Rectengular duct elbow with different size

[tikos.mate]

Hello everyone!

I have a problem about a duct elbow. In real life, there is an option to use an elbow with two different size of the connector (only one, same dimension). My first thoght about to find a family on the web, but I couldn't find anything. Then, I wanted to make the family, and tried the Swept Blend. And it works but not perfect. The two size is different, but the outter arch of the elbow is not a real. Can I just made the sketch path with a line, and an arch? In Swept Blend I can't make it, the Sweep is not an option in that case. I tried to make a void extrusion, that cuts it in the proper way, but it was a dead end. I uploaded the family (it has extra parameters, wich are help to count the fitting area, and a shared parameter for count the angle of the elbows (Idom_szög))

I know it is a specific question, and maybe this kind of elbow is only use, when the Angle is 90°, but if there is an option to make this, it will be useful in other angles too.

If something is not understandable, please let me know, I try to explain properly.

[iainsavage]

Swept blend would be the way to go so it should just be a matter of sorting out the path.

There are reducing pipe elbows available and maybe they would give you a clue as to how to construct the family.

I also found this but haven’t tried it https://www.bimmepaus.com.au/product/bend-reducing-koolduct/

I can’t access your own family at the moment but I’ll try later if no-one else answers in the meantime.

[iainsavage · Accepted solution]

Attached is a family from the out of the box UK metric library and a similar one from the English-Imperial library. Haven't tried them though but might give you some clues. They're maybe slightly off where I've highlighted.

Also attached is a reducing pipe elbow from Victaulic (I deleted the collars) and as you can see there is a slight misalignment of the swept blend at the inner curve on this one but its not bad. 

Maybe you can adjust your family using the principles in these examples (or just use the OOTB families?)

[tikos.mate]

Thanks for the answer @iainsavage . I inspected all the families, and all have the same problem, the bend part is not behave as in the real life. I tried a lot of parameter combination, but not worked at all. 

To solve this problem, I only need to change the Duct Width in the two sides, the height is the same. I made an extrusion, because it is better to handle. The only problem, that I couldn't parameter the angle, but the 90° angle is the most common value, other angle is very rare.

(The problem, if I parameter the angle between the connectors, the inner arch is still the same length. I tried to fix it with reference planes, and lines, but sadly I am not trained enough to made it by myself. I have no time for this.)

I upload the family, if someone need, or can upgrade to a paramerable angle.

[iainsavage]

The core issue here is being able to accurately plot the sweep path for various combinations of widths, so it really needs someone who knows how actual reducing bends are made to explain how that path would be worked out in a workshop or CNC machine.

HVCA/BESA spec DW/144 doesn't give dimensions, it just shows this:

I don't have access to SMACNA standards.

I could maybe reach out to a couple of ductwork manufacturers that I know to ask how they would do it but maybe someone on the forum already knows the answer - having said that I generally find these days that ductwork contractors (in UK at least) won't make one-off reducing bends like this because its cheaper to use a standard bend with a separate taper bolted onto the outlet.

Your fitting family is, by the way, a Generic Model - you need to change it to Duct Fitting and then check that in the model the correct ASHRAE code is applied for pressure drop calculations. When I insert it I get no table is applicable error (I get the same with the ootb fittings):

...so you probably need to use the "Specific Coefficient" option and put in your own loss coefficient.

You've also modelled the family upwards from the reference level so it will only work automatically with bottom aligned ducts. If you want it to work with middle aligned ducts have the extrusion height equidistant above and below the ref level so that the connectors are on the ref level.

Lastly, that looks like a very tight radius on the inner curve - is that deliberate? Are you not worried about flow separation and eddies? 

[iainsavage]

Just for the record, one of my contacts sent me this from an old SMACNA manual:

[tikos.mate]

Thank you for the inspect of my family. I changed it to Duct fitting, and I will change the extrusion, that you mentioned. 

I don't use ASHRAE, only Specific loss for elbows, it make the pressure drop more accurate to our calcualting method, so it is not a problem for me. On the picture you can see an elbow, which we usually use (in Autocad, befor Revit), I tried to made the family like this. 

[j_feeke]

In case anyone still needs it, as I am currently creating plain revit family's that match our production specs, I have created attached base family for Reducing Bend's. It doesn's use a sweep, but is a extrusion, which gives perfect control over the final form. 

It is using a pretty nasty trigiometric formula (I combined the separate calculations into a single formula) to get the valid exterior radius (if you are interested, i added the picture where you can see the different needed steps to get to it).

Anyway, attachted family works, but currently limited to Width 2 <= Width 1 (i will have to make it work for W1<W2 too someday). Let me know what you think of it.

[iainsavage]

---

@j_feeke  schrieb:

It doesn's use a sweep, but is a extrusion, which gives perfect control over the final form

---

That surely only works if its only the width which changes and won’t work if the height changes? The drop check bend allows both dimensions to be different and that requires the use of a swept blend unless you’re using voids to cut parts out of your extrusion?

[j_feeke]

Yes, it does. It's designed for our production specs, we dont want to (and dont) produce bends with differtent heights. That would require larger inner radius to work (due to sheetmetal not really forming on smaller radius) where we only use 100, 120 and 170 mm. But if you would want to do it, void's would be the way to go :-). 

With the larger radius in use, you should have room for an reducer for the height after the bend.