Frog polarisation - any interest from others in a custom PCB?

[Kallaroonian]

Hi. A few responses aligned to above numbering :

 

3) We may be talking at cross purposes I'm not sure. In terms of wiring diamonds I suppose you may have a point but from my perspective the inconvenience of having to swap wires over when the unit is installed under the layout would be best avoided by providing a switch. It really is a right pain re-visiting screw terminals under the baseboard. I don't think reliability is going to be a key factor because the switches are theoretically only going to be operated once or twice.

 

4) I'm using a reasonably substantial laptop power supply to drive the accessory decoders. I only have experience with switchpilots and they don't seem able to drive one point straight after another

 

5) The way I've wired it so far the DCC bus goes to the track and the switchpilot takes its control signal feed off the track. The power for the switchpilot is from a separate power bus driven by the laptop PSU mentioned. I have very little experience of DCC so far as you are probably deducing. I don't really understand your response to 5). There is no accessory decoder bus as such - there is just one DCC output from the ECoS which is picked up by locos or accessories alike. I therefore don't understand how the track might shut down but the accessory decoders would still function. Also the only shorts I've had so far have resulted from having the frog polarity the wrong way round and hence a short occured when the loco drove over it; this shuts down the ECoS and there's nothing that can be done about it until the loco is physically removed

 

6) I'm not saying it's wrong but there seem to be too many connections to the latches

 

7) Can you provide the pcb layout also.

 

8. there was an eight but it got emoticonified. I was just saying I'll send the schematics sketches I have to you. We can resolve the query at 6.

 

10) I don't necessarily agree with this. First of all the cable size required is nowhere near 2.5mm and the issue here is not so much the cable entry but the need to screw the contacts closed when under the baseboard etc. The way I look at it is this. When wiring up I fit all the flyleads to the point and then these present underneath the baseboard alongside the DCC bus and laptop PSU bus for the accessory decoder. This provides all the connections necessary to hook-up the switchpilot and prototype latch assy. The connections between switchpilot and latches (not required if all combined in your design obviously) could conveniently be pre-loomed with sutiable plug and header connectors. However the connections from the point are just wires and not something I would want to pre-wire with plugs due to access through the board for a start. But they are connections that have to be made under the board of course and the issue is trying to do this with screw terminals - a spring clip connection would be much easier and much quicker

I realise too that you are probably talking about straight screw term blocks on the PCB whereas I have been thinking about either plug and header term blocks (aka what switchpilots have now) or spring clip screwless term blocks. For me it would have to be one of the latter two.

 

CDU - I'm probably being super dense here but I don't understand how a single CDU can be wired to drive multiple points. You have a digital bus driving multiple accessory decoders and these decoders interpret control signals in order to determine which outputs they should pulse. Once an output is pulsed it activates a simple solenoid wired directly to that output. It is this pulse that needs a boost from a CDU but for it to achieve this it would need to be connected between each and every output and associated solenoid. That can't be so I don't see how this works. I know I'm missing a key point here somewhere

 

12) The latch side of the schematic seems to have 10 connections for solenoids and 10 connections for frogs. Nominally the 4 solenoids would require 12 connections with a further 4 for the associated frogs. I can see some extra connections points for the frogs would be good in case of related diamonds on a point ladder etc but overall am not sure about this part of the wiring. I'll send you those scematics then we can compare notes

 

13) testing the finished product may be tricky. Simply testing the latch array was straightforward but proving the DCC decoder aspect may be harder.

 

Will send you a PM as well

 

regards

[Suzie]

I have addressed point 3 now. The spring terminals are a superset of the PCB layout for the screw terminals, so I have now created a new component for the spring terminals, and screw terminals can be fitted in their place if desired using the same pads.

