Lesson finished
Unlock all 65 lessons of Geometric Dimensioning and TolerancingFull course ₹40,000 · paid
Tolerance Stack-Up Analysis with RSS, MRSS and Worst Case
A free lesson from Skill-Lync's Geometric Dimensioning and Tolerancing course, on Grow by Skill-Lync · Last updated 16 Nov 2026 · Session: “Geometric Dimensioning and Tolerancing Video”
What will you learn in this free Geometric Dimensioning and Tolerancing lesson?
This live session works through a tolerance stack-up for a gap between two surfaces, building the tolerance chain part by part. It computes RSS, MRSS using a 1.2 factor, and worst case, and explains why datums, form and orientation tolerances do not contribute while position and profile do.
- Building a tolerance chain along the stack-up direction
- RSS, MRSS and worst case calculations
- Why datum surfaces contribute 0.0 to a stack-up
- Which tolerance types contribute and which do not
What comes after lesson 20?
Later in Skill-Lync's Geometric Dimensioning and Tolerancing course (paid): lesson 21, Stack Up Analysis Overview and Basics; lesson 22, Coaxial Tolerances; lesson 23, Understanding Concentric Tolerance. The full course has 65 lessons (13 h 6 min of video); see the course overview or its price and syllabus.
Lesson transcript
Machine transcript of the lesson audio (lightly cleaned). Technical terms may contain errors; the video is the reference.
Brown part, this hole, and this hole. Together, they can move either up or they can move down. Okay? Up, down. So, we have to take a position tolerance of that hole also. So, let's go to our yellow-colored or brown-colored part. It's 1. All right? So, position tolerance on hole. You got why we took 0.0, correct? This one? Yes, sir. Right? Because it's such a small part, we assume that, okay, it doesn't have any position tolerance. The outer surface or this surface over here doesn't have any position tolerance. So, we're going to assume that, hey, it's not going to contribute. Or even if there's a contribution which is so small, I don't really want to, you know, waste paper on that. So, position tolerance on hole on the brown part. All right? And then, basically, the brown part goes in up-down direction. So, like someone said, up-down is going to be the surface.
Okay? Surface-to-surface contact is going to stop the up-down. So, we're going to go to the bottom surface. So, the bottom surface is an 8-atum. Okay? And it has a flatness. Flatness is a form tolerance. Form tolerance never goes into stack-ups. All right? So, and this is an 8-atum. So, it doesn't have any contribution in us stack-up. All right? So, I'm going to just put 0.0 because the bottom surface is an 8-atum. Then, from that, we jump to this one over here. This surface, which is actually going to make the contact. So, surface profile. And there's no hard and fast rule how to write this description.
This is just description. Okay? You can write any way you want. Like, you know, make it as simple as possible to understand. There's no hard and fast. This is just description of the tolerances. So, just make it surface profile on Salman part. It's just like after two years when you're going to open the stack-up, you should be able to, you know, make connection which surface is which one, which profile is which one. So, surface profile on the Salman color part. Let's go to the Salman color part or the pink color part.
Have you called out that surface as, again, one. All right? Pretty easy. So, plus minus 0.5. Okay? This surface is done over here. And then, we jump back to our Q surface. We started with P surface. From the P surface to this hole over here. From this hole to the green part. From the green part to this hole over here. From this hole to this surface over here. From this surface to this surface over here. And then, from this surface to this surface over here. Again, Q is a datum. If you go to slide number 3, like, I'm not going to go back, but, like, you know, you remember this one.
We called it as a datum. So, surface profile. File on Q. It's a datum. So, 0.0. So, these are all the tolerances that go into your stackup. All right? Started with Q. Created a tolerance chain. Came back to P. Okay? Only consider tolerances that are going to affect your tolerance stackup direction. Tolerance stackup direction is up-down. Okay? That's why you take a position tolerance of this hole. Because that hole is going to move up-down. All right? You don't take position tolerance of this hole. Because it's, like, this is a datum. But anyway, if it has a position, it's not up-down.
