Trying to make a 4 JFET Wheatstone bridge

Discussion in 'The Projects Forum' started by Darkstar, Sep 3, 2010.

  1. Darkstar

    Thread Starter Member

    Sep 3, 2010
    104
    0
    Hello, this is my first post here. I should let you know that I am unfamiliar with transistorized circuits so I'm doing a lot of guesswork here
    and I could use some good explanations for what I should do.

    I am currently working with a Wheatstone quarter bridge containing 1 JFET to sense atmospheric static fields, and the circuit is working.
    However, having 4 sensors in a bridge gives a signal 4 times as big. This is called a full bridge. I'm looking for a simple way to connect 4 JFETs
    so they are all operated by a single antenna wire and their resistances will rise and fall appropriately to give me a larger voltage drop across
    the bridge. I don't want to use an amplifier to increase a small, noisy signal, I want more sensitivity to be able to detect smaller changes in
    atmospheric potentials such as when thunderclouds pass overhead.

    I have included a schematic of what I have so far.

    It seems like it should be a simple enough circuit but I can't find anything like it and I don't know enough to figure it out myself.

    I don't expect all the FETs to react to the same extent, but I thought there would be some changes in the right directions. I know the circuit
    is not perfect, but I don't care. Just looking for a varying DC voltage across the bridge proportionate to the voltage sensed by the antenna.
    I also don't care about temperature effects on the outdoor FET.

    My first version of the circuit had no bias resistors. This is common in many static sensing circuits so I kept up the tradition, but it was
    suggested that I needed at least one for T4, so I added one in this second version. The FETs are simply acting as voltage controlled resistors anyway, passing DC only. There are no AC signals involved (except spikes from lightning.)

    Brief explanation of operation:
    The gate of T1 senses a positive voltage, for example, so it becomes more conductive. The drain side of T1 becomes more negative so the
    gates of T3 and T4 become negative and they conduct less. Meanwhile, T2 gate is more positive than T3 and T4 so it conducts more, just like
    T1. And vice-versa.
    The result is 4 changing resistances instead of one so more signal output.
    T1 and T2 work together and T3 and T4 must work together but opposite to T1 and T2.
    The purpose of this circuit is to get away from using resistors. I have a working circuit now with just T1 as the only FET and 3 resistors. But
    if the resistors can vary, like T1, the bridge output will be greater.
    One reason I tried to stay away from using bias resistors is that the extra resistance will increase the total resistance of the bridge, and it is
    most sensitive to small signals picked up by the antenna when the bridge has a small resistance.

    I've checked this using the Wheatstone Bridge calculator I found online at:
    http://www.daycounter.com/Calculators/WheatStone-Bridge/WheatStone-Bridge-Calculator.phtml

    The FET connected to the antenna is outdoors with only the source and drain wires coming indoors. The rest of the circuit can be built
    indoors and connected to my current circuit.

    I would like some opinions about whether or not you think this circuit would work. If not, where did I go wrong and how should I fix it?
    Please keep the explanations simple, this is just a hobby and I'm not familiar with using or biasing transistors.

    Thanks
     
  2. beenthere

    Retired Moderator

    Apr 20, 2004
    15,815
    282
    There are a number of thing wrong with the circuit. The most important one is that measuring bridges do not work like that. The variable elements are arranged differently - swap T2 & T4. The signal is also differential in nature as in the Ebook explanation - http://www.allaboutcircuits.com/vol_2/chpt_12/5.html

    Someone with lots more RF background may prove me wrong, but my understanding of increasing receiver sensitivity works more by narrowing the signal passband than by leaving it open to anything from DC to whatever frequency the JFET will respond to.

    Radio works by energy transfer, so the more the receiver input resonates with the applied signal, the stronger the output.

    You are also more likely to get non-obscured results by staying under licensed frequencies, as in the link - http://www.scribd.com/doc/36014945/...f-Ionospheric-magnetospheric-by-R-Singh-et-al

    My search skills are not turning up much material, so I will hope another member can give more help.
     
  3. DonQ

    Active Member

    May 6, 2009
    320
    11
    That seems to be less a wheatstone bridge, and more a differential input stage. That might be what you want anyway. Some of the advantages include high gain, common-mode rejection, and reduction of output changes caused by temperature and voltage variations.