 

I have now found some nice SPDT PCB mounting toggle switches in my component library and have fitted one to each relay to reverse the polarity of one of the frog outputs. The switches can be replaced with simple jumpers if not required. I am hoping that it is a standard type because I cannot find a link to a datasheet. The three inline pins have a pitch of 0.175". It is pretty close to this one at JPRelec

 

This decoder does not have the low power limitation of the Switchpilot. It will provide at least 10A per output as long as the power supply can provide it, with no recovery time. It will be enough to switch a pair of solenoids on each output.

 

Unlike the switchpilot the DIY decoder is powered only by the DCC signal input. There is no need for any other supply to power the decoder, but the solenoids will need a supply. Here is a little diagram showing how it all goes together in the grand scale of the layout:-

 

 

You do not have the Booster/circuit breaker and therefore in your layout this is just wired straight through. Similarly you do not have a CDU so are using a high current DC supply instead.

 

The latching relays have ten terminals, two coils with two terminals each and two sets of changeover contacts with three terminals each. There are other pinouts available but that is what is on the relays that are linked and the most sensible option.

 

Here is the PCB layout showing the component overlay:-

 

 

1.5mm spring terminals have now been accomodated as an option to the 2.5mm screw terminals originally specified for all three connector blocks.

 

There is a 2-way connector block for the accessory bus DCC input.

 

There is a 10-way connector block for the solenoids comprising of four normal/reverse pairs to feed to solenoids (pink and violet wires in the above diagram), and two terminals for the negative and positive supply from the CDU or DC power supply (green and orange wires shown in the above diagram). The positive supply is also the common for the solenoids (orange wire shown in the above diagram).

 

The other 10-way connector block comprises of the DCC track supply to feed the frog switches (red and black wires to the bottom of the accessory decoder in the above diagram) and four pairs of frog supply corresponding to the four solenoid outputs. One of each frog supply is switchable (the one furthest from the track supply input, shown with a brown wire in the above diagram). You will only need the non switched one if you have a crossover, diamond crossing or reverse loop, I have not shown it on the above diagram in order to not confuse it.

 

I put my decoder together and it worked. There is not much testing to do other than operate it, it is a very simple device. There are some voltage measurements you can do on the circuit before you plug the PIC in to make sure everything is working.

[Suzie]

Here is a 3d impression of the decoder with the switches and spring terminals. 3D models of the spring terminals are not very good, but they are about the right size:-

 

[SHMD]

Hi Suzie,

 

Might I suggest that a gap is placed between U3 and U4?

I notice that the IC sockets have been placed next to each other. Quite a lot of DIL16 packages actually overlap more than the 0.1” pitch pin spacing between the sockets will allow. Might I suggest an extra 0.1” be spaced between the two IC sockets?

 

Also, the switches are very large and will only need to be set once. They are too easy to be hit! Might I suggest that they are replaced by DIP switches/jumpers?

 

Only observations, and by the way, great work.

 

 

Regards,

Kev.

(Can I have a look at the schematic please?)

[Suzie]

The schematic is in post 25.

 

The opto isolators are fine butted up to one another. This is normally the case with 8-pin and 16-pin devices, it is 14-pin devices that need the spacing because the body of the IC is the same as for a 16-pin making it overhang either end by 50 thou. As you can see from the image created from the 3D models of the components - they fit. You can use 4x TLP621-2 (8-pin devices) instead if required, they will fit as well and might help with quantity ordering (you need twice as many so will hit bulk discounts more easily).

 

My personal opinion is that the switches are unnecessary. I suspect that for the most part people will not fit them. Given that the reason for fitting them is that it is too fiddly to change the wire from one terminal to the adjacent one it would make sense that they are easy to operate, dil switches or jumpers will obviously be more fiddly than moving the wire between adjacent terminals. The other aspect is that I have put the switches between the track and the relay which means that they take the track current (2A max with the specified relay) so that I could use simple SPDT switches. If people would prefer a small jumper block to change the polarity of each relay that could be arranged - but smaller switches if used would need to be DPDT type to swap the connections to the relay coils. No one has as yet specified which switches they would like, so feel free to offer suggestions.

