Skip to main content

Stan,

I found my MTH traffic signals, a two signal set that I was thinking was a four signal set? Also, it was in a type of blister pack and I was thinking it was in a box. I also ordered some of the We_Honest traffic signals. It appears all I have ever ordered were the dwarf signals for switch positions? Both of these items are a bargain! Quality is good too, at least on the dwarf signals I have. 

Also the MOSFETs seem to be MIA, still have not arrived so we still have only the transistors to fiddle with. 

It appears that the MTH signals are common cathode so modifications would need to be made on the common anode PCBs we have from the current project. I would be willing to make another PCB for these, if the current ones can not be made to work with both types of signals.

The wire colors on the MTH traffic signals are as follows:

Green - Common Cathode
Blue - Green LED
White - Yellow LED
Yellow - Red LED

One thing MTH thought of that I did not is an 'On-Off' switch on the signal controller PCB/Controller. Might be a good idea for an addition to the PCB.

Something interesting turned up while fiddling with the MTH signals, I lowered my DC power supply to 3 volts and tested the MTH signal LEDs directly without their controller. The current 4 wire circuit using transistors (per Stan's earlier recommended revisions) was still on the bread board and also being powered. It was working on 3 VDC! The lights were not quite as bright, but it was still working! 3 VDC might be a bit low, but I am now wondering if instead of the 9-12 VDC I was previously using, 5 or 6 VDC would be sufficient and I should switch to that permanently? 

Here are some pictures of the MTH traffic signals for you, for reference, if you don't have them already. Also, not sure what type of connectors these are? Maybe someone will be able to ID them and post the info?

IMG_1579IMG_1580IMG_1582IMG_1583IMG_1584

Added later after more fiddling:  I don't know what those large resistors shown in this picture are doing, but they get pretty HOT!

Attachments

Images (5)
  • IMG_1579
  • IMG_1580
  • IMG_1582
  • IMG_1583
  • IMG_1584
Last edited by rtr12

1. I don't get it.  One MTH photo shows 4 flying-lead bare-wires and the other shows the 4 wires in a connector plug?

2. What's the space on that MTH connector - looks like 2.54mm (0.1")?   That looks like a standard connector family but that's easy for me to say!  As you know you go to digikey and try to punch in a few parameters and still come up with a zillion choices.  Maybe someone in the know will identify the manufacturer and series designation.  But in general, as long as you get the pitch correct and leave enough clearance, if/when you find the right connector it should drop into the PCB.

If re-doing the PCB, I suppose you should change the order of the 4-terminals from G-Y-R-com (as is) to R-Y-G-com (MTH style).

You might also look to see if a standard square-pin male connector (non-polarized) securely mates with the MTH plug.  That is, this saves the trouble of tracking down the connector family.  So you lose the benefit of polarity and would have to be mindful to plug in the MTH connector plug in the correct orientation.

3. The Common Anode (CA) vs. Common Cathode (CC) is interesting.  First off, the original design was a common cathode configuration.  That is, you could simply take the 6 signals (RYG E-W, and RYG N-S) BEFORE the 10K resistor and Base input to the transistors and attempt to drive the MTH CC signal head(s).  In other words you could do this today with the hardware you have.

4. But, if this is too dim or you want a dual-purpose (CC or CA) alternative with what I understand is your latest version of the PCB:

traffic light CA vs CC using NPN drivers

a) install the 6 transistors "backwards".  That is, rather than CBE for pins 1-2-3, rotate the NPN transistor 180 degrees.  The B is still the middle pin but you are swapping the E and B terminals of the transistor.  This makes the NPN transistor into a non-inverting buffer using the so-called "emitter follower" configuration.  So you get the benefit of additional current driving capability (i.e., the buffering) that a transistor brings to the table.

b) as shown, the 6 transistor emitter at the top which previously went to GND for CA now go to VDD.

c) as shown, the 4 resistors at bottom which previously went to VDD for CA now go to GND.

Note that unless your PCB layout software did something weird, you should be able to cut a single trace for b) and c) and on the bench simply tack the wires to the opposite power supply rail (VDD become GND, GND becomes VDD).  Or if re-designing a PCB, you'd make, say, a 3-terminal 0.1" header with the VDD on pin 1 and GND on pin 3.  Then use a push-on programming jumper to steer pin 2 to either 1 or 3 depending on the CA vs. CC application.

d) lower the value of the 6 resistors from 10K  for CA to, say, 100 ohms for CC.

e) there is a somewhat analogous method to use N-channel FETs to implement what's know as a source-follower non-inverting buffer (vs. the NPN transistor emitter-follower non-inverting buffer) but it gets tricky so I don't want to go there.