    If you remove the wire from the drain of T1 to the gate of T4 and attach it to a common-mode voltage or a complementary input (like the other end of a loop antenna), it will look just like a differential input stage from any of a number of IC amplifiers.

    Try a search on "Differential input stage" and see if there is anything good there.
     
  4. Darkstar

    Thread Starter Member

    Sep 3, 2010
    104
    0
    Beenthere:
    First, I think you mistook my "antenna" for a radio antenna. As I explained, I'm monitoring atmospheric static charges and the bridge only has to conduct DC from one side of the bridge to the other and monitor the DC difference in potential across the bridge.

    Second, I know 4 sensor bridges do not work like this, but I only have one antenna so I'm trying to imitate the response of a 4 sensor bridge by having the appropriate FET resistances rise and fall as if they were all sensing atmospheric charges. Afterall, if the antenna can change the gate charge and resistance of T1, then it can on the other FETs also. The trick is to have T2 respond the same as T1, and T3 and T4 respond oppositely to T1 and T2.
    -----

    DonQ:
    Actually, I was going for a Wheatstone bridge arrangement. My single FET quarter bridge works, I just want to replace the unchanging resistors with FETs which will change resistance as if they were all sensing a signal like T1. I don't want to use an amplifier to boost a weak and noisy signal that I can already measure, I want more sensitivity so I can detect more subtle changes in atmospheric static charges such
    as when thunderclouds pass overhesd.

    However, I will do the search you suggested and see if I can get any clues. I've been Googling an awful lot and can't find a single example of a 4 JFET bridge. The closest thing I see are "H-bridges" made with MOSFETs for driving motors. I can't believe it should be so difficult.

    Thanks to both for your input.
     
  5. Dyslexicbloke

    Active Member

    Sep 4, 2010
    420
    19
    I have to admit I have never heard of any of this ... what is it for?
    Obviously I get the measuring static charge bit just dont know why, and would like to.

    Have you considdered using 3 J-Fet input opamps in a diferential instrumentation configuration. none inveting input of the circuit to antenna, inverting to ground with floating supply rails? wouldnt that be a full J-FET bridge with an adjustable gain to boot.
    http://www.allaboutcircuits.com/vol_3/chpt_8/10.html
    In fact you could add negative feedback for anything that wasnt 'static' ie radio transmitions and quite possibly get rid of a load of noise.

    Or is that just silly because I am missing the point?
     
  6. marshallf3

    Well-Known Member

    Jul 26, 2010
    2,358
    201
    Often one will monitor atmospheric static condition to predict a coming storm.
    If I recall there are also other reasons to do so but I can't recall them at the moment.

    My grandfather used to put his old Zenith console radio (the one with all the bands on it) down to the lowest frequency band and find a quiet spot, he'd then listen for intermittent static which was often a sign of lightning in the distance.

    I've got his method beat by a mile, my dog will start acting strangely hours before a storm.
     
  7. Dyslexicbloke

    Active Member

    Sep 4, 2010
    420
    19
    Dog ... mine too but that would be only a couple of times a rear in the UK and then lightning is hardly a big issue.

    Now knowing if it was going to rain hard and when ..... much more usefull.
    Al
     
  8. Darkstar

    Thread Starter Member

    Sep 3, 2010
    104
    0
    Hello Dyslexicbloke and Marshallf3,

    I did start this project thinking I might be able to tell when highly charged storm clouds began passing overhead to give me warning that I better shut off the computer. I do pick up the day to night variations in atmospheric static charge and my graphs match the shape of published graphs, but I get almost no change when storm clouds pass overhead. I get plenty of spikes from lightning but it's probably EMF from the lightning itself rather than from rapidly changing cloud potentials overhead.

    I thought this would be something a little different from using an AM radio tuned to about 300 KHz (the best frequency for lightning static.)