[Kallaroonian]

I still need to reply to you on this Suzie. I thought I had provided a link to possible switches in an earlier post but looking back I see I didn't. The scematic I sent you via direct email had a simple wiring arrangement for 2 DPCO switches per latch in order to totally avoid the need to re-wire in situ.

 

Given othes are showing an interest I'll reply here too a bit later on. I am, unforrtunately, currently out of work so dealing with that issue is taking a priority.

 

Here's the switch I was thinking about : http://www.jprelec.co.uk/store.asp/c=1017/Subminiature-PCB-slide-switches. That thought process went no further than finding it online though really.

 

rgds

[Suzie]

Those switches are a bit big, have an undesirable centre off position, and would require a lot of relaying out to accommodate them. I have found some others that I think are more suitable, 800-966, The 800-966 switches are on a standard 0.1" grid so can easily be replaced with a 6-pin header bloc k and jumpers if desired, or a couple of links. here is how it looks now:-

 

 

I think I see what you are getting at with the extra bank of switches now. The function of them to reverse the solenoid connections would be better achieved with a CV in the decoder.

[Kallaroonian]

Hi. Sorry for the delayed response. I’ve been preoccupied with the family going to the UK and my ongoing search for new employment.

 

I have a number of comments and issues below. Long post, sorry.

 

Fundamentally I think we need to iron out all of the issues in this thread and then perhaps start another thread asking about interest in the finalised design. As part of that process we are going to need to provide a high level spec, the design and a parts list. Then we can see what level of interest, if any, we are looking at.

 

None of this feedback is intended to be negative. I think this is a great exercise and it’s certainly got way way further than I could have by myself. Nearly done in fact but I just want to do is get the design as robust and as “value-add” as possible for two reasons : a) because I’ll be paying for the prototype and I want to avoid the chance of doing this twice B) to spur as much interest as possible from others.

 

Here’s the comments and issues/question then, in no particular order :

1. Bearing in mind my experience is solely with the ECoS and switchpilots I found it time consuming, difficult and a bit detrimental to the lifespan of the little plug/header connectors they use to keep altering the connections. Mind you this was made a lot worse by having more than one wire in some of the terminals. It is also debatable whether the use of spring clip terminals and 1 wire per connector would minimise these sort of issues
   
2. That said the cost of adding switches to reverse polarities was minimal. The idea for two switches on the PCB per point motor was to avoid the frustrations above. One switch is intended to reverse the solenoid connections in the event that it operates back to front and the other is intended to reverse the frog polarity in the event that it is back to front. The way this worked in the schematic I sent you was to have the first switch swapover the DCC control signal to the whole circuit and the second swapover the DCC signal to the frog switching latch.
   
3. The advice in the swpilot instructions is that if the solenoid operates back to front then you need to “simply” swap the wires. You commented that this would be better achieved with a CV in the decoder. This is not achievable with the swpilot at least - is this ability built into the decoder firmware we are talking about here ? Otherwise a second switch might be preferable
   
4. So far you have used a single switch to swap the frog polarity. I don’t understand how you have wired this using SPDT rather than DPDT per the schematic I sent you.
   
5. As a side note it amazes me that esu don’t provide a frog polarity connection on the swpilot; you have to buy the swpilot extension. Maybe that’s the point…..
   
6. The way I have the latches connected on my stripboard prototype has the DCC Track +/- each connected to one side of one latch and then the frog polarity output is taken off alternate sides of each latch and connected together for a common frog output. This is why I use only 8 terminals on the latches rather than 10. I think you have used all 10 and included diodes. I don’t think these are necessary
   
7. Happy to see toggle switches rather than sliding switches
   
8. In terms of connectors we need to find options that have the same pin spacing and then provide sufficient physical space as required. Then you just install the one you want. Personally my least preferred option would be screw terminal blocks. I am still torn between plug and header terminal blocks or spring clip terminal blocks. The reason I don’t like regular terminal blocks is that you have to be able to get a screwdriver to the PCB in situ. This might be a real pain. With plug and header the wires can be screwed into the plug away from the PCB and with the screwless term blocks obviously it’s just a case of inserting the wire. The issue here is finding a supplier who has all 3 types with the same pinout. I believe jprelec do however
   
