next topic »

[on:]

Hi guys!
I'm trying to interface an old Seeburg Wallbox with my Arduino. The Wallbox sends a timed signal using 24VAC. My idea was to grab that signal using an optocoupler, since, to me, it seems like the best way to isolate the pure-AC circuit of the Wallbox from the pure-DC circuit of the arduino.

My question would be if anyone of you knows about an optocoupler that can handle 24VAC at 2A max. (I doubt it'll ever reach the 2A, but it'd be nice just to be safe) on the LED-side and the default 5VDC of the Arduino on the transistor side?

Thanks a lot in advance!
Philipp

[Reply #1 on:]

What is a Seeburg Wallbox? You won't get 2A through an optoisolater.
Is it a power supply because you can't transfer power through an opto isolator only signals.

[Reply #2 on:]

The 2A doesn't matter because you will limit that with a resister.

The +24V doesn't matter because of the same resistor.

It's the -24V that may be an issue depending on the max rev voltage of the opto LED. So stop that with a normal diode in series with the LED.

______
Rob

[Reply #3 on:]

Could you run it through a diode and put a resistor on that, even just to feed a regular opto?

Just a quick look shows me some that can take 24-120V AC.
http://search.digikey.com/scripts/dksearch/dksus.dll

[Reply #4 on:]

I'm trying to interface an old Seeburg Wallbox with my Arduino.

Yes, right resistor and a diode in series to block the reverse 24VAc should work... The peak voltage of the waveform will be about 34 volts (24 * square root of 2?); use that in picking the resistor to put about 20 ma through the optocoupler LED. Almost any Optocoupler should work..

These wall boxes (as seen in thousands of USA Diners, years ago) received coins and button pushes and drove mechanical relays in the JukeBox music player.

As I recall they drove "stepper switches" in the jukebox to select the record to be played.

I think there are some discussion groups online that know a lot about these; Google Away...

(I gotta get one of these off Ebay sometime) 

[Reply #5 on:]

Hey guys!
First of all, thanks a lot for your answers!

@Grumpy_Mike: No, it's not a power supply, it's the signal line coming from a Seeburg Wallbox (http://www.flippers.com/images/Prop-Seeburg-wallbox.jpg)

I've just been to my local electronics shop and they've told me that an opto-coupler won't work at all for this.. kinda strange.. The man said something along the lines that

a) There's no opto-coupler that allows AC on it's LED (makes sense now that I think about it..)
b) When simply using a resistor and diode to only get the +24V-0V part he said the signal would get "distorted", which has something to do with the waveform of the pulse.

Thing is.. can't I just "recalculate" the correct signal in my code? I know that the pulses are seperated by something between 300-500 milliseconds, so I guess that there should be enough room to detect a "gap" in the signal and make out where one pulse begins and ends, right?

Greets,
Philipp

[Reply #6 on:]

he said the signal would get "distorted"

Technically that's correct, it's also irrelevant in this case (I think, see below).

no opto-coupler that allows AC on it's LED

Half of the AC is just varying DC, so this doesn't make much sense either unless he though you wanted to get AC out the other side.

Exactly what do you plan to do with the output signal?

If you just need to know that the AC signal  exists that's one thing, if you need an exact replica then your shop guy was right.

Anyway doing as we suggested you should get pulses out the digital side, you can detect them to determine if the signal is there or not.

BTW I'm not convinced you really need the isolation of an opto, but I suppose it never does any harm.
______
Rob

[Reply #7 on:]

Perfect! Yes, knowing that there IS a signal is enough for my project. Essentially, I want to check if there's a signal (haven't decided whether to use interrupts or loops yet) and if so, check the time between the current and next signal. There's an exact combination of signals for each title selection you can make on the box. I'll then send these via bluetooth to my PC/Laptop/etc. to play an according tune

I've just read about diode bridges/bridge rectifiers. Would it also be an option to convert AC to DC using a diode bridge like this one (http://en.wikipedia.org/wiki/File:Diode_bridge_alt_1.svg) and then run the output DC through a regular opto-coupler?

Greets,
Philipp

[Reply #8 on:]

All the bridge does is give you twice as many pulses, might be useful if you need more accurate timing.

What is the frequency?

Is there a coded sequence of AC/nothing you have to decode?

