Experience with Peco Smartswitch

[Colin_McLeod]

Hi 

Came across this on the Peco site, could not find anything on here about it.

Darren

http://www.peco-uk.com/page.asp?id=recall

 

There is a hiccup. Some Smartswitches have been recalled due to a fault/safety issue.

 

Details here http://www.peco-uk.com/page.asp?id=recall

[ejstubbs]

There is a hiccup. Some Smartswitches have been recalled due to a fault/safety issue.

 

Smartfrogs, not Smartswitches.  And your link is identical to that in darren01's post that you replied to.  Apart from that, thanks for making it all a lot clearer!

[Colin_McLeod]

So it is. Silly me.

[ChrisG]

I have spent a couple of days installing Peco Smartswitches to control handbuilt signals on my layout. So far I have very little to show for my effort. I could have given up at the point where the faults included the factory reset not working, any attempt to change the servo settings was causing a track short circuit, servos were not centring properly, with  the arms rotating 180 degrees to the opposite position! Given a number of additional problems it was looking like a dead loss.

 

This was running the system as suggested off the DCC track supply. When I disconnected that and attached a 12v DC supply it was a different kettle of fish altogether. In fact, it looks like it will work rather nicely. Another breakthrough was deciding to move away from the Ratio mechanisms I had been using to  a much more simple mechanical implementation as shown in this video on Youtube:-

 

 

 

Despite my position of relative happiness right now, I have to say that the instructions are not good - they refer in places to the ANE original product rather than the slightly modified Peco one and I am not impressed by the ease with which one can plug things in the wrong way around nor the flimsiness of the attachment of the various connectors to the PCB. I had to resolder many joints which were loosened by repeated screwing and unscrewing whilst fault diagnosing. It's a shoddy product which, at the price, should be a good deal better than it is. 

 

 

Chris Gardner

Alton

Hants

[ChrisG]

I have spent a couple of days installing Peco Smartswitches to control handbuilt signals on my layout. So far I have very little to show for my effort. I could have given up at the point where the faults included the factory reset not working, any attempt to change the servo settings was causing a track short circuit, servos were not centring properly, with  the arms rotating 180 degrees to the opposite position! Given a number of additional problems it was looking like a dead loss.

 

This was running the system as suggested off the DCC track supply. When I disconnected that and attached a 12v DC supply it was a different kettle of fish altogether. In fact, it looks like it will work rather nicely. Another breakthrough was deciding to move away from the Ratio mechanisms I had been using to  a much more simple mechanical implementation as shown in this video on Youtube:-

 

 

 

Despite my position of relative happiness right now, I have to say that the instructions are not good - they refer in places to the ANE original product rather than the slightly modified Peco one and I am not impressed by the ease with which one can plug things in the wrong way around nor the flimsiness of the attachment of the various connectors to the PCB. I had to resolder many joints which were loosened by repeated screwing and unscrewing whilst fault diagnosing. It's a shoddy product which, at the price, should be a good deal better than it is. 

 

 

Chris Gardner

Alton

Hants

 

 

Although I was able to post that the system performed better on 12v DC than DCC track power, it again began to reset itself whilst turned off. Having worked on it for many hours a day now for 3 or 4 days, I have finally thrown in the towel and will try the Heathcote Electronics bouncing signal controller, still using the Peco Servos so as to recoup some of the expenditure....... Judging by earlier posts in this thread I am far from being the only person who has given up on the Peco product. 

 

Chris Gardner

Alton

Hants

[kirtleypete]

I'm sorry so many people are having trouble, but to be honest I'm very relieved that it wasn't just me!

Peter

[John Max]

When I first tested the smart system out I was surprised to see the reference too DC connections and no reference to DCC, they worked ok on DCC. I decided to contact Peco, asking if they worked better with DC, I never got a reply until I made a second request. The reply just stated that they can be used for either. Sent same request to ANC who replied that DCC is ok but seemed to no nothing about using DC, They seemed surprised that someone would want to use DC if they had a DCC system.

My 28 servo system is still working ok, although occasionally the plugs drop apart son the ones without the locking hook that was on the original versions. I repeat that by far the worse part of the system is the frog boards all disposed of at a big loss having installed Tam Valley frog juicers.