5. The on-off switch function should not require a PCB modification - just cut power to the board with a toggle switch.

6. If the lower voltage works I see no problem operating the circuit at 5V or 6V.  You will need to lower the values of the 4 resistors in the common leg of the LEDs.

I think I addressed the issues at hand. Whew!

Attachments

Images (1)
  • traffic light CA vs CC using NPN drivers

1. The MTH signals have a separate harness for each signal and power that plugs into the controller. The signals themselves have their own, much smaller wires coming out of them. I should have included this picture in the above post.

IMG_1589 

2. The spacing is 2.54mm as you said above. I haven't yet looked on Digikey, but it's on the to-do list. Maybe if I stall long enough GRJ will be by to ID the connector type, he's awfully good with connectors, among other things. Of course looking for them on Digikey might also be interesting.

Thank you for the detailed explanations for the rest of this and I really appreciate the time you take to explain it all so thoroughly. It's a big help in learning.

This morning I was thinking about a couple of the things you covered here, like swapping the wires (VCC & GND) and also some way of switching between the common anode and common cathode modes. I wasn't sure it would all work though. You have more than covered all this and then some here. Several things that I hadn't thought of or even knew was possible or knew to do as well. Resistor value changes, turning the transistors 180 deg, adding jumpers, and everything else you pointed out.

Now if those MOSFETs would show up. I did learn something interesting here though... Contacted the seller, he said he would resend or refund if they don't show up soon. But here's the good part and an FYI (if you didn't know this already, I didn't?), he told me ebay was the one that offered the guaranteed arrival date and they would give you a $5 voucher if that date was not met. Not sure how they can do that, but I now have a $5 credit for the next electronics 'gadgets' purchase!   But no MOSFETs... 

I printed out your post and I'm off to the study lab for more learning and experimenting! I'll report back with the findings (or more questions, hopefully not too many).

Attachments

Images (1)
  • IMG_1589
Last edited by rtr12
rtr12 posted:

1. The MTH signals have a separate harness for each signal and power that plugs into the controller. The signals themselves have their own, much smaller wires coming out of them. I should have included this picture in the above post.

IMG_1589

I'm confused.  If the MTH signal heads have 4 bare wires as shown in your photo, is the end-user expected to make 4 connections to the cable with plug for use with their controller?  In your case, since your PCB has screw-terminals to go to the wehonest signal heads, then why do you need to track down that MTH connector?  Can't you just take the 4 bare-wires from the MTH signal head and attach them to the screw-terminals on your PCB?

 

You are correct, it appears MTH expects you to fill in the missing middle wire section or, at minimum, splice the signal and connector wires together. 

Yes, the screw terminals could easily replace the MTH connectors. I hadn't gotten around to figuring out that the PCB's screw terminal block could be the same as their connector, doh. I was still thinking about wiring changes, pins for the jumpers and other suggestions from your previous post. You are always a mile ahead of me, thinking of these things before I even figure out I need it.

And FWIW, I briefly scanned Digikey for connectors matching MTH's. Quickly gave up, there must have been at least a million of them! It was a real nightmare, so not needing them is very nice! 

stan2004 posted:

 

4. But, if this is too dim or you want a dual-purpose (CC or CA) alternative with what I understand is your latest version of the PCB:

traffic light CA vs CC using NPN drivers

a) install the 6 transistors "backwards".  That is, rather than CBE for pins 1-2-3, rotate the NPN transistor 180 degrees.  The B is still the middle pin but you are swapping the E and B terminals of the transistor.  This makes the NPN transistor into a non-inverting buffer using the so-called "emitter follower" configuration.  So you get the benefit of additional current driving capability (i.e., the buffering) that a transistor brings to the table.

b) as shown, the 6 transistor emitter at the top which previously went to GND for CA now go to VDD.

c) as shown, the 4 resistors at bottom which previously went to VDD for CA now go to GND.

Note that unless your PCB layout software did something weird, you should be able to cut a single trace for b) and c) and on the bench simply tack the wires to the opposite power supply rail (VDD become GND, GND becomes VDD).  Or if re-designing a PCB, you'd make, say, a 3-terminal 0.1" header with the VDD on pin 1 and GND on pin 3.  Then use a push-on programming jumper to steer pin 2 to either 1 or 3 depending on the CA vs. CC application.

d) lower the value of the 6 resistors from 10K  for CA to, say, 100 ohms for CC.

