SAC Spinner Electrical Repair - Advice Needed

[Ikrananka]

We all know how flaky the spinners on Coleco's Super Action Controllers are now that they are 35+ years old.  I have one that works perfectly, but I also have one that is interpreted by my ColecoVision as always spinning in one direction regardless of which way I actually turn the spinner.  It is also registered intermittently and so the spinning movement reported by the SAC test program (the excellent one by Bruce Tomlin) is very jerky.

 

I'd love to get this controller's spinner working correctly and so am looking for some advice on what components may be at fault here.  Is it one or both of the spinner switches or possibly one or more of the associated diodes?  Or both?  I'm sure suitable replacement diodes are readily available to buy, but are the switches?

 

PS - my tests are on a real CV with a new controller chip.  Frustratingly, I fried the last chip by simply plugging in the SAC.  Time to try out the Ruggers Customs DB9 ESD protection boards from console5.com.

[retroillucid]

1 hour ago, Ikrananka said:

 

 

... Frustratingly, I fried the last chip by simply plugging in the SAC.

I wonder if the SAC is the main cause of frying IC chips   

[+Yurkie]

I had a couple of glass tubes that contained contacts that when the magnet in the physical spinner was spun would cause rapid mechanical contact.

That doesn't make much sense...anyway it is like flipping a light switch off and on very quickly. I believe there was one set with positive charge and one with a negative charge, so that you could get left and right movement. Been a long time since I looked at these and not sure where they are. I only found a couple mounted to standard controllers boards out of the many hundreds that I refurbished when I used to do controllers. I'm can't remember if the SAC has glass tubes like this or not. I always hated the SAC and thought they where junk. Anyway I found one of those tubes that I took out of a standard controller, here is a pic. Looking at how there are only 2 contacts not sure how you would get 2 directions from it.

 

EDIT: after thinking about it I think there where 2 of these glass tubes beside each other. Spin the knob one way and the contacts on one tube would open/close rapidly, spin it the other way and the contacts on that one would open/close rapidly.

[Tursi]

They're called Reed Switches - they are pretty common for magnetic motion sensing. They have fast response and very little bounce in the magnetic field.

 

http://www.arunet.co.uk/tkboyd/ec/ec1reha.htm

 

The two switches just act like any other contact. If you watch the magnet rotate near them, you'll see that they will trigger one, and then the other. This helps you to determine direction of rotation.

 

For a more recent use - the rock band guitars use them for the strum bar.

 

[Ikrananka]

9 hours ago, Yurkie said:

I had a couple of glass tubes that contained contacts that when the magnet in the physical spinner was spun would cause rapid mechanical contact.

That doesn't make much sense...anyway it is like flipping a light switch off and on very quickly. I believe there was one set with positive charge and one with a negative charge, so that you could get left and right movement. Been a long time since I looked at these and not sure where they are. I only found a couple mounted to standard controllers boards out of the many hundreds that I refurbished when I used to do controllers. I'm can't remember if the SAC has glass tubes like this or not. I always hated the SAC and thought they where junk. Anyway I found one of those tubes that I took out of a standard controller, here is a pic. Looking at how there are only 2 contacts not sure how you would get 2 directions from it.

 

EDIT: after thinking about it I think there where 2 of these glass tubes beside each other. Spin the knob one way and the contacts on one tube would open/close rapidly, spin it the other way and the contacts on that one would open/close rapidly.

Yep, they're the switches used for the SAC spinner.  Here's how I believe they work:

 

 

[Ikrananka]

6 hours ago, Tursi said:

They're called Reed Switches - they are pretty common for magnetic motion sensing. They have fast response and very little bounce in the magnetic field.

 

http://www.arunet.co.uk/tkboyd/ec/ec1reha.htm

 

The two switches just act like any other contact. If you watch the magnet rotate near them, you'll see that they will trigger one, and then the other. This helps you to determine direction of rotation.

 

For a more recent use - the rock band guitars use them for the strum bar.

 

Great - thanks.  I thought they were called Reed switches so good to get confirmation.

 

So, if a spinner is acting flaky, such as jerky movement and always in the same direction am I correct in guessing that one of the reed switches is likely acting up?

 

Is there a supply of suitable replacements (NOS or modern) for these (other than ripping them out of other controllers)?

Edited January 26, 2020 by Ikrananka

[Ikrananka]

2 minutes ago, Ikrananka said:

Is there a supply of suitable replacements (NOS or modern) for these (other than ripping them out of other controllers)?

 

6 hours ago, Tursi said:

They're called Reed Switches - they are pretty common for magnetic motion sensing. They have fast response and very little bounce in the magnetic field.

 

For a more recent use - the rock band guitars use them for the strum bar.

 

I've partly answered my own question by doing a search on Digi-Key.  They have hundreds of different glass bodied through hole mount reed switches.  I now need to narrow this down by checking the dimensions of those use in the SAC.

