Twitchy servos

DougN
Posts: 1250
Joined: Wed Sep 01, 2010 9:57 am

Re: Twitchy servos

Postby DougN » Sat Feb 02, 2013 10:09 am

Reading through the above I too have the same trouble with twitchy servos. I am using Brunel models circuit. ($6 AUD for each circuit) I have a suspicision that the power supply that I used is causing the problem. The power supply is a 5V DC wall wart/ transformer type. I will speak to John tomorrow to see what his recommendation is (well he did design the circuit) It could be also a problem with the soldering or I have got a component wrong! Any how it is interesting to experiment!

Oh forgot to mention that I am trying to get this resolved this week for the next local meeting and also Use Mark Tatlow's mounting method.
Doug
Still not doing enough modelling

Mark Tatlow
Posts: 897
Joined: Tue Dec 02, 2008 11:24 pm

Re: Twitchy servos

Postby Mark Tatlow » Sat Feb 02, 2013 11:01 am

grovenor-2685 wrote:A set of those capacitors you normally cut off locos when converting to DCC should do the trick.
Refer http://www.pololu.com/docs/0J15/9

from Rapid electronics http://www.rapidonline.com/Electronic-Components/100nf-2-5mm-Y5v-Dielect-Ceramic-Capacitor-08-0275
Regards
Keith


Just to say thanks Keith, I will make a purchase over the weekend and let you know how it happens.

I seem to have a further problem now of killed Servo4's on second operation. I have put something on the MERG forum about that.

Life would be no fun if it were easy............
Mark Tatlow

DougN
Posts: 1250
Joined: Wed Sep 01, 2010 9:57 am

Re: Twitchy servos

Postby DougN » Sun Feb 03, 2013 7:02 am

Thanks to Mark and every one else I did get in touch with John of Brunel models and we did diagnose that the problem was the power supply. SO the solution was to use a pack of 4 batterys (I will change to rechargibles eventually) and use momontary on switches. So a big thank you to Mark for his servo mounting ideas and here is the result.
signal & servo.jpg


It has a few refinement that John and I will work on... at the meeting I was hoping to have it ready for! We are thinking of adding signal bounce (yes I know signalman weren't supposed to drop the signals back!) but a little would be nice. The intention, when I get to a layout is to have each signal slightly different in the same way reality was!

Thanks again Mark and to Hobbyking.com for the servos (good value and good service!)
You do not have the required permissions to view the files attached to this post.
Doug
Still not doing enough modelling

User avatar
Jol Wilkinson
Posts: 1111
Joined: Mon Jul 21, 2008 7:39 pm

Re: Twitchy servos

Postby Jol Wilkinson » Sun Feb 03, 2013 5:29 pm

The promised update on my work to overcome servo twitch on London Road, but let me firstly redefine what that is.

The problem I have experienced appears to be interference created by loco motor and pickups causing the servos to randomly move. The movement is always in one direction - away from the servo position when the switch is in the off (open contact) position - and the movement is further than the servo “on” setting stored in the MERG controller (set by the MERG servo setting box). The servo connected to the J5 plug seems to be more frequently affected than the others.

Nigel Cliffe, a member of both S4Society and MERG has provided help with reprogramming the MERG controllers. The MERG controller boards are now all of the latest type and have “G” version firmware. This was changed for version “H” which appeared to fix the problem. However, the controllers wouldn’t then hold the servo position and speed settings when power was switched off, so we had to revert to the earlier "G" version. The inability to hold settings was found with the setting box and also confirmed by setting the controllers from a PC. The MERG controller links are set to power down the servo after movement.

I have now fitted a third MERG controller to reduce the length of the servo leads, now only one has a short extension lead. The power leads to the MERG controllers have been closely “twisted” and the feeds from the baseboard D connectors to the controllers are now through 6 way screened cable. The power and control cables have been rerouted to move them apart and away from the DC section track feeds, etc. Locos running over one particular turnout seemed to cause a regular problem and I only recently found that the nose of the Vee had become detached electrically from the wing rails. The work so far has been restricted to one baseboard as I can only set up one at a time in the workshop. This is the one with the largest number of servos (seven) and MERG units (three).

As I carried out this work the problem seemed to be increasingly restricted to the servos connected to one MERG unit, so I swapped this for a new one which seemed to help further. A good look at the replaced unit doesn’t show any signs of dry joints or other assembly problems.

