Topic: How Accurate can you be?

[Krushert]

A couple of us Archies are wondering a couple of things.

How accurate is your equipment? 

And how accurate can you stake (lack of better term) the new building out in relation to what is drawn on paper?

I have a new building being plunk down between a rock (set backs) and a hard place (existing building) with very little wiggle room.  And of course the new building is going to be at some magical angle and be parallel to the set backs.  For coordination purposes, I have xref in the setbacks and  provide (forced to) dimensions from set backs to tie points on the building.  The we we have it laid out it making us wonder how the heck are they going to do it?

[MickD]

On larger projects I always used a surveyor to at least stake out/mickey pin some major grids/corners in for us, this way they were always spot on. The expense (not that much considering) is well worth it 'cause if the job starts 'square' it has a better chance of finishing square

[sinc]

How accurate is your equipment? 

And how accurate can you stake (lack of better term) the new building out in relation to what is drawn on paper?

This is something that has come up a lot for us recently.

Our total stations are among the better ones in the business, and their positional accuracy is roughly within about 0.03 feet - roughly 3/8 of an inch - assuming shot lengths are kept short (less than 400 feet).  A given instrument may consistently return measurements within 0.01 feet, leading some people to think they are getting more-accurate results, but that isn't the case.  Just because a given instrument consistently returns the same result doesn't mean it is returning the correct result.  Bring another machine out to the job site and take the same measurement, and you may get slightly different values.  And of course, there's all the temperature and pressure effects, which cause their own variations.  Not to mention the other "errors", such as centering errors, survey rods that are not precisely straight, etc.  So 0.03 feet is about the minimum possible in most conditions.

But usually, that's not what matters.  What we've found is a more important point is "How accurate can this item be BUILT?"  In general, we find that we can stake things more-accurately than they can be built.  For example, thermal expansion can cause significant varation in a concrete panel, so even after it's built, there can be more than 0.03' of variation in the measurements, depending on the time of day and the weather.  And then there's all the stuff that happens during construction, after we have staked points.  Workers walk around on forms after they are set, but before concrete is poured.  Forms start to give from repeated use.  Vibration from large machines causes random effects.  There are a whole slew of things.  And that doesn't even count problems like control points that get accidentally or erroneously destroyed and have to be re-established in the middle of the project, etc.

The leaps in technology seem to be having a strange effect.  There seems to be a tendency in recent years for people to forget about the inherent tolerances necessary for real-world construction.  People seem to feel that if it can be drawn in Autocad, it can be built.  We had a project recently where we had been doing lots of work for a company, up until they hit the point where they needed to pour a seven-foot slab of concrete 30 feet in the air, and said they had a tolerance of 1/16 of an inch.  The survey company would be held liable if they failed to meet this tolerance.  We tried to tell them that it's impossible to build anything that close except for maybe under tightly-controlled conditions, which is not the way most people would describe springtime in Colorado.  But they didn't like that answer - possibly they've been watching too much Discovery Channel - so they tried to find someone else to do the surveying.  They couldn't find a survey company in Colorado that would take the job.  So they found an unlicensed LSI who told them that we were basically incompetent, and it is very possible to get that slab built to within 1/16 of an inch.  He started spouting something about carrying all his coordinates out to ten digits and closing his horizons and lots of other BS.  His story was what they wanted to hear, so they hired him to do the surveying for the project, despite the fact that he is unlicensed and using rented equipment.  We're waiting to see how it turns out... 

[Mark]

How accurate is your equipment?

Our equipment is very accurate. But it's a lot more then the accuracy of the equipment in this situation.

And how accurate can you stake (lack of better term) the new building out in relation to what is drawn on paper?

Closer than you might think. Redundancy of measurements plays an important role in accurately staking a structure.

As sinc said, it's can you build it as accurately as I can stake it! I did most of the layout out for a major theme park project here in Florida. We set anchor 1-1/4" bolts, poured in place, for a large steel structure that was part of the project. We were given 1/4" tolerance for these bolts which we accomplished. Some days later when the erection crew got there and started to set the first column and it didn't quite fit our bolt pattern they took their 16lb sledge hammer and made the bolts fit their column, so much for accuracy! This wasn't my first red-iron job so I wasn't surprised but the look on the engineers face was priceless.

The we we have it laid out it making us wonder how the heck are they going to do it?

The existing building is the key. If you have a competent field crew they can stake the building. Sounds like a nice challenge, wish I could give it a shot.

We had a project recently where we had been doing lots of work for a company, up until they hit the point where they needed to pour a seven-foot slab of concrete 30 feet in the air, and said they had a tolerance of 1/16 of an inch.