 

BUT, there are other specifications that I'd need to know before ordering some to try, such as Operating Range, Release Range, Switching Current, Switching Voltage, Operating Time, Release Time etc.  Any guidance on what to look for with these criteria?

[ChildOfCv]

I guess one way to improve on this would be to build a replacement circuit that uses Hall sensors.

 

But yeah, it's easy to test whether a reed switch is "on" because it just conducts like a standard switch.  Simple continuity would work.  Since it's constantly triggering the quadrature interrupt though, my money is on the right side one.

 

Switching current doesn't need to be high.  Most likely, any of them will work there.  Same for voltage.  Operating range (AT) is probably the most critical.  Unfortunately, I don't know what the AT rating of the spinner magnets is, but fortunately they're cheap.  Just look for SPST.

 

A couple of starting suggestions:

https://www.mouser.com/ProductDetail/MEDER-electronic-Standex/GP560-10-15?qs=sGAEpiMZZMvFdY0L2HfHxr7Unu%2B3I5KEIi81FlU0snk%3D

https://www.mouser.com/ProductDetail/MEDER-electronic-Standex/GP560-15-20?qs=sGAEpiMZZMvFdY0L2HfHxr7Unu%2B3I5KE9SJGO%2BhYUA8%3D

 

Also, make sure to orient them so that the contacts are in the same plane as the magnets when you bend the legs.  If the magnet comes up "beside" the contacts because the switch is laying on its side, it won't activate.

[nick3092]

You should be able to test the switches if you have a multimeter/continuity tester. If you look at the end of the sac plug the bottom left hole is 9, then 8, then 7. Pin 8 is the common, and 9 and 7 are each hooked to a reed switch.

 

If you check continuity from 8 to either 7 or 9, you can spin the wheel, and should alternate between beeping or not depending on where the magnet in the wheel is (there are actually 2, 180 degrees apart). If you are in a quiet room and your hearing is good enough, you can hear the switches opening and closing as you spin the wheel.

 

I have some probes small enough to fit in the plug, but otherwise you may have to stick a wire or component leg thick enough to make contact and then put the meter on those.

 

I just confirmed this with my meter and on my sac that is working fine on Phoenix.

[Ikrananka]

5 hours ago, nick3092 said:

You should be able to test the switches if you have a multimeter/continuity tester. If you look at the end of the sac plug the bottom left hole is 9, then 8, then 7. Pin 8 is the common, and 9 and 7 are each hooked to a reed switch.

 

If you check continuity from 8 to either 7 or 9, you can spin the wheel, and should alternate between beeping or not depending on where the magnet in the wheel is (there are actually 2, 180 degrees apart). If you are in a quiet room and your hearing is good enough, you can hear the switches opening and closing as you spin the wheel.

 

I have some probes small enough to fit in the plug, but otherwise you may have to stick a wire or component leg thick enough to make contact and then put the meter on those.

 

I just confirmed this with my meter and on my sac that is working fine on Phoenix.

What resistance are you getting when the switches are active?

 

I've stripped down my problem controller and measured the resistance directly across each reed switch.  On the left one (SW9) it reads zero ohms when the switch it closed.  But on the right one (SW10) it reads 10 to 15 ohms and often very briefly starts with a very high resistance (200 to 300 ohm) that drops fairly quickly down to the 10 to 15 ohms.  The left switch does not exhibit this behaviour and cleanly starts with zero ohms as soon as the switch closes.  So, looks like in this controller SW10 needs replacing.

 

5 hours ago, ChildOfCv said:

But yeah, it's easy to test whether a reed switch is "on" because it just conducts like a standard switch.  Simple continuity would work.  Since it's constantly triggering the quadrature interrupt though, my money is on the right side one.

Wow - how did you know it would the right hand switch?

 

5 hours ago, ChildOfCv said:

Switching current doesn't need to be high.  Most likely, any of them will work there.  Same for voltage.  Operating range (AT) is probably the most critical.  Unfortunately, I don't know what the AT rating of the spinner magnets is, but fortunately they're cheap.  Just look for SPST.

 

A couple of starting suggestions:

https://www.mouser.com/ProductDetail/MEDER-electronic-Standex/GP560-10-15?qs=sGAEpiMZZMvFdY0L2HfHxr7Unu%2B3I5KEIi81FlU0snk%3D

https://www.mouser.com/ProductDetail/MEDER-electronic-Standex/GP560-15-20?qs=sGAEpiMZZMvFdY0L2HfHxr7Unu%2B3I5KE9SJGO%2BhYUA8%3D

 

I have now physically measured the switches and their bodies are ~2.3mm diameter by ~14.2 mm long which is identical to the two switches you suggested.  Was that a good guess or did you know that already?

