Servo Twitch

User avatar
David B
Posts: 1508
Joined: Sun Jul 20, 2008 5:30 pm

Servo Twitch

Postby David B » Mon Feb 21, 2011 8:16 am

In recent discussions, I have heard mention of some difficulties with servos when a loco passes or an electromagnetic uncoupler is used in that the servos 'twitch'. For example, the servos used to control level crossing gates give a twitch when a loco passes.

Twisting together the wires to the servo and putting a suppressor on the loco do not appear to be effective.

It would seem this problem is inconsistent as some people have no problems at all, whilst others get the 'twitch'.

As someone planning to use servos (when he gets around to the layout), this is a concern. Can anyone shed any light? Has anyone else had similar problems?

David

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

Re: Servo Twitch

Postby grovenor-2685 » Mon Feb 21, 2011 9:30 am

David,
Are these "discussions" on line somewhere? If so could you provide a url?

I have not experienced such myself and had not heard of twitching previously except at switch on.
Regards
Keith
Regards
Keith
Grovenor Sidings

User avatar
David B
Posts: 1508
Joined: Sun Jul 20, 2008 5:30 pm

Re: Servo Twitch

Postby David B » Mon Feb 21, 2011 10:37 am

No, Keith. The discussions were with people in the Mid Devon group. There was experience of the 'twitch' there and they had heard of other people who had also had it.

David

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

Re: Servo Twitch

Postby grovenor-2685 » Mon Feb 21, 2011 1:00 pm

OK so we don't have any info on the way the servos were connected, or whether they were sharing power supplies with the other parts of the layout, which are the most likely sources of trouble.
I have flagged up this question with MERG and will feedback any responses.
Regards
Keith
Regards
Keith
Grovenor Sidings

User avatar
David B
Posts: 1508
Joined: Sun Jul 20, 2008 5:30 pm

Re: Servo Twitch

Postby David B » Mon Feb 21, 2011 1:12 pm

Thanks, Keith. I, too, will see if I can find out more.

David

User avatar
Wizard of the Moor
Posts: 191
Joined: Mon Jul 21, 2008 9:02 pm

Re: Servo Twitch

Postby Wizard of the Moor » Mon Feb 21, 2011 9:49 pm

David,

I have had some fun and games with servos twitching. I recently converted all my turnout operation from Fulgurex motors to TowerPro servos driven by MERG boards.

Here is what I have posted on the MERG forum:

<quote>
Well, it seems as though I've solved my original problem.

To recap, My layout has a single turnout at one end which I had controlled from a Servo4 board at the other end. This meant taking the 3 servo control wires the length of the layout, about 9 feet. The servo controlling that turnout twitched a lot when other switches on the control panel were used, especially those that control electromagnets for uncoupling. Adding snubber diodes across the electromagnet coils helped but did not cure the twitching.

I noticed that the servo was also affected by interference form outside the layout. Using my resistance soldering unit would twitch the servo. Also, a visiting loco travelling towards the turnout would cause rapid movement of the switchblades - not what I wanted at all!

Today I have mounted another Servo4 board next to the turnout and the twitching has disappeared. Therefore having long runs of servo control wiring does not seem to be a good thing. Maybe this is obvious, given that the servos are supplied with 6 inch cables, but I hope that my findings are useful to others when planning their layout wiring.

The servos still twitch at power on, but they settle down straight away.

</quote>

The length of the servo control wires seems to be the key factor in my experience.

I'm now very happy indeed with the servo turnout control and wouldn't use anything else.

Cheers,

James
James Dickie

My workbench

User avatar
Russ Elliott
Posts: 930
Joined: Thu Jun 02, 2011 6:38 pm

Re: Servo Twitch

Postby Russ Elliott » Mon Feb 21, 2011 10:19 pm

Use shielded cable for the longer runs?

User avatar
David B
Posts: 1508
Joined: Sun Jul 20, 2008 5:30 pm

Re: Servo Twitch

Postby David B » Tue Feb 22, 2011 9:32 am

I now have more information about the way the servos are installed. I quote directly from the email I received.

'The servo and servo control board are as supplied by Heathcote. There are two servos installed under the layout to work the level crossing but only one twitches when a loco passes. The servos are connected to the board such that when the first one finishes moving it sets the second one going and it is the second one that twitches. The power supply to the board and servos is 9vDC supplied from a regulated power supply that is completely isolated from any other power supply to the layout. Power and control cables to the board are by multi strand wire and all the cables are twisted together. The connection from the servo to the boards is via the supplied and fitted 3 way ribbon cable.

The other problem is the effect that the electro-magnets for uncoupling have on the servos. When pulsed they can cause the servo to move to maximum travel and often need resetting by switching the power off and on again. Again the power supply to the uncouplers is completely isolated from the power to the servos.
'

David

craig_whilding

Re: Servo Twitch

Postby craig_whilding » Tue Feb 22, 2011 10:36 am

I asked Mike Bolton about this in the club last night as we will be installing Servo4 boards on Slattocks for the signal control (we had plenty of Tortoise for the points already). He mentioned that the Servo4 board has been revised and now old stocks are gone new boards would have an additional resistor or something to combat this issue. It'll be a while before we fit them unfortunately but they will be near electromagnets too so that will be interesting to observe.