 

Ok, I am confused now, self inflicted I'm sure.   The dual purpose idea you have in in 4. and the Note paragraph sounds perfect! I think that is the way to go. I added some pin headers and jumpers to the bread board circuit and was trying to make this work with the mods in b) & c) (along with notes on the modified drawing included). If I am interpreting this correctly, I have b) 3 pin header switching the transistor's emitters between VDD & GND and c) 3 pin header switching between VDD & GND at the Common resistors and another. 

I am confused about reversing the transistors, seems like that should be a separate mod (two separate PCBs) and would make using the pin headers kind of difficult, unless that is a permanent reversal for both CA and CC modes?  Without reversing the transistors, it's working in the CA mode with headers & jumpers installed, but no luck so far in the CC mode. Going to try it again tomorrow, retracing my wiring mods on the bread board, the BB is getting a bit crowded along with some self induced confusion while making changes.   

On the brighter side, no smoke has been released!  Also, the mailman hit the trifecta this afternoon, got MOSFETs, JLC PCBs and We_Honest traffic lights in the mailbox! Been fiddling with all that, which could be adding to my confusion...too much fiddling going on at once.

npn flip

In the original CA (Common Anode) design, as you point out, all the NPN emitters are tied together as a single PCB "net".  This so-called net is then hooked to GND.  The NPN collectors go to the LEDs..

What I'm suggesting is you simply install the NPN backwards.  Now, all the NPN collectors are tied together in a single net.  And this net is then hooked to VDD.  The NPN emitters go to the LEDs.

The "trick" is you don't have to change the PCB or breadboard wiring to the transistors.  You simply rotate the transistors when installing them.  Your NPN transistor is of the TO-92 package which has that "D" shaped body which I attempt to show in the diagram.  

It sounds like you correctly figured out what I meant by making a pair of 3-pin headers to steer the VDD or GND accordingly. 

To be sure, this is a DIYer's hack.  You'd likely get fired if you tried this in a manufacturing company.  I figured you had boards made and this would be a way to make them work with either polarity of signal head with only a modest amount of fuss.

As mentioned earlier, do not try this trick with the FETs.  There is an analogous idea but it has too many rabbit holes that you can stumble on.

Attachments

Images (1)
  • npn flip

Ok, I think I get it now on the reversing transistors, just to be able to use the boards I already have. Got sidetracked yesterday, company came over...so I only got drawing revisions started, finished up today. This is the schematic for new PCBs with the jumpers to switch between CA & CC modes. I also changed the RYG & C wiring around as you suggested, good idea and I like that better too.

I think I have everything correct, including the wiring to the header pins. But, I think my jumper placement instructions are somewhat confusing? I was trying to make the jumpers for each mode the same on both headers to simplify things. In trying to eliminate confusion, I may have made it more complicated? I've looked at this a few times and sometimes it makes sense, sometimes not so much?  Words of wisdom from the much wiser would be helpful here.   Telling me I have it all wrong is ok (and appreciated) too...

4-Way Traffic Signals Schematic-Transistors v4B

The PCBs are very nice and the transistor PCBs work just fine, but there are a couple of small glitches in the PCB layout and component spacing. I am not good at that just yet. I crowded the wire terminals a bit and the terminals covered up all the "what wire goes where" writing on the PCB, oops. Some PCB enlargement is in order for the next batch, if the writing is to be seen. JLCPCB allows up to 100mm x 100mm at their $2 for 5 PCBs rate before the price goes up. That allows some wiggle room. (Definitely won't be fitting in any O gauge engines like GRJ's stuff does though.) 

I haven't yet gotten to bread boarding or assembling a PCB with using the MOSFETs, but I will be following your advice and NOT trying the CA/CC switching jumpers on that one. I don't want to end up in any rabbit holes and it's probably too complicated for me to stay out of them. Thanks for the warning before I did get any hair brained ideas about trying that.

Thanks again for all your help Stan!

Attachments

Images (1)
  • 4-Way Traffic Signals Schematic-Transistors v4B

The jumper instructions look correct.  To be clear, we're talking about a DIY hack to use the same PCB board for either the CC or CA.  No doubt one year from now we'll look back and wonder what were we thinking (or smoking)!  

