Model Design Errors

[Ian J.]

...and that's the other systemic problem, 'us'. As long as we are willing to say 'it looks more like 'x' than a brick so I'll buy it' then the producers of the models will work down to that lowest level of spend, regardless of the probability that a better model may make more profit in the long run.

 

What irks me is that the basic checks that I outlined previously wouldn't cost much, and with the reduction (though not eradication) of errors that would follow, could lead to better sales, more profit and better reputation. With my business head on, I'd call it a 'no brainer'.

[Chris M]

1 hour ago, Ian J. said:

...and that's the other systemic problem, 'us'. As long as we are willing to say 'it looks more like 'x' than a brick so I'll buy it' then the producers of the models will work down to that lowest level of spend, regardless of the probability that a better model may make more profit in the long run.

 

What irks me is that the basic checks that I outlined previously wouldn't cost much, and with the reduction (though not eradication) of errors that would follow, could lead to better sales, more profit and better reputation. With my business head on, I'd call it a 'no brainer'.

Have you thought of offering your consultancy services to Bachmann, Hornby and Dapol? I'm sure they would welcome your expertise on making perfect models and improving profits.

[spamcan61]

1 hour ago, Ian J. said:

 

What irks me is that the basic checks that I outlined previously wouldn't cost much, and with the reduction (though not eradication) of errors that would follow, could lead to better sales, more profit and better reputation. With my business head on, I'd call it a 'no brainer'.

As I asked earlier, checking against what - what will be the master reference? Who will check the checker?

 

There are precious few business where making the absolute best (for whatever somebody's definition of 'best' is) is actually the most profitable way of doing business.

[Ian J.]

OK, look, I have not said I'm looking for 'perfection', 'absolute best', etc. I'm just trying to understand why repeated errors occur of similar nature, and trying to think of ways, within reasonable bounds and costs, to reduce the occurence of those errors so that all of us as modellers can benefit, with the added positive that the producers sell more models.

 

The last two replies are still in the 'too difficult, don't bother, civil service' attitude, and make it look like I'm writing in invisible ink!

Edited June 26, 2020 by Ian J.

[MJI]

Just got to make this point again.

 

In the past the crudish models looked like what they were supposed to.

 

Old Triang, Some Lima looked more like what they are than more modern stuff from Bachmann and Hornby.

 

Look at the 31?

 

I have been modifying a lot of Mark2 models recently and the most Mark 2y model I have used are the old Triang TSO and BFK. Glazing is inset, the door trims a bit over done, the vents just blobs, but with work and new window frames they can pass muster with newer models.

 

Lima profile not quite right, Airfix the bottom of side profile, Hornby 2E too many small inconsitancies to bother with compared to Airfix off Ebay, Bachmann 2F price so too expensive to modify.

[spamcan61]

20 minutes ago, Ian J. said:

 

The last two replies are still in the 'too difficult, don't bother, civil service' attitude, and make it look like I'm writing in invisible ink!

Nah; it's the decades of product development in commercial businesses attitude, not civil service.

 

 

[melmoth]

3 hours ago, Ian J. said:

...and that's the other systemic problem, 'us'. As long as we are willing to say 'it looks more like 'x' than a brick so I'll buy it' then the producers of the models will work down to that lowest level of spend, regardless of the probability that a better model may make more profit in the long run.

 

What irks me is that the basic checks that I outlined previously wouldn't cost much, and with the reduction (though not eradication) of errors that would follow, could lead to better sales, more profit and better reputation. With my business head on, I'd call it a 'no brainer'.

 

Ok, the three points that I've highlighted jumped out at me. In order:

 

1. This does not necessarily follow. Obviously, profit is made by selling something for more than it costs you to produce it. A better model does not in anyway automatically lead to more profit. If the increased cost of a better model erodes margin, or, if a higher price is needed to maintain margin, it just adds cost without benefit - running to stand still. Model railways are what an economist would define as an 'elastic good'. The rate of consumption can fluctuate massively dependent on any number of market conditions (food is a fairly 'inelastic good' because everyone needs to eat regardless).

