5, 6 and 7 ramp parclos

[Peter Freeman]

I really dislike the term 'partially-unrolled cloverleaf'. It's not descriptive enough. 'Cloverstack' is much better because it more accurately characterises the interchange as 50% cloverleaf (ie. two loops) combined with 50% stack (ie. two semi-direct connectors).

'Cloverstack' isn't a complete description. For example, are the two loops in adjacent quadrants or opposite ones? And are the connectors central and symmetrical, as they would be in a regular 4-level stack, or are they squashed out to one side, avoiding the loops? So there are these, and other, variants, which might need to be stated.

[jackal]

Using international terminology one would call the design a cloverstack variant (it has one stack-like connector and one turbine like connector running parallel with each other whereas a classic cloverstack uses two stack-like connectors, e.g., https://www.google.co.uk/maps/@51.12081 ... 369921,12z).

Jackal, did you insert the wrong URL here ?

Aye:
https://www.google.co.uk/maps/@48.9732494,2.4883389,16z

[jackal]

I really dislike the term 'partially-unrolled cloverleaf'. It's not descriptive enough. 'Cloverstack' is much better because it more accurately characterises the interchange as 50% cloverleaf (ie. two loops) combined with 50% stack (ie. two semi-direct connectors).

'Cloverstack' isn't a complete description. For example, are the two loops in adjacent quadrants or opposite ones? And are the connectors central and symmetrical, as they would be in a regular 4-level stack, or are they squashed out to one side, avoiding the loops? So there are these, and other, variants, which might need to be stated.

And a classic cloverstack is just as "unrolled" as a "partially-unrolled cloverleaf" - maybe more so. So partially-unrolled cloverleaf is not a very logical name, and certainly more wordy and less descriptive than cloverstack. I still use it simply because, as you note, a name is useful for each cloverstack variant, and this one happens to have one, even if it's unfortunate in some respects.

I included a classification of cloverstack subtypes in my survey of US four-way freeflow interchanges:

Cloverstack refers to design with two loops in opposite corners, including 'cloverstack classic', 'clovercorner', partially-unrolled cloverleafs, cloverturbines, and clovermills.

I don't count interchanges with two adjacent loops as cloverstacks, as they are functionally rather different, having weaving between the loops. I rather call them two loop interchanges.

More broadly there are common components to freeflow interchanges that can be used to classify them. How many arms do they have? Full access or partial? Are merges and diverges all on the nearside or are some offside? How many long turns are there of each type (loop, stack-like, corner cut, turbine-like, windmill-like)? And if there are two loops, are they opposite or adjacent?

A cloverstack is four arm, full access, nearside, with two loops in opposite corners, and two other long turns as follows: two stack-like (classic cloverstack), two corner cuts (clovercorner), two turbine-like (cloverturbine), two windmill-like (clovermill), or one corner cut and one turbine-like (partially-unrolled cloverleaf). These are the main variants at any rate - others are possible.

[Mapper89062]

Before the new slip road was built for the A14 upgrade, Girton was essentially a four-ramp parclo. Obviously any folded diamond or dumbbell could be thought of as one, but Girton was very unusual in that a) all permitted movements were freeflow and b) it allowed two adjacent turns rather than two opposite ones.

[Peter Freeman]

I don't count interchanges with two adjacent loops as cloverstacks, as they are functionally rather different, having weaving between the loops. I rather call them two loop interchanges.

'Single-sided semi-cloverleaf interchange' ?

[jackal]

I don't count interchanges with two adjacent loops as cloverstacks, as they are functionally rather different, having weaving between the loops. I rather call them two loop interchanges.

'Single-sided semi-cloverleaf interchange' ?

I could have called them adjacent interchanges, as the loops are adjacent, and short is good given this is only part of the overall name. For instance this is a 2 loop turbine: https://www.google.co.uk/maps/@41.78149 ... a=!3m1!1e3 "Adjacent turbine" would have been okay, though even that takes up more space and is harder to read in a spreadsheet! "Single-sided semi-cloverleaf turbine" would be a bit much...

