisions.  When you have exit options such as A & B, it requires drivers to think which exit do they want.  By forcing everyone into one ramp, they can make that decision at a later point.  This, personally, is especially true when I see a blue 'Food' or 'Gas' sign on the highway, then realize I needed to concentrate on not only the business I wanted to get to, but also which particular exit I needed.

Those on the highways are going at a faster speed, so they can use a straighter ramp as time to slow it down as they come off the highway.

No left turns are necessary on the two thru roads themselves.  Those on the highway exit right at the single interchange.  Those on the cross street either exit right and loop down 270 degrees, or exit right and swing right 90 degrees.

In other cases, it can be a preferred method of exiting as per state policy, or even just a particular preference of the engineer or designer.

In most cases, the interchange that was built wasn't the only interchange under consideration.  Usually nearly every design option is open for discussion at the beginning of a project; they take a look at the landscape to see what room they have to work with, and then do their analysis to see what works best at each particular location.

If money was no object they would stop building DDIs entirely (since they are not very efficient).

I doubt people who own land next to interchanges would be very happy with that.

VTK argued that other interchange types would perform "comparably well" to the DDI in relation to operational performance.  We aren't looking at cost or the environmental impacts of the interchange.  The focus was on total network delay.  You are muddying the waters by bringing up people crying about owning land next to a freeway??   Thanks for that irrelevant comment.

Actually, your focus is on network delay, and only that.  In the real world, engineers have a lot more to look at.  If an interchange can be built without requiring the taking of land, that is going to be the preferred option.  Again, that's real world, not Engineering 201.  There are tons of examples where single landowners or groups of people have delayed projects for years, and even cancelled them.   Transportation departments just have to work with them as best they can.  The optimal design may not ultimately be what is actually constructed.

[tradephoric]

Actually, your focus is on network delay, and only that.  In the real world, engineers have a lot more to look at.  If an interchange can be built without requiring the taking of land, that is going to be the preferred option.  Again, that's real world, not Engineering 201.  There are tons of examples where single landowners or groups of people have delayed projects for years, and even cancelled them.   Transportation departments just have to work with them as best they can.  The optimal design may not ultimately be what is actually constructed.

VTK made a very specific statement.  You sometimes have trouble following the conversation J&N.  This is what VTK said:

Other interchange designs may operate comparably well, but would have been significantly more costly to build

My point is VTK is wrong in suggesting that other interchange designs would operate comparably well to a DDI.  If you want to spend a trillion dollars on an interchange it's going to outperform the DDI in every measure of effectiveness imaginable.  I gave them a chance to backtrack but they doubled down:

I could agree with that bold statement if you replaced "comparably well"  with "better" .   Maybe you are so biased towards DDIs that you won't even admit that other interchange designs would perform better. 

No, I meant "comparably well" because other designs have other drawbacks. A SPUI has a three-phase signal instead of the two-phase signals of a DDI, so wait times for traffic turning left to enter the freeway may be longer. A parclo has loops that take more time to drive around than direct ramps. Anything that doesn't fit in the original ROW would probably still not be built yet because of the additional paperwork. In the absence of a perfect solution, there's very little room for improvement over "adequate". Or do you have a perfect solution you would like to share?

It's not about the cost of the interchange.  It's about VTK being so biased towards DDI's that they apparently believe a trillion dollar interchange couldn't outperform the DDI (it would only perform "comparably well"  to them).  Maybe I'm taking VTK's comment too literally, but they got to realize the DDI is not the best interchange design out there.  Again, if money was no object we could come up with a more efficient design then the DDI they came up with (and in a smaller footprint to boot). 

[jakeroot]

It's not about the cost of the interchange.  It's about VTK being so biased towards DDI's that they apparently believe a trillion dollar interchange couldn't outperform the DDI (it would only perform "comparably well"  to them).  Maybe I'm taking VTK's comment too literally, but they got to realize the DDI is not the best interchange design out there.  Again, if money was no object we could come up with a more efficient design then the DDI they came up with (and in a smaller footprint to boot).

Maybe I'm not following the conversation here either, but I think "comparably well" only applies to interchange designs which are comparably practical. If there's no ROW for loops or flyovers, your discussion narrows to only a few interchange designs: SPUIs, diamonds, dumbell or dogbone roundabouts, full roundabouts with two under- or over-passes, and DDIs. "Comparably well" in this case, only applies to interchanges that may have, at one point, been considered practical when replacing the 270/Roberts Rd interchange.

