What is a good design?

[spec]

Well put TCM- I know exactly what you mean about design weaknesses.

You can get a car where the parts inside the engine- crank, pistons, big ends.. can last thousands of miles while under extreme mechanical stress, temperature and pressure. Yet the manufactures cant make a door hing that has the most undemanding conditions and is hardly used, last more than a couple of years. Or electrical connections that do not corrode.

One car has to have the whole dash out just to change a bulb- crazy

Then there is the great exhaust scam- why an exhaust cant outlive the car I dont know. An exhaust in only a pipe and it doesn't do much.

I won't mention car batteries.

spec

 

[JimB]

why an exhaust cant outlive the car I dont know

It can, if you make it of stainless steel.
But then it would cost a few pounds more and that would reflect in the selling price of the car.
As most people who buy new cars only keep them for two or three years, they don't care for longevity, they just want a car at the lowest price. There is still plenty of life left in the cheap carbon steel pipes after three years.

A bizarre example of why things are done some very odd ways...
Some years ago I was involved with a project where I was responsible for the control computer for the overall system, including A PLC which controlled an Hydraulic Power Unit (HPU).

Early in the design phase I proposed that we should fit the PLC in an explosion proof box at the HPU. (The HPU was to be located in an area where there could be explosive gasses in the air).
The client guy responsible for design and procurement was against this idea, he wanted the PLC mounted in the computer rack in a safe area.

I pointed out to him that my way would result in a self contained HPU. Useful for acceptance testing involving the HPU where the control computer may not be available.
There would be a minimum amount of cabling to be run on site between HPU and computer cabinet.
Overall this would be a cheaper option for the project.

But no, the client guy was having none of it.
His reason...
Doing it his way, the procurement cost would be lower which would make his budget look good.
The extra costs would be on the installation and commissioning budget. Not his problem, he was not responsible for that.

Joined-up thinking for the benefit of the overall project?
No chance.


Why I bought a Bosch dishwasher.
Many years ago the late Mrs JimB gave me ear ache about wanting a dishwasher.
So off we went to the various domestic appliance stores and looked at what was on offer.

One thing which immediately stood out to me was that most of the dishwashers from the various manufacturers were badge engineered versions of the same thing.
Particularly the door release mechanism which as you opened the door it felt and sounded like tortured cheap plastic being graunched together.
All except the Bosch dishwasher where the door mechanism felt nice and smooth, and still does 15 going on 20 odd years later.

JimB

 

[spec]

It can, if you make it of stainless steel.
But then it would cost a few pounds more and that would reflect in the selling price of the car.
As most people who buy new cars only keep them for two or three years, they don't care for longevity, they just want a car at the lowest price. There is still plenty of life left in the cheap carbon steel pipes after three years.

So true. The people who suffer the faults have no clout.

A bizarre example of why things are done some very odd ways...
Some years ago I was involved with a project where I was responsible for the control computer for the overall system, including A PLC which controlled an Hydraulic Power Unit (HPU).

Early in the design phase I proposed that we should fit the PLC in an explosion proof box at the HPU. (The HPU was to be located in an area where there could be explosive gasses in the air).
The client guy responsible for design and procurement was against this idea, he wanted the PLC mounted in the computer rack in a safe area.

I pointed out to him that my way would result in a self contained HPU. Useful for acceptance testing involving the HPU where the control computer may not be available.
There would be a minimum amount of cabling to be run on site between HPU and computer cabinet.
Overall this would be a cheaper option for the project.

But no, the client guy was having none of it.
His reason...
Doing it his way, the procurement cost would be lower which would make his budget look good.
The extra costs would be on the installation and commissioning budget. Not his problem, he was not responsible for that.

Joined-up thinking for the benefit of the overall project?
No chance.

Yeah, I have seen this sort of thing. When you get different responsibilities on a project as you say it is not all joined up. I have seem people doing tasks that they know are wrong, but they have a specification in their hands to cover themselves and that is all they care about.

Why I bought a Bosch dishwasher.
Many years ago the late Mrs JimB gave me ear ache about wanting a dishwasher.
So off we went to the various domestic appliance stores and looked at what was on offer.

One thing which immediately stood out to me was that most of the dishwashers from the various manufacturers were badge engineered versions of the same thing.
Particularly the door release mechanism which as you opened the door it felt and sounded like tortured cheap plastic being graunched together.
All except the Bosch dishwasher where the door mechanism felt nice and smooth, and still does 15 going on 20 odd years later.

I always buy Bosch, especially tools. They are just so well made and they perform so well.

I bought a Bosch road breaker about 10 years ago- it wasn't that expensive, by comparison either. It has had some real stick but still cuts through concrete like butter. Yet more expensive and bigger breakers I have tried were nowhere near as good.

spec

 

[tcmtech]

Then there is the great exhaust scam- why an exhaust cant outlive the car I dont know. An exhaust in only a pipe and it doesn't do much.

