Jump to content

Photo

Looking for modern replacement of this SENSOR

Colecovision Trackball rollerball

18 replies to this topic

#1 alortegac OFFLINE  

alortegac

    Chopper Commander

  • 193 posts

Posted Fri Jan 4, 2019 12:05 PM

Hi, apologies for this question. I am not an expert in electronics.

I have been searching for a replacement of a photo sensor used on COLECOVISION roller ball controller. Attached a partial schematic of the rollerball controller, one of the sensors has failed. I have confirmed everything else works perfectly by reposition the working sensor in place of the bad one.  I get no movement, no reading from the bad one. 

 

Also I added few pictures of the sensor. I could not find anything on this part number, or the manufacturer.  I then opened it,  and would like to see if someone can help me confirming what kind it is.  What would its internal schematics look like?  Is this a photo transistor and a IR diode?  There are a few sensors out there with similar dimensions, however I am not sure on the correct positioning of the pins vs. the original.  The original has a #1  on the pin that connects to as shown (please consider this is a mirrored image because they are opposite when turning them around to face each other.) I tried to show it better in the other picture, where I wrote the correct pin # orientation. Hopefully someone can confirm with a diagram to be able to find a replacement. I need to know where the cathode/anode should be and the Collector/emitter should be.  (if indeed this is a transistor)

 

Thanks in advance for your input.

 

Al

Attached Thumbnails

  • IMG_26711.JPG
  • IMG_26721.JPG
  • IMG_26641.JPG


#2 ChildOfCv OFFLINE  

ChildOfCv

    Chopper Commander

  • 184 posts

Posted Fri Jan 4, 2019 4:35 PM

The only thing I've been able to find so far using the part numbers and whatnot, is that it was probably made by Lite-On.  They have similar part numbers and they still make optical switches.  Unfortunately, none of their current offerings seem to be pin-compatible.

 

Now that you've ripped the boards out, though, you might be able to replace the individual components.  First you'd need to know which one(s) are failed.  You could use the diode test function on your multimeter if you have it.  For the PT you'd need a heat lamp or a bright incandescent bulb to shine on it.

 

Of course in the picture it also looks like the PTs have a hair-thin wire connecting the top to the other trace, but one of them is missing.  Or is there just a hair on the lense?  :)

 

There are surface-mountable LEDs and photodiodes that can be soldered in place of those.  You just need to know the measurement (in mm) between the traces.  Just guessing at around 1mm based on the fact that that's a DIP-8 pin arrangement, and looking at the gap in the picture.



#3 -^CrožBow^- ONLINE  

-^CrožBow^-

    Quadrunner

  • 6,881 posts
  • Collector of Fine Atari and Sega Antiquities!
  • Location:Ivory Tower, Fantasia (Oklahoma)

Posted Fri Jan 4, 2019 4:44 PM

Is that trace on pin 1 severed in the middle on that one module? Looks like it in all the pics?

 

Wonder if the CX53 uses the same optos on it?



#4 alortegac OFFLINE  

alortegac

    Chopper Commander

  • Topic Starter
  • 193 posts

Posted Fri Jan 4, 2019 8:06 PM

Ok, I took some pictures with the microscope for a better viewing.  Yes, one PT has a little hair connection and the other doesn't.  I am not sure if I had unintentionally damaged it myself while opening the casing up or it is just the failing PT itself.  Which pin are collectors/emitter and cathode/anode on this part? Could you tell me how I suppose to know which side is which? 

I am very curious to receive suggestions on potential stock parts to try out.  I have tried setting 2 switches back to back TCST1103 (pic26761), but I could not make it to work. The signal would never flip on pin 2 of the 4013 IC when changing directions of the trackball.

I have been thinking about 2 other potential switches. attached pics. 26731 and 26741. Not sure on either.

I would really appreciate your suggestions.

 

 

thanks a lot for your assistance!

 

 

Al

 

 

Attached Thumbnails

  • Snapshot 2 (1-4-2019 8-45 PM).png
  • Snapshot 1 (1-4-2019 8-45 PM).png
  • IMG_26731.JPG
  • IMG_26741.JPG
  • IMG_26761.JPG


#5 ChildOfCv OFFLINE  

ChildOfCv

    Chopper Commander

  • 184 posts

Posted Fri Jan 4, 2019 11:14 PM

Well, as wired into the roller controller, the switch looks like this:

 

Screen Shot 2019 01 04 At 10.27.33 PM

 

Now to make sure my entire post is correct with respect to the picture:

 

I switched the halves in order to simplify the schematic.  But pins 1-4 are the PTs and pins 5-8 are the diodes.  Cathodes are pins 6 and 8.  Emitters are pins 2 and 4.