Just looking at a map of the area, I don't think signal progression is important along that stretch of Roberts Road.

[tradephoric]

Maybe I'm not following the conversation here either, but I think "comparably well" only applies to interchange designs which are comparably practical. If there's no ROW for loops or flyovers, your discussion narrows to only a few interchange designs: SPUIs, diamonds, dumbell or dogbone roundabouts, full roundabouts with two under- or over-passes, and DDIs. "Comparably well" in this case, only applies to interchanges that may have, at one point, been considered practical when replacing the 270/Roberts Rd interchange.

I actually think loop ramps would have fit at this interchange.  The advantage of the model below is there are zero traffic signals (reducing driver delay) and it has a smaller footprint than the DDI.  The disadvantage is extensive bridge work and loop ramps would have been needed to build this interchange. 

https://www.youtube.com/watch?v=B8DUMjlKqx8

A similar interchange was built in Florida but it was a  Parclo B4 with exiting loop ramps as opposed to entering:

https://www.google.com/maps/@28.34539,-81.5333086,579m/data=!3m1!1e3

[vtk]

It's about VTK being so biased towards DDI's that they apparently believe a trillion dollar interchange couldn't outperform the DDI (it would only perform "comparably well"  to them).

Again, I'm not "biased toward DDIs".  There are at least a few different, practical, interchanges that would work adequately at I-270 & Roberts Rd, considering their advantages and drawbacks, the traffic patterns, and the adjacent intersections.  There wouldn't be much difference between the operational efficiencies of the different designs, though each would have its specific disadvantages.  I drive through this interchange not infrequently, and at various times of day, and it is my opinion that the general deficiencies of the DDI design are manifest at this specific interchange to a very limited degree – so limited that it's not worth worrying about.

And while I'm not prepared to agree to narrow the conversation to operational efficiency only, with no concern for cost or environmental concerns, I am having a very hard time imagining what kind of interchange could be built here, given unlimited budget and dictatorial power of eminent domain, that would have no operational weaknesses at all, considering the various other crossroads that intersect Roberts Rd.  (Cutting off any of them so their traffic has to go a different way is an operational weakness.)

(Don't have time to review tradephoric's most recent post at this time. Will look at it later.)

[jeffandnicole]

Actually, your focus is on network delay, and only that.  In the real world, engineers have a lot more to look at.  If an interchange can be built without requiring the taking of land, that is going to be the preferred option.  Again, that's real world, not Engineering 201.  There are tons of examples where single landowners or groups of people have delayed projects for years, and even cancelled them.   Transportation departments just have to work with them as best they can.  The optimal design may not ultimately be what is actually constructed.

VTK made a very specific statement.  You sometimes have trouble following the conversation J&N.  This is what VTK said:

Other interchange designs may operate comparably well, but would have been significantly more costly to build

Huh?  I was responding to what you said, not what he said, especially in reference to "You are muddying the waters by bringing up people crying about owning land next to a freeway??". 

Don't worry about me following the conversation.  You're having a difficult enough time yourself.

[vtk]

Maybe I'm not following the conversation here either, but I think "comparably well" only applies to interchange designs which are comparably practical. If there's no ROW for loops or flyovers, your discussion narrows to only a few interchange designs: SPUIs, diamonds, dumbell or dogbone roundabouts, full roundabouts with two under- or over-passes, and DDIs. "Comparably well" in this case, only applies to interchanges that may have, at one point, been considered practical when replacing the 270/Roberts Rd interchange.

I actually think loop ramps would have fit at this interchange.  The advantage of the model below is there are zero traffic signals (reducing driver delay) and it has a smaller footprint than the DDI.  The disadvantage is extensive bridge work and loop ramps would have been needed to build this interchange. 

https://www.youtube.com/watch?v=B8DUMjlKqx8

I can imagine problems with traffic exiting SB I-270 and cutting across the westbound Roberts Rd lanes to turn left onto Westbelt Dr.  Note, the as-built left turn lane for Westbelt begins right about where your design has the ramp from SB I-270 joining Roberts Rd.  This could be an issue in PM rush hour, as workers from Hilliard head to their second-shift jobs along Westbelt while office workers coming from Downtown head for their subdivision homes west of the railroad yard.  I'm not sure if this can be quantified in terms of vehicle-seconds of delay, or just how bad the problem might get, but it's enough that this design is not clearly better than the DDI that was built.  Maybe better, but not clearly so.

