Issue link: https://iconnect007.uberflip.com/i/857644
24 The PCB Magazine • August 2017 pressure. We have a local under-pressure, and we have over-pressure in other areas to com- pensate. Everything balances. Holden: That's the one process I don't find in North America, this ammonium sulfate one. But also this one is tied up in a very rigorous closed-loop system, and that is highly unusual. Goldman: How long did it take you to get that to balance? Stepinski: Six months. It was one of the longest items, because the process was brought to us by a supplier, and we relied on the supplier a lot initially to get it set up, and then when it didn't meet my expectations we had to go and real- ly dig into it ourselves. We changed everything and ended up where we are now. Half the equip- ment is not even there anymore. It was highly overengineered and overcomplicated, and sim- ple is the way to go. Holden: That's one area I'd like to look at, the chemical dosing or chemical control stations. Stepinski: Yeah, you can't even really see it. It's so simple. You can see it, but there's nothing much to see. We have a specific gravity control- ler on the galvanic cell, and one on what we call the complexer, with the cascade inside, which oxidizes. This is all specific gravity based, but it's just a huge, long tube. Holden: Most of them are controlled by specific gravity, which you could really throw away be- cause the specific gravity controllers are abso- lutely terrible. Stepinski: Terrible. I agree. We made all our own specific gravity controllers. We just use a den- sitometer. We put a pressure transducer on the bottom of a tall pipe. The taller the pipe, the more sensitive it is. Holden: That's the differential pressure (DP) method. Most of the guys that sell these things with the chemical use a hydrometer or some- thing like that, and if you break a beam or close a microswitch… Stepinski: Any turbulence causes a lot of prob- lems. Holden: Yeah, they bounce around and things like that. It's better if you have a totally sealed one where that center mass never changes but you can change the calibration. It's an inter- esting set of equations and debate between re- membering what Archimedes' principle really says. Most of them work off the changing vol- ume, and that is the displaced mass can change, because it floats up and down, but if it's totally immersed then that mass isn't going to change. Or it works off a little titanium chain, and as it goes up and down, it changes the weight rather than changing the volume. There's a radiation- based densitometer and there's differential pres- sure-based specific gravity controllers. Are there any other etchers or pre-etches in the line? Stepinski: The only pre-etches are associated with final finish, so HASL pre-clean or ENIG pre- clean, and then we have an alternative oxide process we use prior to solder mask, which also does double duty as our multilayer treatment. Goldman: I was going to say, your innerlayers must get some oxide treatment. Stepinski: We have one process for both. Goldman: In the very beginning, you've got cop- per coming in. You have to clean that. Is that a microetch up there on your laminate copper? Stepinski: No, we deburr it—mechanical deburr right now for double-sided products. For cores, we don't clean it; we go straight into inkjet. Holden: The inkjet has much better adhesion. Although, changing current density or puls- ing in the panel plate for the inkjet, how is that compared with the innerlayer inkjet? Stepinski: The unannealed grain structure and annealed grain structure are very different things. We're going with an unannealed copper on all the electroplated stuff, because we're go- ing too fast, we're not giving it time to anneal. It's a very different grain structure than what we WHELEN ENGINEERING, TWO YEARS LATER