It's only in the flat surface. Either it's going to go in this direction or it's going to go in this direction. Okay? Any, just take tolerances. That's going to affect your direction of the tolerance stackup. So, now what I want you to do is take out your calculators and do an RSS of these numbers. So, the way you do an RSS is that square each number, then you sum it, and then you take a root of the sum. That is the RSS. Go ahead and you can just either tell me or you can just put it in your chat box.
All right? But do an RSS of all these numbers. I guess we have a good chunk of replies. Everyone has replied, which is nice. So, yeah, everyone got the right answer, obviously. Pretty simple. 0.5 square plus 0.5 square plus 0.5 square plus 0.5 square plus 0.5 square plus 0.5 square plus 0.5 square. All right? When you do, it's basically 0.25 plus 0.25 plus 0.25 plus 0.25, which is basically 1. All right? And then when you take a root of 1, it's obviously 1. So, the RSS, let me start a new, right?
So, the RSS is equal to plus minus 1. The MRSS is equal to plus minus 1.2 because MRSS is basically RSS into 1.2, which is the safety factor or factor of safety. And then finally, the worst case is basically you add up all those numbers. All right? So, 0.5 plus 0.5 plus 0.5 plus 0.5 gives you plus minus 2. So, this is the worst case. So, now let's see who was doing the stack up. I guess Vignesh was doing the stack up, right? So, Vignesh's boss has asked him to do a stack up tolerance for that gap.
Cat gap is 50 millimeter. And he asked him, okay, how much that gap will either increase or decrease. So, Vignesh will go back to his boss and he'll say that the stack up tolerance is plus minus 1. Okay? The mean gap is 49. And the max gap is 51. Based on RSS, plus minus 1, 49, 51. Okay? Mean. And then nominal. Nominal is basically a 50. Then MRSS plus minus 1.2. This would be, how much is 48.8? 51.2. And then finally, 48, 52. All right? So, Vignesh is going to go to his boss as he's going to present this table as a tolerance stack up.
And then basically, you will either take RSS, MRSS, or worst case, depending on how much safety you want. Okay? So, this is the tolerance stack up that you guys are going to come up with this specific example. So, doing an actual RSS is pretty easy. It's simple arithmetic. You just square it, you know, add it, and take a look of it. Tolerance chain is the most important thing in any stack up. Okay? The tolerance chain should be perfect. Okay? You should not miss out any parts.
You should make sure that all components have been taken into account. Okay? And even after you take all the components into account, so the first you're going to do is you're going to write all the components. Okay? Component 1. Component 2. 1, 2, 3. Then, you know, you're going to put all the tolerances that go into this tolerance chain. Because, component 1 will have some tolerances that's not going to affect your direction. Your direction is up-down. Okay? If you have a hole that's going to move in cross-car.
Okay? Or it's going to move in x direction, it's not going to affect your tolerance stack up. So, don't take that. Okay? Take the tolerances that are acting only in up-down direction. So, for example, profile of this surface is up-down. Okay? Because this is your tolerance envelope. Take this one. If you take a hole over here, which is like this, and it's going to move up-down, take that one. Okay? So, first, you create a tolerance chain. You write all the components. That's going to go. Then, you write all the tolerances.
And then, you're going to do the arithmetic. All right? You're going to square it, sum it, take root of it. There are ready-made sheets available, Excel sheets available online. If you want, I can send it to you. What we just did, basically, it's just a formula behind an Excel sheet. Nothing fancy. Okay? What we did, they just put a formula, and then they publish an Excel sheet. So, if you want, I can send that Excel sheet to you. You can find it online as well, where you just put in the numbers 0.5, 0.5, 0.5, 0.5, and it will give you the answers, rather than using a calculator.
Just fancy way of doing what we did. So, you can use those Excel sheets also. You can use your calculators as well to do an RSS. Major thing is keep an open mind. Whenever you do a tolerance stack-up, make sure you take all the components. Make sure you take all the tolerances. Okay? And then do the math. Any questions on this example? Sir, I have a doubt. As we are measuring the distance between P and Q, we have considered the surface profile of T, which is 0.5.
and then do the final look at a look at. So, I want to see the surface of the surface profile, the other 1-some position, and then do the derivative, Then do the next step. But did we consider the surface profile of the paying part? We did not create surface profile on Q as 0.0 because that Q surface is basically a datum, right? So when we go over here, that surface is called as a datum. A datum just has a form tolerance. So datums never play a role into stack-up tolerances, okay?