9. The connection through the latches potentially carries the current draw of the loco traversing the point. While the current tracks may be adequate I don’t see any point in not minimising the risk. The PCB tracks into and out of the latches should be as wide as can be fitted onto the PCB
   
10. It is not necessarily that easy for me to source parts for this from here due to minimum order quantities, shipping etc. We therefore need to identify a parts list that can be sourced from a single supplier. I have used jprelec because I know they supply the Hongfa latches. I would assume the complete parts list will end up being a) DIY decoder kit + B) balance of parts all from one supplier.
    
11. In terms of the PCB layout it is true that it should be kept as small as possible in order to reduce cost. That said a slightly larger board doesn’t add excessive price. Therefore I believe we should ensure the spacings of components and sockets are, if not generous, then at least not minimal. And…..
    
12. We need to establish a mechanism for mounting the PCB. I take your point about the potential for the latches to fall out of the sockets. This suggests the latches should face upwards or the use of specific latch sockets that I believe exist. My preference is to stick with regular DIL sockets but provide for the PCB to be mounted horizontally. I suggest three holes down the centre line of the PCB so that it can be mounted using screws through spacer standoffs rising vertically past the components. Hope that makes sense. Three holes simply to provide some mounting options. I believe 1 central fixing would be adequate but 3 gives the option of using holes 1+2, 2, 2+3, 1+3.
    
13. “This decoder does not have the low power limitation of the Switchpilot. It will provide at least 10A per output as long as the power supply can provide it, with no recovery time. It will be enough to switch a pair of solenoids on each output.”. How does it provide 10A?
    
14. Great diagram showing how it all goes together is the grand scale of the layout. It has messed with my understanding or wiring the solenoids however. Am I correct is saying that the way I wire things one side of each solenoid is ground/common and the decoder provides the voltage to complete the circuit …..whereas ….. in the arrangement you have shown the power to the solenoid is direct from the PSU and the decoder switching is providing a switched return path to ground ? And this answers my earlier question about how to use a single CDU across a whole layout?
    
15. A question arises about how to connect the orange wire to the solenoids in practice. It would be easier to have 3 wires per point motor connected at the PCB ie a 12 way connector for the solenoid connections. Meaning it would be more straightforward to connect the CDU/PSU once to each PCB and then separately make the point motor connections as a group
    
16. Ideally all the connections for one point would be grouped together ie 3 connections to solenoid and 1 to frog ie 4 x 4way connectors
    
17. I only saw the need for 4 frog polarity connections. Can you expand on the diamond, crossover reference
    
18. On my prototype board I have one power connection for the PSU and one DCC control connection taken from the track feed. I don’t think we need – per your diagram – two red and black connections to the PCB. Why not make only one connection?
    
19. Rates I have mentioned are for a 2 layer board so if this is what is necessary to resolve above layout issues then no problem
    
20. Need to get labels for each external connection. Nobody wants to get frog output 2 hooked up to solenoid 1 for example. I did this while trialling my stripboard and it took me a while to figure out what was going on
    

 

When we’ve finished discussing all this I’ll want to check the PCB design and complete schematic to make sure it’s all straight before committing dollars to the proto board

 

Thanks

 

Mark

[Suzie]

All feedback is good. Lots of things to look at,

 

2+3. I will add the second bank of switches. On reflection setting CVs at to change the polarity will not be practical since the decoder will already be installed and operations mode programming will not be possible (decoder does not support it and probably not worth adding because most command stations do not support it either).

 

4. The relay does the DPDT function, the switch was just swapping the frog between the complementary outputs from the relay.

 

5. In the same way that most modellers don't bother fitting auxiliary switches to the bottom of their solenoids, most people probably do not want auxiliary frog switching. Frog switching is just for us discerning modellers, and I guess ESU see this as a value adding exercise.

