I've written and debugged firmware for many years and I can really relate to this story. Solving problems like those described in the story take time and a very good engineer. Unfortunately there's no way to measure the difficulty of a problem so there's no way to measure the value of the solution. Managers usually just look at how long you took and draw their own conclusions.
Interesting that in a new system -- new hardware and new software -- what ultimately turned out to be a signal integrity problem was blamed on software. "Mechanical and hardware assemblers said it was obviously software." And management simply took their word for it, without any data to back up that claim?
There are two reasons for that:
1) It is relatively easy to see how hardware works and much harder to see how software works. That means people will always tend to believe the problem is where they can't see it (ie. in the software).
2) It is also a matter of hope. Software errors can be fixed on short deadlines, but not hardware errors. Thus management/sales people always hope it is a software problem. When a product is on stop ship they will readily believe anything that gives them hope that a solution is at hand.
Even if it is not really a software problem, there is often a software solution. Whenever this happens it becomes a software problem.
I once had to fix a lubrication problem in software. Without that, many idetms would have had to be recalled and scrapped costing the company a few hundred thousand dollars.
On another occasion a mechanical stop was repositioned. This could have cost approx $100k and 4 months to fix mechanically meaning the company would have lost a few million in sales. I fiddled for a day and found a software solution.
At the end of the day, all that matters is that a solution is found. The customer doesn't care what is broken or who's fault it is.
Just work together to find a solution. Everyone is trying to earn money for the same company.
The trouble is, when you find a software WORKAROUND (emphasis important) to a hardware PROBLEM, and all that people remember is "we had to rev the software to fix it", the software side loses respect - including at review time.
Why should the software guys lose respect?
Make sure that it does not get framed that way. Make sure that the software guys get the credit for turning around a multi-million dollar loss.
If you want recognition in a company then the best way to do that is to make an association between what you do and the performance of the company. Demonstrate **value**.
In your performance review, point out the value of what you did.
The software guy looses respect, because he caused the product to be late in the first place.
It doesn't matter if it was a HARDWARE problem, we get the short end of the stick; because we fixed the problem after everyone said the product was ready to ship.
This reminds me of a time when a watchdog timer would sporadically time out when a processor took too long to execute a command/response sequence on a control bus. This was a new product in both HW and SW, and the late night sessions in the labs were not burdened by "it's a HW fault / no it's a SW fault". We just admitted that nobody knew yet where the fault was and we had to work together to find it.
Finally, using an analog scope (digital scopes had not been invented yet) the SW guy and myself saw an event whiz by that the timeout occurred within the 1 second allocated time of the hardware. I took another look at the hardware, a long-chain ripple counter and realized that the guy who designed this had done the stage count based on a complete cycle at the final stage. He forgot that the timeout actually occurred on the rising edge HALFWAY through the cycle.
Solution: knife and green wire.
Lesson learned: Never assume HW or SW. Test, test, test....
Well, I've worked with enough boneheaded software types who didn't know anything about hardware to say that you are the exception, Tim. Embedded engineers who know both are worth their weight in gold, even if boneheaded managers don't realize it. I totally agree with zeeglen; forget the finger-pointing and just figure out what's not working. Sometimes it's both H/W and S/W!
Yes, sometimes it is both. But we mustnot start pointing fingers before exploring the problem and what the cause is. I recently spent two days troubleshooting a problem and have worked with the HW engineer to try and determine the cause. So far, we have elimiated SW as being a possible cause to the issue.
Sometimes this is all in the software domain. Reviewing DOS driver code while writing new interface code on an ISA plug-in board, I found some code that seemed to have a race condition and asked about it. Since this had been written by one of the original product group, now a technical leader, and had been working for years, my question was disdained and dismissed out of hand. A mere few months later, intractable customer problems were blamed on my interface code. I finally traced it back the the same DOS driver code. Turned out this was the first customer PC system we had seen at whatever speed it was (500 MHz?) and it was the first time that the PC was fast enough to catch the race condition. Now it was still my fault for not having pushed harder when I found the problem before.
Been there, done it, got the company golf shirt. Years ago we had a hydraulic problem, excess pressure when braking. The redesign fix was messy, even for hydraulics. The $1M machine was down and people were not happy. I suggested a SW workaround to control deceleration, which we did and were soon back up running. However, word got around that I had to fix the SW to get the machine running. Unfortunately now I have to think twice before helping others.