John

[Gordon H]

It's a pity Peco don't seem to provide a downloadable version of the Smartswitch instructions, so that potential customers can evaluate the system beforehand - and the quality of the instructions that come with it.

It is fairly obvious they have tried to produce a system which can work with a DC or DCC system, otherwise why would they have produced a special decoder to go with the system? If it was going to be usable with DCC only, surely the decoder would have been included in the main board?

One of the main considerations with any multiple servo system is the power supply, which needs to be capable of providing plenty of current without dropping out when several servos move simultaneously.

[ejstubbs]

It's a pity Peco don't seem to provide a downloadable version of the Smartswitch instructions, so that potential customers can evaluate the system beforehand - and the quality of the instructions that come with it.

 

Er, I think they do: http://www.peco-uk.com/imageselector/Files/Instruction%20sheets/Smartswitch%20Manual%20v2.pdf

Edited July 25, 2016 by ejstubbs

[John Max]

Firstly a correction I should have said ANE not ANC.

The ANE versions of the instructions are available on the ANE site and are much clearer than the Peco instructions and probably more uptodate.

I don't understand the DC idea as I thought the system is meant for DCC only.

[Suzie]

The core Smartswitch system is just a DC servo controller. Add-ons are required for the DCC decoder, and for the frog switches.

 

I suspect ANE aimed the product at a purely DCC audience, while Peco have targeted the analogue market.

[ejstubbs]

Link for the ANE manual: http://www.anemodel.com/Upload/Download/201503312203272305.pdf

 

There also offer a fairly detailed servo programming manual: http://www.anemodel.com/Upload/Download/201410212000373524.pdf

 

The ANE Smartswitch web page does talk about both DC and DCC, as does the manual.

[bertiedog]

I had heard of the Peco system when it first appeared, but rather thought it had been dropped or little advertised, as this post seems one of the few mentions.

 

It now seems it is the implementation of servo control that it the issue, the wiring up and quality of the connectors.

 

But there also seems to be a bit of a high expectation that the servo will follow the commands to the degree required for the use in MR.

 

Unless the servos are modified in some way, they are merely analogue slaves to the applied mark space signal applied to the signal wire, and in normal RC uses, depend on a neutral signal ratio being applied to send the arm to a nominal neutral mid position, which can be fine tuned by the transmitter signal. The extremes of the motion are set by the operator.

 

I assume the controller is a processor to provide the signals and to be able to handle a quantity of signals at the same time.

 

But there seems a fundamental flaw, the servos are dumb followers, they provide no feedback as to what they are doing or have done.

 

Normally in a modern control system there is feedback, both from DC motors and steppers to provide the signal confirmation as to what is happening at the servo end.

 

But here we seem to have a "dumb" servo, the assumption seems to be that the correct signal will always drive the servo to the same rotational position.

 

But servos for RC are not designed to do this, the electronics is not accurate enough or stable enough to get the angular accuracy needed for repeatable operation.

 

I notice comments on resetting the circuit, yes it may well need constant recalibration to maintain the positions, servos contain electronics that warm up, vary values, and shift about in use, none of which matter in R/C, as the feedback is provided by the operator on a second by second basis.

 

There is no such loop in servos, unless Peco or the Korean makers have re-invented the wheel..

 

I can't quite see the complaint about complex wiring, it is just repeated looms for each servo, but then I was used to wiring up BT exchanges. It is merely more complex as power plus signal goes to each servo.

 

The instructions seem clear enough as they stand.

 

But getting back to the principles, it seems the whole idea is trying to be too clever by half.......

 

Servos are great for MR, they are flexible, able to give variable rotary positions to order, and are cheap, even the cheapest will work, but they only work under similar conditions to R/C nor model rail use.

 

Points require to go from one end of a stroke to the other, the stroke is defined mechanically by the point and adding a bit for omega loops or springs. I assume the Peco controller tries to set itself to generate the required signal to achieve this, but how is stability to be provided at each end of the stroke?

 

Mention was made of jittering, as yes,all servos jitter, the better systems just don't do it so much, and they wander from the set position with time and heat.

 

But Peco seem to have used a modest servo type, and in my opinion unable to maintain the parameters required for accuracy.

 

Now I have used servos for many years on MR projects like my P4 layout, especially where slow motion was required. works fine, but with several caveats that I am not sure Peco have addressed.

