6N137 low output

[ltourist]

Hi there,

I'm building a MIDI enabled device. The first prototype worked very nice, so I went on and build the second iteration that corrects all sort of mistakes on it. That time, the MIDI signal is not pickup and after some digging, I've found that the ouput of the 6N137 optocoupler was very low and "dirty" compared to the first iteration. Moreover, the signal low is not always at the same level. I'm wondering if anyone can provide any advice where to look for ? Could it be the 6N137 itself that could be failing ?

I've attached two screenshots from my oscilloscope. Channel one is the 6N137 input, channel 3 the ouput.

Also, I've attached the schematics used for MIDI I/O, based on that page. Although there is some difference, that wiring has always worked through all the development stages and on the first prototype without a glitch.

Thanks,
LTourist

[David_AVD]

For a start you need to connect pin 7 of the 6N137 to Vcc ( pin 8 ).

[ltourist]

Thanks. I can jumpwire that  . But according to the datasheet on page 2, that would limit "Input Voltage Not to Exceed VCC by more than 500mV". As I understand it, it can offers some extra protection against voltage spike, is that corrent ?

[Tac Eht Xilef]

Thanks. I can jumpwire that  . But according to the datasheet on page 2, that would limit "Input Voltage Not to Exceed VCC by more than 500mV". As I understand it, it can offers some extra protection against voltage spike, is that corrent ?

It's a little odd that V_E is specified in the "Emitter" section - I'd suspect that placement is an error. However, that table is also the Absolute Maximum Ratings, so you need to read it in conjuction with the Recommended Operating Conditions table below.

I guess what they're trying to get at is that V_E should not exceed V_CC+500mV, with an absolute maximum limit of 5.5v.

It's not for spike protection - V_E is  an 'output enable' signal; check the truth table on page 1. Though don't check it too closely, since it kinda suggests that leaving it NC is the same as pulling it high. That's bad practice since it can be prone to picking up noise, which leads to the kind of symptoms you're seeing.

[ltourist]

Thanks for the clarification. I've obsviouly read that document way too quickly, putting the context of that statement away, sorry. Indeed I checked that table a while ago during the breadboard stage and the fact that NC has the same truth table as pulling Ve high might explain why I didn't wire that pin 7. Can't remember, it's been a while.

Now I can understand why I'm getting that noisy signal. But the fact that the first prototype doesn't produce any noise worries me as it could mean that my second design is worse than the first attempt on that point.

Anyway, I'll try to fix it tonight and keep you posted with the results.
Thanks to the both of you

[David_AVD]

Different layout, leakage currents or just a different brand of opto can make enough difference sometimes.

[Kjelt]

If you look at both input and output of an optocoupler with a scope what is your reference ground? 
If on the secondary side Vcc to Vss is 5V then just and only connect one probe correctly to the output of the optocoupler and look again, probably looks very good 

[ltourist]

Different layout, leakage currents or just a different brand of opto can make enough difference sometimes.

That's what worrying me. It is a different layout.

If you look at both input and output of an optocoupler with a scope what is your reference ground? 

That's a very good point I didn't think about until now. But as you said, Vss = Vcc here.
Anyway, I'll try to wire up Ve to Vcc and come back to you 

[ltourist]

Wiring Ve to Vcc totally did it ! See the attached picture 
Thanks a lot for helping out and the explanations !

[LukeW]

I've had successful results leaving the 6N137 enable signal unconnected before - but as pointed out above, it's not perfect in terms of noise immunity. Not sure if it has some sort of internal pullup or how strong that is.

[David_AVD]

Do you have the required 100nf capacitor between pins 4 and 8 of the 6N137?  Is is nice and close to the chip?

[ltourist]

Do you have the required 100nf capacitor between pins 4 and 8 of the 6N137?  Is is nice and close to the chip?

Yes I do, right next to it

I've had successful results leaving the 6N137 enable signal unconnected before - but as pointed out above, it's not perfect in terms of noise immunity. Not sure if it has some sort of internal pullup or how strong that is.

That's also my case in a previous device I built, on the breadboard and on the first prototype. That's why I've never took care of that Venable before 

[David_AVD]

I've always connected pin 7 to Vcc, even though the 2 data sheets I checked show it as optional.  The TI data sheet mentions it specifically ("has internal pullup").

Maybe there are other brands (or copies) out there that don't have the pullup ?