Are All Medical Power Supplies the Same?

April 1, 2007
A customer recently came to Fairchild Semiconductor and asked us to help them with a make versus buy decision on a power supply for a medical application.

A customer recently came to Fairchild Semiconductor and asked us to help them with “a make versus buy” decision on a power supply for a medical application. The power supply the customer had been using in the application was one with universal input and 12-Vdc output. The customer was considering a change because the company was having EMI/RFI problems, as well as reliability and safety issues.

We asked if the company wanted to take on the responsibility of certification of a “medical” power supply internally, as well as good manufacturing practice (GMP), special EMI/RFI certifications and so forth. The company was unsure because it did not fully understand these issues.

So before entertaining further discussion on whether to build a supply, we asked whether we might take a closer look at the company's existing power supply. The customer agreed, and we disassembled the store-bought medical power supply. Sure enough, it had the requisite paperwork, although it looked like a copy of a copy of a copy.

Now, in reality, this was a “no-name” power supply. When it was taken apart carefully, we found numerous cold solder joints. For example, the heatsink was not grounded due to bad connection — one of the many reasons for the high EMI/RFI.

In this supply, the spacing for safety regulations would not have even met commercial specifications, not to mention medical spacing requirements. The filter appeared to be the normal X-Y filter set up without any special regard to leakage. In fact, when the leakage current was measured, it was on the high side, even for a commercial power supply, and nowhere close to medical specs.

The semiconductors were unrecognizable brands with odd markings and obscure or missing date codes. The mounting of the power semiconductors to the heatsink was done in such a manner as to supply high isolation. However, this high isolation was at the expense of severe thermal isolation. The semiconductor devices themselves were running quite hot and the heatsinks were quite cool. The insulation was electrically sound but thermally ineffective, which further compromised long-term reliability.

The plug provided on the input of the power supply was ungrounded, and the ground actually used the neutral of the supply line. A traditional three-wire grounding system was not even used. The optoisolation barrier did have a slot cut in the pc board; however, the optoisolator did not seem to be type rated for an application in such a harsh environment, nor did the spacing. Normally in a medical-grade supply, I would expect to see VDE lead spacing or the like.

So, after reporting these details to the customer, I asked why the company had bought this particular supply. The answer was cost — this supply was only $15.

Cost? The application for this power supply is a $30,000 medical instrument that cannot fail. So why not use a “good” power supply? The customer responded that if the manufacturer says it's medical, isn't it all the same?

Of course, the answer to that question is a resounding NO! This point leads us back to the customer's initial inquiry regarding make versus buy. As we explained to the customer, you do not want to build your own power supply, because the volumes are insufficient to recapture development costs and you do not have the necessary resources or expertise.

Instead, we told the customer to buy a power supply from a proven supplier with a track record in medical power supplies, GMP and FDA certifications. And perhaps the customer should also visit the supplier's design team and manufacturing facilities. Ask what other medical electronics companies they are qualified to supply to and ask for documentation that proves the power-supply company meets these requirements.

Oh, and yes, buy a good power supply and add an isolation transformer on the input just to be sure that you meet the isolation requirements. The happy ending is that the customer implemented the advice and experienced no further safety, reliability or EMI/RFI incidents, and the customer's product is selling well.

Customers remember quality and long-term value forever, while price is only a temporary memory. Why risk it all on a no-name power supply when price is not an issue and lives are at stake? Buy from someone with a proven track record that invests in meeting the market requirements so that your end product can meet its end-market requirements.

Kevin Parmenter currently leads the FAE team in the Americas region. Previously, he held various technical and management positions with ON Semiconductor and Motorola. Parmenter serves on the board of directors of the Power Sources Manufacturers Association and as program chair for APEC 2008.

Sponsored Recommendations

Comments

To join the conversation, and become an exclusive member of Electronic Design, create an account today!