Issue link: https://iconnect007.uberflip.com/i/1083315
90 I-CONNECT007 I REAL TIME WITH... IPC APEX EXPO 2019 SHOW & TELL MAGAZINE to talk to the expert in fluid dynamics and do some experiments, modeling, or simulations to see whether our speculation was valid. We went to Jason because he has a Ph.D. degree in fluid dynamics, and he's very familiar with simulating flows and doing experiments to understand how air flow affects dust deposi- tion. We started to work with him a few years ago, and he did most of the experiments in Ire- land when he was with Nokia. Then, he left to join Imperial College London. Holden: Was there a practical result of all of your research and work in terms of how Alcatel- Lucent and/or Nokia designed or redesigned some of their equipment? Xu: Yes. If you look at the indus- try, there are multiple ways developed to deal with the dust issue because it's a problem for everyone in the electronics industry. However, it has become more of an issue recently due to more deployments in chal- lenging environments and the increasing circuit density and decreasing pitch size, making the electronics more susceptible to dust. When you have a little bit of dust deposition, due to fine pitch, the effect of dust becomes even more pronounced. Also, as an end customer, there's a trend to use less-controlled environments. It takes costs and energy consumption to have good envi- ronmental control. You need to do the heating, you need good air circulation, and you need to filter the air because dust affects everything such as cooling efficiency, for example. To save ongoing operating costs, the end user would like to reduce the control of the environmental conditions. So, the equipment is increasingly exposed or deployed in the environment with high dust concentration. There are multiple ways to mitigate the dust effect on the electronic equipment, but no one has looked at air flow and the impact of the air flow pattern on the dust deposition rate. What we found is that if you can optimize the air flow pattern, you can decrease the dust deposi- tion rate by a factor up to 100 for PM 2.5 par- ticles. That's quite a large impact. Holden: I read through your paper and recog- nized that everybody knows that because of cooling, you have to have good air flow and cir- culation, so you put in a fan and it's finished. It's interesting that you explain how air is not just air. Air has things in it; not just humid- ity, but particles, which we call dust. And it comes into your device, it has an impact. You illustrate a lot of different parts of the electron- ics and where dust settles or attaches, so it's not just air. Of course, dust can mainly be an insulator, but it could be an electrolyte, which is probably even worse. Xu: Exactly. Holden: It's not a trivial prob- lem because all of that will interfere with its cooling effect, and that's what you're trying to do. Out of all of this, what were some of your conclu- sions? Xu: Basically, what we're try- ing to do is optimize the air flow to minimize dust deposition. At the same time, when we optimize the air flow, we don't want to reduce cooling efficiency. That's what we have been able to achieve in our work by optimizing the air flow to reduce the dust deposition rate by a factor of 100. The impact on cooling efficiency is as little as 1–5%; our simulations have dem- onstrated that, which is the major result of our experiment. As I mentioned before, some of the key con- clusions include that we observed localized dust deposition at a certain location. Through our work, we have demonstrated this highly increased preferential deposition is related to the air flow pattern. Specifically, if you have air flow parallel to the circuit board, dust depo- sition is very low, but if it becomes perpendic- ular to the circuit board at an inclined angle,