 

My servos are driven by the usual variable mark space signal, generated in my case by a 555 timer, simple and reliable, but I would not expect the arm to repeatable mimic the position of the control knob, or a computer DC/Analogue signal.

 

They have to be used with overshoot on the movements, with somr mechanical compensation, (omega loops), to rectify any problem, and also must cut out at the destination of the stroke, they cannot be trusted to hod a position under the mark space signal.

 

I have used Merg control circuits and my own, and really cannot see how Peco can trust the servos much as to what position it is in.

 

I finally found the solution was limit switches, crude but effective, the stop provided is accurate, and maintains power off positions. There appears no other way to do it.

 

The cost is lower than the Peco control, but theirs is not very high for what they supply, apart from the cheap servos. People are not electronics experts , they want out of the box solutions that work, and it looks like Peco works, but needs caveats added to combine the electronics with the mechanics of the servos.

 

I must add I recently tried servos with the electronics removed, straight DC control, with micro switch stop, works even better as it is slow and silent, losing the growl of the signal noise. It is cheaper as well, about a servo + £1 per point, or Signal or Level crossing, and it is a once set, then forget system.

 

These comments are that, comments, not complaints, just queries as to operation, and set up of a basically good idea.

 

Stephen.

[Suzie]

Servo control can be done well with the correct components and good design, or less well.

 

Firstly, take the servo choice. For what we need as modellers we need a servo that will stay in position and not make too much noise doing it. We also need a servo that is robust. To this end a servo like the TowerPro SG92R with digital electronics and carbon fibre gears serves our purpose, where older analogue versions like the SG90 with its plastic gears and analogue electronics is not quite the same beast. The SG92R can be stalled without stripping the gears, and moves quite quietly. The digital electronics stops it juddering wildly if a rogue pulse appears due to induced noise in the control wire. The SG92R will sit quietly all day without any need to shut off the power or desist with the pulses as some systems do to get round the shortcomings. I am sure that replacing the (very expensive) Peco SG90 based servos with SG92R will reap rewards.

 

Secondly, pulse width stability is key to keeping the servos quiet. The SG92R will ignore a bit of pulse jitter, but it cannot remove it all because it has to decide what is jitter and what is a request for a small movement. I have not put a Smartswitch on the oscilloscope to look at the output so it might be OK, there is no reason for it not to be, but other makes like ESU Switchpilot Servo do make a bit of a hash of keeping the pulses consistent. When a manufacturer espouses the advantage of switching off the pulses, you can bet that their pulses are not very uniform.

 

Thirdly wiring. The Smartswitch is a potential birdsnest of wiring with anything up to ten PCBs wired together to provide all the control options for four crossovers. Main unit, decoder and eight frog switches is a lot of kit and making sure all of those connections are reliable in a confined under baseboard space is asking for trouble. If all you need is to operate four dead frog points it might make sense, but the simpler wiring of the plug together modules of the ESU Switchpilot Servo + Switchpilot Companion or the single board Signalist SC2 reaps a huge benefit when you come to install in a DCC live frog scenario.

 

Fourthly mounting. The servo is device with a rotary output, typically 90°, and the point has a tiebar with a linear movement, typically around 3mm. To make life easy the mounting needs to be able to translate that rotary 90° to an adjustable 2-4mm throw without too much messing about. The mounting also needs to maintain a small profile. The smartswitch motor appears to be optimised to take the minimum board area under the baseboard which is commendable, but it does make it it stick down below the baseboard a bit more than it needs to.

 

While we are using a components originally intended for RC models, we are using them in a very different way. Positioning has to be precise, there is no gyroscope feedback to overcome jitter!

[Gordon H]

Er, I think they do: http://www.peco-uk.com/imageselector/Files/Instruction%20sheets/Smartswitch%20Manual%20v2.pdf

 

Fair enough, I couldn't find it in my haste at the time.

However on looking at the document, it is noticeable that there are plenty of words and warnings regarding supply voltage, but no mention of the current the system is likely to draw from the power supply.

This may be why DCC users are likely to get better results, as most DCC systems can provide several amps.

Try using a 12V 0.5A DC 'Wall Wart' instead and the results are likely to disappoint.

