Model Design Errors

[Matt C]

3 hours ago, Talltim said:

There is a difference between detail(s) and basic shape. I'd be much happier with a model that lacked some details that were add-ons, but had the basic shape right than vice versa. Details can be added (or removed) to suit your requirements/clumsiness level. The basic shape is far harder to change.

 

 

My point exactly. Get the basics right and make sure it runs properly. 

[Ian J.]

7 hours ago, Mattc6911 said:

 

 

I think we will have to agree to disagree. The thread is about design errors. My arguement is that by pushing the manufacturers into ever more complex and detailed 'perfect' models at the price those who want 'perfection' are prepared to pay, something has to give, which is quality control. Out of a run of 2000 models (number picked from the cloud) how many want that 'perfect' display cabinet model and how many want a good representation of the real thing, that looks right, has the right colour scheme and runs well straight out of the box ?  My own  thoughts ( with no factual evidence) is that the former group is smaller but VERY vocal in pushing manufacturers towards that 'Perfect' display case model to the detriment of those who want some that looks right and gets used. My suggestion about less detail isn't a return to Triang, its to make a good '  accurate model with less micro detail ( that falls off, breaks before you get it out of the box or disapears  on the first run out) that sells for a reasonable price to the majority of purchasers ( ball park pie in the sky figure say £150) and then sell an addon detail pack for (again pie in the sky figure) £100 that the end user who wants the display case model, can add themselves, or commission to be added if their skills aren't up to it.  Some manufacturers are all ready including some detail pack in with models, I'm saying take this to the next level. In Hornby's case they could do the basic accurate model run of 2000 and at the design stage commission Airfix to do the detail pack ?

 

My point is scale back on SOME of the detail, get the bloomin basics right, sell at a price the majority are happy with and then add detail for the smaller number who want 'display case' models via addon packs at extra cost.

 

Edit to Add, its almost a case of say Hornby, instead of using the railroad version as they do at the moment ( confused mishmash that no one really understands) make it the 'ready to run' standard version for all new models. . For the majority to purchase then do a 'Delux'  detailed version at the higher price range. Consign all the old tooling runs to a lower priced 'classics' range

 

I think I'm seeing your position now - with a fixed CAD budget, you're wanting the money to cover basic shape first, and not waste money in the CAD work with lots of little details? Yes, I would disagree, but only to a point. Basic shape is still the number one priority, but I just feel it shouldn't rule out decent detail as well. To not go the extra mile, is, to my mind, as false economy.

 

1 hour ago, The Johnster said:

It needs to look good and to work, but it doesn't need to be hyperdetailed.  I would be happier with a model that was as close to scale as 00 will allow in dimensions and shape but was not highly detailed than I would with an all singing all dancing model with inside working motion and posable reversing gear, and gauges in the cab that had needles that moved correctly, and working brakes that was not dimensionally correct and not shaped well.  I can work up the basic model to any degree I want, but the hyperdetailed model is still wrong, and irredeemably so; there is nothing that anyone can do to improve it.

 

Good running is highly important to me, and the better and smoother the slow running is the more I'll forgive in other design faults if the model captures the general 'look' and character of the prototype.  Finish is a major factor in this respect.  To take as an example the awful Triang 8750; no way would I give this house room.  But I saw a photo of one a few months ago that had been very skilfully weathered, in late BR black dirty condition, and was fooled; the incorrect coupling rod not noticed and the loco looking like a worn out 1960s 8750; it captured the 'look'.  However, there is nothing that anyone can do about the Gaiety 57xx; it's just out of proportion and too wrong!

 

A case in point - the Heljan 35 remains a firm favourite of mine, and has none of the extras models seem to get these days.

 

1 hour ago, Andy Reichert said:

 

Then the thread has the wrong title.

 

Andy

 

??

 

 

43 minutes ago, Andy Reichert said:

 

...If you really want "perfect" models, I suggest you have a go at designing and manufacturing one yourself. With of course, no later discovered errors.

 

Andy

 

That's still an unfair argument, and it's bad it's still fallen back on when issues with a model are found. I can't design and build a car, I can't design and build a smartphone, or a house, or who knows how many things. But does that mean I have no view on the quality of those things? Really, that argument needs to die.

 

[woodenhead]

1 hour ago, Chris M said:

A choice of class 45s!!! Us poor N gaugers don't have even have one available.

And being on topic, I got my Farish 46 out of it's box after 3 years in storage - went from excellent runner to stuck gear.

 

I had to remove the errant gear completely to get to loco running, at least it has all it's wheels one of my 40s went to 1-co-bo-1.

 

I think in the next few years there will be another set of N peaks but this time DCC enabled with a speaker.

[Andy Reichert]

1 hour ago, Ian J. said:

 

??