Nobody has an exclusive right to make the mistakes, that is for certain. The problems arise when either the product definition is fuzzy, or the designer makes a mistake, or the builders don't follow the drawings. The funniest one that I experienced was a fairly simple circuit that I had designed and then built. I carefully selected the resistors to all be in the middle of where they needed to be. But when the production model was built, it was miles out of specifications. The problem eventually was traced to the arrogant detailers who did the circuit drawing. They corrected the "Mistake" made by "That dumb engineer", and added the "K" that I had neglected to put on several low value resistors. So instead of 22 ohms and 470 ohms there was 22,000 ohms and 470,000 ohms. The shortcoming of my designs are that they don't work as intended unless they are built as designed. This was reported to my manager at my next performance review, when the problem with that product was brought up. The response was "OH".
Mr. Fyock, welcome to the monkey house. Your best strategy would have been to quit, although for sure that is not always a feasible solution. Since idiotic management and moronic sales types have a tendency to butt into engineering (software and hardware) you might want to consider this advice that I was once offered: "the time to start looking for a job is the day you land your new one." Sounds cynical? Wait until you've been in the field for, let's see, 11 minus 74, ah, 37 years, and consider it again. Well, in any event, we've all been there...
In the groups I manage, when there is a bug both hardware and software are considered guilty until proven innocent. Hardware and software engineers are expected to work as a team. As mentioned above software often fixes hardware problems. Also hardware debugging techniques are sometimes helpful in tracking down a software bug. Management should not allow the different disciplines to blame each other.
I look at it a bit differently.
Without hardware, software is useless and without software, hardware is pretty lame.
It is the cooperation of both the hardware and software that makes a well designed system; which is our goal ultimately for software and hardware.
Software and hardware are often trained by management to blame each other :-)
(As in: when the "real" mission to create and ship and SUCCEED [as a team] is derailed into the "fake" mission to grab credit and avoid fault, this blame-game usually results...)
This is the second story about the hardware induced noise I read here. Since noise mixed in a signal is so common and most of the time it is random, problems happen here and there somehow give a hint, even not strong and clear.
A specific software in general will hardly produce random troubles here and there.
Fresh real-life example: A few weeks ago I released code that worked exactly as expected on prototype hardware. When used on the next board-spin of prototypes, it didn't work, which of course meant it was a software fault. Yesterday it was proven that not only was the circuit changed, but a wrong part had been used on the dozen proto builds, so it wasn't even the proper changed circuit. Put in the right part, and the software works again. Gee, what a concept.
I worked for a company for 12 years designing a range of fibre optic switches. (check out the spelling of fibre!)
My bit was the opto-mechanics. I can sympathise deeply with the sentiments in these posts. Guys, I feel your pain. One of the posts earlier mentioned that it is obvious when mechanics isn't working, but not so software and electronics. In the fibre switch, the mechanics was small and delicate, and small mechanical disturbances (thermal, vibration, electrical noise) could make a difference but it wasn't obvious by observation. When there was a problem the blame was assigned thus:-
1. it's a mechanical problem, the mirror is in the wrong place (it was once many years previously but not on the five plus subsequent problems)
2. Right, the mechanics is OK, must be an electrical problem
3. Right the electrics is OK, must be a software problem
Usually the 'head in charge would decide on what the root cause was and waste time and money trying to fix something that wasn't broken. The three disciplines had enough respect for each other to get on and sort the problem (eventually when the management "help" had evaporated when the solution was looking intractable)
Oh, and I relate to the story above about the saw project that was estimated to take 2 months scheduled to be done in one month and quoted to the customer as 2 weeks.
I don't work for that company any more :-)
Sweet. Ten years later after I worked for a certain solution “froward” company that my EEtimes story was based. In a recent 2015 interview one of those special people, who is no longer with said company, told an interviewer something that was strange and likely not true, if what was said was really said and not redacted. I asked for clarification from the interviewer about what was said or if they even had the right person and the interviewers are being tight lipped about who spewed. Just goes to show that if you do the right thing for an employer and they ultimately benefited, the ex-sycophants still years later whine that their obtuse behinds which they roasted in the fire for years was scorched and was my fault because I came by and pointed out they may be happier turning around. Makes me wonder how many other of my interviews/jobs these people have ruined over ten years.
Well the company seems to be going down for the third time. Talked to a former company engineer recently laid off and he said the company has been doing the same things as in the past. Not listening to engineers, cutting/turnovers, not buying needed parts/tools, and cheaping out. The technical debt eventually caught up and lead to low yields. So 200+ workers are now going to lose their jobs unless someone buys and bails them out as before. Sad.
Most of us work in some sort of chain, (or tower, if you prefer), producing work that depends on others, and is in turn depended on. (E.g. business or functional logic encoded in software that depends on hardware to work. I've always felt it desirable to have at least an outline understanding of the neighbouring fields. Hardware people need to know something about physics and chemistry on one side, and software on the other. Similarly, software types need to know a bit of hardware and some business, and so it goes on.