What has been the result of all this? The twitching is almost but not entirely eradicated and seems to depend on which loco which is running on the track, the worst being a Precursor 4-4-0 chassis fitted with a Mashima 1424. Fitting a .1uF capacitor across the motor leads has had little effect but I’ll now add two more from the leads down to the motor case. Running the light chassis without the body obviously causes some pickup issues, but is worthwhile as a test vehicle! Track/wheel/pickup cleanliness certainly seems to help.

Two 0-6-2 Coal Tanks with Mashima 1224s don’t now seem to cause a problem (all the other locos are in the display cabinet in the house). Again, either Pentroller or Modellex controllers seem to make no difference.

It has been suggested that the problem is pickup/motor interference being transmitted by the track acting as an aerial and that the MERG controller is susceptible to this. One thread I found elsewhere suggests mounting them in a tobacco tin to act as a Faraday cage, the lid being removable for access to the settings connector. That’s also something else to try, plus shielding the servo cables.

The N&SAG layout Coldfair Green uses the same MERG controller and Towerpro servos and doesn’t suffer from the problem. The only obvious difference AFAIK being that the track on Coldfair Green is divided into 60’ (scale) sections which are electrically separated at rail joints but connected by the feed cables underneath the baseboard. The plain track on London Road has soldered fishplates at most joints but the worst problems occur on a section of track that is divided up by turnouts.

I still have a few more things to try and will report back on what they are and if they work.

Jol

Kos

Re: Twitchy servos

Postby Kos » Mon Sep 16, 2013 3:17 pm

Hi Grovenor,

I've just had a severe re-emergence of servo twitching. Could you please elaborate on items e) and f) in your list of Dec 12. I use MERG Servo4 boards which have 3 sets of connections: 12V DC power, ON/OFF switches (0 and 1-4) and the 3-wire board to servo connection. Which are e) and f)?

Thanks, KoS

User avatar
Jol Wilkinson
Posts: 1111
Joined: Mon Jul 21, 2008 7:39 pm

Re: Twitchy servos

Postby Jol Wilkinson » Wed Sep 18, 2013 8:48 am

Hi Kos,

further to my post on 3rd Feb, I've now mounted the Servo4s in small aluminium boxes and used sheathed cabling for the switch and power feeds on the baseboards. The cabling has been rerouted to get it away from the track feeds. This has gone a long way to reducing the problem although it is still not entirely eliminated.

The cables from the control panel to the baseboards run through umbilical cables, but I haven't yet rebuilt these to include shielded cables (these are fitted with D connector plugs and sockets as it irs a transportable exhibition layout). The servo leads aren't shielded as I haven't yet found a reasonably priced woven metal cable sleeve.

I believe that there is another firmware upgrade for the Servo4 pic, but as I am no longer a member of MERG (I am not sufficiently technically educated to understand most of their developments) I aren't sure what that provides.

Jol

User avatar
grovenor-2685
Forum Team
Posts: 3917
Joined: Sun Jun 29, 2008 8:02 pm

Re: Twitchy servos

Postby grovenor-2685 » Wed Sep 18, 2013 6:24 pm

Kos wrote:Hi Grovenor,

I've just had a severe re-emergence of servo twitching. Could you please elaborate on items e) and f) in your list of Dec 12. I use MERG Servo4 boards which have 3 sets of connections: 12V DC power, ON/OFF switches (0 and 1-4) and the 3-wire board to servo connection. Which are e) and f)?

Thanks, KoS

Hi KoS, that was not actually my list but I did post it so I'll try.
Item e) These are the servo wires connected to one of the 3 way headers. As servos normally come with a pre-installed flat cable its not often possible to do this, the wires specified to be twisted are the 0V and the signal, usually black and yellow respectively. On those odd occasions where the servo cable needs extending then you can include this twist in the extension cable you make up. I have some done like that but I don't know if it did any good as I don't seem to have the problem.
Item f) Has nothing to do with the Servo4 board but is the track feeds between controller and track, worth a try if you are desperate but not practical if you have wired up on the common return system.
Regards
Keith
Regards
Keith
Grovenor Sidings

Kos

Re: Twitchy servos

Postby Kos » Wed Sep 18, 2013 7:11 pm

Thanks Jol and Grovenor. I just wanted to be sure I'd understood what Grovenor was recommending. With the help of an electronics friend I have finally resolved all (fingers crossed) issues with my servos and signals. This has involved a number of modifications, no one single thing did the trick. I'll do a bit more testing then report my findings.