 

2. How exactly do you know this? As my maths teacher used to say "please show your workings".

 

3. There are people who have responded to this thread who appear to have some grasp of what is involved in industrial production, project management and quality control, to name but three relevant areas. I have absolutely no experience with any of those things, nor can I do anything CAD related (although I can and do run a business). You may have skills in all of those fields and more. If so, I'm sure you'll be able to outline the ways and means (in some detail) as to how to achieve what is a thoroughly laudable aim. If not however, you do run the risk (IMO) of appearing like the man who thinks a particular task is easy because they are blissfully unaware of the various complexities of said task.

[scottystitch]

2 hours ago, Ian J. said:

. I'm just trying to understand why repeated errors occur of similar nature,

 

Without wishing to take sides, how about this ?:

 

"Boss we scanned the peak, and here's the CAD"

 

"Him, it doesn't look right here, here and here.."

 

"Okay, I'll take another look"

 

later

 

"Okay boss, here's where we are"

 

"hmm, okay, that front end still doesn't look right.!"

 

"yes, I know, but we can't get the material we use to go to the required thin-nis/angle/shape that would improve it, without throwing something else out"

 

"Okay, can we use something else?"

 

"Sure, but that costs three times as much, and will add £50 onto the model cost"

 

"Okay, well I can't see that being an option, it takes it out of our target price bracket.  Do the best you can with what we have.  Now there's something about the curve of the roof and that door line."

 

"Yes the laserscan doesn't do curves very well when there's a lot of reflection from the sun and stuff, so we've really had to do that manually.  It's a problem, because we don't have the real machine in front of us we're kind of working from photographs from different angles.  It's tricky.  But I asked Dave and he took a look and he says it looks pretty good."

 

"hmm, I'm not sure. If you look at this photograph..."

 

"Yes, I know what you mean, but look at this photograph and this drawing."

 

"Yes it looks different.  But that's an original works drawing, so I'm not sure we can take that as gospel."

 

"I know."

 

"The thing is the slot we've got for this, the CAD needs to be complete by tomorrow"

 

"Okay, so what do we do?"

 

"We do the best we can, with what we've got."

 

"But I'd really like to get this 100% correct"

 

"Yep, me too but we have to deliver this within a timescale and within a budget, we could spend a year on the CAD, but your time designing this isn't free, you're a cost.  The more we spend on you doing this, the more gets added to the model price."

 

 

and on....

 

 

 

I'm not suggest it's what happens, I'm not suggesting it's right if it does happen, but this type of thing might account for at least some of the common areas where errors are found.

 

Or maybe not. 

 

Just thinking out loud.

 

Best

 

Scott.

[TEAMYAKIMA]

Having read all 5 pages of this thread this morning and heard about all the angst people are suffering from these (relatively) minor issues I realise how lucky I am to have built and exhibited a foreign layout - a VERY foreign layout !

 

In fact in my case it's sooo foreign that 99.9% of exhibition viewers probably don't know the prototypes and certainly don't know the models - they just look at the scene and enjoy it (or not) for what it is  - a  model railway which gives the viewer a flavour of the real thing.

 

Perhaps it is true that knowing too much about a subject can be a bad thing.

Edited June 26, 2020 by TEAMYAKIMA

[bécasse]

"Yes, laser scans don't do curves very well ............" Nor did original drawings, and therein lies the real problem for a manufacturer attempting any "modern"-image model, replicating how sheet metal was bent up by skilled craftsmen on the shop floor without any reference to drawings (which never made it to the shop floor in most loco works).

I might add that even if the manufacturer does manage to create a perfect scaled-down replica, it can still look wrong because the pair of eyes viewing it haven't been scaled down in proportion and therefore the perspective of curves as seen is wrong.

[Ian J.]

2 hours ago, melmoth said:

 

Ok, the three points that I've highlighted jumped out at me. In order:

 

1. This does not necessarily follow. Obviously, profit is made by selling something for more than it costs you to produce it. A better model does not in anyway automatically lead to more profit. If the increased cost of a better model erodes margin, or, if a higher price is needed to maintain margin, it just adds cost without benefit - running to stand still. Model railways are what an economist would define as an 'elastic good'. The rate of consumption can fluctuate massively dependent on any number of market conditions (food is a fairly 'inelastic good' because everyone needs to eat regardless).