It must be a world wide thing.

As you know I work with more commercial and industrial application stuff than most so I know the details about certain material that relate to them better than most as well.

I had a new exhaust put on my pickup about a year or two ago. standard 2.5" galvanized muffler pipe cost me ~$65 a 10 foot section and it's thin at that. Standard 2.5" stainless steel pipe of the similar wall gauge is about the same price at the steel yard ~ 3 blocks from the muffler shop I got my exhaust system work done at. Coincidentally they also carry a thin wall galvanized pipe that looks almost exactly like my muffler pipe and it's about $1 - $1.25 a foot. Hmmm.

Guess who's supplying his own stainless steel pipe for the next exhaust system job?

 

[tcmtech]

This summer I had to replace the main PTO drive gears and bearings on our old Massey Ferguson 2745 tractor largely due to a stupid simple design flaw that lead the failure. Big messy labor intensive two day project to do.

The whole issue comes from what came to be seen as a common weak point in that series. It's by far one of those 95/5 designed machines where the 5% part and its crazy hard and costly fixes gave it the hugely bad reputation that killed its production and much of its market popularity back in the day.

The problem it has with the PTO design is there is an inner bearing buried way inside the rear end that is inaccessible without doing a major rear end teardown to get to. The problem with the bearing is that it doesn't have an inner race but rather runs directly on the PTO shaft itself which by design is not a case hardenable shaft component so as time goes by the raceless bearing starts to chew into the shaft until it goes so deep it loses its rollers and falls apart. Once that happens the shaft will start to sit crooked and that causes the teeth on the big drive gear and the one that drives it to mesh cooked until they start to chew themselves to bits.

Now what gets me is that special bearing now costs around $150 but there is another mass produced bearing of near identical dimensions that has an inner race and exact same OD and thickness but it has an ID of something like 10 thousandths of an inch smaller (32 mm Vs 1.25" a ~ .00984" difference) plus it costs like $25 because it's so widely used.

Not machining that shaft down another .00984" so that it would use a cheaper but far superior designed bearing that has an actual inner race that wouldn't be chewed up by normal use is just one of a dozen tiny stupid engineering oversights that gave that series of tractor a horrible reputation.

The second major issue with that series is they were notorious for having their overly complicated multi stage transmission and related integrated shift lever mechanisms jump out of position inside the transmission and leave the tractor locked either in one gear or between gears which when that happens is a grimy 3 - 4 hour job to undo.

However as I figured out, and later redesigned it after doing that several times, the main plate that overly complicated shifting mechanism is built on (the 3 - 4 hour job to get off and back on) can have three holes, two 3/4" and one 2" pipe size, drilled and tapped into it that allow a guy to get full access to the out of place internal linkages for the outside (without spilling a teaspoon of transmission fluid) using nothing more than a channel lock pliers and a 3/8" x 14" metal rod to put them back in place in about 5 minutes. (I've done it twice now so it is that easy.)

That's the stuff that drive me nuts.

 

[KeepItSimpleStupid]

For the most part, I built stuff for in-house use and for the most part, when I built it, i never wanted to see it again in the same form. I started to build things more like appliances and they never got thrown away. They were repurposed. Control system could have been stripped to the re-useable component, but they never had to. A temperature controller case I designed was there 25 years later, with a different inside.

One was a fairly small box with a huge 30 A power switch. A place to put a temperature controller and an SCR unit. An EXT/INT and ON/OFF type switch where OFF set the set point to zero and allowed you to still monitor temperature. So, the front panel had full disconnect (rarely used), and On/Off.
A front panel mounted load fuse and an internal semiconductor fuse on the pull-out SCR unit.

All of the temperature controllers were converted to pluggable and the pins standardized and they all had Int/external setpoint and measure value out.

The back had a hole for a panel mount thermocouple connector AND a grommet for a length for say some exotic extension wire, like a "C"" type thermocouple. So, if the chassis mount wasn't available, use a pigtail.

The back also had a standard 120 VAC outlet AND a cord gland. Hardwire if you want.

The back brought out the set point and measured value stuff on a terminal strip and also offered a lockable potentiometer to set the current limit.
Insides were wired for 30A.

Eventually our production stuff was upgraded to a design by me and those little "desktop" boxes became homes for the obsolete stuff.

==

Repair was always fun. TCM's right. in a particular power supply that broke, I asked why did you use a 10 A rectifier in something that needs to supply 22 Amps continuously? Your right! it will require a major engineering change and we'll do it for you for free. Thank you for finding it.