 

Whatever replacement you find needs to be arranged in a DIP-8 footprint.  DIP-8 means that pins in a row are 0.1in apart, and the two rows are 0.3in apart.  So the first datasheet has the wrong dimensions.


Edited by ChildOfCv, Fri Jan 4, 2019 11:34 PM.


#6 ChildOfCv OFFLINE  

ChildOfCv

    Chopper Commander

  • 184 posts

Posted Sat Jan 5, 2019 1:02 AM

So, the most promising replacement by pinout, dimensions, etc. seems to be this one.  But as you'll quickly notice, it won't be a drop-in replacement either.

 

https://omronfs.omro...n-ee_sx1340.pdf

 

To attempt to solder on this one, I think what I'd do is solder some small wires into the PCB, lay them flat, then surface-mount it on the wires.

 

reason this one comes out on top is that it's thinner than the 1103 that you tried, so it might be able to achieve the proper slot spacing between the two detectors.  The slots have to end up 0.2in apart in order to work with that code wheel.



#7 BigO OFFLINE  

BigO

    River Patroller

  • 4,399 posts
  • Location:Phoenix, AZ

Posted Sat Jan 5, 2019 9:59 AM

I don't have any real or particularly useful answers, but ChildOfCV's comment triggered a thought. I figured I'd throw it out there in case you run into a dead end.

 

The two channels of the sensor don't need to be physically attached together with the original spacing. You could do two separate single channel sensors placed anywhere around the wheel. They just need to be triggered 90 degrees out of phase. The two sensors don't have to be triggered by the same spoke/hole, just the relative timing between the two has to be correct.

 

It would certainly be fiddly to position them appropriately and hold them in place, but in a worst case scenario two separate sensors could be used. Looking at the internals of the controller, there's probably sufficient space to do something like that.

 

Good luck.


Edited by BigO, Sat Jan 5, 2019 10:02 AM.


#8 alortegac OFFLINE  

alortegac

    Chopper Commander

  • Topic Starter
  • 193 posts

Posted Sat Jan 5, 2019 10:26 AM

So, the most promising replacement by pinout, dimensions, etc. seems to be this one.  But as you'll quickly notice, it won't be a drop-in replacement either.

 

https://omronfs.omro...n-ee_sx1340.pdf

 

To attempt to solder on this one, I think what I'd do is solder some small wires into the PCB, lay them flat, then surface-mount it on the wires.

 

reason this one comes out on top is that it's thinner than the 1103 that you tried, so it might be able to achieve the proper slot spacing between the two detectors.  The slots have to end up 0.2in apart in order to work with that code wheel.

Greatly appreciated. Very clear.  I understand now about the critical element of distance between slots.  I thought the most critical and only concern was to match the voltages, gap width, etc.. I was also thinking it was very important to find a dual channel sensor.

I have ordered the ee-sx1321 already last week.  I will see what I can do once I get them according to your comments.  In any case, I will now go ahead ordering 1340's as you suggested too.  Soldering is not a problem, I have very good skills and some pads from other SMT IC's that I could probably adapt to work with wires to this board.  I also moved the working original sensor into the main board and left the missing sensor side in the separate satellite board that is connected through a 4 wire cable to the main board. It is very easy to handle this way because is a small board on the side, and I do not need to flip or handle the main board anymore.

 

https://omronfs.omro...n-ee_sx1321.pdf

 

Funny, I do not have Vernier here in my tool box...…..to take precise measurements.  I will buy one.


Edited by alortegac, Sat Jan 5, 2019 10:32 AM.


#9 alortegac OFFLINE  

alortegac

    Chopper Commander

  • Topic Starter
  • 193 posts

Posted Sat Jan 5, 2019 10:30 AM

I don't have any real or particularly useful answers, but ChildOfCV's comment triggered a thought. I figured I'd throw it out there in case you run into a dead end.

 

The two channels of the sensor don't need to be physically attached together with the original spacing. You could do two separate single channel sensors placed anywhere around the wheel. They just need to be triggered 90 degrees out of phase. The two sensors don't have to be triggered by the same spoke/hole, just the relative timing between the two has to be correct.