Anyway, to perpetuate the hack, I suggest you place ALL instructions for the conversion in a neat table on the 1-sheet schematic.  The table would itemize the differences between CA and CC.  That is, identify the 3-pin jumper settings, the change in the 6 base resistors, and the orientation of the NPN transistors.  As I think about it, since you must "commit" to CC or CA by soldering different components anyway, there's no added value to having the 3-pin jumpers with those push-on programming connectors.  In other words you need the pads and solder a jumper wire from 1-2 or 2-3 accordingly.  Or, if we're in learning about DIY PCB mode, another technique actually used more than you'd think is to make the 3 pads close enough to make it easy to place a blob of solder to bridge 1-2 or 2-3.

Again, if messing around with PCB design, a very common practice for multi-use boards is to use the white silkscreen to identify which board type you assembled.  So you'd print a white circle or square labeled CA and another marked CC.  When you assemble the board, mark it accordingly.

 

While reading this, the hair brained idea of a dip switch (or similar) popped into my head.   Probably too many chances for error in the settings with dip switches (having two modes selected at the same time). But, the wheels kept turning and I found a 6PDT switch at Mouser for 89 cents. They have a few more to choose from while keeping the price under $2 or so. (Not to worry, I didn't order any. )

I think having the separate PCBs is just fine and will probably never do anything different (no 6PDT switches). It's just fun to consider the options and fiddle with the schematics. Also helps in learning a thing or two and I usually figure out how to do something new in Diptrace each time if I fiddle with it.

I still haven't perfected silkscreen info and getting it properly placed on the PCB, but I do like your suggestion of marking the PCBs with the assembled board type. I'll work on that one some more, maybe something useable will be achieved.

The PCBs I have fit my needs for now, but I plan to modify the design files for a couple of different types of PCB for more practice and learning and will post the complete files when finished, tested working properly. This is a really fun project and great learning experience.  

I was afraid the dip switches would cause problems in the way they could be set. So I was thinking a C form type, being either one or the other, might eliminate those errors. However, after fiddling with this for a bit it seems a tad more complicated than I first thought.  Maybe not such a good idea...

Now it looks like one DPDT switch per transistor...Ran out of switch contacts. The parts are multiplying! Makes that a worse idea. So much for my thinking of things on my own...

Last edited by rtr12

Well, suppose the objective is to have only one control toggle (or whatever you want to call it) that selects between CA and CC.  Here's one method:

3-aspect signal controller

This applies a so-called Exclusive-OR or XOR gate which is about 10 cents per gate in small quantity.  These are 5V logic IC's so obviously we are restricted to 5V operation - here's the 74AC86 from DigiKey that has 24 mA of high or low level drive (74AC family has symmetrical capability) so it can directly drive a pair of signals.  I'm figuring each LED will be plenty bright at, say, 5 mA or so.

74ac86

They come 4 gates per 14-pin DIP so you'd need 2 for this application.  I show a SMT version which might be a good way to start messing with SMT devices.  They also come in regular thru-hole DIP.

The diodes on the left are as-is from existing circuit.  There is a single 3-pin "jumper" as before shown on the bottom. 

When in the CC mode (pins 2-3 connected), GND is applied to each XOR gate.  This makes the gate a NON-INVERTING buffer so the outputs drive a CC LED head.  The common resistor is also tied to pin 2 of the jumper so this means the LED common returns to GND as it should be for CC.

When in the CA mode (pins 1-2 connected), 5V is applied to each XOR gate.  This makes the gate an INVERTING buffer so the outputs drive a CA LED head.  The common resistor is also tied to pin 2 of the jumper so the LED common returns to 5V as it should for CA.

I show this simply in the spirit of discussion.  That is, you can't beat the price of an NPN transistor for a penny or two.  This method would cost maybe 75 cents more per board in parts.  It does have the advantage that all the parts can be soldered down and the polarity can be changed with just one jumper.  But it can only operate on 5V DC and cannot drive the 100+ mA per LED output of the NPN method.

Attachments

Images (2)
  • 3-aspect signal controller
  • 74ac86
Last edited by stan2004

Stan, this sounds very interesting, I might just give it a go. Pretty sure I have some CD4070s  XOR in a 14 DIP package, would that work? There's a chance I might have some of the 74xx86 also, but not sure of the AC part of the number as above? I'll look around. Why I am not sure, but I have been adding a logic chip or two here and there for a while now so I have a few on hand, but I am not sure of exactly what all I have so far? If I live long enough I might even learn what one or two of these are used for. 

