Megapoint Controllers

[leopardml2341]

I only partially agree.

With what?

 

The Analogue/Digital pros and cons or the ferrite core suppression?

[Vecchio]

This might be useful info ScRSG:

 

Already tried it, no noticeable effect.

But as I said - never used a digital servo for model railway application.

[Vecchio]

As it looks we write at the same time.

My answer is with both. I tried the ferrit rings and I do not see the need of a much more expensive servo for a model railway application. I know I will hear now how good the prices are at hobby king but normally what you pay is what you get.

[Revolution Mike B]

I can't speak for the MERG stuff but as per the original title of this thread, the Megapoints stuff we have does exactly what it says it does.

 

We currently have two systems set up:

 

One is on our club N Gauge layout measuring 18x9 and has one multi panel (only used to cut down on board locations), 8 servo control boards, 84 cheap Hobby King servos, about 27 miles of cheap servo extension cables (300mm, 500mm, 1m all connected together in random sizes), 1.2mm control wire (way oversize and currently being changed for 0.6), micro switches for the frogs and a 13.8v, 100w, 8amp power supply that all works perfectly - no twitches, no headaches and a bit of chatter that we don't even notice after 5 minutes.

 

The other is a home layout using 3 stand alone servo boards, cheap Hobby King servos, 0.8 control wire, autofrog relays and a 12v, 6amp power supply, again, no twitches and minimal chatter.

 

If I'm honest, I prefer the Megapoints system to the noisy Cobalt motors I have on my exhibition layout and once you get your head around setting it up, it's a walk in the park.....and great value for money!

[Jol Wilkinson]

I can't speak for the MERG stuff but as per the original title of this thread, the Megapoints stuff we have does exactly what it says it does.

 

We currently have two systems set up:

 

One is on our club N Gauge layout measuring 18x9 and has one multi panel (only used to cut down on board locations), 8 servo control boards, 84 cheap Hobby King servos, about 27 miles of cheap servo extension cables (300mm, 500mm, 1m all connected together in random sizes), 1.2mm control wire (way oversize and currently being changed for 0.6), micro switches for the frogs and a 13.8v, 100w, 8amp power supply that all works perfectly - no twitches, no headaches and a bit of chatter that we don't even notice after 5 minutes.

 

The other is a home layout using 3 stand alone servo boards, cheap Hobby King servos, 0.8 control wire, autofrog relays and a 12v, 6amp power supply, again, no twitches and minimal chatter.

 

If I'm honest, I prefer the Megapoints system to the noisy Cobalt motors I have on my exhibition layout and once you get your head around setting it up, it's a walk in the park.....and great value for money!

DC or DCC?

[Revolution Mike B]

DC or DCC?

 

Both!

[steammad]

 

These are some of my servo mounts.

Edited February 10, 2017 by steammad

[ScRSG]

Ok folks, getting a bit fed up now! I am using MERG servo4 boards mounted in the control panel rather than next to the servos. At first they just wouldn't work at all, twitched and jumped about almost uncontrollably - this was attributed to too thin wires for the 5v line to the servos, which are up to twenty feet from the panel.

So, we tried thicker wire and this did help, but when we started to add more servos into the circuit the problem re-established itself.

So (2), it was suggested we needed a "beefier" power supply, so this was acquired and installed. This helped.

So (3), we then, from another recommendation, applied capacitors across the 5v line and also, from another recommendation placed a diode in the 0v line, whether this has helped we can't really answer, but has not done any harm.

So (4) we added a further device to supply the 5v at a constant 5 amps and from a combination of all the above, we seemed to get all the servos to work OK. ( we also removed the link on the boards which stops the signal being sent after a few seconds)

THEN, we finally applied track power to test whether a loco could run over the points we had motorised, and to test the wiring of feeds etc. All seemed fine until the loco got to within about ten inches from one of the points and...... the servo jumped so far that the arm of the servo jumped out of the operating arm of the servo mount (MERG kit).