Kos

Re: Twitchy servos

Postby Kos » Fri Sep 20, 2013 3:55 pm

Following on from my posts earlier this week, here's my contribution to the servo debate.

It has taken me nearly two years to finally be able to use servo-driven signals without loco interference. With the help of a knowledgeable friend, an ex-professional electrical/electronics engineer, I have worked out a strategy for the use of MERG Servo4 boards, albeit in a simple, test-bed type situation.

KoS Mk 2 16 130920.JPG
Servo test bed

The test bed (actually one of the two scenic boards on my layout Kyle of Sutherland) consists of a single line entering a loop, controlled by a signal in each direction. The point is operated by a Fulgurex point motor, the servos are TowerPro SG90s, the controllers are Pictroller and Pentroller, the traction, point and servo power comes from Maplin wall-mounted transformers, and the whole is operated from a small control panel off to the left of the picture.

The five locos I used to test the set-up have Mashima 1224, Mashima 1626 and Portescap 1219 motors, and whatever motors a currect Bachmann and an old Triang are provided with.

I followed a four-step process of implementation, as follow:-

1. Short 3-wire connection from Servo4 to servo
2. Route traction wires away from Servo4 and from servo
3. Protect servo boards from 12V DC input
4. Suppress locos

At each stage I evaluated progress with each of the five locos before moving on. I wasn’t able to entirely eliminate twitching until I’d been through the whole process. Until stage 3 the servo power input and switching remained local to the board and, while twitching had been reduced by stages 1 and 2, it was still present with some locos. When at stage 3 I first connected the 12V DC power from a transformer plugged in remotely from the Servo4, the twitching became significantly worse again.

The Servo4 boards were then protected from the 12V DC input by wiring a 0.1 microfarad capacitor across the terminals and adding 10 ohm resistors in line with each terminal, as shown top left in the photograph

KoS Mk 2 17 130920.JPG
Protected servo board

By the completion of Stage 3, the twitching was down to only two of the five locos. At its worst twitching occurred with all but the Bachmann diesel. Of the two remaining locos, the Mashima 1224 motor of the 0-6-0 was suppressed using the same technique and components as the board protection and that left only the Triang shunter.

The Triang shunter had been provided with suppression from new: an unknown capacitor across the motor leads, but no resistances. I replaced this capacitor with a new 0.1 microfarad capacitor and all was well. Of course, this means the Mashima 1626 and the Portescap 1219 are still unsuppressed. I don’t know what goes on inside the Bachmann diesel. I have been told how to decide on what R-C combination, but I'll deal with that another time.

Stuart
You do not have the required permissions to view the files attached to this post.
Last edited by Kos on Sat Sep 21, 2013 3:17 pm, edited 1 time in total.

User avatar
grovenor-2685
Forum Team
Posts: 3917
Joined: Sun Jun 29, 2008 8:02 pm

Re: Twitchy servos

Postby grovenor-2685 » Fri Sep 20, 2013 6:25 pm

Stuart,
re photos, when writing your post in the composition window use ths "Upload attachment" tab below the window, find your photo, then click "add the file". Once it is uploaded it will appear in a list and if left like that will be at the bottom of your post. If you place the cursor where you want it to fit in the text then click "place in line" that will do as indicated.
This works in edit mode as well as a new post so you can still add the photos.
Regards
Keith
Regards
Keith
Grovenor Sidings

Kos

Re: Twitchy servos

Postby Kos » Sat Sep 21, 2013 3:19 pm

Thanks, Keith, useful photo info which I keep losing.
BW Stuart

User avatar
grovenor-2685
Forum Team
Posts: 3917
Joined: Sun Jun 29, 2008 8:02 pm

Re: Twitchy servos

Postby grovenor-2685 » Sun Mar 23, 2014 11:36 pm

Just an update, one of our MERG members has cured his twitches by fitting Capacitors across the track in the sections where the loco was running when the twitch was produced.
Easy to test and implement if it works for you.
I have added a single 10 nF ceramic capacitor across the rails of each circuit of track as advised by Hardy (thank you), and the problem is solved! Even the 61xx causes no problems. The loco still needs a major overhaul though.

Keith
Regards
Keith
Grovenor Sidings


Return to “Electrics”

Who is online

Users browsing this forum: ClaudeBot and 0 guests