_______
Rob

[Reply #9 on:]

Tbh, I have no Idea.. my multimeter isn't accurate enough to measure the pulses in time, so my best chance of finding out when exactly the pulses fire is by hooking it up to my arduino constantly checking its input.

It basically works like this:
There's letters (A-V, omitting some like O or I) and numbers (0-8), making a total of 160 possible selections. When a combination of Letter+Number has been pressed, a motor moves a contact arm first over numbers 1-8. The contacts are wired in series, so if I'd press 5, pulses would be fired when the contact arm went over 1,2,3,4 and 5. Same thing then happens again for the letters, with a longer pause in between to detect the transition from number to letter.

Documentation on these machines (especially on the parts you usually wouldn't have to worry about when hooking it up to a regular jukebox) is quite hard to find, so unfortunately I haven't yet found anything about the pulse timing..

Greets,
Philipp

[Reply #10 on:]

So you really need some more data before you can decide exactly how to do it but it sounds like in the end you don't need do anything clever, just count a few pulses.

Before you can do that you need the timing information so yes you're on the right track if you don't have any proper test gear.

All you need to know is the gap between each number/letter in a burst of AC pulses and and gap between the two groups of bursts.

So I would rig up the opto as we've been discussing (with or without a bridge, as this is mechanical it will be slow so the double resolution shouldn't be needed but won't do any harm) and run the pulses into an Arduino external interrupt.

You could also just use a diode or bridge and two resistors forming a voltage divider.

Either way it may also be necessary to clean up the sine wave (assuming that's what it is) with a Schmitt trigger or a comparator with some hysteresis.

In the ISR simply save the millis() value into an array.

When the sequence is finished print the values to the screen and analyse them. Each reading will be the (say) rising edge of a pulse so you will easily see the patterns. Put the values into Excel and do a graph or calculate the time between each reading .

_______
Rob

[Reply #11 on:]

Hi Rob!
Thank you so much for helping me out this detailed

Apparently so, yes. From the looks of it I can't seem to find anything more complicated to the mechanism than this - considering it's a consumer machine from the late 50's, this might even hold true.

Since it's a cheap part, I'll additionally use the rectifier bridge. If the Schmitt-Trigger is needed, I'll add one, but for now I'll try with as few parts as possible to not make it too complicated for me.. been a while since I've fiddled with electronics

Yea, that's pretty much what I was planning to do. Then once I have the correct sequences, it'll just be a matter of deciding on which device to decode them. I'll probably do it on the arduino, since I'm not too sure about sending time-critical data via bluetooth, although after some tests it shouldn't really matter.

As for the opto-coupler itself, I've been looking at this one:

http://www.produktinfo.conrad.com/datenblaetter/175000-199999/187038-da-01-en-PC817.pdf

Though I'm not sure which ratings to use to calculate my resistor values, etc.. Reading these data sheets is indeed quite hard.

Greets,
Philipp

[Reply #12 on:]

That opto seems to need 20mA forward current which is a lot but that doesn't matter here.

Terry had the voltage numbers above, I'd use the Vf of 1v2, just under the max of 50mA (say 45) and the peak V for the LED current calcs, it doesn't matter at what point in the waveform the LED turns on.

As you say try without cleaning the pulses, if there's a problem either add the Schmitt trigger or debounce in software.

BTW I just love the idea of a 50s jukebox controlling a PC/iPad etc.

______
Rob

[Reply #13 on:]

Sorry my DigiKey link went garbage. Yes, there are AC capable "Optoisolators - Transistor, Photovoltaic Output" at DigiKey and probably elsewhere.

Manufacturer is Weidmuller.  24VAC, 120VAC.  

Weidmuller parts #'s -- 24VAC/DC p/n 0609860000, 24VAC/CD p/n 8234580000, 120VAC/CD p/n 8234600000

Maybe these links will work:
http://search.digikey.com/us/en/products/0609860000/0609860000-ND/491692 ---- only $60 and change
http://search.digikey.com/us/en/products/8234580000/8234580000-ND/492853 ---- a steal at $62-some
http://search.digikey.com/us/en/products/8234600000/8234600000-ND/492852 ---- likewise

Holy cow!

[Reply #14 on:]

H11AA1, -AA2, -AA3, -AA4  AC Input/Phototransistor optocoupler, < $1
UL-recognised

With a nominal 24vac, use a 4700Ω resistor (for 5ma), 1/4W.