 

That's still an unfair argument, and it's bad it's still fallen back on when issues with a model are found. I can't design and build a car, I can't design and build a smartphone, or a house, or who knows how many things. But does that mean I have no view on the quality of those things? Really, that argument needs to die.

 

 

Re: "??"

 

It's an unfair argument to criticize relatively minor body dimension faults (which may be necessary for economic production) while personally deciding the absurdly narrow 16.5 gauge and almost doubly over scale fat wheel widths and flanges is OK by you and shouldn't be talked about. Your title should be:
"Other model design errors than the track and wheel standards elephant in the room"

 

As has been mentioned previously, these are mass produced Toys at volume prices.   If you want perfection and museum accuracy, then be prepared to pay thousands, not hundreds, for your items. Then express concern if they are not to your ordered specification.

 

Andy

 

 

[AY Mod]

Give it a rest Andy rather than attempting to propogate gauge wars. It's the hobby equivalent of trolling.

[Andy Reichert]

10 minutes ago, AY Mod said:

Give it a rest Andy rather than attempting to propogate gauge wars. It's the hobby equivalent of trolling.

 

Sorry.

 

Andy

[ColinK]

It looks near enough a Peak for me.  What’s important to me is that is runs well.

[Lazy]

I have designed various models for 3D printing (for my own use rather than for sale) over the years, so here are some of my own design bungles :
1) Rushing to beat a Shapeways price rise, I put my design file through a repair program prior to uploading. The repair program removed a few details from the model. With no time to correct the "correction", I submitted the design anyway, reasoning that I would manually restore the missing bits when I received the print.
2) Baulking at the apparent complexity of a bogie design, I decided to create a simplified model which I would detail manually after printing. Not only did the simplified bogie not look like anything, I forgot that bogie frames are "handed" and ended up with four identical frame prints rather than two opposite-handed pairs .
3) Having designed a wagon body and chassis separately, I mated them to form a one-piece design. Unfortunately, I got the chassis the wrong way round. It wasn't possible to manually correct the delicate details on the print so I left it as was. Not only that, I decided not to correct the design so that all subsequent models would be consistent (ie wrong) with the first .
4) Owing to a measurement error, a coach print ended up 0.8mm too wide. This might not seem like much (for 3.5mm scale) but it gave the model a somewhat pear-shaped profile . As the print was expensive, I decided to salvage it by manually sanding and filing away the excess and then reprofiling the original design (a more involved process that I might have wished).
5) Part-way through designing a model of a locomotive, I realised that it was actually starting to look like another (but very similar-looking) type instead. I saved a copy of the design as that other loco before trying to correct the original to represent the intended one.
 

[NittenDormer]

Is it still the case that Ordnance Survey include a deliberate error in each map so that unauthorised copies can be spotted? Maybe the same approach here. 

[The Johnster]

Persian carpet weavers traditionally include one misplaced stitch, because perfection is only for Allah, and blasphemous for humans to attempt. 
 

A map can never be perfect, as there wiil always be both natural and human caused changes between the survey and publication.  The map is static but the ground is dynamic.  A deliberate error known only to the makers could be used to detect unauthorised copies, though!

[PJT]

1 minute ago, The Johnster said:

A deliberate error known only to the makers could be used to detect unauthorised copies, though!

 

41 minutes ago, NittenDormer said:

Is it still the case that Ordnance Survey include a deliberate error in each map so that unauthorised copies can be spotted?

 

I don't know if that's the case with Ordnance Survey maps but a deliberate error certainly was (still is? Don't know) deliberately put in A-Z street maps as a copyright trap.  Every technical illustration I've ever drawn (e.g. exploded diagrams for car parts catalogues etc.) have included a copyright trap, too.

 

Pete T.

 

[jcm@gwr]

Certainly the A-to-Z London guide had a few roads that don't exist, 

or had the wrong name.

Plus, both my sat-navs had deliberate errors for the same reason! 

[Ian J.]

Deliberate errors as a copyright trap, in the case of the models we're talking about, would hopefully be in a minor detail or two, not in a significant and noticeable body shape.

 

To sum up my feelings for this thread so far, I feel that there are systematic and systemic issues that lead to CAD errors.

 

The systematic is a lack of checking vital dimensions and risky areas of a model against a carefully thought through list of checks before committing to tooling. These checks should probably be done against CAD, and against 3D prints. Painting of the 3D print should be possible as a rough guide, but colouring up a CAD model is certainly possible.

 

Systemically, I think we're seeing producers put too little budget aside for the CAD work, and are committing to tooling either knowing that errors exist, or due to the systematic failure above, being unaware of the errors.

 

But there is another issue that I feel is influencing these decisions: 'us', the buyers. 'We' seem to be somewhat too willing, in my opinion, to accept an error on the basis of 'it looks more like an 'x' than a brick, so it'll do in my book'. Obviously I don't feel that way.