By the way, your visualization has a lot of vehicles exiting I-270 only to re-enter it in the opposite direction.  That may be a thing that happens on occasion in real life, but not several per minute.  Perhaps it's not relevant to our disagreement, and you probably already noticed it, but I felt a need to point it out anyway.

[tradephoric]

^^^
If the adjacent signals led to problems during rush hour we could tap into our trillion dollar budget and build loop ramps to bypass those signals as well.  The fact is my trillion dollar interchange would operationally outperform the DDI (I would hope so, it cost a trillion dollars!).  It wouldn't be "comparable"  it would be "better" .  But you can't even admit that.

[vtk]

^^^
If the adjacent signals led to problems during rush hour we could tap into our trillion dollar budget and build loop ramps to bypass those signals as well.  The fact is my trillion dollar interchange would operationally outperform the DDI (I would hope so, it cost a trillion dollars!).  It wouldn't be "comparable"  it would be "better" .  But you can't even admit that.

I would have to see a specifc design to estimate its operational efficiency.  The issue that concerns me is adding loops and braids to the interchange tends to expand its footprint along the surface road(s), likely introducing more weaving issues as the growing design creeps closer to existing intersections on the fringe of the interchange's scope.  Then, to solve those weaving issues, the scope of the interchange is expanded to include those intersections, more loops and braiding is added to fix the weaving issues identified, the interchange design has expanded again, and there are new weaving issues to address between the increasingly enormous interchange and the next nearest intersections.  This is why I don't believe a perfect service interchange design exists for an urban location such as this.

[jakeroot]

Here's a drive-through of the first DDI in Denmark (my home country 💗), in the community of Odense: https://goo.gl/tkKDTt (E20, junction 52).

https://youtu.be/MDYR6Y1HH-g

It has several characteristics of the American designs, as one might imagine. I think the most notable is that the cross-over movements are completely straight-across. On the first page, near the bottom, I posted an example of a DDI in Durban, South Africa. The cross-over movements are very compact, and require a slight turning maneuver through the intersections. This is necessary when trying to fit the DDI in where a previous diamond interchange already existed; when DDI's are built from scratch, however, the cross-over movements typically don't require any "corning".

[NE2]

Am I doing this right?

[jakeroot]

http://i.imgflip.com/2egkk5.jpg
Am I doing this right?

The man in flannel needs to be a diamond, and the bird in blue needs to be a DDI.

EDIT: if this is from trade's perspective, flip the women. Red is a DDI, blue is PARCLO B4. Flannel remains a diamond.

[johndoe]

Here's a drive-through of the first DDI in Denmark ...
It has several characteristics of the American designs, as one might imagine. I think the most notable is that the cross-over movements are completely straight-across.

When we lay DDI out this is always one of the big debates.  Some designers (and DOT) prefer the tangent section to extend from stop lines through the intersection while others don't think it's so crucial.  It's sort of funny how opinionated engineers get on some topics even if there aren't that many data points on safety records.  Have ANY DDI had significant wrong-way problems regardless of tangents and crossing angles?  I'm not aware of any. 

https://www.google.com/maps/@35.807396,-80.875909,209m/data=!3m1!1e3
Here is another one that seems very "tight", it's in NC.  I don't think I've seen any others with such small radii and lack of tangent for the crossover movements (in north America...the French examples are pretty tight), but if anyone has examples I'd be interested to see them.

The other thing that caught my eye on your Danish example is the "extra" intersection just beyond the main crossover.  It seems they're treating the off-ramp left as it's own intersection (or maybe timing it all together but giving extra signal heads and stop lines for reassurance).  The driver goes through the main crossover at about 1:25.  Then at about 1:30 they are at the intersection for the off-ramp.  Off the top of my head, I don't know of any locations like this.  For the signal timing nerds: this would impact the all-red time required since the size of the intersection would change.

[jakeroot]

The other thing that caught my eye on your Danish example is the "extra" intersection just beyond the main crossover.  It seems they're treating the off-ramp left as it's own intersection (or maybe timing it all together but giving extra signal heads and stop lines for reassurance).  The driver goes through the main crossover at about 1:25.  Then at about 1:30 they are at the intersection for the off-ramp.  Off the top of my head, I don't know of any locations like this.  For the signal timing nerds: this would impact the all-red time required since the size of the intersection would change.