 

6. You have used the relay as a pair of complimentary SPST contacts, not a very conventional way of using a DPDT relay in a changeover function, but a perfectly valid way of achieving what you wanted.

 

8. The spring terminals are a superset of the pinouts for the screw terminals and removable block screw terminals, so laying out for the spring terminals allows for all options to be accommodated. If looking for alternatives which will fit pitch between terminals is 0.2" (or 5.08mm) and pitch between the two rows of pins is 0.4" (10.16mm).

 

9. 50thou (0.05" or 1.27mm) is the maximum practical track width and is more than adequate for the 2A rating of the relays carrying the track power to the frog. The short distances involved should offer little voltage drop. Track lengths will increase of course if neat groups of terminals for each point are required.

 

10. I will try and put a parts list together for at least Farnell because they will be able to supply everything, and I will try and do Rapid and JPRelec as well since they are popular. RS is another good supplier with a global alliance like Farnell but their website is hard to navigate, the parts are probably in there somewhere!

 

11. We know how big the parts are and 3D modelling shows what fits and what does not before the prototype is built, so it will all fit together fine. See pictures above of the 3D virtual model. I can spin this around in real time to get a good look at how it all fits.

 

12. PCB will have holes in the four corners for mounting. It will need more than one fixing realistically. I think most people will not socket the relays because there is no reason for them to be unreliable. Using sockets in this situation is over engineering - it is creating real unreliability (the socket interface) in order to cure a perceived unreliability (that the relay will not meet its manufacturers stated reliability) that does not in reality exist. the relays are not switching a load and are unlikely to be handling current anywhere near the rated maximum so should last well in excess of the manufacturers rating anyway. I would never socket these relays unless I was anticipating reusing them in the short term. I have never had a relay in this format fail.

 

13. The decoder just passes the power that is supplied by the DC power supply or CDU. I have tried it with a peak current from a CDU of around 16A and it was fine. The manufacturers rating for the output transistor is 8A so I would recommend that you do not exceed that of course, but in practice you can operate a pair of Peco solenoids with short pulses from a CDU OK. If using the decoder for continuous output (which it can do if required) you should not exceed 5A (per output) or it will overheat.

 

14. Yes. The solenoid common is the positive supply from the PSU, the decoder switches the negative side.

 

15. An extra set of four connections can be provided for the point motor common terminal to make wiring easier.

 

16. I think 4x 5-way connctions is best with the two frog connections.

 

17. Here is a little diagram showing the typical types of scenario where dual frog outputs will be required:-

 

 

As you can see with or without the diamond the two points A and B should be operated together. There is no point in reversing point A if point B is not going to be reversed and vice versa. In which case it makes sense to operate them both from the same accessory decoder output. When doing this you will note that frogs A and B are in opposite rails and will therefore need complementary frog outputs if being operated from a single decoder output.

 

The same is the case with a diamond. When point A is set normal, the two frogs of the diamond will need to be set for the main line, frog C is the same as frog A, but frog D is in the opposite rail. When point A is reversed the diamond will need to be set for the crossing. Frog C is still the same as frog A but in the opposite rail this time, and the polarity of frog D will need to be reversed too.

 

Here is a reverse loop diagram:-

 

 

As you can see, as well as the frog (and inner rail that it is connected to) you need to switch the outer rail of the loop too, and it is the opposite polarity to the frog. This is a much simpler solution than using a reverse loop module.

 

There are plenty of situations where you need both frog polarities.

 

18. Using the accessory bus to feed the frogs will result in the accessory bus being taken down in the event that a short circuit occurs on a frog. If the accessory bus is not available it will not be possible to change any points, including one that may require changing to clear a short if a train has gone up against an incorrectly set point. I suspect that on a large layout with 120 points there is a possibility that this might happen occasionally, so it would be a good idea to follow best practice and separate the track feed from the accessory bus with a booster or circuit breaker. You need to be aware that when using latching relays to power the frogs changing the point by hand will not clear the short, it will have to be changed by the decoder using DCC or the train will have to be moved by hand. This will not be easy in any hidden areas for example. Best practice really should be followed in this case. I guess that it is possible to put optional jumpers on the PCB to allow the use of a single pair of red and black wires.