That is just crazy! Sounds like the engineer has never been on a project site.

[numa]

Being super accurate setting points is doable, it just takes lots of time.  My crew has in the past, set verifably accurate column lines, bolt anchors, etc at accuracies of BETTER THAN 0.01 feet.  Did we use a total station, nope.  we didn't

When you have to be REALLY accuate, get out your chain, plumb bob, and 1" theodolite (yeah, you know, the old ones), and factor in temp correction, take your time, and boom, you have a building that closes in the thousandths. 

We sometimes monitor large/long walls for movements, and often times have to detect 0.01 movements. 

so sure, we can make it accurate, but more than likely, your shots of the 'surronding' buildings aren't going to be good enough .  It can be very difficult to survey existing structures because they are very rarely straight up and down.  Also, make sure your geotechnical engineer is involved early on.  If you have building really close together, you should tie them togetther so they settle as a unit, as opposed to 'floating' and running into each other. 

Cheers.

NUMA, PE

[sinc]

We sometimes monitor large/long walls for movements, and often times have to detect 0.01 movements. 

I can see that being possible in certain situations.  But if you are using equipment setup on tripods, I doubt you are really even getting 0.01 accurately.  Situations like this generally involve permanent mounts for the equipment, so the variance between setups does not affect the results.

And then there's a large question as to whether there is any point.  If the construction crew can only build it to within 3/4 of an inch, then using a chain and 1" theodolite (if you can even find them these days) is a waste of time.

Anyone see that Discovery Channel thing about the new football stadium in Arizona?  It has twelve rails on which the entire field slides in and out of the stadium.  They had ridiculous tolerances for the rails.  One of them was a fraction of a degree out of tolerance, so they spent three weeks shimmying and jogging and trying to get that rail into the engineer's tolerance.  When they were done, they had the same amount of error in the opposite direction.  Finally the engineer relaxed the tolerance, and everything works fine.  But they wasted three weeks trying to achieve a ridiculous tolerance that in no way affected the performance.

[Josh Nieman]

We sometimes monitor large/long walls for movements, and often times have to detect 0.01 movements. 

Anyone see that Discovery Channel thing about the new football stadium in Arizona?  It has twelve rails on which the entire field slides in and out of the stadium.  They had ridiculous tolerances for the rails.  One of them was a fraction of a degree out of tolerance, so they spent three weeks shimmying and jogging and trying to get that rail into the engineer's tolerance.  When they were done, they had the same amount of error in the opposite direction.  Finally the engineer relaxed the tolerance, and everything works fine.  But they wasted three weeks trying to achieve a ridiculous tolerance that in no way affected the performance.

haha yea I saw that... I laughed out loud.

[pmvliet]

One of them was a fraction of a degree out of tolerance, so they spent three weeks shimmying and jogging and trying to get that rail into the engineer's tolerance.  When they were done, they had the same amount of error in the opposite direction.  Finally the engineer relaxed the tolerance, and everything works fine.  But they wasted three weeks trying to achieve a ridiculous tolerance that in no way affected the performance.

I think it all comes down to cost and time. As with anything in life, more accuracy will require more time which means more money...

This is something I'm learning more about doing structural steel detailing and installing some steel here and there in the field. The next round we are setting anchor bolts for columns, then I have a 102'-8" truss that will be attaching to columns...

I believe the anchor bolts are being set by the concrete contractor... (will probably double-check them before they pour though)

Pieter

[Greg B]

And still cut with a bull dozer.

[jpostlewait]

Interesting discussion.
I've spent some time a while ago doing a few precision jobs.
Locating anchor bolts on bridge piers, checking columns in a power plant for Xing, etc.
There is always a tolerance, no matter what the contract says.
Nobody knows for sure how far it is from here to there.
Some just get closer than others.

[Keith™]

In all of our plans we allow 1/16" tolerance for all assemblies, but in reality it will generally work if it is within 2" for most hard set dimensions. I am one of those types that design with reality in mind. So ... I need this corridor to be 48", there ya go ... 52" it is ...
Contractor - Do we really need 52" in this corridor?
Me - Yep, we need to make it that to keep within the guidelines in the contract.
Contractor - Ok

2 weeks later

Contractor - Someone screwed up on that corridor and made it 52" from outside to inside, that means it ended up being 48 1/2" ... is there any way we can get the owner to budge on the requirements .. I really don't want to spend the time and money to change it if I can get out of it.
Me - Well, I'll schmooze the owner a bit and get him to see it like that.