I've actually designed a few DDI's (literally on paper) with a second signal for the off-ramp, to reduce the all-red time for the off-ramp. The off-ramp would need a couple extra seconds of red to allow cars to clear the intersection, but if it had its own signal, it could start earlier. Any cars that made it through just barely on the first signal, would end up stopped at the second signal. In theory, the lights for through-traffic would go red at the same time. However, I would be worried about trucks getting trailers stuck in the crossover point, so ideally it would be at least 25 meters from intersection to stop line. Any less than that, and I think one big intersection would be best. My paper design does not follow my suggestion, but I was like 15 when I drew that, so I didn't give the design much consideration.

Or, you could do it the South African-way, and just have both off-ramps be yields (even if two lanes): https://goo.gl/jJ1gc2. Given that the yield is basically a sharp turn, I would be less worried about drivers ignoring the yield. Not sure if anywhere in the US actually has this as a setup. Many even use NTOR signs for the off-ramps, even if single lane, so they can't even act like it. Of the two examples below, only the first permits turns on red (which is cool since it's a double left turn).

Here's a drive-through of the first DDI in Denmark ...
It has several characteristics of the American designs, as one might imagine. I think the most notable is that the cross-over movements are completely straight-across.

When we lay DDI out this is always one of the big debates.  Some designers (and DOT) prefer the tangent section to extend from stop lines through the intersection while others don't think it's so crucial.  It's sort of funny how opinionated engineers get on some topics even if there aren't that many data points on safety records.  Have ANY DDI had significant wrong-way problems regardless of tangents and crossing angles?  I'm not aware of any. 

https://www.google.com/maps/@35.807396,-80.875909,209m/data=!3m1!1e3
Here is another one that seems very "tight", it's in NC.  I don't think I've seen any others with such small radii and lack of tangent for the crossover movements (in north America...the French examples are pretty tight), but if anyone has examples I'd be interested to see them.

Yeah, there doesn't seem to be significant (any?) evidence that one design is better than another, just some guessing on the part of the lead engineers. I can see why they might not want any tangential movements in the crossover, but I don't know of any evidence that suggests its necessary. Traffic is generally moving quite slow, and since the other direction of the crossover is red, and traffic is probably waiting at it, I don't think drivers would be likely to misconstrue that leg of the intersection as being legally traversable anyway. While neither the US nor Canada is generally known for using an abundance of pavement markings, that's definitely not the case at DDI's; nearly all examples that I've seen use a ton of edge extension markings, those giant one-way arrows painted in front of the stop lines, raised markings for additional reflection, and extra signals and signs. You'd have to be blind to drive the wrong way.

Your example is definitely quite tight! Especially notable is how close the eastbound off-ramp's left turn is to the crossover point. Nevermind the near complete lack of any sort of curve prior to entering the DDI. I honestly prefer this design, as I'm not a fan of big curves (except in my women...). Engineers really seem to put in some big curves prior to DDI's, mostly so drivers enter the crossover point at a sharper angle, but as we just discussed, there doesn't seem to be any evidence of this being necessary, since wrong-way collisions don't seem to be an issue.

Metro Atlanta seems to have quite a few tight DDIs: I-85 @ GA-140 (first image), and I-85 @ Pleasant Hill Road (just north of the first example; second image). I think the second example is probably the tightest DDI that I've ever seen.

[inkyatari]

They're in early stages of construction of a DDI at I-55 and Weber Rd, in Bolingbrook / Romeoville, Illinois

[webny99]

Rochester (well, technically Brighton) has a DDI at I-590 and Winton Road.
It seems to work OK. My main beef is that it resulted in the install of a bunch of overhead assemblies which are absolutely gargantuan - way bigger than they needed to be.

While neither the US nor Canada is generally known for using an abundance of pavement markings, that's definitely not the case at DDI's; nearly all examples that I've seen use a ton of edge extension markings, those giant one-way arrows painted in front of the stop lines, raised markings for additional reflection, and extra signals and signs. You'd have to be blind to drive the wrong way.

Yeah, this is the case in the above DDI as well, and that makes it kind of ugly. Maybe once drivers become more familiar with DDI's, DOT's won't have to be as excessive about "Wrong Way" signs and pavement markings on new installs.

