Biography has not been added
- On reference code
Is Embedded.com moderated? I have constantly seen abuse by spammers. Blocking URL posting from recently registered users could help, though I am not sure it is feasible (in a simple way, I mean).
Hi, Aubrey, if it works like TI, it will be fine. TI folks are very responsive most of the time. The problem is if it works like Analog's forum. :( (Linear Tech folks tend to be more responsive though).
- Saving power with relays and solenoids
Hi, Aubrey. I saw that idea. To be honest, I don't see why in these days one would not use a ready to run monolythic solution such as TI's DRV120. It's been a while since I worked with solenoid valves (including "linear" :) ) but I am pretty sure there must be some part that could do the job at a low price and with less hassle (i.e. less parts). FWIW, many aeons ago I used part of a step motor controller (xxx3717 - I don't remember the prefix) to do the job. At that time I think Allegro had some specific ICs but we were already using the 3717 on the board so we avoided yet another part on the BOM and it worked fine, reducing power consumption and heat.
- A rebuttal to “Why every embedded software developer should care about the Toyota verdict”
I think the question is not the resource/construct but the way it is (ab)used. IMO global variables fall in the same category as goto: because of the way they were badly used in the past they became anathema, a case of “shoot the messenger” (the contruct/resources) instead of the sender (bad design). I have seen beautiful software coded in Assembly and hideous one in C++ and I am pretty sure hideous code can be done in any programming language no matter how good it is. It might be possible to write a software without global variables. However, I wonder how much effort it would take and what would be the quality of the code in terms of readability, cost/benefit etc. I mentioned goto because it is deemed evil (more than globals I suppose) but there are circumstances where using it produces clearer and better code than avoiding it, though the instances where its use is justifiable is pretty limited so the amount of each should be minimal but hardly zero. Same with globals. Thus, the number of globals in the mentioned Ballista project is an indication of a lousy design. IMO the same applies to Toyota’s code, regardless of the extrapolation making sense: too many globals even for a limited sample, even if the use of globals is not the cause of the accidents. As for the expert witness in court, I feel some ad hominen attitude. I do agree though that claims based in hearsay are not best thing (including for the expert’s resumee). But in courts not always the truth or the right wins, unfortunately.
- Switch interfaces
Aubrey, we are never old enough that we cannot learn new things (wetting current I mean) :) . Also thank you for the heads up on the ICs. I have to admit I would probably not think of looking for specific parts like these for this kind of application. Cheers! Elder.