 

Digi-key have thousands of these in stock and they are only CA$0.73 each.  Would you suggest I buy a few of each in case the AT rating has much of a bearing on performance, or do you think in this range it'll make no difference (there's also a 20-25 AT model in the GP560 range)?

 

5 hours ago, ChildOfCv said:

Also, make sure to orient them so that the contacts are in the same plane as the magnets when you bend the legs.  If the magnet comes up "beside" the contacts because the switch is laying on its side, it won't activate.

Looking at the orientation, SW9 has the flat surface of the switch internals close to horizontal (parallel to the PCB).  However, SW10 looks to be offset axially by 45 degrees - does that make sense?  I'll strip down one of my other SACs to check that it is the same.

[nick3092]

32 minutes ago, Ikrananka said:

What resistance are you getting when the switches are active?

 

I've stripped down my problem controller and measured the resistance directly across each reed switch.  On the left one (SW9) it reads zero ohms when the switch it closed.  But on the right one (SW10) it reads 10 to 15 ohms and often very briefly starts with a very high resistance (200 to 300 ohm) that drops fairly quickly down to the 10 to 15 ohms.  The left switch does not exhibit this behaviour and cleanly starts with zero ohms as soon as the switch closes.  So, looks like in this controller SW10 needs replacing.

Measuring mine through the cable, both switches go from 218 ohms to infinite and back to 218 as I spin the wheel. 

[ChildOfCv]

3 hours ago, Ikrananka said:

Wow - how did you know it would the right hand switch?

The right-hand one triggers the interrupt which tells the system to read the spinner.  If the left hand one is on at the time, then you were spinning left.  If it's off, then you were spinning right.  But both are only active while reading the keypad (or is it the joystick?)  Anyway, that means that if the right switch is closed when the particular mode is selected, it keeps triggering an interrupt.

3 hours ago, Ikrananka said:

I have now physically measured the switches and their bodies are ~2.3mm diameter by ~14.2 mm long which is identical to the two switches you suggested.  Was that a good guess or did you know that already?

Just a guess, though it appears that all the glass ones I saw were about that size.

3 hours ago, Ikrananka said:

Digi-key have thousands of these in stock and they are only CA$0.73 each.  Would you suggest I buy a few of each in case the AT rating has much of a bearing on performance, or do you think in this range it'll make no difference (there's also a 20-25 AT model in the GP560 range)?

Yeah, I think that's a good idea.  It's just a guess, but if the magnet is stronger than the rating of the switch, it may pull too hard and destroy it.  On the other hand, if it's not sensitive enough, then it will never close.  So both range numbers may be important.  I'd go with the least sensitive one (largest AT number) that can be triggered by the magnet.

 

[Ikrananka]

17 hours ago, nick3092 said:

Measuring mine through the cable, both switches go from 218 ohms to infinite and back to 218 as I spin the wheel. 

That's reassuring.  On one of my good controllers I get readings of around 230 ohms both times.  My test leads are really thin and add some resistance so that could easily account for the difference between 218 and 230 ohm.  At least we're in the same ball park.

 

Did you say that the Phoenix detects both directions from your spinner?  Which test program are you using?

[Ikrananka]

Here's an interesting observation.  I stripped down two of my SACs.  One is the problematic one where I get jerky movement and it always registers as spinning to the left.  The other is one that works perfectly on a real CV in both directions.  What's interesting is the orientation/rotation of the right-hand (SW10) reed switch.  See the difference in the images below.  Shouldn't they be orientated roughly the same?  While in both cases the left-hand switch has the reed near horizontal to the PCB, why is the right-hand switch not orientated similarly (they are after all located equidistant from the spinner centre point)?  I'm wondering what the impact might be if I simply re-orientate SW10 in the jerky spinner to a more horizontal position?????  Or maybe axially rotate it roughly -90° to match the position of SW10 in the perfect spinner????

 

Jerky "Left Only" Spinner

 

Perfect Spinner

Edited January 27, 2020 by Ikrananka

[emmanuelf]

RED work the same regardless their orientation. They are in the magnetic field or out.

They are simple, cheap and in this kind of application better than micro switches (rated for 10^8 cycles).

So here the bad one need to be replaced.

Edited January 27, 2020 by emmanuelf

[Ikrananka]

3 minutes ago, emmanuelf said:

RED work the same regardless their orientation. They are in the magnetic field or out.

 

On 1/26/2020 at 12:12 PM, ChildOfCv said:

Also, make sure to orient them so that the contacts are in the same plane as the magnets when you bend the legs.  If the magnet comes up "beside" the contacts because the switch is laying on its side, it won't activate.

 

Seems to be a difference of opinion here.  So, which is it?

[emmanuelf]

Try yourself with a small magnet, you will see. I am confident. (and correction: REED not RED, or ILS in French).

