At the conclusion of our webinar, How PCB Design Choices Affect Overall Cost from Your PCB Fabricator, we had several questions submitted to our presenter, Ed McMahon, CEO at Epec. We have compiled these questions into a readable format on our blog.
Q&A From Our Live PCB Webinar
Quick Links:
- What are the biggest cost-adders when you go from a Class 2 to a Class 3 or mil-aero type specifications?
- How does type lines and spaces affect the cost of manufacturing?
- How does material yield on an 18 by 24 panel affect the cost?
- Do the same cost adders apply to rigid-flex?
- Is back drilling OK to do vs. blind vias?
- Regarding the tolerances for embedded resistors, if I want an accurate ratio and don't care so much about absolute resistances, does the larger resistor size still remain important?
- What is the smallest via can you drill?
- What is the added cost of planarization when using via in pad with non-conductive fill and plating over?
- What is a microvia?
- Microvia is a traditional via just smaller?
- I have a product that needs to operate in a 200°C environment. What PCB material would you recommend?
Watch the Recording Below:
Question: What are the biggest cost-adders when you go from a Class 2 to a Class 3 or mil-aero type specifications?
Answer: That's a really good question. As a fabricator, there's really two big things that are required of us that are different. The first one is, there's an annular ring specification that's required. For Class 2, both internal and external, it's called the 90-degree breakout. So, if you have a drill bit or a hole where the drill broke out of the actual pad, that's still acceptable from a Class 2 perspective. However, in Class 3, what that requires is on an outer layer you have to have a minimum of 2-mil annular ring, and for an inner layer a minimum of 1-mil. So, that creates some challenges on several different fronts from a cost perspective. One, is you’ve got to increase the pad sizes by almost 40% in some instances, so you're really spending a lot of time going through on the CAM side, making sure that you've got all those pads blown up there, and that you can still make all your required clearances through the rest of the features.
The second thing is that typically when we do drilling in print circuit board manufacturing, we like to stack product up in the drill. So, you'll stack four boards up and you'll drill all four at once, without changing it out. Typically there are a couple things. One is what's called drill wander. On Class 3 boards, you have to be careful, and many times you don't stack it up because you don't want the drill wander to create breakout. And you’ve got to slow it down because there's this feature called rapid run-out if the drill bit isn't sharp enough and the bit hits the actual copper itself and then begins to shoot to one of the directions outside. So, it really costs us a little bit more to produce.
The second thing on Class 3 is plating thickness. For Class 2, you’ve got 0.8 mils in the hole, minimum. For a Class 3, it's 1 mil in the hole minimum. And as you know, if you put a mil in the hole, you're going to have, or if you put up an ounce of copper in the hole, you're going to have about 1.3 to 1.4 mils.
Now, the real issue with this, is that the way you measure this on a Class 3 or on a mil board, is on the coupon, right? You're not going through and measuring every board to see if you have that amount of copper. And the problem is with the coupons around the edges, so you want to make sure that those coupons come out good because if the coupons come out good, the theory is the rest of the boards will come out of fine.
The challenge becomes is that if the coupon's bad, basically the panel's bad. So, that's why it takes longer because we have to make sure that the coupons are good as well. So, that's what adds some cost to Class 3 vs. Class 2.
Question: How does type lines and spaces affect the cost of manufacturing?
Answer: In all honesty, it really doesn't. Whether you've got 3-mil lines and spaces or 6-mil lines and spaces, it doesn't necessarily matter to us. What it does is it matters what kind of equipment you can use to do it, right? The traditional photo-imaging equipment is really difficult when you get anywhere below 6 mils. So, you really need to laser-direct imaging. Really, when you're going down to that level, you've got to have the proper equipment. So, you're more just limiting the number of people that can do it. The other thing to think about when you're talking about that type lines and spaces, the biggest challenge that we normally have is 4 mil and under solder mask dams. That's the biggest one that we find that we're having trouble with. When you have a tape test and you didn't process that with an LDI, many companies don't use LDI for solder mask. Well, if you don't use LDI, and you don't go through the right bake cycles, it's going to be really easy to pool those 4-mil dams up on a tape test. So, that's the bigger thing with tight lines and spaces vs. any real cost adders.
Question: How does material yield on an 18 by 24 panel affect the cost?
Answer: It's interesting. For the last, I don't know, 20 years, 18 by 24 was the panel size that everybody used, and that’s what we used as an industry, and that did well when you were talking about how it came out of the panels that we got it from. But, when you looked at post-edge punch and X-ray registration, that was really good to help us stay with 18 by 24 panels as the tolerances just got tighter. But, in all honesty, when you get down to where we are today with some of these features, an 18 by 24 panel's really just too big, in many instances. So, we go ahead and shrink it down, and in many instances even on some of the Rogers material, even a 12 by 18 is too much to be able to manage. A lot of fabricators basically get the material in a 48 by 48 inch sheet. And domestically, when we're doing it, because we're usually dealing with smaller panel sizes, it's even better if we have to manage a larger amount of smaller panels because you're still not talking about that many.