[Chris5156]

Parclo is the global term for a partial cloverleaf: https://en.m.wikipedia.org/wiki/Partial ... nterchange

Some time after this use of parclo was established, Sabristi starting using the term "partially-unrolled cloverleaf" for a freeflow interchange design that is used at three sites in the UK. I believe this began with, or was at least popularised by, Chris's CBRD website, the ancestor of the site you linked. Using international terminology one would call the design a cloverstack variant (it has one stack-like connector and one turbine like connector running parallel with each other whereas a classic cloverstack uses two stack-like connectors, e.g., https://www.google.co.uk/maps/@48.9732494,2.4883389,16z).

Over time some SABRE users heard the word parclo used in the international sense, misunderstood it as a contraction of partially-unrolled cloverleaf, and started using it with this new meaning. But certainly now, where we have better familiarity with international terminology, parclo is generally understood on SABRE as referring to a partial cloverleaf rather than a partially-unrolled cloverleaf.

I've been meaning to reply to this for ages and am only getting round to it now.

I agree that "partially unrolled cloverleaf" is not an elegant name - it isn't brilliantly descriptive, and it's overly wordy. I don't know exactly where it came from and have no great attachment to it. I don't believe that I coined the term; when I set up the original "interchanges" pages on what was then CBRD, I was trying to document the types of junction that existed in the UK and provide names for them, so the list omitted (and still omits) junction types we don't have in the UK, and applied the best names for them that I could find. That particular name must have come from somewhere. My guess would be that it was used, at least occasionally, on the usenet group misc.transport.road, which I frequented in those days and which also gave us terms like "multiplex" that were in part popularised among UK road enthusiasts by my site.

I'm part way through updating the Interchanges pages with new pictures and text, and am going to add the parclo as a type in its own right, since it's currently missing. As part of that work I've already fixed some other terminology to bring it into line with international practice and I'm open to ditching the name "partially unrolled cloverleaf" altogether too.

At the moment my thinking is that the page should be called "cloverstack" and should clarify, in the text, that the UK has just one common type of cloverstack in use - the Denham/Croft/Winnersh type - but that internationally there are other kinds. But I'd be interested to hear other views on that, along with any other questionable wording on those pages that needs attention.

[jackal]

Parclo is the global term for a partial cloverleaf: https://en.m.wikipedia.org/wiki/Partial ... nterchange

Some time after this use of parclo was established, Sabristi starting using the term "partially-unrolled cloverleaf" for a freeflow interchange design that is used at three sites in the UK. I believe this began with, or was at least popularised by, Chris's CBRD website, the ancestor of the site you linked. Using international terminology one would call the design a cloverstack variant (it has one stack-like connector and one turbine like connector running parallel with each other whereas a classic cloverstack uses two stack-like connectors, e.g., https://www.google.co.uk/maps/@48.9732494,2.4883389,16z).

Over time some SABRE users heard the word parclo used in the international sense, misunderstood it as a contraction of partially-unrolled cloverleaf, and started using it with this new meaning. But certainly now, where we have better familiarity with international terminology, parclo is generally understood on SABRE as referring to a partial cloverleaf rather than a partially-unrolled cloverleaf.

I've been meaning to reply to this for ages and am only getting round to it now.

I agree that "partially unrolled cloverleaf" is not an elegant name - it isn't brilliantly descriptive, and it's overly wordy. I don't know exactly where it came from and have no great attachment to it. I don't believe that I coined the term; when I set up the original "interchanges" pages on what was then CBRD, I was trying to document the types of junction that existed in the UK and provide names for them, so the list omitted (and still omits) junction types we don't have in the UK, and applied the best names for them that I could find. That particular name must have come from somewhere. My guess would be that it was used, at least occasionally, on the usenet group misc.transport.road, which I frequented in those days and which also gave us terms like "multiplex" that were in part popularised among UK road enthusiasts by my site.

I'm part way through updating the Interchanges pages with new pictures and text, and am going to add the parclo as a type in its own right, since it's currently missing. As part of that work I've already fixed some other terminology to bring it into line with international practice and I'm open to ditching the name "partially unrolled cloverleaf" altogether too.

At the moment my thinking is that the page should be called "cloverstack" and should clarify, in the text, that the UK has just one common type of cloverstack in use - the Denham/Croft/Winnersh type - but that internationally there are other kinds. But I'd be interested to hear other views on that, along with any other questionable wording on those pages that needs attention.