 

I can only hope that there is more willingness to do checks in CAD and produce better models, and consequently gain more sales.

 

I don't like naming any producer where errors are concerned without mentioning that all the others also make mistakes. Now I will name producers who have done well, starting with three recent producers who I know I for one appreciate their approach to their models: Accurascale, Rails of Sheffield and Rapido. I mention them because I'm seeing them care about at least trying to get the basics right despite any other issues they might be dealing with. I think also Kernow and Hattons deserve mention. But I will also mention the four majors, Hornby, Bachmann, Heljan and Dapol again here, as for every model they've produced that has a flaw, they have produced some right crackers.

 

For all producers, I say we know you can do models well. I really hope you're willing to look at whatever issues are in your systems that cause errors to get in, and to try and weedle them early on and consequently get us to give you our money.

 

Edited June 23, 2020 by Ian J.
Add Rapido to recent producers list

[Huw Griffiths]

57 minutes ago, The Johnster said:

Persian carpet weavers traditionally include one misplaced stitch, because perfection is only for Allah, and blasphemous for humans to attempt. 
 

A map can never be perfect, as there wiil always be both natural and human caused changes between the survey and publication.  The map is static but the ground is dynamic.  A deliberate error known only to the makers could be used to detect unauthorised copies, though!

 

1 hour ago, NittenDormer said:

Is it still the case that Ordnance Survey include a deliberate error in each map so that unauthorised copies can be spotted? Maybe the same approach here. 

 

47 minutes ago, PJT said:

I don't know if that's the case with Ordnance Survey maps but a deliberate error certainly was (still is? Don't know) deliberately put in A-Z street maps as a copyright trap.  Every technical illustration I've ever drawn (e.g. exploded diagrams for car parts catalogues etc.) have included a copyright trap, too.

 

Pete T.

 

 

37 minutes ago, jcm@gwr said:

Certainly the A-to-Z London guide had a few roads that don't exist, 

or had the wrong name.

Plus, both my sat-navs had deliberate errors for the same reason! 

 

I'll have to remember this stuff the next time anyone challenges me about any of my "homebrew" models.

 

Well, it certainly sounds better than admitting that I'm rubbish at building decent models ... .

[PJT]

9 minutes ago, Ian J. said:

Deliberate errors as a copyright trap, in the case of the models we're talking about, would hopefully be in a minor detail or two, not in a significant and noticeable body shape

 

Yes, sorry, another off-topic digression, not exactly the first I've ever been involved in either.  Just to set the record straight, I would hope there wouldn't be any deliberate errors at all set as copyright traps in a detailed model, where fidelity is so important and infringing copyright would be hugely more difficult anyway than copying and republishing an image without permission.

 

Pete T.

 

[HonestTom]

I feel like there's only a limited amount you can say about why errors occur, other than "someone made a mistake." The more complex a model is, the more chance there is that something will go wrong. I think also that as our standards have raised, we've become less tolerant of errors that we once would have allowed to slide. I can't imagine anyone today producing a one-size-fits-all engine like the old Dapol Terrier and expecting to sell it as a premium model. Look at the criticism Hornby got over Design Clever, or Oxford over their Dean Goods - yet at one time, those models would have been considered state of the art.

 

Speaking as someone who deals with the public on a regular basis, it's my experience that when someone says "Why did [X] happen?" they usually mean "I know what happened, I just want an excuse to yell at someone."

Edited June 23, 2020 by HonestTom

[Steven B]

17 hours ago, Ian J. said:

But there is another issue that I feel is influencing these decisions: 'us', the buyers. 'We' seem to be somewhat too willing, in my opinion, to accept an error on the basis of 'it looks more like an 'x' than a brick, so it'll do in my book'. Obviously I don't feel that way.

 

But if you don't have the skills to correct the model, do you accept what's available, even if it's less than perfect, or do you do with out?

 

If there's a choice between two competing models is which is the lesser of two evils? I sold my old Farish class 50s some years ago as modelling NW England in the 1980s I had no need for one, let alone three. However, I recently bought the new Dapol Sir Edward Elgar. Given the choice between the Dapol and Farish version of the loco I'd still go for the Dapol one even though the headcode box is terrible.

 

 

Steven B.

[Ian J.]

59 minutes ago, Steven B said:

 

But if you don't have the skills to correct the model, do you accept what's available, even if it's less than perfect, or do you do with out?

 

If there's a choice between two competing models is which is the lesser of two evils? I sold my old Farish class 50s some years ago as modelling NW England in the 1980s I had no need for one, let alone three. However, I recently bought the new Dapol Sir Edward Elgar. Given the choice between the Dapol and Farish version of the loco I'd still go for the Dapol one even though the headcode box is terrible.

 

 

Steven B.

