SEGA 3-switch 4-position shifter truth table

[baritonomarchetto]

Hi all, could someone please clarify me which switches are closed (and which are open) when gears are inserted in 4 ways shifter SEGA used to install in Daytona cabinets?
I am toying with Eagle CAD lately and wanted to realize a simple PCB with a 74HC138 demultiplexer as a hardware solution to the "3 input to 4 output" issue.
Yes, I will share gerber files so that anyone can have their pcbs printed and the hardware at disposal at the cost of components

[PL1]

Hi all, could someone please clarify me which switches are closed (and which are open) when gears are inserted in 4 ways shifter SEGA used to install in Daytona cabinets?

there's a rod sticking out the sides of the shifter that activates/deactivates switches 1 and 2 simultaneously when you move it forward or backwards.
Gear 1 and 3 would both be activated at the same time if it weren't for switch #3 allowing only switch 1 or switch 2 (not both) from going to ground.
Switch 3 is activated by the left/right action of the shifter.

The switches aren't actually in the positions on the shifter as they are in the diagram.
It was just a diagram to draw the connections.

Switch 3 is closed when in 1st or 2nd gear and open when in 3rd or 4th gear.

There's also a video near the end of that thread showing something similar to what you're doing.


Scott

[baritonomarchetto]

Thanks Scott, I read that post but honestly it's not clear enought to me. Could you (or badmouth or anyone else with ome of thoe shifters... I don't own one to map it myself) please simplify that ?
i.e.:
First gear: all switch open
Second gear: ...
And so on.
In the post you linked someone wrote that the connections to the demultiplexer where wrong, so I need the info I asked for in order to figure out myself the correct wiring

[PL1]

I don't own one to map it myself) please simplify that ?

The diagram shows how to wire the shifter so you don't need a 3-to-4 demux circuit.
- The downside is that one encoder input is always grounded so there's no "Neutral" gear unless you add a switch that interrupts the ground on Switch 3 Common. ("COM" is labeled "IN" on the diagram)

Shifter in first gear:
- Switches 1 and 2 open
- Switch 3 closed
Ground comes in on Switch 3 COM, out on Switch 3 NO, in on Switch 1 COM, out on Switch 1 NC to first gear encoder input.

Shifter in second gear:
- Switches 1 and 2 closed
- Switch 3 closed
Ground comes in on Switch 3 COM, out on Switch 3 NO, in on Switch 1 COM, out on Switch 1 NO to second gear encoder input.

Shifter in third gear:
- Switches 1 and 2 open
- Switch 3 open
Ground comes in on Switch 3 COM, out on Switch 3 NC, in on Switch 2 COM, out on Switch 2 NC to third gear encoder input.

Shifter in fourth gear:
- Switches 1 and 2 closed
- Switch 3 open
Ground comes in on Switch 3 COM, out on Switch 3 NC, in on Switch 2 COM, out on Switch 2 NO to fourth gear encoder input.

If you still want to make a demux circuit, you can simplify the truth table because Switches 1 and 2 are always in the same state.
- 0=open, 1=closed

Gear     Switch 3    Switch 1 or 2
  3               0              0
  4               0              1
  1               1              0
  2               1              1

As with the wiring diagram approach, there is no "Neutral" gear unless you add a switch and expand the truth table.
- Maybe use a beam break sensor that triggers when the shift lever is in the middle of the "H" pattern if that's possible.   


Scott

[baritonomarchetto]

Ok, I made a first prototype. The schematic is tested on breadboard, the PCB is NOT TESTED YET but has a high probability of success .

Attached is a gerber viewer simulation of the board and a zip file to be sent to the PCB manufacturer of choice 

[PL1]

1. Consider adding a ground connection on the switch input side so users don't have to get ground for the switches from a separate connection.

2. Is there enough space between the screw hole and the G4 pad?  It looks way too close.

3. What is the spacing between the switch input and gear output pads?  Looks like 6mm.
- Looks like the 5v and ground pads are 5.08mm pitch.
- Users might want the option to use normal 5.08mm pitch screw terminals, small 2.54mm pitch screw terminals, or 2.54mm pitch Dupont pin headers for the switch and gear output lines.
-- IMO the 2.54mm pitch options are the better choices for this application because 5.08mm screw terminals are overkill.
- Bruno designed the KADE miniArcade 2.0 daughterboard so users could choose a mix of 2.54mm pitch screw terminals (green, outside) and Dupont pins. (blue, inside)

 

4. Consider adding toggle/rocker switch connections that set all outputs high by applying ground to 74ls138 pin 6. (input G1)
- Wire COM to pin 6 and the other terminals to 5v and ground.
- When the switch is in the "normal" position, 5v is applied to pin 6 and the demux works normally. (one button always pressed)
- When the switch is in the "disable" position, ground is applied to pin 6 and all outputs are set high. (no button pressed)
- You might be able to get the same result applying 5v to pins 4 or 5 (inverted inputs G2A/G2B), but check the datasheet to be sure.   