GRJ, I did order a PICKIT 3 (I think it was?), the cheaper one. I don't know how to use it, but there are directions available and youtube (see above line about life span and learning all this).

After that I looked at the selection of PIC chips available and got completely lost as to which one to choose? I'll take the chip number you posted and maybe give those a go too. I'm definitely saving that PIC number. Lots of learning to be done on the PIC stuff too, as Sgt. Schultz would say, I know nothing!

I really do enjoy the LOGIC ICs though. It's both confusing, exciting and challenging all at the same time! (Not that the PICs wouldn't be as well.) Plus I learn more about using Diptrace at the same time and it's fun.   

Sure, if you have thru-hole CD4070B XOR gates gathering dust then why not give them a spin.  As discussed earlier, the CD4000B series devices have, say, 10 times less current drive capability than the 74AC series in a apples-to-apples comparison (operating at 5V).  Of course the CD4000B series will operate at 12V DC where it has more current drive capability so that's where I'd run them.  You should be able to drive a pair of LEDs in parallel at, say, 5 mA each with each CD4070B output.  The good news is if that doesn't work you can install 74AC86 devices and just lower the operating voltage using the same board.

I see the PIC chip GRJ suggests is an attractive $1.85 qty 1 at DigiKey.  Note that its output pins have 25mA symmetrical drive capability (5V DC supply) so they can directly drive a pair of CA or CC signal heads.  You'd then use an input pin to instruct the PIC to operate in CA or CC mode.  Note that the PIC chip would displace the 555 timer chip and all its attendant components.  And GRJ points out, you'd be able to deep-six the diode-logic with a few lines of software.  And so on.  The PCB would devote more real-estate to the connectors than to the electronic components!

Should probably start a separate thread if going the microcontroller route.  The Arduino crowd will want to throw in their 2-cents.  And away-we-go! 

Last edited by stan2004

I like the Arduino, and it would work as well.  However, for a one-PCB solution, I lean to the embedded uP model.  The downside of the PIC is you do need to invest around $50 for the Microchip PICKIT 4 to program them.  However, having a single board with all the electronics on it is a powerful draw, at least for me.  Another plus is the boards are smaller, have a lot less components, and thus are cheaper and easier to make.  Finally, if you screw up the logic, you can fix it with your computer, no chopping traces to alter the logic.

Stan, I have some SN74HC86N and CD4070BE ICs. I'll follow your instructions above and see how it goes. However, in my massive shop and parts re-organization project (currently underway) my entire assortment of CMOS ICs are temporarily MIA, but the 74xx ICs are right where they are supposed to be. Half the battle lately has been just finding things. Not sure the difference between the AC and HC versions, but I am guessing the HC will work too. I'll be looking at the datasheet anyway.

Some of GRJ's PIC chips listed above have been added to my next Digikey order. Somewhere around here I have the traffic signals working on an Arduino Uno from previous fiddling a few years ago. I'll save that and the PIC trials for another thread, good idea. It might be a while on that one though, I need to do some studying and learning with the PIC.

GRJ, as I said to Stan above, I have this for Arduino, but I like the idea for the PIC and one PCB solution as you suggest. I am guessing that the PIC code will be similar to the Arduino, if so it shouldn't be too difficult, but I haven't tried it yet either. Sometimes the 'doing' is a bit worse than it seemed in my first thoughts. I just need to learn the PIC pin outs and their functions. I don't suppose my PICKIT 3 will do for this? In that case I'll need to get a PICKIT 4, probably a good idea anyway. I don't mind that as I have always wanted to learn more about the PIC chips since I read some articles on the PICAXE a few years ago. All this should keep me busy (and off the streets ) for a while.

Finding my CMOS ICs might occupy quite a bit of time, I can't imagine what I did with them, I had quite a variety too, more so than the 74xx series stuff. I have been in some state of disarray with one thing or another since we moved almost 6 years ago! I have moved things around several times and so nothing has seemed "just right' as it was in our old place. You are probably already in much better shape than I am, but I'm sure you can relate!  

rtr12 posted:
...

Here are some pictures of the MTH traffic signals for you, for reference, if you don't have them already. Also, not sure what type of connectors these are? Maybe someone will be able to ID them and post the info?

IMG_1579IMG_1583

Moot at this point if using screw-terminal connectors on the PCB, but that could be the Molex mini-SPOX connector:

molex spox

If this is indeed the correct connector system, here's the 4-pin PCB-side vertical header from DigiKey.