Thanks for this Chris. Good to have the historical record put right.

I'd perhaps suggest amending the name of the page to cloverstack, but include partially unrolled cloverleaf as an alternative name for the UK type. Btw, the official names in DMRB (CD 122) are 4-way, 2 level or 4-way restricted height interchange, which are no better than partially unrolled cloverleaf! See pp.64-65: https://www.standardsforhighways.co.uk/ ... d218e74ffd

A few further thoughts on the interchanges pages:

Triangle - I don't recall ever hearing this name elsewhere. It's a semi-directional T.

Not sure about these disadvantages for a stack:

-Need for a large number of complex bridge structures.
-Cost is comparable to a moon landing mission.

I'd have thought it has a small number of structures compared to other freeflow designs - classically, three or four bridges compared to ten or more for a whirlpool. Admittedly, they are big bridges!

I've never seen good costings in the UK but I found this from the US suggesting it is intermediate between cloverstacks and turbines, which is what I'd have expected: viewtopic.php?p=1145600

Similarly, I don't think land take or earthworks are that high for what it does. The issue is, as you mention, visual intrusion.

The above perhaps suggests a rethink of this on the whirlpool page: "it's not nearly as complex or expensive as a four level stack would be". And this: "One of the principal reasons to choose a whirlpool is for a location where space might be limited". I think a whirlpool takes up more space than a stack for a given sliproad speed.

It might also be worth mentioning that an octopus is internationally known as a windmill.

CD122, pp. 58-59 is worth a look if you do a parclo page. It only includes four ramp types and calls them half-cloverleafs.

The other main unrepresented interchange types are compact GSJs. See especially CD122, p. 52. Perhaps this could be as variants on the parclo page, along with 5, 6, 7 ramps.

That's all for now. Probably more than you bargained for!

[Peter Freeman]

... the Denham/Croft/Winnersh type ...

Denham and Winnersh are identical, Croft is very slightly different.

[jackal]

^ Denham is much bigger than Winnersh.

[Chris5156]

I'd perhaps suggest amending the name of the page to cloverstack, but include partially unrolled cloverleaf as an alternative name for the UK type. Btw, the official names in DMRB (CD 122) are 4-way, 2 level or 4-way restricted height interchange, which are no better than partially unrolled cloverleaf! See pp.64-65: https://www.standardsforhighways.co.uk/ ... d218e74ffd

Ugh - DMRB names for junctions are rarely very snappy. But I agree all the alternative names should go in. Actually - as I redraft, a collection of "other names" for each one might be useful: the three level stacked roundabout can also be a stackabout or a volleyball (yuck), and so on.

Triangle - I don't recall ever hearing this name elsewhere. It's a semi-directional T.

I'd need to go back and have a rummage to find the origin of that too, but it's an improvement on the original claim of "directional T" which was a misunderstanding and plain wrong. I think "triangle" was the preferred name in planning documents for Ray Hall and/or Thurcroft. It's not an internationally recognised name though.

Not sure about these disadvantages for a stack:

-Need for a large number of complex bridge structures.
-Cost is comparable to a moon landing mission.

I'd have thought it has a small number of structures compared to other freeflow designs - classically, three or four bridges compared to ten or more for a whirlpool. Admittedly, they are big bridges!

The perils of writing something in your teens and having it stay online for several decades. I think all your points on this are valid. That page hasn't been revised at all yet and, reading it again as I write this, it shows.

The above perhaps suggests a rethink of this on the whirlpool page: "it's not nearly as complex or expensive as a four level stack would be". And this: "One of the principal reasons to choose a whirlpool is for a location where space might be limited". I think a whirlpool takes up more space than a stack for a given sliproad speed.

It might also be worth mentioning that an octopus is internationally known as a windmill.

Agree with all. Here's a question - is there any consensus on names for turbines/whirlpools where the two entry sliproads for a given direction merge into the mainline separately, or merge together and enter as a single sliproad? The difference, in other words, between Thorpe and Theydon. I've always just called both a Whirlpool but they are subtly different. I see that CD122 calls the Theydon type a "cyclic" junction, but considers the standard type to have two diverges and one merge on each mainline, which isn't a version that actually exists in the UK!