==

Your brand now fancy meter puts out 100 V spikes on it's input when you changes ranges and it kills our devices. Them: I don;t believe you. Me: OK, do this. Them:Your right! They investigate and immediately send out an external gizmo that cripples the meter in a good way. Once the engineering fix was designed, it was fixed for free.

==
Your software crashes on this machine only. You sure? yep? Ok, we'll send you a version that makes crash dumps. OK. yep it does. We'll send you a debug version OK. They find problem and fix.

++

I don;t see this sort of relationships with consumers anymore. It was a good relation ship.

One of my more annoying issues was with "the boss" having a student change a thermal in an environmental chamber. It was difficult, but I had done it a few times. Actually, i think they needlessly changed the entire controller and removed the IEEE-488 feature because of what i found after the job was done. They had to call me in to fix it anyway.

==

Then there's this HVAC technician that diagnoses an old heat-pump as bad. They get a brand new one installed and it still wouldn;t work. "Sly" asked for my help. I looked at the manual overnight and said I can give you about 10-15 minutes max before my bean counter catches me. "For the greater good" is not in my bosses' vocabulary. I gave him 15 minutes and we did a few things under my direction. I said replace the thermostat wire and left. He did not believe me it was the problem, but did it anyway. During the remodel, the wire got scraped on the metal studs. That should not have happened at all if it was done right the the first time with wire grommets on the studs.


==

Me as consumer. Your Skype/landline handset product has a few bugs can you fix them. The product is at end of life. We are concentrating on the Skype/Mobile market now.

 

[spec]

It is amazing when you report a design error to a manufacturer and they don't fix the problem. You would think that the manufacturers first priority would be to ensure that their product worked well.

These are some of the responses I have had when reporting a problem with a product:

'No one else has had a problem'

'Oh, we don't actually make that product- it is made by a subcontractor in India.'

'It must be the way you are using it- have you turned the on/off switch on'

'We always ensure that our products are to the highest standard and value you feedback on this aspect. We will raise a fault ticket. Here is the reference DDat4-XCT-gyo3_34GDX-MM-cft- xxyask-#

'Do we make that product- sounds like one of our competitors products'

'Yea, the chap who designed that unit died/got fired'

'That electric toaster is not designed for toasting bread'

'If for any reason you are not happy with our product please see your retailer for a full refund'

'Yes we know about the problem- people keep phoning in about it. It is something to do with the bracket on the spin dryer I think.'

'The one-year warranty period has expired.'

'Yes we know about the problem, but there is no way our manufacturer in China will change it'

'Sorry you have the wrong number. This is customer relations for Alergy Consumer and Industrial, you need Alergy Consumer and Non-Industrial. Please phone xxxxxxx-xxxxxxx. Is there any other way I can help?'
'Hello, Larry here from Alergy Commercial and Non-consumer. No, this is Alergy Consumer and Non-Industrial Europe, you need Alergy Consumer and Industrial UK, on yyyyyyyyy-yyyyyy x yyyyyyy. 'Hello this is Martin from Alergy Consumer and Non-Industrial UK. No you want Alergy Consumer and Industrial on xxxxxx-xxxxxxxxx.'



spec

 

[Cicero]

To me, 'good design' in electronics is only learnt through experience. End of.

Book smarts only gets you so far, but it can never replace the knowledge gained through practical implementations in real world environments.

 

[JimB]

can never replace the knowledge gained through practical implementations in real world environments.

Oh yes!
How many times have we all said "I will never make it like that again!"
I guess that some people cannot recognise just how bad something is, and just keep on repeating the mistake.

JimB

 

[Cicero]

Oh yes!
How many times have we all said "I will never make it like that again!"
I guess that some people cannot recognise just how bad something is, and just keep on repeating the mistake.

JimB

And the amount of times I've reassured myself that "Nah, no one would EVER do that to it, or use it like that, don't be ridiculous, you're over engineering...". Only to find out that's exactly what someone's gone and done.

 

[tcmtech]

Relating to the subject one of my bigger peeves has less to do with poor design but with dealing with incompetent arrogant people both in the customer and the service side of things and it seem to be getting worse as time goes on.

To me if a customer brings a machine or device into a service center to have it worked on that is a very strong indication they do not posses the skills nor tools to fix whatever is wrong themselves. Not 100% of the time but the huge majority. The few who could do the work but lack the tools tended to not play a significant part in my expericanes.

The first problem is the customer who assumes that they are always right BS. Sorry but not everything that breaks or fails is in fact someone else's fault. Deal with it and quit complaining. Second just because you spent a lot of money of something doesn't make it a good quality item either. Deal with it. Some of the biggest pieces of junk anyone will ever find on the market come with the biggest price tags on them. Especially if they are targeting the recreational or disposable income markets.

