13/12/11 Simulation: Measure twice, cut once
Or how to save 20-30% or more in a design project
The old adage “measure twice, cut once” is thought to come from the world of wood work. Assuming the first measurement is correctly taken and marked can result in a large loss. The problem is that a “single” error in the process may be a total loss of the piece of wood.
A similar scenario is also present in the electronics design world. The first stages of a design may create schematics and a circuit board. Once this board is populated with parts, it can look like the whole process has been fast tracked and is “nearly done”. If the gamble pays off then it might work. However, if the gamble does not pay off, then the time taken to make it work will cost dearly in parts and time.
The “cut once world”
Measuring twice looks like taking twice as long to do the job. In reality the measuring side is a small proportion of the whole process. In a situation where there is a problem in the “cut once world”, consider the circumstances. The material price has doubled, the measurement time is now repeated and the whole cutting exercise is also duplicated.
Lean and agile systems tend towards minimised levels of operation and design with test in mind. Hawkshead has long operated such methodology and is pleased to find wider support for the lean/agile world. The difficulties of reconciling electronic design with agile software approaches has left many design entities in the “cut once world”.
A rigorous approach
There is an alternative. It is called measuring twice. There are several ways in which this can be done; simulation is perhaps the easiest of these to portray.
Simulation is an opportunity to work in a virtual environment. Like many such processes the concept of “garbage in, garbage out” very much applies. Calculations and spread sheets are probably in supporting roles too so that the input is not garbage. Usefully constructed simulations offer multiple benefits. A (not exhaustive) list of benefits includes:
- component values are optimised and system tolerant
- temperature or component variance is managed
- verification of circuit topologies
- correct closure of closed loop systems
- ability to continue tests into realms where real components may fail
An important outcome of the simulation process is the measured result or graph. It refines and perhaps defines the understanding that is required for a successful project conclusion.
Cost savings
When the prototype board is ready for test, both the designer and the client can share in a degree of confidence. The circuit values assigned will already be in the right range for first tests. The results of practical measurement can be readily compared with the original, virtual predictions and graphs.
If there is a problem or divergence, a return to the simulations will often highlight the relevant issues in minimal time. This avoids a vast amount of hunting in the proverbial dark. Where there are questions of power control or heat build-up, the simulation approach often saves the cost and time of replacing burned out power devices.
The steps from first concept through prototype and then to production and marketing are increasingly expensive. By the time the prototype is considered to exist it is likely that plans are well in hand to market the product. Delays into the market and worse still problems with the product in the market make the “twice measurement” a pocket change cost. It comes as some surprise that a more informed approach is not far more widespread.
The choice is simple:
- Take the fastest approach to cutting your wood and hope for the best
- Measure twice, avoid the gamble and reach the market perfectly sized
Posted by: Peter Hawkins on 13/12/11.