need polarity-insensitive voltage detector

Discussion in 'General Electronics Chat' started by TWRackers, May 24, 2009.

  1. TWRackers

    Thread Starter Active Member

    Dec 29, 2008
    41
    2
    I have a project where I need to sense the output pulses from altimeters such as those used in the recovery systems of larger hobby rockets, so that each altimeter output (two altimeters, two channels each) can produce a TTL-level signal (active-high preferred but not essential) suitable for sensing by a BASIC Stamp. One altimeter produces 500 ms 9-volt (nominal) pulses on each channel, while the other produces 50 ms pulses, same voltage range. Both altimeters have capacitive-discharge output stages to produce a short high-current pulse to fire electric matches in the rocket's recovery system. Oh, I also need the circuit to work correctly no matter the polarity of the input signal; e-matches don't care, but I don't want to have a circuit that will fail if the input signal is flipped.

    My initial attempt was only half successful. I tapped onto each channel's output with a full-wave bridge rectifier of 1N914 diodes, and the output side of the bridge was connected to the input side of an NTE3086 dual optoisolator with a series resistor. I used a pull-up resistor on the open-collector output side of the opto channel and fed the output signal into the BASIC Stamp.

    This circuit worked fine with the altimeter with the 500 ms pulses without interfering with the output signal, but it presented too great a load on the other altimeter, apparently sucking too much of the stored charge on the altimeter's output driver. I have to use these two altimeters, since I already own them and have mounted them in the rocket's avionics bay. The two altimeters form a redundant pair, and I'm using two different models to avoid any design flow catching me twice in the same rocket.

    One important detail: the power supply for the BASIC Stamp is separate from the supplies (9 volt batteries) of the two altimeters. I think a circuit using some LM339 quad comparator chips could be devised, using a pair of comparators for each channel to define a window, say -1 volt to +1 volt, where the output is de-asserted, and any input outside that window asserts the output. I don't want to couple the circuit grounds of the two altimeters and the BASIC Stamp if I can get away with it; that's why the optoisolator solution seemed ideal.

    Maybe a circuit that directly senses the current pulse through the leads to the e-match? Small toroidal transformers?

    Any suggestions?
     
  2. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    Why don't you use a transistor (or Darlington) after the bridge with a rather large-value base resistor to drive the LED side of the optocoupler? An MPSA14 Darlington, 2N6426 Darlington or the like would do the trick. They're both available in TO92 packages.
    [eta]
    You'd be better off putting the resistance (split in half) on the input side of the bridge. That way, if the bridge happens to fail shorted, your "electric match" will still fire. You don't want that rocket coming straight down and trying to bore a hole through the planet.
     
    Last edited: May 25, 2009
  3. TWRackers

    Thread Starter Active Member

    Dec 29, 2008
    41
    2
    Yes, I can see the logic of moving the series resistance to before the bridge. As for the transistor, all I have ready access to are the two leads of each channel's output; wouldn't I need a supply voltage on the collector terminal which shares a ground reference with... oh, I would need a separate battery whose ground (negative) is referenced to the cathode end of the opto input LED. Since the two altimeters are completely isolated from each other now....

    Okay, it's late, I'm going to think on this and then in the morning sketch up what I think you're suggesting. Thanks for the tips. ;)
     
  4. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    Have a look at the attached schematic. Note that the connections to the electric match is before the resistors. V1 is the simulated input signal, but I only used 5v in the simulation. If your actual input signal is 9v, increase R1 and R2 to around 4.7k-5.1k each. This will lower Q1's base current to under 1mA. Even if one or two diodes in the bridge happen to fail shorted, you should still have plenty of power to fire the match.

    Consider the ground on the output side of the optoisolator separate from the grounds on the left side of the bridge and the base of Q1. They needed to be there for the simulation to run.
     
    Last edited: May 25, 2009
  5. TWRackers

    Thread Starter Active Member

    Dec 29, 2008
    41
    2
    Okay, I've taken most of the details of your schematic, fiddled with a few details, and replicated the circuit to show all four channels. The only potential gotcha involves the internal connections of the two altimeters. Since I don't know in advance the polarity of the wire pairs for each channel, and I don't know if the + terminals or the - terminals on either altimeter are connected on the boards, the circuit needs to work properly no matter how each channel is connected, as long as there are no cross-channel connections. Since everything is wired with twisted pairs and 2-pin Mate-n-Lok connectors, cross-channel connections should not be possible. I could drive the opto LEDs with an onboard watch battery or similar, since each LED should be pulsed at most once per flight.

    So, did I miss anything?

    (If my schematic looks a little non-standard, this is my first one drawn with ExpressSCH, and circuit design is not my usual line of work.)
     