I bring it up because it reminded me of another DIY PCB hack.  Note the SPOX is a 2.5mm (0.098") pitch while many connectors are 2.54mm (0.1") pitch.  So if you do the math on the connector pin diameters and PCB holes size, you can easily install either a 2.5mm OR a 2.54mm connector for short connectors like the 4-pin used in the traffic signal.  Some free-ware/share-ware PCB layout program may not have a library of 2.5mm footprints and if not used to creating new components, it can be expedient to use an existing 2.54mm component which every layout program has.

 

Attachments

Images (1)
  • molex spox

The 74HC family preceded the 74AC family...but has asymmetrical driving capability meaning in an apples-to-apples comparison, the common-anode signal will be a tad brighter than the common-cathode signal.  Otherwise, no problem experimenting with any of the 74 logic families - just be sure you limit the voltage to 5V DC!

Tom, the PIC has come a long ways, if you have the current version of the MPLAB-X, there is an I/O configurator that will greatly ease the setup of the I/O ports.  Since the PICKIT-4 costs the same as the PICKIT-3, silly not to get the latest tool.

The Arduino is programmed basically in C, and so is the PIC, so the mainline logic will be very similar.  The difference will be in the I/O configuration.

Stan, I think I got it, especially the 5 volt part. I looked up the 74xx series ICs and there must be 20 or 30 variations, AC, HC, LS, etc...This can be mighty confusing for a green horn. 

GRJ, the PICKIT 4 is on my next Digikey order, they had it for $47.95. The things you point out make it sound well worth the cost, especially for a green horn that needs all the help he can get. 

Stan, I finally got a new schematic with revisions for making the traffic signals dual mode (either CA or CC) per your suggestions above.  I used the CD4070BE ICs as I have about 10 of those on hand and only a couple of the 74HC86 versions. (In case of unintentional smoke release, I will have sufficient backups on hand with the 4070s.) Still working on the bread board test circuit. I am using individual LEDs and this is taking up a lot of bread board real estate, not to mention jumper wires... I'm thinking that using the actual traffic signals might have been a better way to try this, but I didn't want to damage tany of the actual signals if there are schematic or wiring errors (individual LEDs are much more expendable for testing). 

Here's a picture of how it ended up, hopefully it is correct, or at least pretty close. Comments wanted please! 

If this passes the once over test, I think I'll order 5 more PCBs for testing. Also gives me more practice in creating PCBs and 5 PCBs for $2 (plus $5.50 shipping) is definitely reasonable for learning.

If this is successful, I think next up will be the PIC chip GRJ suggested above to see how that goes. I am also imagining a PCB for that setup too, which should be another good learning experience. As another plus here, I'll be helping Digikey and JLCPCB keep their doors open. 

4-Way Traffic Signals Schematic-v4D-CA & CC

Attachments

Images (1)
  • 4-Way Traffic Signals Schematic-v4D-CA & CC

 

1) Good question...I am not sure where they connect, but they are on their own separate Nets. GND is Net 0 and VDD is Net 1 which applies to both J1 and J6 and associated wires. This is also one of the many things I don't quite understand about Diptrace yet? I believe they are connected on the PCB too, I looked at the nets while fiddling with the PCB layout. But, I'll check them to be sure. I guess I should also add an actual 'wire' for each connection on the schematic for clarity, so I'll add those. 

2) Thanks for the tip on grounding the unused inputs, will correct that tonight. 

Learning this stuff is nice, especially with such great teachers as we have around here! Thanks for all the help!

 4-Way Traffic Signals Schematic-v4D2-CA & CC

This should look a little better with revisions (unless I still have something wrong that is). Thanks for checking Stan.

Attachments

Images (1)
  • 4-Way Traffic Signals Schematic-v4D2-CA & CC
Last edited by rtr12

Yes, but it is probably pretty ugly by preferred design standards (which I know very little about). Everything is 'fit' on the board after a fashion, but I don't have any labeling (silk screen?) added yet. No 'optimizing', just fit in and lined up. I'll make a jpg and post it below.

4-Way Traffic Signals Schematic-PCB-v4D-CA & CC

I also added the Diptrace files for you, if you want to do some poking around in there...both probably need some finishing touches before any PCBs are ordered. For sure the PCB layout file still needs work.