Edited January 27, 2020 by emmanuelf

[ChildOfCv]

I don't have enough direct experience with the switches, but people who claim to know more than I do have said that orientation matters.  It would definitely make sense for dual-pole reeds, since pulling in one direction hits one pole while pulling in the other direction hits the other pole.  But if the switch were, say, designed as a barrel with the contact element on the inside so that any magnetic pull touches the rim, then yeah, any orientation works.  But looking at the close-ups in your pictures, they are swinging rigid levers, in the sense that sideways pull will have little to no effect.  As long as a majority of the force pulls them in a direction that brings them together, though, you should be okay.  But the less lined up it is, the less sensitive I can see it being.

 

In the end, it would have to do with the physical construction of the switch, so close examination with a magnifying glass would reveal that.  Or experimentation with a magnet, as Emmanual suggests.

 

But for reference:  https://www.kjmagnetics.com/blog.asp?p=reed-switches-and-hall-effect-sensors

 

It is possible, perhaps, that the 45-degree orientation is intentional though.  I see that both have it on the right side but are nearly vertical on the left side.  But perhaps that's also a design feature that's supposed to prevent constant triggering of the interrupt signal.  I've been scratching my head over the non-intuitive design of the circuitry there, but perhaps the theory goes that once the magnet gets in range of the right one, it's almost impossible to hold only the right one closed without eventually closing the left one too.  In that case, the fact that both are held closed would prevent multiple triggers of the interrupt, so leaving the spinner magnet off to the right of the switch will still (normally) keep it from constantly registering right rolls.

Edited January 27, 2020 by ChildOfCv

[Ikrananka]

Looking at this reference, https://www.sparkfun.com/datasheets/Components/Buttons/AN104.pdf, it indicates to me that with single pole switch as used in the SAC that it is the orientation and position of the magnet that affects performance.  Looking at the way the magnetic field forces the switch to close, the orientation (axial rotation) wouldn't have an affect on the switch performance as emmanuelf stated.

 

Seeing as I'll likely be replacing the right-hand switch anyway, I may as well remove it and solder it back in rotated to a different position for a little experiment.  But based on what emmanuelf said and the above reference I'm now not expecting any change.

 

1 hour ago, ChildOfCv said:

It is possible, perhaps, that the 45-degree orientation is intentional though.  I see that both have it on the right side but are nearly vertical on the left side.

But one is 90° different to the other.......

[emmanuelf]

Quote

the orientation (axial rotation) wouldn't have an affect on the switch performance

Yes that what I would have said if I was more fluent at thinking in english

Edited January 27, 2020 by emmanuelf

[Ikrananka]

I removed the dodgy reed switch and measured it's resistance with the magnet located at various points around its perimeter.  Resistance measured 200 to 1000 ohm.

 

With the good switch, the resistance is consistently a very low 0.2 ohm regardless of where the magnet is around its perimeter.

 

So, the right switch has definitely had its day and appears to be giving different (high resistance) readings due to the fact that its knackered and not performing as designed.  I will be replacing it.  Also, as emmanuelf stated, my tests with the good switch show that axial orientation does not matter.  Thanks emmanuelf ?  By the way, your English is way better than my French ?

Edited January 27, 2020 by Ikrananka

[nick3092]

I believe those switches are sealed in nitrogen to prevent corrosion. Do you see any cracks in the glass that may have allowed air to replace it causing it to corrode?

[ChildOfCv]

Oh, you mean as long as the magnet is next to the good switch....  if the "good" switch were always 0.2 ohms with or without the magnet, I'd toss that one too.  ?

 

So, now that I've read the paper better, it appears that if the magnetic flux puts an N charge on one leg and an S charge on the other leg, it will work.  If the magnet puts the N/S charge to the sides of the reed switch, or vertically, then no dice.  I'll accept that.

[Ikrananka]

59 minutes ago, nick3092 said:

I believe those switches are sealed in nitrogen to prevent corrosion. Do you see any cracks in the glass that may have allowed air to replace it causing it to corrode?

There are no cracks that I can see and the switch parts inside are still nice and shiny.

 

56 minutes ago, ChildOfCv said:

Oh, you mean as long as the magnet is next to the good switch....  if the "good" switch were always 0.2 ohms with or without the magnet, I'd toss that one too.  ?

 

Oops - yes I wasn't very clear, I did mean 0.2 ohm when the magnet was near.  Moving the magnet away breaks the contact as expected and resistance jumps to infinity (and beyond.....).

Edited January 27, 2020 by Ikrananka

[Tursi]

On 1/27/2020 at 2:28 PM, nick3092 said:

I believe those switches are sealed in nitrogen to prevent corrosion. Do you see any cracks in the glass that may have allowed air to replace it causing it to corrode?

Even so, they are mechanical devices and do have a lifespan. I think it's fair that they are due for replacement.