aleopardstail

put the controller nearer the servo, then talk to the controller over something like RS-485 or something else that can handle the noise that and try to limit the noise in the first place

Incidentally if anyone wants a hand using one of these little beasties to date I have example code to: - driver servos via the PCA9685 (creates a servo controller class that can be used directly but also an individual servo class that allows for things like custom min, max and central positioning) - drive a SSD1306 OLED display, goes a bit beyond the adafruit drivers with clipping rectangles, drawing logic modes etc (I was bored) - drive a custom dot matrix display made from Max7219 ICs, treats these as a display, allows custom fonts etc - DFRobot Player Mini MP3 player - audio "tone" control using PWM for multiple voices I've also used them to talk to 74HC165 & 74HC595 shift registers and similar talking to EEPROM chips shouldn't be hard (or indeed anything on the I2C bus) not yet tested the nRF24L01+ radios out, they shouldn't be too hard to work with and will shortly be working on a colour TFT screen and touch panel driver, I've already got for a different MCU a touch screen "window" interface library thats a mess but works

I have been toying with the Pi Pico since not long after they came out, alongside other MCU I wanted something with a bit more ram, the duel core stuff is easy to use - not played with PIO as yet but plan to. they can easily drive servos, use a PCA9685, this will take the 3.3v input on the I2C bus, then accept a 5v driver voltage and easily drive 16 servos - driving this from the Pico means creating a little bit of a driver but its not hard to do - have an example of my desk at work driving a pair of them. they are much better for driving screens than the Arduinos as they can have a proper frame buffer inside and can easily let the second core manage actually updating things (again you need to write your own drivers but again its not that hard, its just shifting bits & bytes about) key though is the things are dirt cheap, project I have some part on order for is a handset for a DCC++ system to use a Pico, operating a rotary encoder a few buttons and a 320x240 colour touch screen - then talking over the serial interface. these things are very powerful, but cheap enough to dedicate one to managing a keyboard if needed (e.g. a pico is cheaper than a dedicated keypad controller IC..) they do PWM very nicely for variable brightness of LEDs, and the low current output is managed easily by using transistors to control the actual LEDs anyway. so far the only drawback has been the way they act as a slave in SPI mode meaning its hard to pretend to be something else. but as a controller for point servos, signal servos, colour light drivers, serail interface hubs, screen drivers etc they are very hard to beat if you don't need wifi or bluetooth. the Windows toolchain though is dire, I finally got a Pi400 where its a lot easier

Always thought Lima N was more "scale optional", still remember the class 55 and its habit of getting stuck under bridges

Hi, many thanks and much to think about. "Da plan" was to use a 10'x2' sheft in the backroom/office/mess/junkyard above the main work desk (when either get built), hence the width restriction. It started out as something with a train turntable but not sure I could get it working right so wanted a more conventional out & back, looping round and under then back. There is an alternative location which is a cellar, where the same basic station can be managed, through to a storage yard of run round roads (no loop), in theory the station gets about a foot shorter but it wasn't using the full length anyway so managable, can then have two fixed bits and a removable curve between them closer to 36" which should work (having run 48" curves in N in a previous house and found it hard to tell they were actually on a curve I suspect this will work) The concept was "N", I wanted to make my own pointwork, I did consider FiiNetrax, but figured if I was going to the effort anyway 2mmFS would likely look better and I gather more modern stuff isn't too hard to convert. my last attempt at close coupling was a rake of N gauge Lima Mk1 coaches, run as a rake, sadly don't have them but was done by taking the front bit off the couplers to leave the spring and the rubbing plate, then a wire loop round them - bit of concertina folded black paper for a connection, avoid looking too closely and it worked the terrible paint job was a distraction... Will soon have an "N" test loop for checking things work while I plan some smaller test track and start measuring the cellar

Ok, experimentation it is, with possibly some other method of looping around needed, still experiments can't hurt. Yes I did think close coupling would be out of the question, I can live with that if it means I actually get to play trains for the first time in over a decade and a half. limited to twin axle bogies for locomotives would also me manageable, the only ones I have so far are a pair of ancient Lima Class 86 - which perhaps indicates a level of flexibility where accuracy is concerned here. experiments and mad science ahead

Hi, new to 2mmFS, I had been planning an N gauge layout, but desired to assemble my own point work for a range of reasons, it seemed that if going that far I may as well at least consider a finescale layout. The layout calls for a reverse loop but is 24" deep - it could be extended slightly but not vastly, I'm aware that curves of 10" & 11" (ish) are visually not very good so such will mostly be hidden (or entirely hidden) at the ends, my question is how practical are such for 2mm finescale. (essentially roughly 2nd radius set track type curves) the rolling stock planned is all 1970s/80s/90s diesel/electric locomotives (mostly twin and some triple axle bogie stock) plus bogie coaches so nothing with an excessive wheelbase and the curves will not be taken at any significant speed - note there will be zero pointwork on the curves - its just getting a twin parallel track to do a 180 degree turn on a 24" deep board (its going on a shelf). its possible a gradual gradient will also feature on the curves but this could be designed out if its a complete "ahahahhhhahhaaaa no" point also aware that such curves will rule out any long wheelbase stuff (e.g. Class 40 are not expected to feature), coaches will be largely Mk1, possibly some Mk2 stock, not sure Mk3 would look right without the ability to run more than the planned four/five coaches and that there is essentially zero chance for any steam era locomotives to get around the corners - this not being a problem for me as I'm not planning on running any I will be trying some experiments with this but if wiser heads have already tried it and know it doesn't work I can try a bit of a re-think. many thanks