They really should give a recommendation for the power supply current as well as voltage.

[Suzie]

Each servo will take up to 0.9A, and they will all try and move a little bit at startup.

[Dungrange]

Each servo will take up to 0.9A, and they will all try and move a little bit at startup.

 

I assume that therefore implies the need for a 4 Amp power supply to cover each of these four turnout units?

 

This thread has definitely put me off trying the PECO Smart Switches. 

[Suzie]

I assume that therefore implies the need for a 4 Amp power supply to cover each of these four turnout units?

 

This thread has definitely put me off trying the PECO Smart Switches. 

 

Yes. To mitigate that a little 4A switch mode power supplies are quite easy to get hold of for less than the cost of a traditional 16V 1A transformer and CDU.

[John Max]

Hi, I have 7 of the Peco Smart boards with 28 servos and the 7 DCC boards added, except for the occational minor adjustments they work fine on my NCE 5 amp DCC system.

The power draw as no effect I can detect on the running of the trains including power hungry large sound locos. I often run 2 at a time and occationally 3.

The do not touch part of the Peco (ANE) smart system are the frog boards which are a disater zone. I use the Tam Valley frog juicers to power frogs, they are amazing.

I also have a lot of building lights attatched to the system, these also don't seem to cause much of a power drain although this is not recormended and shoud have a seperate power supply.

John

[tjorch]

Hi all, is there any reason why shop (eBay) brought servos do not work on a Smart Switch board, I have 2 boards but only 2 servos were supplied with 1 board ( they are the only 2 I can get to work properly ), all the servo's are the same make and model (Tower Pro SG90 ) all servo's work fine when I check them on my servo tester. any help would be appreciated. Terry  

[John Max]

Hi all, is there any reason why shop (eBay) brought servos do not work on a Smart Switch board, I have 2 boards but only 2 servos were supplied with 1 board ( they are the only 2 I can get to work properly ), all the servo's are the same make and model (Tower Pro SG90 ) all servo's work fine when I check them on my servo tester. any help would be appreciated. Terry

 

Hi, sorry but I can't answer this as I bought some off eBay myself but got them mixed up so I don't know which were the ANC/Peco ones or the on line bought. I have 4 in unused and when I had a problem installing one I just grabbed another so it is possible that the 4 I bought off eBay are the ones left unused. I bought most of mine from the track-shack, Isle of Man I highly recommend them. Also often get good prices from Hatton's or look for good offers from listed on eBay. Always comparing total including postage.

[Suzie]

I think that there might be a few fakes of the SG90 out there. Quite a few suppliers purporting to be Towerpro have disappeared from Alibaba!

[Bo P]

Hi,

Some comments from a new member...

Saw this thread as I was looking for more info on the Peco smart switch, and by all accounts here I would probably discount it, on several counts.

 

I have been looking for a robust, simple but small motor for quite some time, to be used in a 3-track hidden siding on an N-scale layout.

I recently came across a new type of point switch which came highly recommended - the czech-made MP1.

Link:

http://www.mtb-model.com/pages/mp1.php?lang=en

I grabbed one, and did test setup for both N-scale and H0-scale points - and both worked beautifully.

It looks to be a stall-type device (means easy to wire up to a simple SPDT switch) - it is extremely compact - and it's designed to be mounted below the trackbed.

It has outputs for frog polarity - alternative other uses such as control panel feedback or whatever. Feeds off AC or DC.

It can be set up with different throw lengths - 3, 6 and 12 mm, of which 3mm works fine with N-scale, and 6mm with H0 scale points.

It moves slowly (but still only 2 secs or so), but very very quietly.

And when I checked pricing just now, it goes for 12-13 euros - I guess around £10-11.

 

I will be installing these for the hidden sidings shortly, having bought 3 more - 

Personally I think these beat everything else on the market - only unknown is how well they will last.

And in case you wondered - no, I have no interest in the company or its products :-)

 

Happy railroading

Bo

Denmark

US prototype fan

[kirtleypete]

I saw them in 'Loco Review' and 'Rail Miniature Flash', it's interesting to hear from someone who has actually used them. Thanks for posting,

 

Peter

[smokebox]

Looking at the pdf instructions, it isn't a stall motor. It has limit switches to cut power at the end of stroke, so no power is taken when not moving.