Apologies if this has been answered somewhere. If so I couldn't find it.
The situation....
I am aware of the risks. I've spent plenty of time servicing and repairing mains (UK) voltage kit professionally. I have a fair knowledge of digital electronics as an amateur but very little analogue.
At home I have a HVAC system. The control panel has a number of neon state indicators on the control panel. I want to interface these to an arduino.
The neons themselves are easily accessible and easy to tap into.
The series resistors for the neons are on the control pcb and I can't get at them (or I could but I'm not going to!). From the circuit It looks like I'm getting 240v, 1ma or thereabout at the neon. (If I've understood how neons work I believe the voltage would drop when lit.)
The problem....
I have no issue once I get to 5v or a dry contact relay output. It is how best to get to that stage I am struggling with.
1ma is nine tenths of naff all in my world. No chance of driving a relay and even an opto isolator looks pretty dodgy. The ac opto-isolator HCPL-3760 (suggested in an archive thread) looks to require 1.2ma.
The "classic" solution seems to involve an ldr, dob of blu-tac and black nail polish. Sophisticated versions replace the blu-tac and nail polish with a cardboard tube and and/or a roll of black tape. I'll go there if I have to, but It does all sounds a bit Blue Peter'ish to me. I know how I'd react if I opened something I had been called in to fix and found that inside!
If that's the only way to go I'll pot up a proper looking neon/ldr solution, but before I do so am I missing something? Is there a better way to detect a low current 240v signal? (eg does anyone make a neon based opto isolator?)
A phototransistor is what is used now, LDRs are slow, insensitive and hard
to find now (many contain cadmium which is very toxic). Phototransistor
can probably rigged up to generate a logic signal direct with a high enough
load resistor, select on test. http://learn.parallax.com/sites/default/files/content/shield/robo_ch6/series_sch.jpg
However a neon is striking and breaking arc 100 or 120 times a second so
you'll get an ac signal out and need some de-bouncing in software.
Thanks Mark - Shows just how long ago I had to deal with a light sensor directly!
If I have to go this way I'll probably add an rc circuit , feed the output into one of the analogue pins and average (in software) the signal over 100ms - that'll be pretty robust.
Still hoping I don't have to get the potting compound out tho!
A neon will clamp to about 90V I think so if you put a lower-voltage device (LED in opto isolator) in parallel with the neon, the neon will probably no longer light once the LED takes the current.
Some opto-isolators have very high current gain, like 5x or more. They should allow you to detect a mA or so with no trouble, particularly if you load up their output side with about 20k. Even a 6N138 should work OK with 10k load to reliably detect 0.5mA with >5V swing.
Note that because of the diode input, you will get an output at 50/60Hz with a bit under 50% duty cycle so the software needs to be aware of that.
Thanks Polyglot. I just had a look at high current gain opto-isolators and it looks like one will do nicely.
I did a bit more reading yesterday and figured out that I would lose the ability to drive the existing neon directly whatever I did. That being the case I'll give the opto-isolator a go. (I can always drive an led (or even the existing neon) from the output signal easily enough.) I'll try the HCPL-3760 ('cos it has 240v ac input and internal bridge rectifier) and it'd be useful to get to know it anyway. If that doesn't work I'll try a higher current gain one. Thanks for putting me on to that.
The 3760 needs 1.2mA nominal to turn on. I reckon a cheap old 6N138/6N139 (0.5mA threshold) would server you better, even if you only get half-wave from it. They have a minimum gain of about 4, typical gain of 11. Or an HCPL-2731 has a typical gain of 35 at 0.5mA!
You won't be driving the neon (directly) from the output of one of these; a neon needs a lot more voltage. LEDs, sure.
OK sir, I will take your advice and try the 6N138/9. (I'd better get a couple in case I screw up!)
No I wasn't thinking of driving the neon directly (!). Many of the systems I work with have dry contact on/off outputs (relays). It makes it pretty easy to hook up whatever you want - mains, 24v, 5v and gives you a decent current capability. I know in this case it looks like overkill, but it would both future proof it and be familiar to anyone else who has to work on it. I was thinking of dropping in a transistor driven small DPDT relay, one pole for the output and one for the existing neon.
Many thanks for your help and advice. Very much appreciated.