And the thing in Asia is that because there's so much automation, the fact of the matter is that the handling isn't a big issue to them anymore. So, even though they're using smaller panels and they're getting a better yield out of the 48 by 48 inch sheet, that's more important these days than what you can get on an 18 by 24 inch panel. It used to be 18 by 24 was the Holy Grail, but that really isn't so much anymore, especially when the technology gets up.
Question: Do the same cost adders apply to rigid-flex?
Answer: Yes, they do. And for all the features that we talked about here, all those features for the most part are available on rigid-flex parts. But there's some additional ones when we do rigid-flex. The number one rigid-flex challenge you have is the construction. And the stack-up is probably the biggest challenge that we have with customers. The one big misconception about rigid-flex is that, "Oh, you just make a flex board and then you stack it up in the rest of the rigid board, and that's it." Well, that's technically not how it's done. The way that it works is that the covering portion of that flex circuit doesn't go all the way through the rigid parts of the board. So, that's where many times, if a customer is looking for something quickly and sends us in the design that has the covering all the way through, that can cause us to slow down a little bit as we go through and do a redesign, because now your stack-up, your overall thicknesses can change, and will need to be addressed . That's one of the biggest things we see. And then the last thing that we see on that, is really understanding with the customer what the bend radius is. Because, the thinner the material, the more expensive it is to process, right. It's just the way that it is. So, if you're not bending this that much, well, you can actually get away with some thicker material and some thicker copper, which is actually easier to process. If you're trying to take a 2-mil or a 1-mil piece of polyimide and run it without a leader through old rigid processing equipment, you're going to wrap that around a roller. That's why it's important that if we understand what the application is we can help you create the most cost-effective solution from a processing perspective.
Question: Is back drilling OK to do vs. blind vias?
Answer: Back drilling has become a common alternative to blind vias and is more cost-effective (about 10% additional price). The blind via process, depending on the number of layers and laminations, can be 50% or higher for additional cost. Back drilling can save cost but is still only used seldomly.
Question: Regarding the tolerances for embedded resistors, if I want an accurate ratio and don't care so much about absolute resistances, does the larger resistor size still remain important?
Answer: It is better to go larger, even if only the ratio matters. This is due to etch loss. Proportionately speaking, width decreases while length increases, and because of this, it becomes more disproportionate as the resistor gets smaller (i.e. a 0.010” x 0.010” becomes 0.011” x 0.009” and a 0.100” x 0.100” becomes 0.101” x 0.099”). In other words, a 1-mil variance affects a smaller resistor more than it affects a larger resistor.
Question: What is the smallest via can you drill?
Answer: The smallest mechanical we produce is 0.006, and this depends on other attributes such as PCB overall thickness, copper thickness, aspect ratio 1:1. When processing through vias, typically we size up the drill 0.003-0.005 to process through plating. Allowing room to process a larger via such as 0.008 or 0.010 is advantageous to production, the larger drill sizes the more panels can be processed at one time.
Question: What is the added cost of planarization when using via in pad with non-conductive fill and plating over?
Answer: The cost adder is approximately 30%, however, it is driven by volume of panels and quantity of vias not as an adder per hole but time to drill as well.
Question: What is a microvia?
Answer: Microvias are defined by quantity and <0.006 diameter, blind and buried, as well as stacked vias. Usually in HDI – high-density PCBs interconnecting layers.
Question: Microvia is a traditional via just smaller?
Answer: Traditionally, a via was a 0.010 finish, as technologies have pushed the limits of available land on PCBs, the via size was one item to quickly change. Vias were defined as 0.010+.00/-0.010 to keep the hole size smaller. Plating shut was allowed or partial shut. Microvias were introduced as blind/buried vias were the new process, adding cost but allowing to keep the PCB smaller in size. HDI pushed the via sizes even smaller and introduced < 0.008 or 0.006 to processing as the PCB increase in size, layer count, and technology.
Question: I have a product that needs to operate in a 200°C environment. What PCB material would you recommend?
Answer: When looking for materials to use on your PCB, review the TD and the TG of the type. Avoid calling out a particular brand of material such as Isola or Shegyi. This will allow your PCB supplier to use quality products to meet the criteria without using a particular brand. Many good brands are available. Shengyi, Isola, ITEQ, TU are all quality products.