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

Re: Servo Twitch

Postby grovenor-2685 » Tue Feb 22, 2011 11:26 pm

David,
Uncouplers can be a source of interference, I had problems with mine interfering with my DCC but this can usually be sorted. Do you know the circuit for the uncouplers? Are they AC or DC? what voltage? are they wired direct from buttons on the panel? do the wires run close to the Servo control and power supply wiring? Are there any back emf suppression diodes in the circuits?

I think the twitching that Mike is referring to is one experienced by a number of MERG members mostly on switch on and not related to interference from other circuits. It is related to the MERG board design which will be different from the Heathcote one.

We are getting some discussion on the MERG site but its not yet reached the point where I can sensibly distil it for you.
Regards
Keith
Regards
Keith
Grovenor Sidings

User avatar
David B
Posts: 1508
Joined: Sun Jul 20, 2008 5:30 pm

Re: Servo Twitch

Postby David B » Wed Feb 23, 2011 8:07 am

Keith and everyone who has contributed,

Thank you for your help so far. I will see what else I can find out to answer Keith's questions.

I am, though, away for a few days, so will not be able to put anything up until the weekend.

David

User avatar
David B
Posts: 1508
Joined: Sun Jul 20, 2008 5:30 pm

Re: Servo Twitch

Postby David B » Sat Feb 26, 2011 9:14 am

More info about the servo set up that twitches:

1. Cables to the servos are as supplied and therefore quite short (a point made already).

2. Screening the cables made no difference and I also tried encasing one of the servos in lead and again there was no difference.

3. However I tried various locos past the crossing where the twitch applies and not all made them twitch. The cleanest and most free running locos caused no problems but when a loco stuttered, failed or came off then the servos twitched. It would seem therefore that their must be some sort of 'discharge' related to faulty pick-ups or dirty wheels which affects the servos. I know all my locos should be perfect runners but such is life. I wonder if their is any 'smoothing' that can be applied to the track feed that could solve this.

4. BTW I use one of the new Pictroller controllers which does have some feedback to overcome sticky locos and it certainly makes them go better.


David

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

Re: Servo Twitch

Postby Jol Wilkinson » Sat Feb 26, 2011 12:38 pm

David,

rather off topic, but I would welcome your views on the Pictroller. Do you have the Hand Held version orthe larger panel mount type.

I have both versions of the Pentroller at present but may need an additional controller soon.

Jol

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

Re: Servo Twitch

Postby grovenor-2685 » Sat Feb 26, 2011 8:36 pm

David,
I'll report back the suggestions from the MERG forum next week, meantime do you know if the problem with the uncouplers arises specifically when the operator takes his finger off the button?
Regards
Keith
Regards
Keith
Grovenor Sidings

User avatar
David B
Posts: 1508
Joined: Sun Jul 20, 2008 5:30 pm

Re: Servo Twitch

Postby David B » Sun Feb 27, 2011 1:52 pm

In reply to Keith's last post and enquiry:

Keith is probably on to it. The problem did seem to occur when the button was released but also occured if it was held down for a long time. The main reaction though was on quick repetitive on and offs.

Interestingly since I re-sited one of the boards away from all the uncouplers and such that the cabling to the servos was as short as possible I do not get a problem with that particular board. The one at the other end of the layout which controls the level crossing again doesn't seem to be affected by the uncouplers. However this is the one that 'twitches' when a loco is on that particular section of track. I experimented yesterday in trying to re-site the servos as far away from the track as I could but they still twitch.

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

Re: Servo Twitch

Postby grovenor-2685 » Mon Mar 07, 2011 12:12 am

Keith is probably on to it. The problem did seem to occur when the button was released but also occured if it was held down for a long time. The main reaction though was on quick repetitive on and offs.

For the uncoupler, assuming it is on DC, connect a diode across the uncoupler coil, this short circuits the back emf caused by letting go of the button and should deal with that part of the problem.
Uncoupler.png

The diode should be rated for the same current as the coil.
Regards
Keith
You do not have the required permissions to view the files attached to this post.
Regards
Keith
Grovenor Sidings

User avatar
David B
Posts: 1508
Joined: Sun Jul 20, 2008 5:30 pm

Re: Servo Twitch

Postby David B » Mon Mar 07, 2011 9:09 am

Thanks, Keith. I'll pass this on.

David

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

Re: Servo Twitch

Postby grovenor-2685 » Mon Mar 07, 2011 7:23 pm

Following suggestions from the MERG forum.
Regards
Keith

This sounds to me like a current pulse/voltage drop problem rather than a magnetic effect. The sudden application of several amps to make an electromagnet energise could cause a change in the voltage drop along any common wiring that might exist between the two systems (i.e. servo power and electromagnet power).
If the power supply to each system was totally independent such an issue should never arise. It always pays to know which ways your currents are flowing!
Gordon H.

