Sometimes you need to ask for help when designing a Power Supply
Power supplies are often thought to be a low level design consideration,tasks usually given or delegated to inexperienced or new engineers, and this can lead to catastrophic consequences.
Such consequences can and do, significantly impact warranty costs; from
cost blow-outs to scheduling delays to excessive field failures. The underlying problem is usually a failure to fully understand
the complexity of design in relation to power supplies, particularly switching-modes.
This underestimation of power supply design can be traced back to when power supplies were often just a transformer-rectifier set, with a shunt regulator or dissipative series
regulator. I t was not an accurate reflection then, and it still isn’t now; especially when most power supplies these days are switchingmode.
Switching-mode power supplies can be defined as non-linear circuits which are not easily stabilised, and frequently have real and
complex poles in the right-half plane – which are averse to magnetically complex chaos.
This needs serious contemplation of eddy currents and proximity effects, and usually needs planar construction of analog and digital
power circuits. This needs the application of discipline sources of electromagnetic interference for example – they are certainly
not easy designs.
This issue of ignoring or underestimating design complexities of power supply, and its resulting consequences, applies to virtually all
electronic systems – as pretty much all need a power supply.
To solve any such problem first requires widespread awareness. Initially the subject was buried or ignored – the perception was
that power supplies were simple and therefore not of any notable significance. In reality the
problem is addressed by a shift in approach and taking the issue more seriously. First this needs engaging someone from the start who
has the necessary skillset and understanding.
Secondly, involving the vendor or manufacturer at an early stage - and providing adequate skilled resources and tools. Finally by using
manufacturing screens and regular in-depth design reviews.
The key to solving this issue is awareness; mistakes are bound to be repeated where a new group comes along and does not have
visibility of problems and solutions undertaken by members of a previous group. History, in such cases, is bound to repeat itself.
During the 1 980s there was great awareness of power supply reliability and a solution was identified. But there is still a significant design
and application issue going unaddressed. Ignorance or lack of awareness of the implications and the need to manage this risk,
leave any current program manager significantly exposed. This can lead to large problems such as cost blow-outs, scheduling
delays and failures in the field.
In contract selection the lowest bid should not the main factor. The main factor is that the contractor selected is capable of doing the
work. This is normally established by providing evidence of their manufacture and design of other, similar power supplies. In addition to
their history of previous bids is the time and cost contained in those bids. One time I asked a successful bidder for the three priority
recommendations he would give to others.
When problems arise, always look at how things could be improved, and avoid repeating previous errors. I f you award the contract to
the lowest or unqualified bid; or retain it inhouse, this just means you will get an unqualified design or inexperience with power
supply design.
Call for help at the first sign of problems, and then use enough force to complete the job – in time you may be able to leverage your internal expertise and end up with a well-designed
supply – but it will likely be a high cost of scheduling delays and project cost blow-outs.
When a scheduling issue or delay occurs, program managers will request full assistance of the company, and in addition look to their
customer for resources to help too. Calling for assistance, and determining what needed to be achieved, the job gets done in the right
time.
The lack of program managers calling for help, is increasing today, not falling. Such product teams and their managers take great pride in
being able to be seen to get the job completed, however they frequently get out of their depth because specialised skills are not applied
when and where needed. By the time this comes to light, the program schedule is usually already running behind and therefore costing
you extra money.