DDI's basically create a scenario in which making a turning movement, in and of itself, never requires crossing traffic. Therefore, turning movements are always free flowing (or at least free-flowing a greater percentage of the time). This is only optimal when there's a lot of turning traffic, especially left-turning traffic getting on and off the highway, and little through traffic. This basically applies to Winton and I-590, so it's probably a net positive compared to a standard diamond.
It wasn't horribly congested before, and it still isn't, so it was probably a good test location, while simultaneously having desirable traffic patterns for DDI implementation.

[ipeters61]

Delaware's diverging diamond at DE-1 and DE-72 was given an award recently: https://www.wdel.com/news/video-deldot-s-diamond-gets-an-award/article_e29bac14-8aab-11e8-8687-032cea41d2a5.html

[tradephoric]

I've actually designed a few DDI's (literally on paper) with a second signal for the off-ramp, to reduce the all-red time for the off-ramp. The off-ramp would need a couple extra seconds of red to allow cars to clear the intersection, but if it had its own signal, it could start earlier. Any cars that made it through just barely on the first signal, would end up stopped at the second signal. In theory, the lights for through-traffic would go red at the same time. However, I would be worried about trucks getting trailers stuck in the crossover point, so ideally it would be at least 25 meters from intersection to stop line. Any less than that, and I think one big intersection would be best. My paper design does not follow my suggestion, but I was like 15 when I drew that, so I didn't give the design much consideration.

This is a little different but I have noticed some SPUIs are designed with off-ramps that have signal heads for both the off-ramp and the main arterial while others only have signals for the off-ramp.  Here's an example of each:

Off-ramp only signal: 
https://www.google.com/maps/@43.5669106,-116.1963375,3a,82.8y,40.27h,89.92t/data=!3m6!1e1!3m4!1sCWQYFuuGVobMqmEmpME0SQ!2e0!7i13312!8i6656

Main+Off-ramp signal:
https://www.google.com/maps/@35.9096677,-78.935117,3a,63.7y,23.23h,85.72t/data=!3m6!1e1!3m4!1sbBthmzhZ3GX6YdPlhmKJoA!2e0!7i13312!8i6656

The example of the off-ramp only signal seems like a potential safety issue to me.  How does the arterial know to stop when the off-ramp has a green indication when there are no signal heads along the arterial?  Now presumably there is enough all red for the left-turn off-ramp traffic to clear the opposing right turn off-ramp with time to spare, but what if the arterial is backed up through the SPUI for some reason?  In that case  the main arterial traffic has no indication that the right-turn off-ramp has a green. 

[tradephoric]

I've actually designed a few DDI's (literally on paper) with a second signal for the off-ramp, to reduce the all-red time for the off-ramp. The off-ramp would need a couple extra seconds of red to allow cars to clear the intersection, but if it had its own signal, it could start earlier. Any cars that made it through just barely on the first signal, would end up stopped at the second signal. In theory, the lights for through-traffic would go red at the same time. However, I would be worried about trucks getting trailers stuck in the crossover point, so ideally it would be at least 25 meters from intersection to stop line. Any less than that, and I think one big intersection would be best. My paper design does not follow my suggestion, but I was like 15 when I drew that, so I didn't give the design much consideration.

This is a little different but I have noticed some SPUIs are designed with off-ramps that have signal heads for both the off-ramp and the main arterial while others only have signals for the off-ramp.  Here's an example of each:

Off-ramp only signal: 
https://www.google.com/maps/@43.5669106,-116.1963375,3a,82.8y,40.27h,89.92t/data=!3m6!1e1!3m4!1sCWQYFuuGVobMqmEmpME0SQ!2e0!7i13312!8i6656

Main+Off-ramp signal:
https://www.google.com/maps/@35.9096677,-78.935117,3a,63.7y,23.23h,85.72t/data=!3m6!1e1!3m4!1sbBthmzhZ3GX6YdPlhmKJoA!2e0!7i13312!8i6656

The example of the off-ramp only signal seems like a potential safety issue to me.  How does the arterial know to stop when the off-ramp has a green indication when there are no signal heads along the arterial?  Now presumably there is enough all red for the left-turn off-ramp traffic to clear the opposing right turn off-ramp with time to spare, but what if the arterial is backed up through the SPUI for some reason?  In that case  the main arterial traffic has no indication that the right-turn off-ramp has a green. 

