Following on from my last blog post about “Foundations of Research and Development for SMEs", I now want to share my thoughts on the workflow needed to get from a Schematic to a Printed Circuit Board (PCB).
For companies developing electronic products, an important part of that journey is understanding a circuit in detail before committing to physical hardware. This is where SPICE simulation tools play an important role.
SPICE, which stands for Simulation Program with Integrated Circuit Emphasis, is used to model and analyse electronic circuits. Without needing to build a prototype, you can investigate how a circuit is likely to behave under different conditions and with different component values.
This is a valuable way to reduce uncertainty, explore alternatives and make better-informed decisions.
From Circuit Idea to Simulation
From an idea/concept, projects start with a Schematic design, but then the challenge is understanding whether the proposed design can actually deliver the required performance.
SPICE models allow investigation of questions such as these virtually:
- What voltage and current will the circuit produce?
- How will it respond to different input signals?
- What happens at different frequencies?
- How quickly does the circuit respond?
- How sensitive is it to changes in component values?
- Will it remain within its required operating limits?
Methods such as Monte Carlo Analysis (where values are varied within set limits and the simulation is re-run) can be used to make the process systematic and ensure no unexpected surprises as component values vary.
Exploring Different Designs
One of the key benefits of simulation is the ability to effectively explore different design options such as component values which can significantly affect performance.
This can help identify the best approach while also revealing the trade-offs involved. A design might improve efficiency but increase cost, or improve signal quality while making the circuit more sensitive to component tolerances.
Understanding these relationships is an important part of R&D because the objective is not simply to find a circuit that works, but to understand why it works and under what conditions it remains suitable.
Understanding Uncertainty, But Simulation Does Not Replace Prototyping
However, real electronic components are not perfect with all devices having tolerances and characteristics that can change with temperature, frequency and operating conditions.
Whilst simulation can help investigate and understand these uncertainties, it is important to understand that SPICE itself is not perfect, especially for electronic/semiconductor components where the quality of the models can be very variable!
So simulation is not a complete replacement for physical testing! In particular, physical hardware introduces additional factors, including PCB layout, parasitic effects (though some can be included in more sophisticated models), component variation, temperature and manufacturing tolerances.
For this reason, simulation should be considered part of a wider development process:
Model → Simulate → Prototype → Measure → Compare → Improve.
But this should be iterative not linear and importantly looping back after Simulation is much cheaper and quicker than after Comparing!
The comparison between simulation and physical measurements can itself become an important source of R&D knowledge. If the real circuit behaves differently from the model, the model may need improving, or the physical design may need changing.
Supporting the R&D Process
For SMEs, SPICE can be more than simply an engineering design tool, providing evidence of the technical investigations during an R&D project. Documenting what was being considered, assumptions made, alternatives considered and outcomes from each iteration.
This fits closely with the wider principles of effective SME R&D: defining the technical challenge, investigating uncertainty, testing possible solutions and recording the knowledge gained along the way. As Biztech’s previous guidance highlights, successful R&D benefits from disciplined planning and an understanding of both the technical and commercial challenges involved. (biztech.org.uk)
Turning Simulation into Better Engineering Decisions
The real value of SPICE is in helping engineers answer questions before committing significant time and money to hardware. Especially for informing decisions about when physical testing is most needed and valuable.
The goal is not to eliminate experimentation, but rather to make it more informed and result in designs that are better understood, better tested and ultimately more likely to succeed.
There are many SPICE tools available with various pros/cons and these are some of the freely available ones :
LTSpice - Analog Devices (formerly Linear Technology).
https://www.analog.com/en/resources/design-tools-and-calculators/ltspice-simulator.html
QSPICE - Qorvo - By the original developer of LTSpice, Mike Engelhardt who is a renowned expert, that started developing simulators 50+ years ago!
https://www.qorvo.com/design-hub/calculators-simulation/qspice
KiCad - Open Source used by CERN - Integrates PCB Layout with SPICE (ngspice, etc.).
These tools have varying capabilities to help optimise designs, which include: automating the running of the simulations with different values for Monte Carlo Analysis, targetting particular required values, etc. However, they need for goals to be understood and to be specifically set up or configured! - Inevitably there is work being done on integrating AI with these tools, but I think there is a long way to go before this will be truly usable for complex circuits with many constraints?
If you are working on a particular design or concept, then please get in touch and I can explore with you the possibilities in detail.
Conclusions
As with most aspects of Design, R&D, etc. the earlier details are understood and catered for in the lifecycle, the lower the overall cost and time involved. For instance, launching a product and then realising any of the following can break a product or indeed a business….!:
- The Firmware cannot easily be updated to fix a bug is problematic.
- A vital additional feature cannot be implemented because there is no spare input/output available.
- Automated assembly/testing is not possible and the volumes have grown.
Understanding and working on the foundations outlined in this post will ease the Research and Development process and avoid major issues.
From my experience, a key consideration is to know how much simulation to do and when enough has been done versus actually building and testing the next prototype version!
What next?
That is all I can cover in this post. The next one will cover what to include and consider in a PCB design, how to get PCBs made, assembled and tested, etc.
I could not find a single good source of non-technical explanations of Monte Carlo Analysis, so contact me if you would like to know/understand more as it does not need to be complicated or mathematical!
I will follow up in future blog posts about the further stages towards being ready to launch a product and beyond…
Please contact me or Biztech if you need any specific assistance or contact details for any of the above.
If you are new to this, I hope that I have provided some insights that are helpful. If you are experienced then please let me have your comments or thoughts. Tell me about anything I have missed or should elaborate on, etc. Either way please do engage in discussion to provide more value to our community.