  6. TWRackers

    Thread Starter Active Member

    Dec 29, 2008
    41
    2
    I just remembered another detail: each altimeter has a continuity-check circuit for each channel, to alert the user upon power-up whether or not the output channels have continuity through the e-matches. With a little research I found out a load as high as 100 kΩ will still be sensed as good continuity for one or both altimeter models, while 1 MΩ shows no continuity for either model. THAT is why I've had comparator circuits rattling around in the back of my mind for this application for the past couple of months, ever since I determined that my version 1.0 wasn't working. That was an unexpected side effect I saw when I tested my original design, that even with no e-matches connected, at least one altimeter showed a false positive on continuity because of the current passing through the bridge and the opto LED. The circuit in my previous post will exhibit the same behavior. :(

    I've got a bag of small toroidal cores with a single wire looped 10-12 times through each one. Maybe I can still think up a circuit that can use the cores as current-sensing coils. Then the altimeter outputs will be completely isolated from any other circuitry. These e-matches, depending on the type, have resistances around either 1 Ω or 10 Ω, so either way the current pulses are going to be somewhere between 0.9 A and 9 A.
     
    Last edited: May 25, 2009
  7. TWRackers

    Thread Starter Active Member

    Dec 29, 2008
    41
    2
    Okay, now that I look at my schematic again, maybe it won't exhibit the false-continuity behavior, since the current path from the altimeter output goes through the transistor gate to its emitter. Problem is, how do you keep the current low enough to prevent false continuity and still have enough current for the diodes to rectify the signal? If I remember correctly from physics electronics lab, don't diodes have to have some minimum current drawn through them before they'll work? Or is it a moot point at the low voltage it must be applying to sense continuity?

    Perhaps this is an application where you need the minimal base current of a transistor rather than the much smaller gate current of an FET?
     
  8. TWRackers

    Thread Starter Active Member

    Dec 29, 2008
    41
    2
    ping....

    Well, it's been three months since the last post to this thread and I still need the answer to a question. As shown in the schematics in posts #4 and #5 above, the proposed circuit has an input signal passing through a full-wave bridge rectifier, with the output connected to the gate of a Darlington. Considering the small amount of base current that a Darlington would draw, will the diodes actually go into conduction when they're forward biased?
     
  9. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    Well, the diodes have a Vf of around 0.7v each, and the base-emitter junction would be another 0.7v or so. Basically, you'd need about 2v before any base current would begin to flow. Then, there are two 2.2k resistors in series with the rectifier diodes and the base-emitter junction. Base current will then depend upon the 4.4k total resistance after the input exceeds 2.1v, or (Vin - 2.1v) / 4400 Ohms.

    The Darlington when saturated will have a Vce of around 1v. The optoisolator will likely have a Vf of around 1.6v. A typical fresh 9v battery will output around 8.6v under light to moderate load. So, (8.6v - (1.6v + 1v)) / 330 Ohms = 18.2mA current through the optoisolator's LED.

    As far as whether the rest of the telemetry has common grounds - I don't have a clue.
     
  10. TWRackers

    Thread Starter Active Member

    Dec 29, 2008
    41
    2
    Well, that sounds hopeful. The continuity test circuitry of the altimeters require something near 1 MΩ across their outputs before they won't sense continuity, so it's probably a test current on the order of 10 μA. My earlier circuit took the output side of the FWBR directly across the opto's LED and its series resistor, but there was apparently enough leakage current that the altimeters sensed continuity even when only my circuit (and no electric match) was connected. Maybe the base-to-emitter junction of the Darlington will make enough of a difference so that I won't get false continuity indications. Looks like a trip to Mark Electronics is in the offing today to pick up a Darlington for test purposes. ;)
     
  11. Ron H

    AAC Fanatic!

    Apr 14, 2005
    7,050
    656
    A Darlington will not reduce current from the altimeter unless you run it as an emitter follower. Otherwise, the Darlington will saturate, and current will flow through the forward biased base-collector junction.

    You have to be careful in your choice of Darlingtons. Some have integral resistors which may give you false continuity signals.

    What is the peak voltage on the altimeter pulses?
     
  12. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    Well, there is a total of 4.4k Ohms between the altimeter, and the Vf of the rectifiers in the bridge. If the altimeter's output is greater than 3v or so, that should saturate the Darlington pretty well.

    Early on I suggested that our OP use something like an MPSA14 or 2N6426, neither of which incorporate resistors. Since the telemetry will now be isolated from the "electric match", there is not currently a provision for such a test.
     
  13. Ron H

    AAC Fanatic!

    Apr 14, 2005
    7,050
    656
    I thought that there was concern for false continuity readings on power up.
    I was pointing out that the Darlington, when saturated, will draw enough base current to cause continuity to be indicated, whether or not the electric match is connected. It sounds to me like TWRackers is concerned about this.
     
Loading...