 

19. If you want ergonomically grouped outputs it will be necessary to use a double sided PTH board.

 

20 The terminals can be labelled with what they are quite easily if you have the printing done when the PCB is manufactured.

[SHMD]

Hi Suzie

 

2&3/

Not having “Online” CV programming will also help reduce/eliminate the number a Decoder memory corruptions happening.

 

8/

Terminals with 5.00mm spacing are becoming more “popular”. Just a caution that's all, along with ensuring that the holes are big enough!

 

10/

Agree about the RS website. The only one worse is the B&Q one!

 

12/

Totally agree about NOT socketing the Relays – especially since some of these Accessory Decoders will be mounted upside down under the baseboard!

 

19/

In my experience, the cost difference is small between single and double sided PCBs is small or even not available!

It will, however, add much to the simple usability, (ergonomics), by members who are not, shall we say, electronically gifted!

But it will add extra work, not just in laying out, but also in checking the PCB before submitting the (Gerber?) files.

 

20/

I can not stress the importance of this one too much!

Given enough space, I label each terminal on the silk screen AND in copper on the bottom layer (reversed). As big as you can(!) bearing in mind that a lot of us are older and/or with failing deteriorating eyes!

 

 

As an aside, I am interested in this thread, but I have no need of the finished product. OK, a couple as a platform for DCC programming and interfacing.

 

 

Good work – and fast!

Kev.

(Here to help, hopefully!)

[Suzie]

I am not going metric for through hole components, only surface mount, so 5.08mm (0.2") is what it will be, and that will provide three different terminal options, pluggable block, spring terminal or normal screw terminal.

 

Latest itteration is 4.5" x 4.7". It now includes both banks of switches, and four identical connection blocks for the outputs. There was a spare corner on the PCB so I popped in a DCC track powered CDU. It only needs to be populated on one unit for the whole layout. If the CDU components are fitted the CDU terminals then become an output rather than an input.

 

Each point now has a 5-way terminal block providing normal solenoid pulse, reverse solenoid pulse, common for solenoid, frog connection, and second frog connection. If the second frog connection is not desired a 4-way block can be fitted instead.

 

There is a dedicated 6-way block for the CDU, track bus and accessory bus connections. Again a 4-way block can be substituted if the seperate accessory bus is not desired - there is now provision to fit a pair of links to use the track bus to provide the drive for the decoder if a 4-way block is used.

 

Here are the pics of the latest CAD model:-

 

 

 

[StuartM]

All of the above looks rather complicated just to switch a frog

I don't know it this was the sort of thing you had in mind (see photos below)

I made a pcb up to control two dpdt relays (non latching) which are controlled by a Picaxe microchip http://www.picaxe.com/

I made it to control the polarisation of a diamond crossing and a double slip but the chip will switch the relays on/off as you wish and could switch 4 frogs The relays and the chip all plug in to dil sockets so are easy to replace, the small capacitor filters unwanted interference across the chip power supply and the led just proves everything is working as its switched on via a one line code instruction controlling one of the chips output pins

The red wire on the left is 5vdc, the blue is 0v and the purple wire is the data bus. The connections to the track are at the bottom.

The pcb could be redesigned to omit the chip altogether and just have power supply's to switch self latching relays, but having the ability to operate the chip and therefore the relays via a data bus suited my needs.

http://www.rmweb.co....m/page__st__100 if your interested

[Suzie]

Something tells me there is not quite as much functionality in your little circuit.

 

The requirement is to provide non volatile frog switching with both simple connectivity and have reversing switches to simply change the frog polarity and the solenoid direction. Adding the DCC decoder has added little to the cost but replaced the need for a seperate ESU Switchpilot decoder and all of the inter-wiring that would be required.

