And it is often said that an engineer (indeed this is true for most professional fields) is a net cost for the first 3-6 months because they require other engineers' time and don't produce much work in that timespan.
…There is a company going very strong now, selling three main products like hot cakes, which I fixed up in less than three months. One was an EMC fix, 2, A Mains Transient improvement, and 3, A Comms system which wasn’t working.
I think with SMPS its different, as they are generally very quick to design or fix, and often have existing engineers who can easily take it forward once someone has “gotten them over the threshold”.
At one company, I solved a long term problem in a few days…..they were developing a 600W current source power supply. (LT8705 based Buck-or-boost converter) They had omitted to add open_load protection to it. (This is not good because having output overvoltage protection for the case of open_load is a basic standard procedure for a current source power supply).
The problem was that each time the power supply was inadvertently started up on no_load, the output voltage rose up too high, and subsequently ‘killed’ the FETs which went short. They were using current limited lab power supplies to supply it, so this damage happened with the FETs “dying” silently. Anyway, they kept on finding themselves with damaged FETs and wondering why. The FETs were QFN style and only one person in the company was permitted to replace them. This board had spent most of its time on the “to be repaired” shelf. (6 months)
Therefore, I implemented open load protection for this power supply, and this solved the problem.
At another company, they were developing there first major product, they had a great engineer, who could easily do the job himself….they didn’t want another engineer coming in as they wanted to keep it secret…however, Vicor wouldn’t support them because they were putting more than 8 modules in parallel……so they brought me in for the development, just in case……as it happened we did it no problem, so they parted company with me after the 2 months that it too to piece it together.
Ditto many others
I had one company that were shipping a product in which the micro was sometimes browning out because its R/Z/NPN regulator had too high a base resistor…that was diagnosed within the first hour.
For one company, i fixed their problem at the interview...before even starting work....
A company had had some LED driver boards designed by a consultancy. These were Buck LED drivers, with 20V absolute maximum input voltage pins. There was an LC filter upstream of the LED drivers. This board was powered by a 12V OffTheShelf Power supply.
The boards were suffering a high field and production failure rate.
During the interview, I realised that the problem was possibly that every now and then , the production staff were “hot-plugging” the power supply in to the LED driver PCB. This could result in a >20V peak voltage ring on the Buck converter ICs. This turned out to be the case, and indeed the cause of the problem.
As a solution, Initially, it was decided to ensure that no “hot-plugging” took place in production. (The power supply’s soft start then ensured that no overvoltage occurred at startup)
I also provided an alternative design schematic with protection such that even when “Hot-plugged”, no overvoltage ringing would occur.