97xx

Sorry about that, and agree you certainly don't want to sign up. This might work, although you'll have to download it to view it, which I know won't feel ideal either! https://app.box.com/s/gru6hjongqomrrprhr3bs8w7oo7bn7s6

Note sure how well this link will work - perhaps someone can confirm? This is the underside view of mine. I am so disappointed as this is possibly the best model I have come across in nearly all other respects. https://www.dropbox.com/scl/fi/mxzao38uconq3xgi2s9iv/Wobble.m4v?rlkey=y9jhhb283bu89tsg6ht32gh7p&dl=0

Afraid mine will need to go back - even after a 30min running in in each direction, it has an increasingly pronounced wobble. Taking my own advice and videoing from above, the front wheelset is hopelessly out of true on both wheels. The middle looks poor too, but as that is sprung it's harder to tell by how much. Very slow running is showing a very tight spot - either as a result of out-of-true , or something else. Another potentially really fine model ruined by the usual poor quality manufacture from the usual place. I also have a damaged front coupling, which judging by the clamshell has been done BY the packaging - the packing piece bears down on the NEM pocket and has broken the tiny flexible section between triangular pocket location and remainder.

Sadly this sort of issue afflicts most/many models using any sort of plastic wheel centre/insulator - the modulus of the plastic just always seems to cause the wheel, after being pressed on 'true' , to walk back to a relaxed state where it is no longer. (read on before commenting...) I've had this endlessly with Gibson wheels (plastic centre pressed onto steel axle - even done them in milling machine press with jig!), but far less so on Romfords (latter where the tyre is insulated from the centre which is all metal and there's a good shoulder to axle facing). Nonetheless, we are all paying an expert manufacturer to do this job properly, and 'work out' the idiosyncrasies of manufacture. To determine whether it's short WB effect or true irregularity, I'd suggest upturning the loco in a suitable cradle and applying power to a set of tyres - see which wheelsets wobble and which do not. (this can help spot tight spots too). On my 15xx, it would appear the driven wheelset is as true as I'd reasonably expect and the other two have varying degrees of out of parallel. In practice I think I can live with it, but will see it run in well first - simply as any tendency to cog (down to tightspots I'm meaning, not the motor here which seems OK) will exacerbate poor visual behaviour.

Wires are slightly crossed it seems - Car No. 54 is approaching £600k in total.

I do wonder if we are getting "£300k over budget" incorrectly added to the £500k recently reported - i.e. perhaps the budget (if there ever was one) was £2-300k and it's gone way over, but I would doubt £800k is really likely. However, just guessing!

£75k came from from the Department for Transport’s Heritage and Community Rail Tourism Innovation Competition, principally dedicated to provision of wheelchair access. Surprised to see a figure of £800k - where did this number come from? I recall reading nearer £520k so far.

How old is it, and did you buy it from a retailer?

I seem to have found myself with an Airfix 14xx - a complete but very dodgy runner, a High Level 14xx chassis kit I intended for the prior, and now a K's 14xx complete kit. I tried various fixes on the Airfix as I'm a bit attached to it, being the first 'detailed' RTR I ever bought, but concluded a decent chassis was worth the effort. So, do I stick to Plan A - HL into Airfix, and try to build the K's but with probably improved wheels/motor? Or Plan B - HL into Airfix, and another HL into K's? Umm...?

Examples of suitable caps at 0.1µF, 100V (a) TDK FG28X7S2A104KRT00 (b) VISHAY K104K15X7RH53H5G (c) TDK FA14X7R2A104KNU00 (d) TDK FA16X8R2A104KNU06 (e) MURATA RDER72A104K1P1H03B What you are looking for is: (1) Ceramic or multi-layer ceramic capacitor (2) Minimum 100V DC working (usually tested to 250% of that) (3) Around 0.1µF or 100nF (4) Small enough to fit (which almost any cap meeting this spec will be, but you have your model and can see the available space) (5) A brand like TDK, Murata, Kemet, Kyocera, Vishay, Samsung And per (2) yes anything over 100V is fine, but again see (4). They have no polarity for the avoidance of doubt, so you can fit either way. Another excellent supplier is Mouser (.co.uk). RS Components also but quite limited selection these days. If you wish to use a well known auction site, search for "0.1µF 100nF 100V ceramic capacitor" and then see if you can spot any made by a decent manufacturer. (If you see polypropylene ones show up - electrically suitable - you will see the substantial size difference). Note that you could choose polyester or plastic film (e.g. polypropylene), so long as rating is the same. The reason we tend to use ceramic is that they're small - their undesirable characteristics are of no concern in this application. If anyone is really stuck, PM me. NOTE: you fit at your own risk (I need to add that!) and also be aware a correct cap will not fix a fault which exists in model or controller or layout. EDIT: do NOT buy TANTALUM, or ELECTROLYTIC - these are polarised and may explode if used.