Okay. But if the form tolerance, like the flatness, is too high, in this case, like it's just 0.25, but if it's too much, so will it affect the distance between two surfaces, right? Right. That's a very good question. Answer is no, it will not. Reason is that even if it's like a very high tolerance, okay, even if it's something like this, okay? Now, this is, let's say, a flatness of 2 mm, which is very, very high. Once you give the flatness, then after that, it becomes a datum, okay?
Then you're going to put this, and then you're going to call this as a datum. Now, this is just a microscopic view of the hills and valleys. Then this becomes a datum. And then everything is related to this one, all the others. So this becomes your reference, okay? Okay. So the hills and valleys is just local. After the flatness is done, then you take the top three points, 1, 2, and 3 over here. And then you kind of make a plane over here. That becomes your a datum.
And after that, this a datum is used reference for everything else. So it is the origin, all right? Everything is related to a datum. So it doesn't play a part into the tolerance stackup because that is, like, the origin. In simple words, I'm thinking how I can explain this one. So I'm trying to put an analogy. This is your coordinate system over here. This is your origin, okay? Which is 0, 0, 0. This is your point P over here. And then this is your point Q over here.
And you're trying to find out the variation or the stackup tolerance between point P and point Q. Point Q is x1, y1, z1. And Q is x2, y2, z2, okay? And you're trying to find out this gap is 50 plus minus x, all right? So now, even if you move the origin to over here, all right, will this P to Q going to change? This delta between P to Q, it's not going to change, right? Even if you move the origin anywhere, this P to Q, this point to this point, so Q minus P, is going to remain the same.
Even if you move the origin anywhere, correct? Because if you move the origin, it's going to affect both P and Q equally. So Q minus P is going to be the same. So the same analogy you can use over here. So A is your origin. Like when you give flatness, what you're giving is you're moving the origin, all right? You're going to move the A datum, all right? But after moving, then that becomes your reference. And then you're going to, from A datum, so I'm going to call this A datum, you're going to measure P and you're going to measure Q.
So even if you move A anywhere, the P to Q distance is going to remain the same. So it's kind of difficult to understand, I know. But just consider that it becomes a datum after the hills and valleys go away. After you have put the flatness, then it becomes a datum. And then everything is referenced back to the surface. So yeah, this is how the analogy will go. So even if you move the reference, the P and Q is going to remain the same. So here, basically, when you change the flatness, you're moving the reference.
All right. Any more questions, guys? Hello, sir? Yeah, yeah. Go ahead. Go ahead. Yeah. So one doubt, sir. So like for the feature of size over here in this example, you didn't give like MMC conditions or LMC conditions. Will that affect the whole stack up? So you're saying that if I give over here as like BM and CM like this, right? Not for datum, sir. For the feature of sizes, like for the holes. I see. Okay. Yeah. No, it doesn't. All right. It's not going to affect the tolerance stack up.
All right. Okay. It may affect, but like in industry, what we do is even if you have, we tend to ignore that. All right. So if we start putting it together, it's going to be a huge stack up because for example, if you make this 10.1 plus minus 0.1 at, and if you give 1M, okay, then what if the hole becomes 10.2? Yeah. Then you have to change the tolerance to 1.1. Okay. Then what if the hole becomes 10.3? So it becomes 1.3. Okay. So you will have N number of iterations for the same tolerance stack up.
Yeah. So we tend to ignore it. All right. Okay. The answer to your question is it, it slightly affects very slightly. We tend to ignore it because for the same tolerance stack up, we have to do so many again and again. I ask because like in industries, mostly they use with MMC conditions and all right, because bonus balances are there. So if there's a mixture of LMC's and MMC's in this thing, then do we have to calculate for each or we'll just ignore that and we'll calculate directly with the value itself?
