I'm goofing around with having a PIC generate a signal that is then feed
into a tachometer. Normally, this tach is feed by a hall sensor (gear
tooth). I scoped the signal coming off the sensor and have been able to
mimic it with a PIC. Fortunately it's a nice square wave that's easy to
generate. My next problem is that the tach expects a 0-9v range. OK. So I
tried messing with an opto that I had lying around (6N136). Unfortunately,
the transistor is a bit slow to switch off so I get a great leading edge but
a sloped trailing edge and I never really get back to 0 before the next
pulse is generated.
Should I try one of the logic gate optos or would it be better to use an
op-amp to change the voltage offset to a 0-9v range? Whatever I choose, I
need it to run at no more than 4KHz and as square wave as possible.
>Friends,
>
>I'm goofing around with having a PIC generate a signal that is then feed
>into a tachometer. Normally, this tach is feed by a hall sensor (gear
>tooth). I scoped the signal coming off the sensor and have been able to
>mimic it with a PIC. Fortunately it's a nice square wave that's easy to
>generate. My next problem is that the tach expects a 0-9v range. OK. So I
>tried messing with an opto that I had lying around (6N136). Unfortunately,
>the transistor is a bit slow to switch off so I get a great leading edge but
>a sloped trailing edge and I never really get back to 0 before the next
>pulse is generated.
>
>Should I try one of the logic gate optos or would it be better to use an
>op-amp to change the voltage offset to a 0-9v range? Whatever I choose, I
>need it to run at no more than 4KHz and as square wave as possible.
>
>-DO
>
The 6N136 is spec'ed at 0.2us typ, and 0.8us max. turn time and 0.6us typ. and
0.8us max. turn off time with a 1.9K load resistor with 16ma of drive.
This should be fast enough for your needs. Is the output of the opto driving a
large amount of capacitance such as a cable?
I'll add a quick rule of thumb about opto's:
Photo diodes have about a 1us response time, photo transistors have about a
10us
response time, and photo darlingtons have about a 100us response time. These
times assume a "reasonable" resistor value. There are amplified optocouplers that
are in the 100ns range as well.
> Hi!
>
> If it is a simple square wave... try a single transistor and two
> resistors... in an invertor configuration...
>
> Email me if you want more details.....
>
> Your OPTO should work too... but use a smaller pull up resistor.
>
> Cheers,
> Don
>
> At 12:13 PM 2/20/00 -0500, you wrote:
> >Friends,
> >
> >I'm goofing around with having a PIC generate a signal that is then feed
> >into a tachometer. Normally, this tach is feed by a hall sensor (gear
> >tooth). I scoped the signal coming off the sensor and have been able to
> >mimic it with a PIC. Fortunately it's a nice square wave that's easy to
> >generate. My next problem is that the tach expects a 0-9v range. OK. So I
> >tried messing with an opto that I had lying around (6N136). Unfortunately,
> >the transistor is a bit slow to switch off so I get a great leading edge but
> >a sloped trailing edge and I never really get back to 0 before the next
> >pulse is generated.
> >
> >Should I try one of the logic gate optos or would it be better to use an
> >op-amp to change the voltage offset to a 0-9v range? Whatever I choose, I
> >need it to run at no more than 4KHz and as square wave as possible.
> >
> >-DO
> >
Through some help from your guys, I got it figured out - sorta.
After mulling about the responses, I started to snoop around my circuit.
First of all, bonehead move on my part, the GND was floating - I tied it one
"hole" off on my breadboard. That gave my first funky wave.
Secondly, I could generate the right signal without being coupled to the
tach. Using the 1.9k resistor (which was my second mistake), I got what I
wanted. However, when I connected the signal lead to the tach, the wave goes
flat. For some bizarre reason, I removed the 1.9k load resistor and I get my
wave and the tach starts to swing.
Now, it comes down to the fact that the tach signal input measures about
2.4M Ohms. I should have measured this first and then adjusted my circuit.
My fault in assuming I could generate a nice wave and then plug it in.
I've had it running for about 12 hours now and there's now ill effects on
the Opto.
I've been able to drive this opto up to about 4KHz with enough definition to
make the tach register. After that, either the tach can't count it or the
trailing edge slopes too much.
> The 6N136 is spec'ed at 0.2us typ, and 0.8us max. turn time and
> 0.6us typ. and 0.8us max. turn off time with a 1.9K load resistor with
16ma of drive.
