Just the inverts and the rims. Why would you want the center? The manhole should slope from invert to invert maintaining the same slope as the pipeline.
The center of the manhole is going to REFLECT the inverts. this is assuming no significant drop across the manhole. if you dip multiple inverts, and have the standard eccentric cone, there will be some variance in the dip. now, you will assume the surveyor accounts for this. but now, if youre doing prelim, then why do you need all the inverts (assuming there isnt a significant drop in or across the manhole)? for design all you need is the average. your likely going to replace that manhole and only need a reference to base youre new design. obviously, if there is a line dropping 12" into the manhole, or an outside drop (or inside I suppose), you will need those shots too. The reason this is a concern is because, in an urban environment, existing lines may be at grades that do not meed todays standards. Now, if youre surveror does not accurately estimate the variance on his shot due to the angle of the rod, you may lose a couple tenths of grade, which, in a relay is a lot.
just inverts and rims. I prefer all the inverts and realistically, the error induced by holding the rod at a slight angle is generally less than a 1/10 of a foot.
In some areas that are pretty flat, say Florida, a 1/10 of a foot error in a short sewer run can be the difference between a Utility accepting the sewer system or not...I've seen it happen. The rodman is SUPPOSED to measure the rod angle for every measurement. Also, there is a fitting you can put on the bottom end of the rod where you don't have to tilt it to get all the inverts.
Anyway, back to the original question, in my opinion, since they've popped the manhole lid off, why not get all the inverts while they're there? Why take chances to assume the drop across the manhole? The drop could be anywhere from 0.1' to 2.0'
Placing the rod at an angle may only produce an error less than 0.1', but this is only the case if the manhole is very shallow.
We work with deep manholes all the time, (>8' depth) and when manholes get this deep, this error is magnified, and it does become significant (10' deep manholes will be approx. 0.2' off).
It's hard to get real accurate measurements without actually crawling down to the inverts, but guys that crawl into sanitary manholes this deep have special codes and requirements they need to meet, and most surveyors are not certified to do this. So, we usually just shoot the middle of the invert and assume a 0.1' rise/drop to both inverts (obviously, only after a visual inspection). While this also has some error associated with this methodology (usually in the hundredths), I've seen much bigger errors when surveyors take vertical measurements when their level rod is not plumb.
I will say, the most accurate way to get these inverts is to use the attachment that CEE98 mentioned. Great idea! That's why I come here, to learn something new, and hopefully offer some positive feedback. Thanks.
The error in deep manholes is usually less then 0.02' or a quarter of inch. A rodman and instrument person giveing TWO TENTHS of a foot error should not be retained!