Sure enough you can run into similar situations at DDIs.  Here is a streetview of a DDI in Atlanta.
https://www.google.com/maps/@33.9171717,-84.337966,3a,60y,334.31h,82.55t/data=!3m6!1e1!3m4!1s8CZ0B0Qei_URNcYkyzTGvw!2e0!7i13312!8i6656

Once drivers on the arterial pass the main signal, they have no indication if the off-ramp is green or red.  So if the arterial is backed up for whatever reason inching through at 2 or 3 mph, there could be 7 cars lined up between the main signal and off-ramp signal (all assuming they can just keep going since they already passed through their green light at the main... even as the off-ramp signal is displaying a green). 

[tradephoric]

Once drivers on the arterial pass the main signal, they have no indication if the off-ramp is green or red.  So if the arterial is backed up for whatever reason inching through at 2 or 3 mph, there could be 7 cars lined up between the main signal and off-ramp signal (all assuming they can just keep going since they already passed through their green light at the main... even as the off-ramp signal is displaying a green). 

Here's an aerial of that DDI basically showing the scenario (but with only 1 or 2 cars queued between the main signal and off-ramp signal as opposed to 7).  That silver car in the aerial (queued up through the DDI behind the red car) would have no idea if the off-ramp turned green.   The driver of the silver car is going to want to clear the intersection since they already passed through a green at the main signal.  Meanwhile the off-ramp traffic that gets a green indication are going to want to go too.

[doorknob60]

I've actually designed a few DDI's (literally on paper) with a second signal for the off-ramp, to reduce the all-red time for the off-ramp. The off-ramp would need a couple extra seconds of red to allow cars to clear the intersection, but if it had its own signal, it could start earlier. Any cars that made it through just barely on the first signal, would end up stopped at the second signal. In theory, the lights for through-traffic would go red at the same time. However, I would be worried about trucks getting trailers stuck in the crossover point, so ideally it would be at least 25 meters from intersection to stop line. Any less than that, and I think one big intersection would be best. My paper design does not follow my suggestion, but I was like 15 when I drew that, so I didn't give the design much consideration.

This is a little different but I have noticed some SPUIs are designed with off-ramps that have signal heads for both the off-ramp and the main arterial while others only have signals for the off-ramp.  Here's an example of each:

Off-ramp only signal: 
https://www.google.com/maps/@43.5669106,-116.1963375,3a,82.8y,40.27h,89.92t/data=!3m6!1e1!3m4!1sCWQYFuuGVobMqmEmpME0SQ!2e0!7i13312!8i6656

Main+Off-ramp signal:
https://www.google.com/maps/@35.9096677,-78.935117,3a,63.7y,23.23h,85.72t/data=!3m6!1e1!3m4!1sbBthmzhZ3GX6YdPlhmKJoA!2e0!7i13312!8i6656

The example of the off-ramp only signal seems like a potential safety issue to me.  How does the arterial know to stop when the off-ramp has a green indication when there are no signal heads along the arterial?  Now presumably there is enough all red for the left-turn off-ramp traffic to clear the opposing right turn off-ramp with time to spare, but what if the arterial is backed up through the SPUI for some reason?  In that case  the main arterial traffic has no indication that the right-turn off-ramp has a green.

I haven't used that first SPUI enough (at least in that direction) to say if there's any problems caused there. I have used this similar one pretty often like a mile away though: https://www.google.com/maps/@43.5717278,-116.2146678,3a,82.9y,287.9h,84.74t/data=!3m6!1e1!3m4!1sOQXxi3nfFiSvwdHe2cbpYw!2e0!7i13312!8i6656

I have never seen any problems here, except 95% of the time I am turning right on red anyways, so hard to say for sure.

But there is another interchange nearby (not a SPUI) where I saw this kind of thing all the time: https://www.google.com/maps/@43.5976763,-116.3543267,3a,75y,276.48h,89.27t/data=!3m6!1e1!3m4!1swHhQb3Of67oxAr3QrQLF3Q!2e0!7i13312!8i6656

Basically, the thru traffic northbound on Eagle Rd would go through the "main" part of the intersection on green, but then would immediately get stopped due to backed up traffic from the St Luke's St light. Then, the thru traffic light would turn red, and the right turning traffic from the off ramp would get the green. Traffic would still be backed up, with cars waiting in the "no man's land" between the crosswalk where the thru green light is, and the curve where the right turners turn on. All this is on a 50 MPH speed limit road lol, driving that area is awful. People always seemed to handle the situation pretty well, though (the right turners were usually good about waiting for traffic to clear before entering).