 

A standard servo version would be a lot simpler (and cheaper)! But the requirement is for solenoids and easy configuration hence all the complexity of latching relays, high current drivers and switches. Lots more functionality than a switchpilot and companion, but £40 less per unit.

 

p.s. You are pushing it running those relays on the 25mA PIC output. They really should have 40mA (from two outputs in parallel for example) or the output voltage is likely to drop to around 3V.

[StuartM]

Something tells me there is not quite as much functionality in your little circuit.

 

The requirement is to provide non volatile frog switching with both simple connectivity and have reversing switches to simply change the frog polarity and the solenoid direction. Adding the DCC decoder has added little to the cost but replaced the need for a seperate ESU Switchpilot decoder and all of the inter-wiring that would be required.

 

A standard servo version would be a lot simpler (and cheaper)! But the requirement is for solenoids and easy configuration hence all the complexity of latching relays, high current drivers and switches. Lots more functionality than a switchpilot and companion, but £40 less per unit.

 

p.s. You are pushing it running those relays on the 25mA PIC output. They really should have 40mA (from two outputs in parallel for example) or the output voltage is likely to drop to around 3V.

The relays are TTL compatible and work fine.

If you want cheap, make some diode matrix's instead

keep things simple

Anyway that's my contribution, I'll leave you to it

[Gordon H]

The relays are TTL compatible and work fine.

They might be OK with actual TTL open collector buffer drivers with sufficient current sink capability, but they are certainly not a good choice for long term use with standard PIC port pins for the reasons that Suzie stated.

The high sensitivity 150mW coil variant would be better, but even this draws 29mA which is still beyond the PIC pin spec. You might find that the 6V 150mW coil type will switch OK, and that only needs 21mA.

[Kallaroonian]

Folks

 

I started this whole thing. The original design which I built on stripboard uses 4 DPCO DIL latches to effect the frog polarity swapping of 4 points alongside the associated esu switchpilot.

 

It works fine. However building stripboard for multiple units and the hassles of wiring in situ under the board led me to look into getting PCBs made. I found this was not too expensive. I then sought interest from others because the unit price goes down a lot when the order is bigger.

 

The level of interest was limited but a dialogue started with Suzie about adding value to the design by combining it with the actual decoder. It's early days for me with quite a large layout so I am not wedded to any particular brand. This idea looked good to me because :

a) it reduced the overall costs even further

B) by combining the devices together we reduce the number of interconnections

c) it provided a mechanism whereby somebody else stepped up to do the PCB design when I was struggling to get motivated to learn some arcane PCB design software

 

The current status is trying to make the combined unit as functional as possible. It is of course highly complicated to just switch a frog but that's because things have moved on a long way since the thread started.

 

The refinement process is essentially still underway and the input of others it also highly valued. It will be good if the device can control frogs in aldders of course. Others may have other ideas that may or may not get adopted (indicator LEDs might be another thing but you have to draw the line somewhere..........). If you are wondering about the switches these are included because they add little cost but lots of convenience; nothing more annoying, I found, than wiring everything up, finding something is back to front and then having to rummage beneath the boards all over again.

 

I daresay it will be after Christmas before this is finalised.

 

At that point two things will happen. I will source the PCB prototype, build it and test it. With Suzie's remote support I daresay. I will also start a new thread to determine the level of interest in the PCB so that when an order is placed others may benefit. Some kind of capability statement will be provided to help people understand the capabilites and limitations (primarily that you assemble it yourself and there's no tech support)

 

For my part I am expecting it to be highly cost effective. I am still nervous about the actual decoder part of it because I have no experience with the donor design and kit this is taken from vs the simplicity of just buying a switchpilot but we'll cross that bridge when it arrives.

 

Regards

 

Mark

[Suzie]

I have added red and green indicator LEDs now to show the current state of each output relay. they are powered by the track, so also show if track power fails. I have added a LED to the CDU as well to show that CDU voltage is present.