I wonder if the control input of the Heathcote board is suffering from the EMI.
I would try fitting caps (maybe 100nf ceramic) with short leads directly between inputs and ground right on the servo control board.
Everything twisted up may contribute to close coupling, the above might be a fix.
If their board design is good it will have a series resistor for protection and filtering followed by a cap like above to form a low pass filter right at the board input.
That's the sort of thing you need to get CE marking legally.
David Osborne

...There are two servos installed under the layout to work the level crossing but only one twitches when a loco passes. The servos are connected to the board such that when the first one finishes moving it sets the second one going and it is the second one that twitches...


This sounds like noise on the control input line. Shouldn't really be happening given in the configuration described it ought to be just a short link from an output to an input. I suppose noise on the supply could cause a misread but that would seem unlikely as some mechanism to debounce the input should be being applied.
...The other problem is the effect that the electro-magnets for uncoupling have on the servos. When pulsed they can cause the servo to move to maximum travel and often need resetting by switching the power off and on again...


This definitely sounds like noise on the supply causing the microcontroller to 'brown out', the need to reset by cycling the power is a classic symptom. Given (from the pictures on the Heathcote web site) the module appears to have both capacitors and a regulator on board and presuming a nice healthy 9V supply (it does provide adequate current doesn't it?) I'm surprised by this.

If the power supply is good I'd suspect a construction fault with the module and initially I'd carefully inspect it for dry joints. If it hasn't already been done I'd recommend an enquiry be sent to Heathcote given they know the details of the module's design and tampering with the module (excluding of course attachment of external components using the terminals) would obviously invalidate any form of warranty claim.
Chris White

This seems to be a problem with my signals but it only seems to affect one out of the eight in service at the moment and more so with a feedback controller than a "normal" one both are Gaugemaster ones.
I haven't tried it yet but was planning to put a slave servo into the socket of the one with the twitch and see if it is a board problem or a servo problem.
I have used the vv cheapy servos from Hong Kong (£2 a piece) and another one has a little problem finding it's end point and will twitch for no reason intermittently when the signal is off but is OK when on, this is definitely not loco movement related .... on for my railway is servo input to the servo board open.
I also have the signalling supply from a transformer and bridge rectifier at 12V dc with no smoothing separate to the traction supplies so the only connection is via EMC connection not conducted as the return is not common either.
Will try a different servo in the slot shortly and report further.
Gromit

Well I can definitely blame EMC and it is my own fault !!
I had to extend this particular cable to reach the relocated Servo4 board and the extension was run neatly into the main loom .... I use cab control as I am a power electrical engineer by profession and this electronic trickery is like liquorish allsorts glued on to cardboard with bits of foil on the bottom as far as my design ability goes .... give me relays any day.
So it is definitely radiated pick-up from 2ft (erm 600 odd mm) of extension leads. I would suggest if it is really annoying either screened cables and earth the screen to ......now there's a good one as we don't usually have an earth as such on a model railway layout.... too much wood about.
Or move the Servo4 as close to the servos as possible and shorten the cables to min length and run them well away from the power leads and divorce any returns, i.e. don't use the common return for the servo4 12V supply... is it dc or new fangled DCC ?
Either way pound to a penny it is pickup from the track supply as the problem either radiated or conducted.
Advice from the S&T people required as to where to earth the screen and how required at this juncture ? capacitor to neg rail perhaps.
If it was at work I would say each vehicle is earthed individually as far as screens are concerned in one place and only one place and do not carry screens over jumpers otherwise you get earth loops. But that's on the "real" railway.
Regards Gromit
Regards
Keith
Grovenor Sidings

User avatar
David B
Posts: 1508
Joined: Sun Jul 20, 2008 5:30 pm

Re: Servo Twitch

Postby David B » Mon Mar 07, 2011 10:26 pm

Keith,

Thank you very much for this. Please pass my thanks on to your MERG colleagues. It is much appreciated.

David

User avatar
David B
Posts: 1508
Joined: Sun Jul 20, 2008 5:30 pm

Re: Servo Twitch

Postby David B » Fri Mar 18, 2011 11:18 am

Further to the recent correspondence on the twitchiness of some servos, I have received the following from the originator of the problems discussed. I post his reply here which may be of interest:

I have been experimenting. I set up a test track with the servo and control board alongside and initially used the servos that came with the 'bouncer' board when I purchased it from Tony Witts. These servos are made by Conran and are larger than the ones I was using on the level crossing. I couldn't get a twitch from them. I then removed the other servos from the layout and installed them next to the test track and low and behold they twitched like mad when I ran a loco past!!. I tried adding the capacitor as suggested by Keith and this made no effect. It would seem therefore that the problem is related to the servos supplied by Heathcote. These are Tower Pro Airplane 9G SG90 and seem to be designed as light as possible to go in foam model aircraft according to the Tower Pro web site.

I am going to e-mail heathcote and ask their opinion but I would also be interested in what the forum think of this.

As far as the layout is concerned I have now installed the beefier servos and all seems to work ok.


David


Return to “Electrics”

Who is online

Users browsing this forum: ClaudeBot and 0 guests