 

It would certainly be fiddly to position them appropriately and hold them in place, but in a worst case scenario two separate sensors could be used. Looking at the internals of the controller, there's probably sufficient space to do something like that.

 

Good luck.

Hi, thank you so much for your input.  Could you please describe a bit more the 90 degree out phase comment?.  I think I get it, but I want to make sure I fully understand it. Currently,  the slots are parallel to each other, equidistant. So, 90 degrees means, when one is in open, the other must be in close relative to the wheel?  Thanks a lot!  Al


Edited by alortegac, Sat Jan 5, 2019 10:31 AM.


#10 alortegac OFFLINE  

alortegac

    Chopper Commander

  • Topic Starter
  • 193 posts

Posted Sat Jan 5, 2019 10:33 AM

Is that trace on pin 1 severed in the middle on that one module? Looks like it in all the pics?

 

Wonder if the CX53 uses the same optos on it?

I will check, I do have a couple CX53.  It is ATARI part # CO20290,  however could not find it in BEST website. This is the schematics for the CS53 section. Basically, it looks like 5v go directly to the A diode then to the resistor R23, where CV goes into 100ohm resitor and then to A diode.. Appears also E and C PT are going in opposite direction to the one in CV.  Am I looking at this correctly? thanks for your input. Al

Attached Thumbnails

  • cx53.PNG

Edited by alortegac, Sat Jan 5, 2019 11:33 AM.


#11 BigO OFFLINE  

BigO

    River Patroller

  • 4,399 posts
  • Location:Phoenix, AZ

Posted Sat Jan 5, 2019 10:53 AM

Hi, thank you so much for your input.  Could you please describe a bit more the 90 degree out phase comment?.  I think I get it, but I want to make sure I fully understand it. Currently,  the slots are parallel to each other, equidistant. So, 90 degrees means, when one is in open, the other must be in close relative to the wheel?  Thanks a lot!  Al

 

What you're describing would be 180 degrees out of phase. You're saying that they would both change states at the same time (and would always be in opposite states). With 90 degrees out of phase, a channel changes states as the other channel is half way through it's current state.

 

In your description the output of the two channels would be 01, 10.

In a 90 degree phase difference, the output of the two channels would cycle through these 4 (hence "quad") values 00, 01, 11, 10. 

(That bit pattern is known as "Gray code")

 

Taking the perspective that one of the channels "leads" the other in phase, by knowing the state of one bit you can tell which direction the quadrature encoder wheel is turning. Here's a diagram I put together when I was trying to wrap my head around how to decode quadrature signals in a microcontroller program I was working on.

 

Quadrature Decoding Method

 

For more information, look at the "Quadrature Encoding" section of this link: https://www.phidgets.../Encoder_Primer


Edited by BigO, Sat Jan 5, 2019 10:58 AM.


#12 ChildOfCv OFFLINE  

ChildOfCv

    Chopper Commander

  • 184 posts

Posted Sat Jan 5, 2019 11:03 AM

 

What you're describing would be 180 degrees out of phase. You're saying that they would both change states at the same time (and would always be in opposite states). With 90 degrees out of phase, a channel changes states as the other channel is half way through it's current state.

 

In your description the output of the two channels would be 01, 10.

In a 90 degree phase difference, the output of the two channels would cycle through these 4 (hence "quad") values 00, 01, 11, 10. 

(That bit pattern is known as "Gray code")

 

Taking the perspective that one of the channels "leads" the other in phase, by knowing the state of one bit you can tell which direction the quadrature encoder wheel is turning. Here's a diagram I put together when I was trying to wrap my head around how to decode quadrature signals in a microcontroller program I was working on.

 

 

 

For more information, look at the "Quadrature Encoding" section of this link: https://www.phidgets.../Encoder_Primer

The "quick and dirty" way to think about it, which is actually how the CV takes its samples, is this:

 

When the right side is triggered, it samples the left side.  If the left side is also triggered, you rolled left (or down).  If the left side is not triggered, then you rolled right (or up).  On the Super Action Controller, this is true because the magnet coming down from the right side will trigger the right reed switch first, then the left.  Coming from the opposite direction, the left reed switch will already be triggered.



#13 alortegac OFFLINE  

alortegac

    Chopper Commander

  • Topic Starter
  • 193 posts

Posted Sat Jan 5, 2019 11:13 AM

 

What you're describing would be 180 degrees out of phase. You're saying that they would both change states at the same time (and would always be in opposite states). With 90 degrees out of phase, a channel changes states as the other channel is half way through it's current state.