 

2. How exactly do you know this? As my maths teacher used to say "please show your workings".

 

3. There are people who have responded to this thread who appear to have some grasp of what is involved in industrial production, project management and quality control, to name but three relevant areas. I have absolutely no experience with any of those things, nor can I do anything CAD related (although I can and do run a business). You may have skills in all of those fields and more. If so, I'm sure you'll be able to outline the ways and means (in some detail) as to how to achieve what is a thoroughly laudable aim. If not however, you do run the risk (IMO) of appearing like the man who thinks a particular task is easy because they are blissfully unaware of the various complexities of said task.

 

1. The more you sell of something, the more profit can be made? I know it's anecdotal, but I often read here of people buying one or more additional examples of a model because it really works for them, and fewer when it's not quite right (examples that spring to mind in the more category being the SJW24, the Hornby Peckett, and in the less category, the original Heljan 47).

 

2. My workings being if it takes an hour to run through the basic checks (and it really shouldn't take much longer than that) every couple of weeks, at say a professional like rate of £50 per hour, that's maybe £75 a week, give or take and remembering that employer's tax and NI have to be taken into account. £75 a week for a set of basic checks shouldn't add even 1.0% to a budget of £100k.

 

3. I am not, and have never claimed to be, inside the industry (I think I may have mentioned that in my original post, but can't see it right now). I am not even in any production industry. However, I do in my job (database and application design) have to review and check on a regular basis and look for efficiencies in effort as well as usage of resources in my, and occasionally others', work. Such skills aren't unique to software, or to any industry, and any system, be it virtual, mechanical or biological, can always stand to be reviewed for improvements. Normally a measure of an improvement will have a beneficial cost element somewhere along the line. Even if such an improvement can't be applied cost effectively at the time, it can be batched with other review findings for later application.

Edited June 26, 2020 by Ian J.

[pirouets]

45 minutes ago, Ian J. said:

The more you sell of something, the more profit can be made?

The challenge with this is its not that simple. As few thoughts as I wonder if point 1 is actually the key:

 

- a run of 500 models may fit a single pallet to ship, but a run of 750 needs 2 pallets. The problem is shipping works per pallet so the cost doubles for only a 50% increase. 

- Handling extra models when received costs more and at what point does it mean either more space or people are needed locally?

- At what point does the size of the run become speculative and end up with models not sold even though some people now do buy more models or a model they may not have bought.

 

All of the above on their own may not make a huge difference, but when you add the 3 together probably ruin the margin, and thus less money to reinvest in new models. Also if there is 1 thing I can be confident of, the manufacturers will know their numbers.

 

[PaulRhB]

7 hours ago, Ian J. said:

and that's the other systemic problem, 'us'. As long as we are willing to say 'it looks more like 'x' than a brick so I'll buy it' then the producers of the models will work down to that lowest level of spend,

That’s the problem, probably 90% won’t realise the error as long as the shape looks generally right. They won’t notice the cantrail is 2” above the door not below as they look at the general shape not the details. As that is 90% of the market there is no pressure relative to a few in the know cancelling and the manufacturer writes it off down to you can’t win. 
A flawed mechanism is more of a decider to most. 
Like I asked before, how many write up a polite detailed review of errors with photos and drawings to illustrate their points and send it to the manufacturer? Unfortunately I know some do but send such rude rants without evidence that they get binned as a waste of time. The passion needs dialling back a notch and you need to think how would I send this to my boss?

I’ve sent polite details to Heljan in the past and had no reply, it’s not great pr  but the subsequent revision did address some points so maybe someone did listen or it confirmed their own findings? I suspect they are wary of people then claiming to talk for them? 
I can’t be an expert on every model I like, I can detail exactly what’s incorrect on my G and HOm Harz 2-10-2’s but accept those compromises because of the unviable cost of totally different tools to correct a 2mm discrepancy on one version. So I do settle for ‘looks right’ on some models and as things like the mini 009 caricature below shows you can capture the essence with massive liberties in accuracy because of how the human brain works. 

 

I can only suggest the polite contact method and ‘hope’ that the increasing competition on duplicate models eases them all toward more accuracy where feasible.
It must be very difficult as seen by Cavalex to share your early CAD and try to get it bang on only to be stomped by a competitor taking advantage of the early notice to react and see the feedback you got. So you hold on until a late stage where no one can gazump you and then errors are too expensive to correct for the 5-10% who notice. 
Catch 22 really. 