Attachments

Thanks! I appreciate the kind words. I should have added above that Diptrace did most of the routing with a couple of minor adjustments, I only arranged the components. Diptrace isn't too good with the component placement, I started over on that part after Diptrace got finished. 

I made it larger for the mounting holes and to leave some room for naming things. I goofed up the mounting holes on a couple orders. I think I could shrink it a bit, but probably no where near as small as you get them. I also had problems with the terminals and another component or two not fitting too well on some of the previous PCBs, so I probably have some things spaced a bit too far apart, just for safety. 

Also for size, OSHPark is too expensive for all the PCBs I have had made for this project so far. As I recall they were $17 and up (some over $20) for all that I checked there. JLCPCB is only $2 for 5 PCBs up to 100mm x 100mm with 24 hour turn around, but they add $5.50 for standard shipping (2-3 weeks or so). Still not too bad for price. With JLC, DHL (2 or 3 day I think it was) is also an option. Drawback there is the price, instead of $5.50 I think it's around $16-$17 or so for that option.

As for any credit, that should all go to Stan and you for all my training. I could have never done any of this on my own without assistance from both of you.  I think it was the TIU tester project that really got me going. Ordering the PCBs and parts and things for that project prompted me to try and make something from scratch. Of course, that is with a lot of help you two. 

 

Last edited by rtr12

It's fun watching someone learn a new skill, and you've done it quicker than most.  I love helping people that help themselves.

As long as you don't mind the shipping delay from China, there are a lot of speedy choices for blank boards.  Note that if you get assembled boards from China, there is now a 25% tariff on them, keep that in mind!

One thing I like about OSH Park is they have very quick turnaround nowadays, they're shipping from the US.  I ordered this board on July 5th, and I got them back on the 13th.  If you're in a hurry, they're the place to go.  Of course, they were only 0.53 x 0.64 inch (13.4 x 16.2 mm), and $1.65 for three boards.  I ordered nine of them, and it was $4.95 shipped.  For many of the modules I do, especially for prototypes, they're the cheapest and fastest game in town.  I like that I can have small purpose-built boards in a flash for a project.

 

Attachments

Images (2)
  • mceclip0
  • mceclip1

Thanks, but please keep watching. I think I have a long way to go yet, and I love getting all the help!  I had a bit of background in wiring up electronic sensors and controllers, control panels, what did what, etc. Even learned how to release smoke! But, I knew nothing about the actual component level stuff or making PCBs until you guys got me interested here. I am still fascinated by all this and with today's modern design software, ease of ordering parts & PCBs, it really makes it all very possible even while not knowing a whole lot. That is with some good help of course.

It'll probably be a while before I get anything assembled, if ever. I need some more schooling in that department. And I still can't actually design anything without help from you and Stan. Plus you two have all the good ideas designed and built already.

Digikey parts are the only thing I have gotten with tariffs added. So far they haven't been a whole lot per part, just a few cents or so. But it could add up. If one can justify quantities you can make up for some of that added cost. As I am sure you know, Digikey has some pretty steep discounts when you get up to larger quantities.  

I would use OSHPark if I ever come up with anything small enough. As a test with OSHPark, I just tried the PCB above, $35.00 for 3 PCBs. They really go up with larger sizes! The PCBs could be downsized somewhat, but not enough to make that big of a difference. I have tested with every PCB so far at OSHPark so far, and they were all pretty pricey.

I suppose learning about surface mount components should be coming up one of these days soon, Stan suggested that above, and I have been thinking about it. Then maybe I can really shrink things up some? I have only tried soldering one thing SMT, it wasn't pretty. Some practice in that department is on the learning schedule as well. 

Not sure if this is of any interest, but you might consider if the led outputs can support additional loads.  For example an application may want to drive logic for actual traffic control.  I used a 3rd party traffic controller to drive a 4 way intersection using the MTH traffic lights.  I added additional logic to control traffic.  The vehicles start and stop on green and red.  On yellow, it depends.  If the vehicle is close to the light, the vehicle runs the light otherwise the vehicle stops.  Actually, my vehicles are trolleys.

If I understand Stan correctly so far (emphasis on the 'if' and 'correctly' here), I don't think there is much extra headroom with the current version I am fiddling with using the CD4070 ICs. There is more load capacity with the 74xx86 ICs Stan first mentioned, but as he also pointed out those are only 5 volt. Seems to be a slight drawback to every plan, without adding a bunch of extra components. Again, as I am understanding this anyway.  