CD122, pp. 58-59 is worth a look if you do a parclo page. It only includes four ramp types and calls them half-cloverleafs.

Thank you!

The other main unrepresented interchange types are compact GSJs. See especially CD122, p. 52. Perhaps this could be as variants on the parclo page, along with 5, 6, 7 ramps.

Yes, I've ummed and ahhed about compact GSJs - the highway design orthodoxy in the UK is that they are a "type" in their own right, but in practice most compact GSJs are some combination of diamond and parclo elements, and the unifying feature is that they have tighter geometry and two-way link roads. It's one of the difficulties in this field that one junction type gradually morphs into another as you make iterative changes, and it's a fool's errand to try and draw a clean line that says this is a parclo but that is a compact GSJ. Equally, though, they are a thing, so perhaps they deserve to join the list.

That's all for now. Probably more than you bargained for!

All feedback is good feedback! Better to get it right.

[Bryn666]

"Triangle" has probably arrived on a small boat from France - the OG three-way interchange for them, the Rocquencourt Triangle, opened in 1950 to connect the now A12 and A13. It originally was a simple interchange but has since been given additional slips to remove weaving conflicts with the adjacent parclo.

[jackal]

Good stuff Chris!

Here's a question - is there any consensus on names for turbines/whirlpools where the two entry sliproads for a given direction merge into the mainline separately, or merge together and enter as a single sliproad? The difference, in other words, between Thorpe and Theydon. I've always just called both a Whirlpool but they are subtly different. I see that CD122 calls the Theydon type a "cyclic" junction, but considers the standard type to have two diverges and one merge on each mainline, which isn't a version that actually exists in the UK!

My view is that the number of merge or diverge points on the mainline is irrelevant to interchange typology. Compare Thorpe and an example in Jacksonville:

https://www.google.co.uk/maps/@30.25035 ... 164278,15z
https://www.google.com/maps/@51.4011024,-0.5371032,15z

They are absolute textbook turbines/whirlpools, and could even be mistaken for each other at first glance. But the Jacksonville one merges the long (left in US) turns directly into the mainline. To me that does not change the fact that they're the same kind of interchange.

I consider a turbine/whirpool to have turns that sweep around the interchange, while a windmill/octopus has at least some tighter, less cyclic turns. Worsley has two windmill-like turns (E to N and S to E), and Hamilton three (N to W, E to N, and W to S), so I'd call them windmills.

The awkward case is Theydon, because its turns somewhat cycle around the interchange in turbine fashion, but E to N and W to S become tight and less cyclic. For that reason I think of it as a windmill though it's a borderline case.

This example in Amarillo, Texas is similarly borderline with all four long turns a little compromised: https://www.google.com/maps/@35.1921214 ... ,16z?hl=en

The Wikipedia interchange page agrees with the basic sentiment: "A windmill interchange is similar to a turbine interchange, but it has much sharper turns, reducing its size and capacity."

[Chris5156]

My view is that the number of merge or diverge points on the mainline is irrelevant to interchange typology. Compare Thorpe and an example in Jacksonville:

https://www.google.co.uk/maps/@30.25035 ... 164278,15z
https://www.google.com/maps/@51.4011024,-0.5371032,15z

They are absolute textbook turbines/whirlpools, and could even be mistaken for each other at first glance. But the Jacksonville one merges the long (left in US) turns directly into the mainline. To me that does not change the fact that they're the same kind of interchange.

Sounds sensible to me.

I consider a turbine/whirpool to have turns that sweep around the interchange, while a windmill/octopus has at least some tighter, less cyclic turns. Worsley has two windmill-like turns (E to N and S to E), and Hamilton three (N to W, E to N, and W to S), so I'd call them windmills.

The awkward case is Theydon, because its turns somewhat cycle around the interchange in turbine fashion, but E to N and W to S become tight and less cyclic. For that reason I think of it as a windmill though it's a borderline case.

This example in Amarillo, Texas is similarly borderline with all four long turns a little compromised: https://www.google.com/maps/@35.1921214 ... ,16z?hl=en

The Wikipedia interchange page agrees with the basic sentiment: "A windmill interchange is similar to a turbine interchange, but it has much sharper turns, reducing its size and capacity."