> This should be fast enough for your needs. Is the output of the
> opto driving a large amount of capacitance such as a cable?
>
> I'll add a quick rule of thumb about opto's:
> Photo diodes have about a 1us response time, photo transistors
> have about a 10us response time, and photo darlingtons have about a 100us
response
> time. These times assume a "reasonable" resistor value. There are
amplified
> optocouplers that are in the 100ns range as well.
>I'm goofing around with having a PIC generate a signal that is then feed
>into a tachometer. Normally, this tach is feed by a hall sensor (gear
>tooth). I scoped the signal coming off the sensor and have been able to
>mimic it with a PIC. Fortunately it's a nice square wave that's easy to
>generate. My next problem is that the tach expects a 0-9v range. OK. So I
>tried messing with an opto that I had lying around (6N136). Unfortunately,
>the transistor is a bit slow to switch off so I get a great leading edge
but
>a sloped trailing edge and I never really get back to 0 before the next
>pulse is generated.
The opto should be more than fast enough.
I suspect that you are using it with a large load resistor and that
capacitive effects are predominating.
If you use a resistor R from Vcc to opto collector and place opto emitter to
ground then it will turn on rapidly. When you turn off the resistor R must
pull the opto output high by itself.
If you have a capacitive load here the time taken will be several times the
RC product.
How big is your load resistor?
You don't say how fast your tach needs to run - lets say it is 6000 rpm =
100 Hz.
With a 1 megohm resistor you shouldn't be driving more than about 0.002
microfarad capacitive load.
For starters try a much smaller load resistor - say about 10k. This will
draw under 1 mA max at 9 volts.
Ensure you don't have a large capacitive load.
You could drive perhaps about 0.1uF load with this but you wouldn't want to
and shouldn't try to.
As its off a gear tooth the frequency may be even higher than this. How many
pulses per second are you trying to handle? - should still be in range of
opto.
If you MUST drive a substantial capacitive load use a buffer (transistor)
between opto and load and preferably use an active circuit that drives the
load in both "up" and "down" directions.
If you still have problems show us the circuit or describe it.
Maybe you can help me understand the phenomenon. I did post a message to the
list on what I've found so far, but I'll fill in some answers for your
questions....
> The opto should be more than fast enough. I suspect that you are using it
with a large load resistor and that
> capacitive effects are predominating.
Actually, one of my first mistakes as no load resistor because it drove Vo
to a constant high (9v). Now, this was also when I had the output GND
accidentally floating. Once I *properly* grounded, I needed about 4.7k to
get the wave to go from 0-9v. Using the 1.9k recommended gave me an offset
of 5 volts - so the wave was 5-9v. When I scoped the hall sensor, there was
a clear 0-9v swing (I wanted to mimic the output as best as I can).
> If you use a resistor R from Vcc to opto collector and place opto
> emitter to ground then it will turn on rapidly. When you turn off the
resistor R must
> pull the opto output high by itself. If you have a capacitive load here
the time taken will be several
> times the RC product.
I'm going to have to research this. Unfortunately, I've had a hard time
xrefing the parts on the tach to any known part (it's a '94 vintage unit).
Not knowing what the tach is doing with the signal, I'm extremely lucky that
I got it to work. I'm a little unclear if the tach is a capacitive load - I
do know that the tach input signal to ground is about 2.4Meg ohms.
I do know that if I scope the hall sensor without it connected to the tach,
I get an incredibly dirty signal (of course, no load). Once I connect it to
the tach and then scope the input, I get a nice wave. Without ripping apart
the tach I'm not sure what they've done for signal conditioning.
> How big is your load resistor?
Nothing now. If I add one, I get a flat line.
>You don't say how fast your tach needs to run - lets say it is 6000 rpm =
> 100 Hz. With a 1 megohm resistor you shouldn't be driving more than about
0.002
> microfarad capacitive load. For starters try a much smaller load
resistor - say about 10k. This will
> draw under 1 mA max at 9 volts.
470Hz = 2000
1.4 Hz = 6000
I'm horrible as ASCII, here's a text example
PIC18F84:
RB1 - to 250ohm resistor to Opto anode
Opto:
Anode - from PIC
Cathode - to ground
.1uF cap from anode to ground
9v in to Vcc (transistor base)
Gnd to GND (transistor emitter)
.1uF cap from Vcc to GND
Vo out to tach
Tach load is about 2.4Meg ohms.
That's it. It works. I'm assuming the tach load replaces the load resistor
???