[melmoth]

Is the, er, baby Deltic going to be a Racehorse or a Regiment? Shetland Pony or King's Own Light Midget Infantry?

[Chris M]

1 hour ago, Ian J. said:

 

 

3. I am not, and have never claimed to be, inside the industry (I think I may have mentioned that in my original post, but can't see it right now). I am not even in any production industry. However, I do in my job (database and application design) have to review and check on a regular basis and look for efficiencies in effort as well as usage of resources in my, and occasionally others', work. Such skills aren't unique to software, or to any industry, and any system, be it virtual, mechanical or biological, can always stand to be reviewed for improvements. Normally a measure of an improvement will have a beneficial cost element somewhere along the line. Even if such an improvement can't be applied cost effectively at the time, it can be batched with other review findings for later application.

I spent 30 odd years managing IT projects and I don't recall ever seeing an application produced without errors . That's why there is always a Beta test. I do recall pressure to go live even though there were still a few minor glitches unresolved, that's why there is always a version 1.1, 1.2 etc. So in my experience the IT world is no better than the model railway world when it comes to being error free.

 

One very,very minor glitch in a project I was managing caused a front page headline in the Financial Times - that made for an interesting day at work! The thing is the correct people had checked out what the app was doing and passed it off as correct. 

[YesTor]

On 21/06/2020 at 18:04, Enterprisingwestern said:

 

Stop belittling the achievements and skill levels of other modellers, being unable to produce Downes like buildings, Hewitt style signals and such doesn't mean you have to !ower your expectations when it comes to rolling stock

That modeller with his badly built layout is probably working to the limit of his capabilities and would expect manufacturers to do likewise.

 

Mike

 

I have to agree with the above.  I mean, if I were to enjoy treating myself to a meal in a fancy restaurant at the weekend and pay the full asking price on the menu, should I be expected to lower my standards and accept being served an overcooked main, simply because I live on microwave meals Monday through Friday? 

 

Furthermore, experiencing the relative luxury of a restaurant-cooked meal might actually inspire me to improve my own skills in the kitchen.  Sometimes we need a symbol of relative 'perfection' in order to inspire/progress.  Similar scenario with modelling really.

 

Best

Al

 

Edited June 26, 2020 by YesTor

[MJI]

34 minutes ago, Chris M said:

I spent 30 odd years managing IT projects and I don't recall ever seeing an application produced without errors . That's why there is always a Beta test. I do recall pressure to go live even though there were still a few minor glitches unresolved, that's why there is always a version 1.1, 1.2 etc. So in my experience the IT world is no better than the model railway world when it comes to being error free.

 

One very,very minor glitch in a project I was managing caused a front page headline in the Financial Times - that made for an interesting day at work! The thing is the correct people had checked out what the app was doing and passed it off as correct. 

 

I work in IT and over the last couple of years have refused to rush work through UNLESS it was very simple (eg captions or forgotten callbacks),

 

Since then our comeback on updates has almost disappeared.

[Andy Reichert]

I'm sad that most mass produced models are produced in plastic. Metal is generally far stronger for tiny details. And I think far more long lasting, repairable and/or modifiable. Of course I realise that would make most fully RTR metal models of similar detail level to modern plastic injection moulding phenomenally more expensive that current plastic models. But that's life and fortunately there are kits that can reduce the pricing considerably by using one's own labour.

 

My EFE tube stock has awful and huge plastic moulded end hand rails. I have no idea whether that was a mistake or a cost saving issue. But replacing them with realistic wire parts risks not being able to match the rest of the quite good overall paint finish, or re painting and decalling the entire set of cars. So, given my present workload,  the decision has not yet been made.

 

Andy

[melmoth]

4 hours ago, Ian J. said:

 

1. The more you sell of something, the more profit can be made? I know it's anecdotal, but I often read here of people buying one or more additional examples of a model because it really works for them, and fewer when it's not quite right (examples that spring to mind in the more category being the SJW24, the Hornby Peckett, and in the less category, the original Heljan 47).

 

2. My workings being if it takes an hour to run through the basic checks (and it really shouldn't take much longer than that) every couple of weeks, at say a professional like rate of £50 per hour, that's maybe £75 a week, give or take and remembering that employer's tax and NI have to be taken into account. £75 a week for a set of basic checks shouldn't add even 1.0% to a budget of £100k.