This feature is very useful when mapping keys during setup since one of the four gear outputs will always be low (one button always pressed) with the current setup.

5. Does this circuit need pullup resistors as implied by this blog post?   
http://www.learnerswings.com/2014/07/controlling-74138-3-line-to-8-line.html


Scott

[PL1]

On second thought, the easier and more efficient way to do #4 is to use shifter switch 1 or 2 --not both-- since both are always the same state.

That frees up the third input for a "disable" switch.

The current logic setup uses outputs 0, 3, 4, and 7 for gear 1 - gear 4.

If you use outputs 0 - 3 for gears 1-4, setting the "disable" switch high (blue) will set outputs 0 - 3 to high (purple) so there's no button pushed.


Scott

[baritonomarchetto]

Thanks Scott: I can implement most of your suggestions. Will do, stay tuned

[baritonomarchetto]

Version 0.1 

as per Scott suggestion, I have removed one of the (redundant) switch inputs in favour of a disable input. In addition, you can disable the demultiplexing by removing a jumper (I must suggest the use of a removable jumper here, like those used to define master/slave in HDDs).

I don't think an additional ground pad was needed (shifter and PSU ground must be shared and the terminal is sufficiently large to accomodate both), but there was some spare space, soo...   

I also added pullup resistors... they are almost mandatory in such circuits, but I did not thought at that in first place

Pads are now equally spaced. This way one can decide to use 5.08 mm screw terminals (sorry but I don't like those tiny 2.54 mm spaced) or solder directly the cables on pads.

Notice that any "138" demultiplexer will work (74LS138, 74HC138, 74HCT138).

Notice that to work correctly, inputs on your receiving device (i.e. Arduino...) must be set to "input pullup" (input forced to HIGH, activated by LOW input signals). If your encoder force you to use input puldowns (shame on closed hardware) you can use a 74LS238 in place of the 74LS138.

I used the IC disable pin (G1) to disable it ( ), forcing pin A2 to ground and thus limiting the truth table to the "green region". By using A2 as disable input (as Scott suggested) could let us introduce an optional neutral output, but I don't think it is really necessary. Let me know if it could be of any real use.

Both circuit and board are NOT TESTED at the moment (not even on breadboard).

[baritonomarchetto]

In this alternative version I added a neutral input/output. When neutral (N) button is pressed (LOW), all 4 gear outs are inhibited (high) and "N" output is LOW (direct connection to the input).

During the mapping phase, you can have NO outputs (all HIGH) by removing the jumper "enable" (or, if "N" output is not connected, by keeping the switch connected to input "N" pressed). You can use a latching button instead of a jumper if the board is placed far from you, but thats a matter of personal preference.

Notice that IC outputs used in this version are different from those of the previous version, so the table I uploaded is not 100% applicable here.

[BadMouth]

My diagram above was a solution to use the switches that are there, but I'm not so sure that's how they were originally wired.

Bit late, but the real solution is here:
http://forum.arcadecontrols.com/index.php/topic,157369.msg1653601.html#msg1653601

[baritonomarchetto]

The solution I posted, developed with the help of Scott and based on the one from segarepair, solves the "a gear input is always sent while mapping" issue without the need to drill the shifter case for a fourth switch. Also, gerber files I posted can be sent to one of the well known cinese PCB manufacturers in order to have everithing wired up at the cost of a dollar or two (+postage) and only some solder to do to populate the board.
Hopes this will help someone (I had some fun anyway, even if no-one will ever use it )

[PL1]

you can disable the demultiplexing by removing a jumper (I must suggest the use of a removable jumper here, like those used to define master/slave in HDDs).

Not sure why you're disabling the demux now that you're using Outputs O₀ thru O₃ (see section about pins 4, 5, and 6 below), but having the option of jumper pins for grounding input A₂ (disable switch) sounds good.   

Pads are now equally spaced. This way one can decide to use 5.08 mm screw terminals (sorry but I don't like those tiny 2.54 mm spaced) or solder directly the cables on pads.

You may not like them, but some users might.   

Consider adding 2.54mm pitch through holes for Dupont pins or small screw terminals in parallel with the current 5.08mm holes for large screw terminals so users can choose whichever they prefer.

Notice that any "138" demultiplexer will work (74LS138, 74HC138, 74HCT138).

The LS and HCT series chips will be fine, but will the HC (High-speed CMOS) logic levels work with typical IPac/Arduino setups?
http://forum.arcadecontrols.com/index.php/topic,163889.msg1728853.html#msg1728853

Notice that to work correctly, inputs on your receiving device (i.e. Arduino...) must be set to "input pullup" (input forced to HIGH, activated by LOW input signals). If your encoder force you to use input puldowns (shame on closed hardware) you can use a 74LS238 in place of the 74LS138.