In the end, the project is within tolerance and everone is happy ... so what if I lied to the contractor ... I am now a miracle worker as far as he is concerned

[Dent Cermak]

Our State Minimum Technical Standards require positional accuracy within 0.05'. It is not uncommon to stake out to 0.01'. Our GPS equipment and network are supposed to deliver sub-centimeter accuracy. So far, it is.
We are running a project now to check building movements due to Yazoo Clay. We are using digital levels and supplying readings to 0.001'.
The equipment of today is VERY accurate. The key is well trained personnel using proper field procedures. But then that was true back in the day of Dumpy Levels and theodolites.

[numa]

When we do wall monitoring, we usually have a permament mount.  It helps to minimize errors in setup (height) etc. 

What can I say, I've been more accuate than 0.01 feet in the past, god I loved workin in the field.   

3/4 of an inch?  wish I could get away with that.

NUMA

[sinc]

3/4 of an inch?  wish I could get away with that.

When I said 3/4 of an inch, I was talking about what eventually gets built, not what gets staked.  Like I said, we can generally stake things more-accurately than they get built.  A variance of 3/4" from plans in a finished building is quite common.

[sinc]

Our State Minimum Technical Standards require positional accuracy within 0.05'.

I have to wonder exactly what that means.  What exactly is required to be within 0.05'?  How do you determine that it is actually within 0.05'?

The ALTA/ACSM Land Title standards, which AFAIK is used in all states in the country for the sale of land, only requires 0.07'+50ppm, and that 50ppm can be significant in large sites.

It is not uncommon to stake out to 0.01'. Our GPS equipment and network are supposed to deliver sub-centimeter accuracy. So far, it is.
We are running a project now to check building movements due to Yazoo Clay. We are using digital levels and supplying readings to 0.001'.
The equipment of today is VERY accurate. The key is well trained personnel using proper field procedures. But then that was true back in the day of Dumpy Levels and theodolites.

Simply because you are staking out to 0.01' or supplying readings to 0.001' doesn't mean that's how accurate you are.  You can stake out a point that your total station tells you is within 0.01' of the calc point, but if your machine only is only rated to be accurate to within 0.03', you can only say you know you have it staked to within roughly 0.03' of the actual point.  And even though you can read to 0.001' with a digital level, I'm sure you adjust your level loops, and don't close flat on every loop you run.  And since, by law, a recovered monument is what holds over all else, GPS coordinates only do so much good.

Surveying is not an exact science.  There are rather elaborate procedures defined to deal with error, because we know it will always exist, regardless of the technology.  Property pins move around due to everything from weather to earthquakes to plate tectonics, and are often obliterated.  Critical documentation can be lost or misinterpreted.  And then there are the silly errors like transposed digits or mislabeled courses that creep in there occasionally, even when extreme care is taken.

[mjfarrell]

Well said Richard!

Now how do you feel about showing Record & Measured vs showing Positional Error of a discovered monument or pin?

[sinc]

Well said Richard!

Now how do you feel about showing Record & Measured vs showing Positional Error of a discovered monument or pin?

I think it depends on what I'm doing, what sort of records already exist, and the exact circumstances.  I don't think there's an absolute rule.

[sinc]

But to try to get back to the original point, I think that no matter what, the important question is not "how closely can it be staked?" but "how closely can it be built?"

That depends largely on the type of building and how it is constructed.  If it is one of those prefabricated metal buildings, then it will probably end up being pretty close to plan.  If it's built entirely in-place and has a masonry facade or something like that, then you can get rather significant variations.

Then keep in mind that the same issues apply to the asbuilts of the existing building.  Sometimes it is really hard to tell exactly where the building edge or building corner lies.  It is often quite easy to fudge 0.1' or more just with that.  Add to that the fact that building walls can easily drift 0.2' or more either way along their length, and it becomes quite easy to have significant error in the asbuilt of the existing building.

How tight exactly are you talking?  Is it a situation where you have less than one foot of wiggle-room?  When placing buildings, we generally try to put them at least 1 foot from setbacks and easements, just to be safe, but that's a pretty generous buffer.  I'd still want to leave at least 0.5', again just because the repurcussions are so bad if the building is built wrong.  But if you don't even have 0.5', then the most important thing is how it is being constructed, out of what kind of materials, and does it have a facade.  And know that you are in dangerous territory.

[Krushert]

First off I want to say thanks to all for your Input.  I have been purposely keeping my mouth shut and sitting on my hands and just reading your post so I learn something.   