Therefore the interference of the loco motor seems to have caused devastating problems, all of which we thought we had solved. We don't want to move the boards from the control panel if we can avoid it because the layout will be transported and we felt the boards were safer inside the panels, so this will be a last resort. If we are forced to do this and it still doesn't work then we will be back to point motors at no little expense.

Anyone got any ideas?

 

Realise this thread was started regarding the Megapoints boards and the above are issues with the MERG boards, but as it has been discussed here before then putting this in this thread, apologies to mods! The problems may be servo related whichever controls are used.

[Suzie]

_{Put the controller boards near to the servos, don't stop the pulses, and switch to using digital servos. SG92R are not expensive.}

 

 

_{The problem you are seeing is caused by interference pickup in the signal (usually white) wire. The longer it is the more interference it will pick up. Servos have a fundamental problem in that the control wire is unbalanced and high impedance so you need to be careful about inducing voltage or current in to it. Also by turning off the pulses you are only allowing interference in to the servo, if continuous pulses are enabled any interference will be corrected fifty times a second significantly limiting the effects of any interference.}

 

 

_{For long runs use three dedicated wires twisted together to feed the servo. This will not eliminate the effect of voltage spikes, but will eliminate the effect of current spikes so may eliminate the problem if it is caused by induced current. Try and keep track feeds away and twisted separately too.}

 

 

_{Long wires, separating the signal wire from the power wires, switching off the pulses, analogue servos - all of these things will contribute to increasing susceptibility to interference pickup.}

[AndyID]

_{Put the controller boards near to the servos, don't stop the pulses, and switch to using digital servos. SG92R are not expensive.}

 

 

_{The problem you are seeing is caused by interference pickup in the signal (usually white) wire. The longer it is the more interference it will pick up. Servos have a fundamental problem in that the control wire is unbalanced and high impedance so you need to be careful about inducing voltage or current in to it. Also by turning off the pulses you are only allowing interference in to the servo, if continuous pulses are enabled any interference will be corrected fifty times a second significantly limiting the effects of any interference.}

 

 

_{For long runs use three dedicated wires twisted together to feed the servo. This will not eliminate the effect of voltage spikes, but will eliminate the effect of current spikes so may eliminate the problem if it is caused by induced current. Try and keep track feeds away and twisted separately too.}

 

 

_{Long wires, separating the signal wire from the power wires, switching off the pulses, analogue servos - all of these things will contribute to increasing susceptibility to interference pickup.}

 

These are all good suggestions. Unfortunately, I doubt they can assure a level of reliability that ScRSG requires, and I don't think he's interested in conducting another science experiment. I think his priority is to get the layout running.

 

The bottom line is that servos do not have a lot of noise immunity. That's not a problem in a relatively benign environment like a model aircraft but it is a major liability in a electrically hostile environment like a model railway, particularly if the servos are controlling turnouts on an exhibition layout. Servos seem to work fine on some layouts, but there are too many examples of people experiencing major issues with servos on layouts to simply ignore the problem.

 

Sorry to be the bearer of bad news, but short of performing major surgery on the servos themselves (i.e., throwing away the electronic bits) or "front-ending" them with with a high-immunity interface of some sort that, AFAIK, does not yet exist, I think it might be time for ScRSG to pull the plug and adopt a more traditional, albeit much more expensive, approach.

[ScRSG]

Thanks all for your inputs to our dilemma. It would seem that the thrust of advice, from here and the MERG forum, that the boards need to be moved nearer the servos so that the long leads are merely power supply and switching of the input.  This can and, I think, will be done. Will discuss this with the other group members to get the best way forward. One fortunate thing is that the panels were designed with this possibility in mind, so the signal (white) wire currently will become the switch wire and the power supply can easily be changed from 5 to 12v. I would hope that as the switch wire is purely "on/off" then no interference will be transferred into the board/servo. We haven't yet tested the Megapoints boards which, at the moment are using the same system (i.e. in the control panel). At least with the fiddle yard, the points are all at the extreme ends and grouped together, but with the scenic part (which will use the Megapoints boards) the points and signals are spread along the whole length, making any placement of the boards on the layout somewhat more tricky!