 

PCB size is now 4.7" by 4.8"

[Kallaroonian]

Hey.

 

:-)

 

I'm not actually sure LEDs are worth the bother. I was just chucking the idea out there to see what others might have ideas about. I think if the frog polarity is the wrong way around you just flick the switch. An LED is too much.

 

An LED to show the board has power might be useful? I don't know. Let's think about what will helkp in practice

 

rgds

[Suzie]

Here is the latest picture showing the LEDs.

 

 

There are three potential sources of power:-

 

1. CDU/DC Power Supply for the solenoids (CDU terminals).

 

2. Track supply for the frogs (Track terminals).

 

3. Accessory bus which powers the decoder and CDU (Track terminals).

 

The white LED currently shows that the CDU supply is present (and the accessory supply on any board that has the CDU components fitted), while the red and green LEDs show that the track supply is present. It would potentially be beneficial to add another indicator to the 5V supply to show the presence of accessory bus power on decoders which do not have a CDU fitted and are using a separated track and accessory bus. Any thoughts on LED colours is welcome, I only have 3D models of Red, Green and White 5mm LEDs at present! I suspect that Yellow or Blue might be in order. I have specified 5mm LEDs because the red and green ones need to be bipolar. If you do not want LEDs they do not need to be fitted - it will work fine without them.

[Kallaroonian]

I need to digest the updates since my last mammoth post and, more to the point, type up a response. I'll try to do this in the next couple of days.

 

In the meantime though let's talk LEDs a little more re the post above.

 

Firstly as far as I'm concerned there should be two power supplies to this device : i) the CDU or DC power supply ii) the DCC signal [which in my case at least I will just take straight off the track]

 

As far as LEDs are concerned I can see a small benefit of one to confirm each connection above ie power supply and DCC signal. That should be just two LEDs and they should be optional fitment anyway. There's no point having anything to show frog output polarity because it wouldn't mean anything and could be flicked over with the toggle switch anyway.

 

I'm also not sold on the idea of including the CDU on this PCB. Bearing in mind it started as just an array of latches it is now turning into an uber device that performs too many functions.

 

More later !!

 

Regards

[Suzie]

If you only want to feed both the frogs and the decoder from the same input, you just fit a 4-way block instead of the 6-way block at J5 and install the jumpers at J6 (just below the 6-way block at the top). This links the track bus and the accessory bus.

 

All of the LEDs are optional, and have just been put in the blank space that would otherwise have been unused between other components. The red and green LEDs show the state of the relay rather than just the state of the switches so will change as the points are changed, and will change if local push buttons are used to control the points outside of the decoder which might aid fault diagnosis.

 

The CDU is only five components and just made use of an otherwise blank area of the PCB. It was silly not to fit it! It is again purely optional. It will save the cost of buying and wiring an external CDU and power supply for very minimal cost by just installing the bulb, two capacitors and two diodes.

 

I think if you want to sell a few PCBs it will be a good idea to retain the flexibility to support a separate accessory bus, and retaining the pads for the CDU components offers the potential of a big saving on a points power supply.

[Suzie]

Here is today's picture. I have improved the 3D models of the Camden Boss screwless connector blocks and the G4 bulb in the CDU and added the blue LED to show the presence of the DCC track bus or accessory bus signal to the decoder, the white LED indicates the presence of the DC power for the solenoids.

 

 

I have put a yellow ring around the optional CDU components and a blue ring around the decoder components to give an indication of what is what on the PCB. If just a frog switching board is required to work with another decoder both of these sets of components can be left off. Leaving them off will not reduce the cost by much since they are all relatively inexpensive components - it is the switches, relays and connector blocks which cost.

 

I have compiled a parts list for JPRelec, Farnell and Rapid. It looks like JPRelec are the probably the cheapest and with a few compromises they can supply all the parts. The optional red and green Kingbright LEDs are only available from Rapid along with the G4 bulb holder, and the 10W G4 halogen capsule bulb should be obtainable locally in your lighting shop.