We're probably going to be slapped for being so far off topic, but it's all interesting and relevant to the quality of electronics. Given LiPo represents about as powerful and stable a power supply as you could ever want I assume the caps are present to provide transient stability to the controller (back to my point about logic and controllers needing to be protected from transients to avoid hiccups/resetting) given the huge surge currents that can be drawn from the LiPos, suddenly pulling down the voltage. I'd love to know what make the caps were. But, yes, a very serious failure of integrity for an electrolytic cap to develop a dead short under rest. I only witness this with 60 year old ones I come across in old radios, and rarely a dead short. More common is open circuit in china made ones in modern kit, like this one in small power supply which split and failed open circuit - you can see a bulge in the top indicating overpressure and there are lines etched into the end to facilitate 'pressure release' which have opened. This make is CapXon - bottom of food chain. And what effect did this have - huge broadcast of RF noise plus a temp. of 108 degC recorded in unit, under zero load, which is dangerous.

A good question, but no.

Exactly. This is an expensive item, and there appears to be a significant number with faults, and/or serious safety issues. Hornby possibly do care, but the actual producers do not. A parallel here is that if you want a multimeter you can buy a Won Hung Lo branded one for £20 that will be fine on ELV (e.g. 12V model railway work) but be crap quality and more seriously will lack any adequate protection and may quite literally blow your hand off if you have a mishap with it on 240V mains. But there is no possible come back. Yes, it was UL and CE labelled, but that's just screen print on the case. Contrast with a Fluke multimeter. Even though some of their cheaper ones ARE made in china, they apply rigorous quality control and having had them open I can attest that they're designed and made well, and consistently. And if it did go bang Fluke would have a massive problem. This APT is a Fluke-priced item with Won Hung Lo build. Hornby are simply NOT applying adequate QC, and sadly as said above if no-one sends this rubbish back, nothing will change. And why would/should it? If Hornby actually have any quality specs with their manufacturers, us not pushing defective products back leaves Hornby believing that all is well, and the manufacturer doing what they do best - cheating and getting away with it. Yes, it's not unique to china, but it is absolutely endemic there. I work with electronics goods and the truly horrifying examples roll in day after day.

No, even worse really as it actually looked like a BS1363 plug. However, I know what you mean...

Omission does not necessarily mean wrong voltage, but they are by inspection the lowest of the Won Hung Lo type of component. As was the one in the picture I posted. Might even be sold as 100V but may not actually be able to withstand it as print is cheap. Ideally, if you can buy a named and respected brand like Murata, Vishay, Kemet, TDK, Kyocera then it is highly likely they will be excellent, unless they're fakes. Buying from somewhere like RS Components, Farnell, Rapid, Mouser gives you greatest confidence. In case anyone thinks I'm being unfairly critical of chinese products, I would refer them to something I looked at recently which had a moulded 3-pin UK plug on the flex, with a removable fuseholder in the back as you'd expect, except that (a) the fuse was not even in circuit, and (b) the fuse was a fake of a known brand with no sand in it, and (c) the earth wasn't connected. Quite literally it has been deliberately designed such that it might kill you, in order to save fractions of a renminbi.

