Kenneth MacCallum, Principal Engineering Physicist, StarFish Medical01.22.18
Inside nearly every medical electrical device today are multiple DC-DC converters (DCDC). These circuits convert from one voltage or current to another by sequentially switching energy between capacitive or inductive storage elements. They can be extremely efficient and made relatively compact. But every time a DCDC is needed, the decision must be made whether it would be better to source it as an off-the-shelf module or design it from scratch.
Why Design from Scratch?
The most common reason to design from scratch is to accommodate space constraints. Although many modules are volumetrically extremely efficient, they can often be inconvenient shapes when trying to shoehorn a PCB into a small space. Another good reason is when seeking to optimize for values or modes not available as a module. For instance, if a constant current output is needed (rather than a constant voltage), the choice of modules can be significantly limited.
Why Not?
When designing a from-scratch DCDC, it is often a challenge to minimize EMC emissions—whereas modules are usually well optimized with published data. Designing a PCB layout for a DCDC that minimizes emissions and coupled noise takes more effort than expected. Another design challenge is to ensure the circuit will perform efficiently and remain stable under all load conditions. This is usually a given when using a module, although not always the case. If the DCDC is not the focus of design efforts—especially for early prototypes—risking trouble with such a fundamental part of the system when under severe timeline or budget pressure is likely a poor choice.
Why Not Not?
A somewhat common sentiment is using a DCDC module is akin to cheating. (“Designing from scratch is the proper way to engineer!”) I argue any thoughtfully analysed solution—whether with a module or from scratch—is a good solution. It is most important that the appropriateness of the solution has been sufficiently evaluated and tested from whatever perspectives make sense for the particular application.
Another misconception is a from-scratch DCDC will be less expensive. This will likely be highly dependent on what the projected production volumes are. It can also be tricky to evaluate properly given the dependence on assembly cost, PCB cost, component price points, NRE and so-on. I have seen the cost go both ways depending on the circumstances.
Choosing the right inductor often catches junior engineers off guard. Don't ever be lulled into blindly using a part suggested in an app-note. In my opinion, every from-scratch DCDC design needs to be simulated in all typical and extreme load conditions to ensure the chosen inductor will not saturate. Similarly, when using electrolytic capacitors on the output, the RMS current should be set with sufficient headroom. Don't forget to simulate under worst case component values. Of course, it must actually be tested on the bench as well.
So?
When designing a medical device, there are many good reasons to design from-scratch DCDCs as well as use off-the-shelf modules. Both can be a great solution for a project’s particular problem. The key is to properly engineer that solution—whatever choice is made.
Find out more about StarFish Medical here.
Why Design from Scratch?
The most common reason to design from scratch is to accommodate space constraints. Although many modules are volumetrically extremely efficient, they can often be inconvenient shapes when trying to shoehorn a PCB into a small space. Another good reason is when seeking to optimize for values or modes not available as a module. For instance, if a constant current output is needed (rather than a constant voltage), the choice of modules can be significantly limited.
Why Not?
When designing a from-scratch DCDC, it is often a challenge to minimize EMC emissions—whereas modules are usually well optimized with published data. Designing a PCB layout for a DCDC that minimizes emissions and coupled noise takes more effort than expected. Another design challenge is to ensure the circuit will perform efficiently and remain stable under all load conditions. This is usually a given when using a module, although not always the case. If the DCDC is not the focus of design efforts—especially for early prototypes—risking trouble with such a fundamental part of the system when under severe timeline or budget pressure is likely a poor choice.
Why Not Not?
A somewhat common sentiment is using a DCDC module is akin to cheating. (“Designing from scratch is the proper way to engineer!”) I argue any thoughtfully analysed solution—whether with a module or from scratch—is a good solution. It is most important that the appropriateness of the solution has been sufficiently evaluated and tested from whatever perspectives make sense for the particular application.
Another misconception is a from-scratch DCDC will be less expensive. This will likely be highly dependent on what the projected production volumes are. It can also be tricky to evaluate properly given the dependence on assembly cost, PCB cost, component price points, NRE and so-on. I have seen the cost go both ways depending on the circumstances.
Choosing the right inductor often catches junior engineers off guard. Don't ever be lulled into blindly using a part suggested in an app-note. In my opinion, every from-scratch DCDC design needs to be simulated in all typical and extreme load conditions to ensure the chosen inductor will not saturate. Similarly, when using electrolytic capacitors on the output, the RMS current should be set with sufficient headroom. Don't forget to simulate under worst case component values. Of course, it must actually be tested on the bench as well.
So?
When designing a medical device, there are many good reasons to design from-scratch DCDCs as well as use off-the-shelf modules. Both can be a great solution for a project’s particular problem. The key is to properly engineer that solution—whatever choice is made.
Find out more about StarFish Medical here.