Sounds like you can convert the board for an "active high" encoder by using a 74__238.
- That should make this board work with "Zero Delay" encoders, right?

I used the IC disable pin (G1) to disable it ( ), forcing pin A2 to ground and thus limiting the truth table to the "green region".

That first half doesn't sound right.   

Looking at the logic symbol on your truth table, there is a boolean AND gate with 3 inputs (pins 4, 5, and 6), two of them inverted.
- Ground on pins 4 and 5 (E₁ and E₂) and 5v on pin 6 (E₃) satisfies all three conditions for the AND gate which allows access to lines 4-11 of the truth table which includes the "green region".
- If you have a logic high on pin 4 or a logic high on pin 5 or a logic low on pin 6, you will be confined to lines 1-3 of the truth table.

There are some typos in your Daytona Truth Table:
- Gear 2 should be Output O₀ (pin 15)
- Gear 4 should be Output O₂ (pin 13)

Looking at the board, you can see that:
- G1 is connected to pin 14. (good)
- G2 trace is on the back of the board, it should already be connected to pin 15.
- G3 is connected to pin 12. (good)
- G4 is connected to pin 11. (should be pin 13)

By using A2 as disable input (as Scott suggested) could let us introduce an optional neutral output, but I don't think it is really necessary.

I agree that a separate output for neutral is not necessary.


Scott

[baritonomarchetto]

Yes Scott, you are right: that version had some pin rounting wrong 
Anyway, the following was correct. Sorry for posting wrong info.

Now, here is the latest version. It's almost like the one from reply #9, but with an additional resistor and 2.54 spaced pinholes in series with those 5.08 spaced 

Scheme and board untested

[Howard_Casto]

Please don't hit me guys but I've got to do a sanity check.  You could get a pcb printed for about 5 bucks, spend about 10 bucks in parts (because every online electronics store seems to charge 6-8 bucks shipping) and take the time to solder it together so it'll output  signals that can be fed into an ipac..... or you could buy the cheapest arduino clone on ebay for 5-10 bucks, ready built and write a bit of code and just have it show up as a keyboard and parse all the data to the appropriate keypresses. 

I've found as arduinos get cheaper and cheaper it just makes sense to use them for most projects  even if they are overkill. 

But I'm just suggesting here.... you guys might have a perfectly valid reason to do it the other way, I just want to know about it so I can better understand the purpose of this type of  multiplexer circuit. 

[baritonomarchetto]

I agree with you Howard that arduino is the way to go nowadays (you should know that I agree! My racing cab is interfaced with arduino only!), but it is nice (and informative) to toy with logic circuits when something simple like this pops up (..."simple" it is, and I made some mistake anyway... Dumb).
About the costs of this project: you can have 5 of this boards printed for 2 dollars from i.e. JLCPCB (+shipping costs. To Italy those are 6 dollars, but you can print more projects at once and shipping costs are almost contant). In parts will cost 2 dollars at most (no costs from china if you are ready to wait for a month... Which is what i always do).
You could wire on perfboard the whole project, but since I discovered Eagle CAD and how easy it is to realize boards I have no reason to look back honestly!
Even if the pourpouse of this topic was merely informative (and a way to spend some valuable time on eagle for me), it leaded to a piece of hardware that someone could find useful. Isn't it better to have one more option?
I could also develop the arduino code if of any interest for someone, just for reference

[PL1]

I agree that simple wiring (like the diagram in the first reply + a switch to interrupt ground) or an Arduino is a logical choice for a project like this,  but figured that since baritonomarchetto was learning how to use Eagle to make Gerber files it would be a good thread for people to learn about and discuss topics like these:
- Datasheets
- Truth tables
- Organizing/assigning inputs and outputs
- Component selection (terminals/pins/IC families/alternate ICs)
- Board design/layout
- Simplifying circuit design (you don't need to use both switch 1 and 2 if you are using a demux or Arduino, you don't need to mess with enable/disable if you choose the right inputs, etc.)
- Hardware mods
- Active high vs. active low circuits
- CMOS logic levels (HC series) vs. TTL logic levels (LS or HCT series)
- Boolean logic
- Schematics
--------------------
1. It took a while to see what you were doing with the 5v and ground Dupont pins. (alternate power input, right?)
- Consider adding markings to clearly identify those two pins as power input or move those two pins closer to the 5.08mm power inputs.  i.e. a 1x2 Dupont for power/ground and a 1x4(1x5?) for switches/ground. (2 grounds?)

2. We already agreed that there's no known use for a "Neutral" output, right?
- I understand having a variant design that uses ground on pin 6 to disable the demux and pull the gear outputs high.
- The "Neutral" input here does the same thing that the "Disable" switch does on the other board, just using different pins/labels/lines of the truth table.
- Consider using similar terms/markings for both boards.
- You might want to eventually add a small truth table like the small Daytona one in reply #8.