Yes 104 is the value in picoFarads, which are million billionths of a Farad (being far too large for most use). 104 means 10 and 4 zeros = 10 0000 = 100,000pF = 100nF = 0.1µF. What is MISSING is the VOLTAGE rating, for which I would select 100V. This is higher than one might select if you were buying quality components but as eBay may be a likely supply for most people, give yourself some margin. After all, the obvious source of crap components has given rise to what appears to be a fire risk here... So, look for a 0.1µF (or 100nF) 100V ceramic capacitor. Any general purpose one will do here as we are not concerned with temperature stability.

Not sure my RF ramblings are that interesting but here is the difference between an unsuppressed and suppressed motor. This is the radio frequency interference sent back into the rails under power. The actual levels don't vary that much with motor speed. The screen goes from a frequency of 1MHz on the left (bang in the middle of Medium Wave) to 100MHz on the right (bang in the middle of FM). The Y axis is the magnitude; The green trace is unsuppressed, the yellow suppressed. The 'difference' (although it varies by frequency) is at least 25dBm which is approximately a factor of 20. BTW, the red trace is the background hash generated by the plethora of little 'wall wart' power supplies and LED bulbs...

I'd expect it to be a pretty similar shape, but superimposed as you say on the AC DCC rails.

Using an RF 'sniffer; I find about a 50% reduction in radiated RF noise with an appropriate suppressor, but as some have said no-one really bothers with or about it these days. With the virtual death of MW listening this has become a background issue, although move a typical cheapo wall wart PSU near to a MW radio and you'll hear lots of noise... If you have a small MW radio, tune if off station and listen to your train.

Agree Paul, and the example I showed is a scratchbuilt model of mine from 40 years ago with no suppressor. Faulty/dangerous components are the issue, but just thought I'd make the point about noise suppression. I imagine/hope that DCC encoders have proper filtering built in.

Agree fully, in either case a cheap component will fail. Just with DCC some not-so-cheap component (decoder, sound system) may well go pop too.

Here is a very quick and dirty look at the voltage on the rails with a motor at 50% speed on a Gaugemaster controller. EDIT: at 65% with 9V measured. Note that this isn't detecting radio frequency interference, but noise directly on the actual supply. Without suppression you can see peak-to-peak voltages nearing 100V. With suppression they're closer to 50V. This is not in any way a perfect test, as you need faster responding kit to capture the true peaks, but you see what I mean. I'll see if I can do some RF tests, but that is secondary to the noise put back into the supply for our debate here.

The suppression caps do three things: (a) Suppress potential radio frequency noise which might affect nearby receivers (b) May slightly improve commutator life as they reduce some of the sparking (c) Prevent any local microcontrollers getting spuriously reset due to high frequency electrical spikes on the power rails. Typically on small DC motors, they are rated in the 1 to 100nF range (0.001 to 0.1µF), and should have a DC voltage rating of 10x the operating voltage - because the 'back EMF' can be very high voltage. (This is why DCC introduces a higher requirement). Some will say they are fitted simply to meet IET/EU/CE/whatever 'rules' but see point (c)... What the capacitors actually do is literally short out high frequency voltage - hence if they're incorrectly made/rated they will fail. The normal supply to the motor is DC, which is left untouched by the caps, but every time the brushes break one and make one new commutator connection a high voltage spike is created as the magnetic field collapses. Personally I would not worry too much on DC, but would on DCC as that's another slew of low quality far eastern electronics on board which might be tripped/reset at best or blown up at worst.. To be really fussy, given the can of the motor may not be connected to the chassis (and so acts as an aerial without ground return) the ideal arrangement is three similar caps: one from each motor terminal to can, and one across the terminals. EDIT: for those interested, this is why a diode is ALWAYS put across a DC relay coil in any sort of sensitive circuit to short the brief but high back EMF when the coil de-energises. A 12V coil will easily develop 500-1000V for literally milliseconds when de-energised and if no diode is present this spike will destroy controlling electronics.

These must be the World's worst ceramic capacitors. I can't imagine they're less than 100V rated and so what sort of motor is producing enough back EMF and energy to destroy one is new to me. EDIT: probably 50V rated in which case they're unsuitable for this duty. Q: are the cap failures all on DCC?

I feel a bit underwhelmed at the H-D announcements. In many ways I would like to see models produced that were NOT H-D mainstays, else I'm buying what I've already got... I'll duck for cover now.