 

We do have some digital servos in stock, these were going to be used on delicate signal mechanisms, but we may try some on the points as well to see if they will work better.

[Jol Wilkinson]

This may be of interest:

 

https://www.scalefour.org/forum/viewtopic.php?f=26&t=2227

 

One reason I haven't considered the Megapoints controller instead of the MERG Servo 4 units I already have is that my signals are fairly well spread about and therefore long servo leads would required. Despite the assurances of Mr Megapoints that this wouldn't be the case I am not willing to risk the investment, given my experiences.

 

I am about to test a Heathcliff two servo controller, to see if that offers any benefits. I have also bought a Tam Valley Octopus, but can't get on with it's remote set up controller, although the main unit does seem quite stable, so that will be going on eBay. I will also try a bigger power supply, although this never came up in all the previous discussions I had. Adding a simple suppressor circuit to the feed voltage input to the Servo 4 dis seem to help though in addition to the other suggestions re feed wires, etc. in Suzie and other's posts. The jury is still out on servos, although cheaper analogue ones do seem to suffer more problems.

 

Ultimately I come back to the view, as AndyID says, that we are trying to use a technology that isn't fully suited to our needs at present.

[Revolution Mike B]

This may be of interest:

 

https://www.scalefour.org/forum/viewtopic.php?f=26&t=2227

 

One reason I haven't considered the Megapoints controller instead of the MERG Servo 4 units I already have is that my signals are fairly well spread about and therefore long servo leads would required. Despite the assurances of Mr Megapoints that this wouldn't be the case I am not willing to risk the investment, given my experiences.

 

I am about to test a Heathcliff two servo controller, to see if that offers any benefits. I have also bought a Tam Valley Octopus, but can't get on with it's remote set up controller, although the main unit does seem quite stable, so that will be going on eBay. I will also try a bigger power supply, although this never came up in all the previous discussions I had. Adding a simple suppressor circuit to the feed voltage input to the Servo 4 dis seem to help though in addition to the other suggestions re feed wires, etc. in Suzie and other's posts. The jury is still out on servos, although cheaper analogue ones do seem to suffer more problems.

 

Ultimately I come back to the view, as AndyID says, that we are trying to use a technology that isn't fully suited to our needs at present.

 

How spread out? We're using leads that are 15 ft long with no issues at all.

[Suzie]

If you are using long leads without interference it is not that your long leads are not susceptible to interference - it is that there is no interference present to pick up.

[Jol Wilkinson]

Up to five feet long. There are three semaphore and four ground signals on one board, one semaphore and two ground signals on another. This is on a portable exhibition layout so each board is fed by an separate multiple way cable (or two) and the servo controllers are mounted on the boards. One reason for choosing the MERG Servo 4 controllers is the excellent remote set up controller, which makes adjustment easy.

 

Twitching may not be such an issue with servo point motors, but it is very visible with semaphore signals and can be damaging to mechanisms.

 

One thing that became apparent during my early discussions is that DCC equipped locos were claimed to have far fewer problems than DC kit built ones. Pickup created interference was considered one issue, so I experimented with suppression, both on locos and power feed to the servo controllers. As the layout often has guest locos, I tried to take out the problem by adapting the layout, so that we would avoid it whenever possible.

[Michael Delamar]

I havent read this thread only skimmed theough it.

 

To overcome chatter. Would wrapping the servo wire around a ferite core cure it?

[Suzie]

Use of suitable servo point motors can mitigate the effects of interference to some extent by limiting the amount of movement. The Signalist SB1 motor uses a cam mechanism which prevents wild movements outside a 90 degree range from having any effect.

 

 

Use of suitable products fitted in line with good practice (digital servos, short signal wires, good power supply, free movement, good quality servo driver with a continuous pulse stream) should result in a completely reliable installation. Get any one wrong and there is scope for disaster.