The most likely SPUI in the area you would run into this issue possibly is this one: https://www.google.com/maps/@43.5931344,-116.3941467,330m/data=!3m1!1e3
It's not on Street View yet. I've used it a handful of times without any issues (but never seen it super backed up or abnormally slow moving).

Going back to your first example, I wonder if they use cameras to see if there are any vehicles coming. Might be a bit advanced, but the technology could exist (ACHD already uses cameras for vehicle detection county-wide), and this is a brand new SPUI. I don't see any cameras that appear to be pointing in the correct direction to do that, though, so probably not. There are no real bottlenecks on either side of the SPUI though so it's unlikely to run into backed up traffic that would be possible to cause those kind of issues.

To bring this DDI related, I just realized the recently upgraded Gowen Rd/I-84 interchange would have been a really good place to try out a DDI in Boise. Shame they didn't, maybe they had good reasons not to (proximity to Eisenman Rd?). At least the new diamond works pretty well (though I haven't been through during Micron rush hours).

[NE2]

That silver car in the aerial (queued up through the DDI behind the red car) would have no idea if the off-ramp turned green.   The driver of the silver car is going to want to clear the intersection since they already passed through a green at the main signal.  Meanwhile the off-ramp traffic that gets a green indication are going to want to go too.

How is that any different from this situation?
http://www.google.com/maps/@28.3780009,-81.5046222,3a,75y,4.05h,91.91t/data=!3m6!1e1!3m4!1sHAYBhjB_JvlD2O2NQTnLNQ!2e0!7i13312!8i6656
Traffic can back up here from the next light (Hotel Plaza Boulevard). How is traffic stuck just beyond this light to know when the right turn from the ramp on the right turns green? It's not; traffic on the ramp is supposed to wait until the intersection is clear, just like at every other intersection.

[tradephoric]

That silver car in the aerial (queued up through the DDI behind the red car) would have no idea if the off-ramp turned green.   The driver of the silver car is going to want to clear the intersection since they already passed through a green at the main signal.  Meanwhile the off-ramp traffic that gets a green indication are going to want to go too.

How is that any different from this situation?
http://www.google.com/maps/@28.3780009,-81.5046222,3a,75y,4.05h,91.91t/data=!3m6!1e1!3m4!1sHAYBhjB_JvlD2O2NQTnLNQ!2e0!7i13312!8i6656
Traffic can back up here from the next light (Hotel Plaza Boulevard). How is traffic stuck just beyond this light to know when the right turn from the ramp on the right turns green? It's not; traffic on the ramp is supposed to wait until the intersection is clear, just like at every other intersection.

I'm not suggesting this problem is only limited to DDIs.  The initial example i gave was at a SPUI after all. I just think DDI's (and SPUI's for that matter) have geometries that are more conducive to this type of "no man's land" issue than a typical intersection.  At Median U-Turns, where you have large queue space between the medians, there will be both near and far signals to help ensure that drivers crossing the "far side" of the intersection will stop when the opposing side-street is being displayed a green indication.  Like this:
https://www.google.com/maps/@42.5626525,-83.1287596,3a,60y,1.98h,88.84t/data=!3m6!1e1!3m4!1sfbB6v3zLzfZ3MES9Q8PlXA!2e0!7i13312!8i6656

[jeffandnicole]

Once drivers on the arterial pass the main signal, they have no indication if the off-ramp is green or red.  So if the arterial is backed up for whatever reason inching through at 2 or 3 mph, there could be 7 cars lined up between the main signal and off-ramp signal (all assuming they can just keep going since they already passed through their green light at the main... even as the off-ramp signal is displaying a green). 

Here's an aerial of that DDI basically showing the scenario (but with only 1 or 2 cars queued between the main signal and off-ramp signal as opposed to 7).  That silver car in the aerial (queued up through the DDI behind the red car) would have no idea if the off-ramp turned green.   The driver of the silver car is going to want to clear the intersection since they already passed through a green at the main signal.  Meanwhile the off-ramp traffic that gets a green indication are going to want to go too.

Not a totally unusual situation.  I've seen this elsewhere at relatively normal intersections where timing may not allow for full clearance, especially in the event of traffic congestion.  Normal people don't intentionally hit each other, so they just let the remaining vehicle go, or merge in as necessary.

[roadman65]

I-4 Exit 58 is slated for one soon.  With all the left turns that take place there, it will work perfectly once in place.  Though not a fan of them, I do commute through here daily and its a big issue with CR 532 turning left onto I-4 and those exiting I-4 WB turning left onto CR 532.