Second is the service people and something all of us here know it too well. Just because you went to school at XYZ and took some classes doesn't actually mean you got a proper and correct education relating to why you were suposed to be there. You were just there that's all. Second just because you worked for or with company XYZ for XX number of years doesn't mean you actually ever paid any attention let alone learned about how what it is you worked on and with all that time was built and the why and hows behind the particularities of its overall design. Very few service technicians ever do. Show up, spin wrenches for X hours then go home. That's what they do and did for their whole life. Not all but the vast majority fit that description.

Its how you can have someone who fixed a certain brand of devices for 40 years and may know how to take them apart and put them together better than the average person but that itself is still zero indication of their having any degree of applied engineering knowledge or skills behind what they worked with all that time which means that despite having 40 years of hands on experience they are likely going to get there ass handed to them on the technical aspects and reasonings of the design and construction of something by an engineer who has only been at it for a year or two.

Just because your the customer and you paid way too much for junk doesn't make you right and just because you spun wrenches for your whole working life doesn't actually mean you know more than any technical engineer. Deal with it.

 

[ronsimpson]

Good Design: One I did.
Bad Design: One that I think I can do better.
Humble submitted, RonS.

 

[large_ghostman]

Hmm for me it would be, a design that does what the person wants, not what you think they want.

Or the opposite of, if it isnt broken then you havnt added enough features.

 

[ronsimpson]

A good design:
A design where the costumer pays money for it.
A great design:
Is where the costumer comes back for more.

 

[spec]

A good design:
A design where the costumer pays money for it.
A great design:
Is where the costumer comes back for more.

Quite true, but how do you do that?

spec

 

[ericgibbs]

hi spec,
I would say 95% of my business was by customer repeats or referrals.
Selling a reliable product at a fair price and on time is far more effective than advertising.
Eric

 

[spec]

hi spec,
I would say 95% of my business was by customer repeats or referrals.
Selling a reliable product at a fair price and on time is far more effective than advertising.
Eric

Yes, a good ethic- it is a great shame that many companies/organizations don't have the same view. In many cases the customer is treated with contempt, but worse still the customer simply does not come into the equation.

I live in a holiday resort which is packed with cafes and fish and chip shops- most are just struggling along and many have closed. But there is one cafe and one chippy that are always packed, even in the graveyard day, Monday.

Although the fare in the cafe is good, many customers go there because of the ambience and friendly staff, not to mention some splendid Polish, Hungarian, Lithuanian... waitresses.

The fish and chip shop food is excellent and, once again, the staff are friendly and very efficient and even though the prices are comparatively high, the place is always packed: on Fridays there is a queue on to the street.

spec

 

[spec]

My wife used to always do our big weekly shop at Sainsburys supermarket, which was slightly upmarket, and over a year this amounted to a big spend.

But slowly the standards started slipping. One day I was standing by the checkout with my five yer-old son and this girl supervisor started being abusive and implied that I had turned on the belt that carries the food past an unmanned till (I don't think that is even possible).

Anyway, after informing her of my displeasure, we never went to Sainsburys again. That was 1975 so just imagine the business that they lost in 42 years, due to poor staff selection/training.

Waitrose is now our favorite supermarket and we drive out of town to go there.

spec

 

[spec]

On this thread there is, quite rightly, great focus on giving the customer what he wants, but there is a subtle implication. Do you give the customer what he asks for, or do you give him what he wants- there is a subtle but fundamentally important difference.

To give an example, our company was tendering for a massive radar contract and, as usual, there was the marketing side and the engineering side (I was not involved). It was in the days when digital was not the defacto choice and components, especially digital integrated circuits, were not only scarce but very expensive.

The engineering department found that a digital approach was by a mile the best approach in every aspect, but the marketing department firmly opposed this saying that a traditional cursive approach was what the customer wanted. The engineering department put forward case after case showing how a digital approach was superior but the marketing department were intransigent.

Several proposals were submitted, but all but ours, were for a digital approach.

After a few months our company was given the nod from the customer, who implied that they particularly like our traditional cursive approach. A few months later we were awarded the contract.

Then the detailed discussions started between our engineers and the customer's engineers. The customer's engineers wondered why we had not propsed a digital approach but we had been briefed by marketing on no account to mention a digital approach.

As time went on it was clear that the customers engineers were not happy with our cursive approach, and to cut a long story short, not that I know the details anyway, we were requested to quote for a revised technical specification using digital.

We went digital and the contract became the biggest, most successful and long lived of the group. And the spin off systems were sold world-wide.

Had we have proposed a digital approach on day one would we have got the contract- who knows, but I could make a guess.

Were our marketing department right- you bet.

Were engineering just pawns in a bigger game- you bet again.

spec

 

[large_ghostman]

Good question Specs. Shouldnt you strive to give them what they need though? Which may be different to what they ask for, or what they want....