 

In your description the output of the two channels would be 01, 10.

In a 90 degree phase difference, the output of the two channels would cycle through these 4 (hence "quad") values 00, 01, 11, 10. 

(That bit pattern is known as "Gray code")

 

Taking the perspective that one of the channels "leads" the other in phase, by knowing the state of one bit you can tell which direction the quadrature encoder wheel is turning. Here's a diagram I put together when I was trying to wrap my head around how to decode quadrature signals in a microcontroller program I was working on.

 

 

 

For more information, look at the "Quadrature Encoding" section of this link: https://www.phidgets.../Encoder_Primer

Very clear!!   thank you so much.  I think I may be able to adapt something this way if the other options do not work. I have learned a lot today, this is greatly appreciated.  Al



#14 alortegac OFFLINE  

alortegac

    Chopper Commander

  • Topic Starter
  • 193 posts

Posted Sat Jan 5, 2019 11:16 AM

The "quick and dirty" way to think about it, which is actually how the CV takes its samples, is this:

 

When the right side is triggered, it samples the left side.  If the left side is also triggered, you rolled left (or down).  If the left side is not triggered, then you rolled right (or up).  On the Super Action Controller, this is true because the magnet coming down from the right side will trigger the right reed switch first, then the left.  Coming from the opposite direction, the left reed switch will already be triggered.

thank you!  I think I got it. But now, even more.  Al


Edited by alortegac, Sat Jan 5, 2019 11:22 AM.


#15 alortegac OFFLINE  

alortegac

    Chopper Commander

  • Topic Starter
  • 193 posts

Posted Sat Jan 5, 2019 11:25 AM

Another member claimed this part worked on this application with no pin relocation, basically almost plug and play. I have not been able to find it anywhere or the manufacturer.  I thought maybe you may know something about it. Here the images from the original post.  Obviously, this listing is not longer there on EBAY. 

 

Any thoughts, information, will be greatly appreciated.

 

Thanks in advance!

Al

Attached Thumbnails

  • post-63971-0-73636000-1521840709_thumb.png
  • post-63971-0-92779300-1521840693_thumb.jpg


#16 alortegac OFFLINE  

alortegac

    Chopper Commander

  • Topic Starter
  • 193 posts

Posted Sat Jan 5, 2019 4:23 PM

Final Update:  I would like to thank you all for your support and sharing your expertise. This was a great micro rescue project for me.  I learned quite a bit and had fun.

 

Attached pics of the solution I was able to implement after receiving your feedback.

 

 I used 2  TCST 1103  switches in a 90 degree out of phase position as suggested. One switch placed on the original board first 2 rows,  and the other one placed and wired in a adhoc board attached with little metal sheet arm to the screw on the side.  It is quite firm and does not move a bit, and since there is no contact I do not expect it to move. However, after taking the pictures, I used some hot glue around it and on the wires just in case of any vibration.  I am also including a video of the controller working perfectly with SLITHER.

 

thanks again!!! AL

 

 

PS: All fire buttons work perfectly, in the video I was only testing the movement up/down/left/right.

Attached Thumbnails

  • IMG_26781.JPG
  • IMG_26791.JPG
  • IMG_26801.JPG
  • IMG_26821.JPG

Attached Files


Edited by alortegac, Sat Jan 5, 2019 4:27 PM.


#17 BigO OFFLINE  

BigO

    River Patroller

  • 4,399 posts
  • Location:Phoenix, AZ

Posted Sat Jan 5, 2019 5:13 PM

Nice.

#18 alortegac OFFLINE  

alortegac

    Chopper Commander

  • Topic Starter
  • 193 posts

Posted Sat Jan 5, 2019 5:24 PM

Nice.


Thank you, much credit goes to you too, this was your suggestion.

Edited by alortegac, Sat Jan 5, 2019 5:48 PM.


#19 BigO OFFLINE  

BigO

    River Patroller

  • 4,399 posts
  • Location:Phoenix, AZ

Posted Sat Jan 5, 2019 9:29 PM

Happy to help.

 

Since you've proven the concept, others might be encouraged to save classic hardware from the trash heap even without a drop-in sensor replacement.







Also tagged with one or more of these keywords: Colecovision, Trackball, rollerball

0 user(s) are browsing this forum

0 members, 0 guests, 0 anonymous users