How tight exactly are you talking?  Is it a situation where you have less than one foot of wiggle-room?  When placing buildings, we generally try to put them at least 1 foot from setbacks and easements, just to be safe, but that's a pretty generous buffer.  I'd still want to leave at least 0.5', again just because the repurcussions are so bad if the building is built wrong.  But if you don't even have 0.5', then the most important thing is how it is being constructed, out of what kind of materials, and does it have a facade.  And know that you are in dangerous territory.

To us Archies it seems tight and a tough layout, but from reading your post, it may not be .
Back ground
We are renovating an 4 floor existing 1912 school into elderly housing 1 bedroom apartments.  Then off to one side the new building that is  a  4 floor with 40 1 bedroom units.  Then between the buildings there is going to be a 1 1/2 story connector that will serve as the main entry.  our concern how to maintain this connector with at one corner of the new building and maintaine an 24" setback on the other corner.

and Yes it will be masonry construction.

Attached is a drawing of the buildings profiles and couple pretty pictures.  The existing building profile and set backs are from the surveyor cad file.  Be forewarned I rotated the file to match conventional drawing standards that architects use   

Image Backside
Image Entry

[Josh Nieman]

I have nothing useful to add to this discussion (though thanks to all who participated as well... because I, too, learned a lot)

but I just wanted to say... that's some kick#%@ images.  Nice presentation!

[Krushert]

but I just wanted to say... that's some kick#%@ images.  Nice presentation!

We can't take credit for the images, we farm that out.  This is guy is unbelievable at what he can do.  He has good rates, good turn around time.  This is hobby/second job for him.  He is a plant engineer by day/night (I can't remember which) for a paper mill.   We could do it in house but we would not get the same level of quality in the same time frame that he does it in.

We are getting into Google sketch up for some of the front end discussions with the clients.  In the future who knows.

[drizzt]

The images look like accurender to me. I like that software, although I don't know what the current version is!

[sinc]

Then between the buildings there is going to be a 1 1/2 story connector that will serve as the main entry.  our concern how to maintain this connector with at one corner of the new building and maintaine an 24" setback on the other corner.

That's a bit hard to parse...  But are you basically concerned by that south corner hitting a setback?  There's some 2.5' of room there, so the builders would have to completely ignore the surveyor's stakes to get it that far off.  And they should also have no problem building the connector.  I don't think you have any issues, from what I see.

BTW, I noticed that your setbacks are not parallel to your lot lines on the "west" and "south" boundaries.  That's probably not right.  The setback distance on the south varies by more than a foot.  But again, you have a lot of room, so it's probably not a problem, just one of those things that civies like to complain about when they get DWGs from archies....   

[Krushert]

That's a bit hard to parse...  But are you basically concerned by that south corner hitting a setback?  There's some 2.5' of room there, so the builders would have to completely ignore the surveyor's stakes to get it that far off.  And they should also have no problem building the connector.  I don't think you have any issues, from what I see.

Thanks.

BTW, I noticed that your setbacks are not parallel to your lot lines on the "west" and "south" boundaries.  That's probably not right.  The setback distance on the south varies by more than a foot.  But again, you have a lot of room, so it's probably not a problem, just one of those things that civies like to complain about when they get DWGs from archies....   

Not Me.  I called up the Civil Guy and asked him for his cad file.  I strip all the extra junk you guys throw in the file (so you guys can justify your cost  ), change the layer color, and pre-fix the layer name.  Then scale factor also. after that, I rotated the whole file about a corner of the existing building so that the existing building wall is parallel with the "X" axis.  That is all I did.  That is my story and I am sticking to it.   

Thanks guys for your input.

[Maverick®]

that's some kick#%@ images.  Nice presentation!

True Dat.  Awesome renderings!

[CADaver]

But to try to get back to the original point, I think that no matter what, the important question is not "how closely can it be staked?" but "how closely can it be built?"

Concrete placing tolerances are governed by ACI-117-06
And teel by the A.I.S.C.

[Josh Nieman]

But to try to get back to the original point, I think that no matter what, the important question is not "how closely can it be staked?" but "how closely can it be built?"

Concrete placing tolerances are governed by ACI-117-06
And teel by the A.I.S.C.

Funny, I thought limits on Teal were set by the Department of Wildlife and Fisheries.

[CADaver]

But to try to get back to the original point, I think that no matter what, the important question is not "how closely can it be staked?" but "how closely can it be built?"

Concrete placing tolerances are governed by ACI-117-06
And teel by the A.I.S.C.

Funny, I thought limits on Teal were set by the Department of Wildlife and Fisheries.

You guys are gonna hurt my feelings now...

...

... never mind

[Josh Nieman]

Couldn't resist.