    I know basic electronics, but it's only a hobby and I haven't built anything in about 25-30 years. Also, this is my first transistorized circuit and the first time I'm trying to design a circuit myself instead of just copying one, so I'm learning as I go. If I knew more, or if this was a more important project, I might try your suggestions, Dyslexicbloke, but I've been working with a Wheatstone bridge and 1 FET for months and is what I understand best right now (though obviously not enough to incorporate 4 FETs into a bridge) so I'll continue to try to make this rather than
    start using more complex parts I don't understand.

    If you'd like, I'd be happy to share my full schematic of the 1 FET quarter bridge circuit that I'm using now. It does appear to be working as it should. I don't think the day to night signal variation is due to temperature effects because I get similar diurnal variation whether the daytime temperature is in the 70's or near 100 F. And, as I mentioned, the shape of the graphs match published graphs online. My "antenna" is a well insulated, inverted stainless steel bowl about 18 inches in diameter with the sensing FET enclosed and shielded underneath.

    I'm sure part of my sensitivity problem is due to the fact that I live in a condo so I can't put the antenna on the roof or even in the yard. I'm limited to my balcony where it is near the railing and building and I'm sure they do their fair share of shielding the antenna from the sky.

    Thanks to you both for your input.
     
  9. Dyslexicbloke

    Active Member

    Sep 4, 2010
    420
    19
    I take your point .... I am hardly an expert, I mentioned the IC route because they are cheep, easy to get and someone way more educated and clever than me has already done all the hard work.
    Personally I only use descreets if I have no choice, basically because they require far more knolidge to get right.
    Opamps are great things and unless you are trying to do something realy fast, so easy to use.
    I would be interested in your project detail ... I might have a play, perhaps we could compare notes.
    I dont know if atmosferric charge data is available in the UK.
    I will look

    Al
     
  10. Dyslexicbloke

    Active Member

    Sep 4, 2010
    420
    19
    You may find this interesting, I did.
    http://www.infostatic.co.uk/cache/Atmospheric electric field instruments.pdf

    They seem to be using a diferential amp with aground plane between the sensing serfaces (I think).

    They also have a rotor to effectivly modulate the charge (again I think)
    and they apply a second modulated test charge as a self test.

    Clever stuff.

    DAM .. my generator just failed ... again .. have to go.

    Al
     
  11. Darkstar

    Thread Starter Member

    Sep 3, 2010
    104
    0
    Al,

    Yes, that's called a "Fieldmill" and is the preferred way to measure atmospheric charges. There are even schematics online for the circuitry as well as the mechanics. My project is not so important as to require something that fancy.

    I would like to be able to calibrate my setup though and I have done some measurements but I'm missing something and can't relate the size of the signal I get to a cloud potential.

    I have a schematic ready to send to you.
     
  12. Dyslexicbloke

    Active Member

    Sep 4, 2010
    420
    19
    Hi,
    Sorry I left abruptly ... had my genny in bits all day.
    New bearings and seals needed ... the air was charged here last night !!

    Back on topic ...
    Schematic would be great, I am going to have to build one just for the heck of it.
    Al
     
  13. Darkstar

    Thread Starter Member

    Sep 3, 2010
    104
    0
    Al,

    Here is my full schematic. It's very easy. I just tried connecting the resistor between the FET gate and source line today. It did not stabilize the current. That has been my biggest puzzle. All it did was reduce the sensitivity to 1/4 what it used to be. I was getting 200 mv p-p between day and night signal levels. So don't use the resistor in yellow.

    The resistor in series with the gate protects it from sudden voltage surges. Removing it makes the FET current fluctuations much worse so they almost swamp the signal I want to see.

    Here are some circuits I found online that got me started:
    http://www.vk2zay.net/article/9
    http://amasci.com/emotor/chargdet.html

    They are very sensitive circuits that easily detect static fields.
    I started with them, and gradually added things like LEDs and a Buzzer but they didn't work as expected so I removed them and got a cleaner signal in the process.

    To get a graph of the bridge output, I got a DVM from Radio Shack that came with software called "MeterView" which allows the computer to monitor the DVM output and it automatically graphs the results onto the computer screen. I usually have the sampling rate set at once per second so I can catch some lightning spikes, and at this rate it will run for 5 days straight before the data buffer fills up.

    I've put some notes on the schematic, but feel free to email me if you have questions or comments.

    Darkstar
     
Loading...