However, I think Stan had the additional output loads in mond with the transistor and MOSFET versions above. Those are not CA - CC interchangeable by jumper selection, they would need separate PCBs for CA or CC. I am not sure of how much extra load they can handle, I'll have to look that up on the data sheets. (I don't know (or completely understand) all the components like Stan and GRJ do.) 

Then there is the PIC chip that GRJ pointed out, which I also plan to start learning about and fiddling with. This would provide features more easily than using the ICs and also be easier to modify, possibly without adding too many extra components. But again I need to study the data sheet for this one, as the PIC chips are completely new to me. I have fiddled with an Arduino Uno, but there's a lot left to learn there as well.

I would certainly be interested in the 3rd party traffic controller you used, what all it does, and how it functions. Sounds very interesting to me. Did you completely replace the MTH controller with the 3rd party one or did you use them both together? Since this is really a learning project for me as much as anything, I would be interested in learning as much as I possibly can.  

Last edited by rtr12
rtr12 posted:

If I understand Stan correctly so far (emphasis on the 'if' and 'correctly' here), I don't think there is much extra headroom with the current version I am fiddling with using the CD4070 ICs. There is more load capacity with the 74xx86 ICs Stan first mentioned, but as he also pointed out those are only 5 volt. Seems to be a slight drawback to every plan, without adding a bunch of extra components. Again, as I am understanding this anyway.  

However, I think Stan had the additional output loads in mond with the transistor and MOSFET versions above. Those are not CA - CC interchangeable by jumper selection, they would need separate PCBs for CA or CC. I am not sure of how much extra load they can handle, I'll have to look that up on the data sheets. (I don't know (or completely understand) all the components like Stan and GRJ do.) 

Then there is the PIC chip that GRJ pointed out, which I also plan to start learning about and fiddling with. This would provide features more easily than using the ICs and also be easier to modify, possibly without adding too many extra components. But again I need to study the data sheet for this one, as the PIC chips are completely new to me. I have fiddled with an Arduino Uno, but there's a lot left to learn there as well.

I would certainly be interested in the 3rd party traffic controller you used, what all it does, and how it functions. Sounds very interesting to me. Did you completely replace the MTH controller with the 3rd party one or did you use them both together? Since this is really a learning project for me as much as anything, I would be interested in learning as much as I possibly can.  

I didn't use the MTH controller at all.  I used this one: http://www.galakelectronics.com/VG-303.htm.  It's not expensive but it is working great for my application.

 

That is an interesting product indeed, thanks for the link. They also have some other interesting items and their prices are very good too. For $15.99 I couldn't resist your link above, one is already on the way!  I'll have to figure up costs on a recent PCB and it's components, they might be selling theirs at close to what it costs me to make one here. Of course I never planned to sell anything, it's strictly a hobby and learning project, and quite enjoyable too.

I had planned to post the final results of all the projects here for others to use, but linking to this site might be a better option and much less work. I'll still post the results for those that like to tinker, but the linked board is really a deal, IMO. And I'll be revisiting their site for further study on some of the other items they have.

rtr12 posted:

That is an interesting product indeed, thanks for the link. They also have some other interesting items and their prices are very good too. For $15.99 I couldn't resist your link above, one is already on the way!  I'll have to figure up costs on a recent PCB and it's components, they might be selling theirs at close to what it costs me to make one here. Of course I never planned to sell anything, it's strictly a hobby and learning project, and quite enjoyable too.

I had planned to post the final results of all the projects here for others to use, but linking to this site might be a better option and much less work. I'll still post the results for those that like to tinker, but the linked board is really a deal, IMO. And I'll be revisiting their site for further study on some of the other items they have.

It's strange that the kit and assembled price are the same ?  I purchased the assembled unit a number of years back and it was more than the kit.  In any case my controller is working fine  Good luck with your project.

I'm currently working on a cuing taxi loop.  Taxi's (3) cue up in the rear of Grand Central Station and the lead cue taxi advances to the passenger pickup area in the front of Grand Central Station after the passenger pickup up taxi departs to the 59th street traffic circle. The 59th Street taxi returns to the rear of the taxi cue at the rear of Grand Central Station.  The sequence repeats.

Add Reply

Post

OGR Publishing, Inc., 1310 Eastside Centre Ct, Suite 6, Mountain Home, AR 72653
800-980-OGRR (6477)
www.ogaugerr.com

×
×
×
×
Link copied to your clipboard.
×
×