That's an angle I hadn't thought about very much before, but I think you're right - Hamilton and the part-built one near Dunfermline are plainly windmills ("octopus" appeared in writing for the Dunfermline one, where it appeared to be talked up as a template for four-way interchanges on the future Scottish motorway network, so there's some evidence that name was preferred in the 60s in Scotland at least). Without having given it a great deal of thought I'd considered the single diverge and two merges per mainline to be a defining characteristic, but actually, it's the more awkward geometry and the more compact overall size that makes them different. This long straight into a sharp bend isn't whirlpool/turbine territory at all.

[jackal]

To avoid going further OT I made a thread on whirlpools and octopuses here: viewtopic.php?t=44862

[Conekicker]

Does Redhouse count?
https://www.google.co.uk/maps/@53.57887 ... ?entry=ttu

[jackal]

Does Redhouse count?
https://www.google.co.uk/maps/@53.57887 ... ?entry=ttu

Yes, I'd say it's a variant of a 5 ramp parclo. I'll add it to the list.

[Bryn666]

Does Redhouse count?
https://www.google.co.uk/maps/@53.57887 ... ?entry=ttu

Yes, I'd say it's a variant of a 5 ramp parclo. I'll add it to the list.

It's very frustrating that an at-grade right turn was introduced when there was space for a (tight) loop using the former exit from the roundabout this replaced.

[Peter Freeman]

Ugh - DMRB names for junctions are rarely very snappy. But I agree all the alternative names should go in. Actually - as I redraft, a collection of "other names" for each one might be useful: the three level stacked roundabout can also be a stackabout or a volleyball (yuck), and so on.

My understanding has always been that three level stacked roundabout = 'stackabout', but 'volleyball' is different: its 'roundabout' level is square, and is signalised in a more conventional 'crossroads' style. I agree that volleyball is a really yuck name - it's not descriptive and doesn't tell you anything.

I use 'three level diamond' instead of 'volleyball' for that square-about with double bypassing. As far as I know, there aren't many. Brisbane in Australia has one **. At a stretch, the extra level for service roads that is included in some USA four level stacks is related, but that's not an interchange in itself.

Of course, whether using a roundabout or a square-about, it's technically not freeflow, and so doesn't belong in this whirlpooly discussion.

** https://www.google.com.au/maps/place/SS ... ?entry=ttu

[jackal]

^ There are a fair number of volleyballs, e.g.:

Dusseldorf: https://www.google.co.uk/maps/@51.2529768,6.7716335,16z
Joburg:
https://www.google.co.uk/maps/@-26.2681 ... 30911,16z/
Philly: https://www.google.co.uk/maps/@39.93649 ... 34057,16z/
3 along 8 Mile Rd, Detroit:
https://www.google.co.uk/maps/@42.44634 ... 955476,16z
https://www.google.co.uk/maps/@42.44587 ... 949306,16z
https://www.google.co.uk/maps/@42.44469 ... 031022,16z

There are quite a few in Texas, e.g.:
https://www.google.co.uk/maps/@29.59028 ... 900296,16z
https://www.google.co.uk/maps/@29.78487 ... 449381,16z
https://www.google.co.uk/maps/@29.71215 ... 541169,16z
https://www.google.co.uk/maps/@32.89551 ... 723334,16z
https://www.google.co.uk/maps/@29.55421 ... 861937,16z

Often this is as a step in an upgrade trajectory leading to a stack. Here are some partway along:
https://www.google.co.uk/maps/@30.19470 ... 235315,16z
https://www.google.co.uk/maps/@29.93553 ... 564317,16z
https://www.google.co.uk/maps/@30.12943 ... 291708,16z
https://www.google.co.uk/maps/@30.33839 ... 002598,16z
https://www.google.co.uk/maps/@30.32466 ... 727187,16z

I agree that a stackabout remnant at a full stack doesn't count as a volleyball, as the signalized level is not then used for freeway to freeway. But IMO it does still count if some of the freeway to freeway movements remain via the signalized level, in which case, throw a stone in Texas and you'll hit one.

And I agree the name is unfortunate, though it seems pretty well embedded.