 

3. I am not, and have never claimed to be, inside the industry (I think I may have mentioned that in my original post, but can't see it right now). I am not even in any production industry. However, I do in my job (database and application design) have to review and check on a regular basis and look for efficiencies in effort as well as usage of resources in my, and occasionally others', work. Such skills aren't unique to software, or to any industry, and any system, be it virtual, mechanical or biological, can always stand to be reviewed for improvements. Normally a measure of an improvement will have a beneficial cost element somewhere along the line. Even if such an improvement can't be applied cost effectively at the time, it can be batched with other review findings for later application.

 

1. I appreciate what you're saying, but you've not actually addressed the point I raised. To reverse your anecdote, if the Hornby Peckett hadn't been as good as it was, and had therefore not sold as well, it could still have potentially made more profit that it actually did if the reason for it not being as good a model was the reduced costs involved. This sort of thing gives rise to circular reasoning either pro or anti if no one actually has a grasp of the actual figures involved. Let's face it, we're assuming that Hornby are making a profit on the Peckett...

 

As a similarly anecdotal response (because I can't verify all the facts), I'll offer you the example of New Order's 12" Single release of 'Blue Monday'. At the time, and quite possibly still, the biggest selling 12" vinyl single in UK recording history. Because it was apparently issued to promote the album it came from and wasn't going to be made available in a 7" edit, Tony Wilson (head of Factory Records) pushed the boat out on a complicated and/or expensive picture sleeve. I do not know the exact details, but when all the costs were calculated, Factory Records made a loss on each copy sold. Like with most things, there's probably more to this than a simple anecdote can cover, but it certainly offers an example of where selling more does not necessarily increase profit. There is apparently a similar story attached to the selling of the distribution rights of the file "My Big Fat Greek Wedding".

 

2. You've not actually 'shown your workings' here. My highlights show the possible irony of inexactitude in a discussion about improving modelling standards.

 

3. I think @Chris M 's response above is better and more relevant than anything I could offer. But it seems to me that what you are essentially arguing is that if we throw enough time and money and expertise at the issue of pre-production errors in RTR models, we can eliminate the errors, sell/buy more models, and create a virtuous circle where the increasing profitability of each model feeds back in more space/time/money/skill to develop the next model. What people with some degree of personal understanding of one or other area of the whole process of bringing a product to market appear to be saying is that it's not quite as simple as that.

[Bernard Lamb]

11 hours ago, Ian J. said:

 

 

What irks me is that the basic checks that I outlined previously wouldn't cost much, and with the reduction (though not eradication) of errors that would follow, could lead to better sales, more profit and better reputation. With my business head on, I'd call it a 'no brainer'.

But you are still a denier of human nature. have you ever seen a study of how much different procedures would cost? I have and the figures can be quite alarming to the accounts people.

I would have thought that the nuclear industry had higher standards than most, but when I had a phone call from site about a problem that had arisen, because a building had not been built to the drawing, no body seemed to be surprised. They all just got on with the job of working round things to sort out the problem. In a very tight time scale I might add.

You might well have a business head but unfortunately you do not have an engineers head and we are discussing an engineering problem. 

Bernard

[The Johnster]

55 minutes ago, Andy Reichert said:

I'm sad that most mass produced models are produced in plastic. Metal is generally far stronger for tiny details. And I think far more long lasting, repairable and/or modifiable. Of course I realise that would make most fully RTR metal models of similar detail level to modern plastic injection moulding phenomenally more expensive that current plastic models. But that's life and fortunately there are kits that can reduce the pricing considerably by using one's own labour.

Not sure this is completely true, though I 'get' where you're coming from.  I am of a generation that grew up at a time when you had to back one of two incompatible RTR horses, one called Triang and the other Hornby Dublo.  HD meant die cast mazak, which didn't always end well, and of course Triang were wall to wall plastic, and pretty good at it.  Many Triang and some HD models are still running after 70 years of use; both plastic and die cast can last years and have proved equally repairable.  