Okay. Ignore it. Okay. My experience for the last seven, eight years as a dimensional engineer, ignore it. Okay. To be safe, multiply it by 1.2. Okay. MRSS. To take everything. So this is, I don't know who came up with this number, but it is like an industry accepted number, like Ford, GM, Chrysler, everyone uses the number. Okay. So ignore MMC condition, ignore all conditions. Basically do the stack up multiplied by 1.2. Okay. And so if there's boundary conditions with this maximum and this thing, if the boundaries are affected, I mean, the datums are affected.
No, you don't have to consider that as well. Okay. Okay. I just consider this as a RFS condition to the stack up multiplied by 1.2 and, you know, that should be good enough. Okay. So Gokul has a question. Can you give a quick revision on the tall lenses you have given? Sure. We can do that. So quick revision is, so. Gokul, just reply to this message. Do you want the tolerances that go into the stack up or do you want to go through individual parts? So I assume you, you are asking for the tolerance on the stack.
Okay. That's. So we go to back to slide number eight. The tolerance chain is done. We start with the profile tolerance on the surface P. Okay. Profile tolerance on the surface P. If you go to slide number five, profile tolerance is given. One. Okay. So this is over here on the bottom surface. So that's why we have considered plus minus 0.5. From that, we go to this hole over here. That hole. Okay. On the yellow part, which is basically this hole over here, this hole over here, it can move up down.
Correct. It has a position tolerance of one. So plus minus 0.5. So we have taken 0.5 over there. Then we go to the green part. Green part. We assume that outside surface is perfectly machined. All right. So 0.0, we just did an assumption. Okay. From that, we go to the brown part. So that small hole over here, which would be in slide number three, this hole over here. Again, position tolerance of plus minus one because it's going to move up down. So that is plus minus one.
This hole can move up down. That is 0.5. This is plus minus 0.5. If the surface, we assume perfectly machined, so 0.0, this hole is a datum, so 0.0. All right. So and then this surface is 0.5 plus minus. So 0.5, this hole plus minus 0.5, 0, 0. This is plus minus 0.5. From that, the bottom surface is a datum, so 0. From that, this surface over here, plus minus 0.5. All this has been captured in the previous slide, right? And then from that, back to our surface, P, which is our data, so 0.
So 0.5, 0.5, 0, 0, 0.5, 0, 0.5, 0. Okay. So 0.5, 0.5, 0.5, 0.5, basically. Four times 0.5. And that is how we came up with the tolerance checkup. Okay. Any more questions, guys? No, sir. All right. So why did we take the profile of the surface on the top of it? So because that is the surface that basically we are considering, right? Let me put this into a new slide. This top surface of this one and the bottom surface of this one. Okay. These two are we considering, right?
Bottom surface. Okay. So we have to take the surface profile on the bottom surface, which is plus minus 0.5. We have to take the surface profile on top surface, which is 0, because it's a data. Okay. So that is why we considered the top surface on this one and the bottom surface on this one. Okay. And then this one also we considered, but that is a data. So 0. And then this one we considered because that's going to move up, down 0.5. Okay. All right. Thanks, Gokul.
I'm glad you got it. So any more questions or we can close this session? I hope. Yep, go ahead. Okay, what if the hole that contributes to the tolerance has a composite tolerance box? For example, position along with perpendicularity. Okay, so perpendicularity never contributes. Okay, perpendicularity is an orientation tolerance, then form tolerance, they never contribute to tolerance, only the location. Okay, thank you. I guess we are done. Unless you have any last-minute questions, we can stop the station over here. What I want you guys to do is, as always, look into some smaller sub-assemblies and try to figure out the DATEM scheme and find any gap, try to do a stack-up tolerance on that gap.
Because complex sub-assemblies are very easy to find, but smaller sub-assemblies are kind of difficult to find. Also, I'll try to share the Excel sheet, a ready-made Excel sheet we have. You just put in the numbers and then you get the RSS stack-up, but it's just a fancy way of doing calculation. It's just formulas behind the Excel sheet. So, I guess we can close the session over here. Okay, thanks guys for attending. Thank you.