 

I (mostly) do without. Much to my disappointment, I'm at least doing without 4VEPs, 4TCs, 33/0s, and others.

 

The Hornby 700 handrail error I've managed to learn to live with, but I still notice it. Given enough time and space, I suppose I could try and correct it, but it's a faff and something I don't really want to tackle.

 

Anything sufficiently significant in body shape I cannot fix. The cab fronts of the 4VEP and 4TC are currently too much for me to do anything about. If I take the plunge later this year (funds allowing) and get a 3D resin printer, I may have a go at doing up a replacement cab front, and see if I can get it to fit, but I'm suspicious that there are body shape issues with the models that might mean a 'correct' cab front can't fit.

[Chris M]

It's down to each individual's perceptions. I would really have liked a large prairie for my N layout but the old Farish one was so far from right that I couldn't bring myself to buy one even though there were plenty of cheap ones around . The Farish class 50 wasn't close enough for me to buy one either (close but no cigar) but as mentioned above  the Dapol 50 is very good so far as I am concerned. The headcode box could be better but it doesn't bother me. 

 

When it comes to the accuracy of any model I feel the worst error is to leave it as it comes out the box. A little light weathering, removing the front coupling and adding the separately supplied bits make all the difference. This is far more important than the usually fairly minor errors on modern models.  I'm always surprised at how many layouts at exhibitions and in magazines run locos and stock in "straight out of the box" condition. Surely no serious modeller would run a loco in that condition.

 

 

[The Johnster]

I agree with that, even if it only a washover to tone down the ‘new’.  

[mdvle]

On 21/06/2020 at 05:50, Ian J. said:

Surely the 'manufacturers' know that a model with errors won't sell as well, while one that is good sells better?

 

Difficult question.

 

A lot depends on whether there is another model available that is better.

 

But I think the overall answer (to the annoyance of those who care) is that models with errors do sell well in general - otherwise we still wouldn't be complaining about it.  And so the lesson, such as it is, is from a financial perspective not to worry about errors.

 

On 21/06/2020 at 05:50, Ian J. said:

An example of a model continually selling - the SJW Class 24. An example of one not selling, the Hornby Mark 2E. There are countless examples of both, from the same 'manufacturers' so I can't list them all.

 

As far as my limited knowledge the Hornby Mk2e is the only currently available(ish) Mk2e model - as such is the lack of sales (if true) because of the problems with the model, or because there isn't much demand for a Mk2e?

 

The SJW Class 24 is an interesting choice to use as an example - we really have no idea how it sells given it is direct sale only (thus no way to guess at production quantities - but more importantly they have produced exactly 1 model (with variations) in 5 years.  How many people would be happy if Hornby/Heljan/etc only released 2 models a decade?

 

On 21/06/2020 at 05:50, Ian J. said:

Does it really just come down to setting a budget that's too low, running out of money, and the 'manufacturer' committing to tooling on a design with known flaws, just to save 'pennies' compared to getting it right and making more sales?

 

Some other possible factors - the need to have something to make for a booked factory production slot, the need to have something made so there is something to sell in x quarter of year y so the company has income/cash flow to pay the bills.

 

Unless you are working as a manufacturer as a hobby (aka have full time employment elsewhere) and thus don't have the time/financial constraints of a full time business you can't afford to hold up a project for perfection.

 

Which isn't to say that I find some of the issues acceptable, but rather until someone tries to do this as a business it is easy to sit on the sidelines and say it shouldn't happen.

 

 

[mdvle]

5 hours ago, Steven B said:

But if you don't have the skills to correct the model, do you accept what's available, even if it's less than perfect, or do you do with out?

 

Personally, I do without, but I also accept that most simply do not care so any loss of sales to me is irrelevant to the manufacturer.

 

[Tiptonian]

On 22/06/2020 at 21:15, The Johnster said:

  A deliberate error known only to the makers could be used to detect unauthorised copies, though!

 

I would have thought a deliberate error to avoid copies would not achieve anything. Did "Great British Locomotives" have the original manufacturers' blessing? 

[Geep7]

18 hours ago, Ian J. said:

The cab fronts of the 4VEP and 4TC are currently too much for me to do anything about.

I'll agree about the 4-Vep. The cab fronts are horrible, and i've modified the corridor connection door to be closer to the front, but it still looks wrong. I'll probably eventually replace the front end with an MJT example, or go for a hybrid. With regard to the 4-TC, I actually think Bachmann didn't do too bad a job on it, at least it looks to be in proportion. I'm not sure how much better they could have done it without making the model too fragile....

[MJI]

I find often that Triang models are surprisingly accurate just a bit crude.

 

With some work they are very good.

 

If Triang could get the basic shape right in the 60s and 70s, why not newer models.

 

Hornby 25 - looks right

Hornby Mark 2 - looks right