[Revolution Mike B]

I can only repeat what I've said earlier, our N Gauge Exhibition Layout is 18'3 x 9'3 and has over 80 points controlled by a Megapoints system with no issues whatsoever. We have long leads, short leads, joined leads, home made leads, cheap servos, home made mountings, cheap frog switches, one multipanel board (just for ease of wiring) and obviously a number of servo control boards. We've cut the servo leads and run them through 37 way connectors too!

 

We've also repeated this on a 15' x 9' layout in a garden shed with exactly the same results.

 

The servos have a conversation with themselves (which you become oblivious to after a while) but they don't twitch EVER.

 

These are running both DC and DC with sound chips too.

 

I really can't speak for O Gauge locos but the smaller scales do not effect the servos at all.

[Vecchio]

I read in this thread that several people recommend digital servos. The digital servo doesn't improve the servo chatting. You just get a higher torque at small movements as the servo motor gets a digital input. And for free you get some high frequency noise from them... In my experience they are of no help and unnecessary expensive for the use on a model railway.

We are not flying 1000£ plus fast model airplanes that need fast and precise reactions. We do not need a lot of power. I use the small 10g servos for 0 gauge - no problem at all.

[Suzie]

The TowerPro SG92R 9g digital servo with carbon fibre gears is typically a similar price to the very popular analogue SG90 so cost is not an issue with going digital. Digital servos have signal processing so a single spike will not send them off wildly - they will just ignore it. Most voltage induced spikes will just be a single pulse.

 

I am sure that most people will not notice the difference, it is only when you have a lot of electrical noise or need to track a fast moving small change (like signal bounce) that you will notice the difference.

[leopardml2341]

I havent read this thread only skimmed theough it.

To overcome chatter. Would wrapping the servo wire around a ferite core cure it?

Or use screened cable

[Penrhos1920]

The TowerPro SG92R 9g digital servo with carbon fibre gears is typically a similar price to the very popular analogue SG90 so cost is not an issue with going digital. Digital servos have signal processing so a single spike will not send them off wildly - they will just ignore it. Most voltage induced spikes will just be a single pulse.

 

I am sure that most people will not notice the difference, it is only when you have a lot of electrical noise or need to track a fast moving small change (like signal bounce) that you will notice the difference.

Where?

 

I can get SG90 for £1 each. SG92r are £3.88 to £17.47 each

[Suzie]

Looks like the price of SG90 has gone down a bit and the SG92R have gone up. I paid £2 each for my SG92R from eBay about a year ago, and the SG90 were the same price from that supplier. Cant remember which supplier it was!

[leopardml2341]

Looks like the price of SG90 has gone down a bit and the SG92R have gone up. I paid £2 each for my SG92R from eBay about a year ago, and the SG90 were the same price from that supplier. Cant remember which supplier it was!

That's a shame, I'm in the market for a couple of dozen and been watching/posting with interest to see how the problem is eventually solved. As I mentioned earlier, I think I'll use screened cable (just cos I have some to hand).

[Junctionmad]

 

The bottom line is that servos do not have a lot of noise immunity. That's not a problem in a relatively benign environment like a model aircraft but it is a major liability in a electrically hostile environment like a model railway, particularly if the servos are controlling turnouts on an exhibition layout. Servos seem to work fine on some layouts, but there are too many examples of people experiencing major issues with servos on layouts to simply ignore the problem.

 

 

Ultimately I come back to the view, as AndyID says, that we are trying to use a technology that isn't fully suited to our needs at present.

 

I cannot agree with these statements.  What we are seeing is people using a servo incorrectly , ( we had a posters here with meters of servo cables and the servo board in the control panel ) , we have people that simply dont understand the current requirements  and we have people that simply dont know what they are doing 

 

The result is these people " throwing their hands in the air" and declaring the technology is unsuitable .  All the time , many , many layouts are successfully using the technology quietly.

 

Abide by the technologies limitations , use short servo cables ( mine are no more then 15cm long ) , appropriate power supplies , good controller design, good power wiring and good grounding 

the result is no issues 

 

The technology is entirely suitable , but only when used by people that understand the technical issues associated with it.