3. The board is getting a bit wider so there's more leverage if the mount hole remains in the lower-right corner.
- Consider moving the mount hole to the lower-center.

4. Not sure about how the "enable" jumper functions.
- Install jumper for normal operation, remove to disable outputs?
- Nudge the jumper a bit left so there's more room if you choose to move the mount hole.

5. When you get a chance, please update the Daytona truth table in reply #8.

6. One thing we are missing in this thread is a good photo/diagram showing where switches 1, 2, and 3 are located on the shifter and which tabs on the switches are which -- like the photos in the thread that BadMouth linked to earlier.


Scott

[baritonomarchetto]

1. It took a while to see what you were doing with the 5v and ground Dupont pins. (alternate power input, right?)

Not alternate but THE power input

2. We already agreed that there's no known use for a "Neutral" output, right?

Yes, but as I said the board is big enough to accomodate that pad. I cannot exclude someone could find a good use for that.

3. The board is getting a bit wider so there's more leverage if the mount hole remains in the lower-right corner.
- Consider moving the mount hole to the lower-center.

This can be done with no problems, but its a detail we can take care of after the board has been tested

4. Not sure about how the "enable" jumper functions.

Pressing (grounding) "N" puts all gears outs to "High" and sort of "disable" outputs, as per your (brilliant I must say) suggestion . It is not said, anyway, that one want to connect a switch to that pin. This is why I also inserted the enable" jumper: when removed, the IC is disabled (and then the outputs). 

5. When you get a chance, please update the Daytona truth table in reply #8.

The truth table for the current design is here attached (I have removed the one from post #8 being not correct)

6. One thing we are missing in this thread is a good photo/diagram showing where switches 1, 2, and 3 are located on the shifter and which tabs on the switches are which -- like the photos in the thread that BadMouth linked to earlier.

I agree. I have not a daytona shifter by hand (never had really. My cab has a sequential shifter) so I must ask someone here to please add it to the discussion

[PL1]

1. It took a while to see what you were doing with the 5v and ground Dupont pins. (alternate power input, right?)

Not alternate but THE power input

By "alternate" I meant that users could power the board via either the two clearly-identified 5.08mm pads or the not-so-clearly-identified lower two Dupont pins.
- At first glance, I thought the 5v Dupont pin might be for an illuminated shifter knob mod.   

4. Not sure about how the "enable" jumper functions.

Pressing (grounding) "N" puts all gears outs to "High" and sort of "disable" outputs, as per your (brilliant I must say) suggestion . It is not said, anyway, that one want to connect a switch to that pin. This is why I also inserted the enable" jumper: when removed, the IC is disabled (and then the outputs). 

Thanks for the clarification and the updated truth table.

Good choice to use the lower part of the truth table (lines 8-11) for normal operations so grounding "N" (input A₂) shifts to the middle part of the truth table. (lines 4-7)   

For anyone wondering why baritonomarchetto made that change, it allows all three input switches to be wired the usual way using the Normally Open tab.
- To use the middle part of the truth table (lines 4-7) for normal operations, the switch for input A₂ would have to be wired using the Normally Closed tab.   
-------------------
Don't know why we didn't think of it earlier, but the practical use for the "N" output is ground for a "Neutral" LED indicator.   
- It will work if you ground the "N" input, but won't work if you remove the "enable" jumper.
- To make it work with both, you'd have to change to a "disable" jumper that grounds the "N" input.
- There should be plenty of room for a current limiting resistor and LED on the board so the user can choose between an onboard LED, an external LED, or no LED.


Scott

[Howard_Casto]

I would also think neutral would be kind of an important state to have so that controls can be remapped.... otherwise the gear button would constantly be held down. 

[PL1]

I would also think neutral would be kind of an important state to have so that controls can be remapped.... otherwise the gear button would constantly be held down.

Very true.

The current board design allows two possible ways to achieve that neutral state where all four gear outputs (O₄-O₇) are set to logic high:

  1. Remove the "enable" jumper which removes 5v from pin 6.  Low on pin 6 (input E₃) ==> truth table line 3, all outputs are high.

  2. Ground the "N" (A₂) input ==> truth table lines 4-7.
- The specific line is determined by the other two inputs "S1/2" (A₀) and "S3". (A₁)
- For example, if the shifter is in first gear (A₀ high and A₁ low) and "N" is grounded (A₂ low) ==> truth table line 5, output O₁ is low and all others are high.
---------------
If you want to add a "Neutral" switch and LED indicator to the wiring-only diagram, insert an SPDT switch between ground and switch 3 "IN". (COM)
- SPDT COM to ground.
- SPDT NC to switch 3 "IN".
- SPDT NO to LED, current limiting resistor, and 5v.




Scott