 

Increasing numbers of modern RTR locos are die cast to a very high standard, and seem not prohibitively more expensive than plastic toolings.  The big difference between now and the Triang/Hornby Dublo days, apart from closer to scale models, is the level of separate rather than moulded detail, some of which is a bit delicate.  It seems, again, not to matter if the separate detail is plastic or metal, as the structural strength of it depends on the method of fixing, which is usually glue.  When you look at the problem from a production engineer's perspective, you can see that a model that requires an assembly worker to solder individual tiny pieces to a metal loco body is possible, but the market will not want to bear the cost.  

 

Cast detail on metal bodies is very robust, but not usually very realistic.  Moulded detail is more realistic as a rule, but not usually particularly robust!

 

The big advantage, to my mind, of metal die cast bodies is weight.  All weight is a good on a model loco as it increases tractive effort, holds the model firmly to the track and thus improves pickup, and acts as a flywheel to smoothen the running and get the loco over the odd dead spot/bit of crud.  Plastic models are always, without exception, too light IMO, no humility about it, and are always improved with extra ballast.

 

Seems that the choice of metal or plastic, or a combination, depends on the production engineer's decision, and each material or mix of materials will be chosen according to the needs and market expectations of that individual model.  Supply at a price we will accept, even if we moan about it, depends on the Chinese production and assembly facilities being able to produce the model to the required standard within the limits of a cost umbrella, on time.  They are pretty good at this.

 

But, my main grouses about RTR are still, after more than 60 years in the game, that they are too lightweight, do not haul realistic loads (not that this is an issue on my small BLT), and have gear ratios too low for good slow running or smooth controllable starts or stops.  38:1 is common nowadays, a big improvement over the 20:1 RTR of my childhood, but I would like to see 50:1 as a standard for passenger work and 60:1 for freight or shunting, with flywheels wherever possible but especially on smaller locos.  Flywheels make very good ballast weights.

[YesTor]

4 hours ago, melmoth said:

As a similarly anecdotal response (because I can't verify all the facts), I'll offer you the example of New Order's 12" Single release of 'Blue Monday'. At the time, and quite possibly still, the biggest selling 12" vinyl single in UK recording history. Because it was apparently issued to promote the album it came from and wasn't going to be made available in a 7" edit, Tony Wilson (head of Factory Records) pushed the boat out on a complicated and/or expensive picture sleeve. I do not know the exact details, but when all the costs were calculated, Factory Records made a loss on each copy sold. Like with most things, there's probably more to this than a simple anecdote can cover, but it certainly offers an example of where selling more does not necessarily increase profit.

 

That's correct, the original sleeve had a unique die-cut design which was expensive to manufacture, so they simply omitted the cut-out design in favour of a standard printed picture sleeve for later pressings.  An awesome 12" either way...    

 

Al

[ardbealach]

 

I can understand the desire to achieve accuracy in the products from the model railway manufacturers, and certainly in the past there have been bloopers.  Designers will often take the decision of what the final product looks like, rather than what is correct.

 

Away back in 1953 Lesney - of Matchbox cars fame - produced a Coronation Coach with eight horsemen as a souvenir for the Coronation of HM The Queen.    When the coach arrived in the shops the eight horses should have had a total of  8 x 4 legs = 32 legs. 

 

However, it was produced with 16 legs, the inner legs on all the horses being omitted.  Yet it looked correct.  The designers at the time took the decision that with 32 legs it looked wrong.  And I bet very few even noticed the legless horses. (AM)  

[Andy Reichert]

Actually I have to heartily agree with the OP on this situation. See the "Hunt Couplers" topic under the "Modelling Questions" section if RM WEB. 

 

Andy

[Graham Heather]

On 26/06/2020 at 21:47, TEAMYAKIMA said:

Having read all 5 pages of this thread this morning and heard about all the angst people are suffering from these (relatively) minor issues I realise how lucky I am to have built and exhibited a foreign layout - a VERY foreign layout !

 

In fact in my case it's sooo foreign that 99.9% of exhibition viewers probably don't know the prototypes and certainly don't know the models - they just look at the scene and enjoy it (or not) for what it is  - a  model railway which gives the viewer a flavour of the real thing.

 

Perhaps it is true that knowing too much about a subject can be a bad thing.

Probably the biggest reason I chose not to model my local railway scene.

Really too many people that have opinions have trains from their local area.

Modelling UP downunder.

Graham.