I need a MOSFET that can sub for a VN3515. :-) I really don't need the
350V rating, >200V should be fine. I do need it to turn on at a low gate voltage (~4V) with only about 9uA of gate drive available. This is part of an off-hook detector so low power is paramount. It's basically this circuit:
formatting link
BTW, what is going on with those transistors in the upper right corner. I'm thinking that it's just some kind of current limiter. It looks like if TR4 turns on too hard and starts building voltage across the 180R resistor that TR3 will turn on and steal some of the TR4's base current. Is there anything else happening here?
Voltage is about 52V when the phone is on-hook and about 6V when it's off-hook. Is it considered reasonable design practice to rely on the voltage drop across the 5.6M resistors to protect the transistor from the excessive Vce that would be present if the transistor weren't biased on all the time?
Make sense. However, you can add a resistor in series to reduce voltages across the TRs and MOSFET.
Then why would you need 200V MOSFET? 60V are far cheaper and more common. With a few volts drop across the TRs and LED, low 40Vs are to be expected on the MOSFET.
I wasn't too worried about the MOSFET, it's the BC547 at TR1 that looks like it would be exposed to voltages higher than its rating, if it weren't dragging the voltage down. For example, if it were to turn off when the phone is on-hook, then the Vce would exceed the parts ratings. Seems allot like TV design where everything kinda holds everything else together. As soon as something fails, things go to hell in a hurry. ;-)
You need a lot of current through R1 and R2 to cause a meaningful Vce. Anyway, the whole circuit is designed to drive an LED (20mA?), you can create voltage drop after the recitifers. Instead of large voltage ratings on active devices (TR and FET), some cheap resistors can reduce the cost substantially.
I hadn't heard that before. I thought that the voltage was the voltage, and could cause damage even if little current could flow.
Closer to 3mA in my case.
reduce
The thing to keep in mind is that this runs on parasitic power drawn from the line. So when the phone is on hook, you want to be as miserly as possible with the juice, this is so that the CO doesn't think something is wrong with the line and disconnect it. Although this circuit does keep TR1 on all the while the phone is on hook, doubling the amount of current drawn in standby. When the phone is off hook, the phone company doesn't care how much you try to draw, in fact they "demand" that you present an impedance _below_ a certain specification.
Also, when the phone rings, the voltages get a bit intense. Not to mention the effects of thunderstorms and whatnot so keeping relatively high voltage margins seems like a good idea, especially considering that this part of the line interface circuit is DC connected to the phone line and therefore not isolated. I have an MC62 (found a really old tube of them in my closet) to isolate the output of the MOSFET. The other channel of the MC62 outputs my ring detection signal, though this circuit does a good job of that too if you leave the cap off the MOSFET's gate.
The third part of my homemade line interface is a transformer that gives me a hook into the off hook audio for the DTMF decoder. Listening to an on hook status line is kinda interesting too. It seems like the CO lets them sort of float somehow as the noise is high on the line until just before it rings or just after it goes off hook. It happens that when the CO connects the line, the noise drops way down. When listening to an on hook line, you can hear all kinds of crosstalk from ringing signals on other lines. Perhaps with enough gain and audio processing, one could hear other voices.
Thanks, I put it into Switcher Cad to check it out for anything unusual. It's apparently just a current limiter, and it really doesn't work all that great IMO. It allows for a 50% increase in current over the expected voltage range. Of course the MOSFET is not supposed to be on during the high voltage exposure so nobody notices I guess. I left it out and things are working fine. ;-)
I think you're misjudging the circuit, somehow. What is the expected voltage range? If you leave it out, what kind of current do you get from the simulator?
The voltage can range from a low of around 5V up to about 90V when ringing but usually ~50V when the line is on-hook. But the MOSFET is only on during the time that the voltage is very low (i.e. the line is off-hook).
If the MOSFET were on _and_ the voltage were high, the current would be quite large obviously leading to the sudden death of the LED. I guess it's not a bad thing to include as a failsafe, but I don't see it as really necessary for my one-off project. I'll have to see if the opto-isolator I use in place of the LED leads a short life. It seems ironic that the circuit author would take the effort to include the current limiter that won't really be exercised, but ignore the max Vce on TR1 and let the resistors and bias save it, but it still takes a bit of a voltage hit while waiting for C2 to charge. So far everything is working fine without the current limiter.
BTW, I subbed a 2n3904 for the BC547. Don't laugh too hard, but I used an IRF830 for the MOSFET since that was all I had on hand. It probably shouldn't work since it really needs more gate voltage, but it works fine so far. Back to the topic, can you recommend a common (jelly-beanish) low power (TO-92 is good), logic level MOSFET, N-Channel of course. Thanks for the input.
But it shouldn't kill the led even if the Mosfet was a dead short. That circuit is supposed to provide constant current independent of Vcc. (Obviously there are limits - Vcc could be so low that you can't draw 3-4 mA or so high that parts melt.) Lets say you have 100 volts on the collector of TR4. When there's 4ma current through the 180, it produces .72V, which biases TR3 on. That in turn increases the drop across R5 which was 100K*IbeTR4 to 100K*IbeTR4 + 100K*IceTR3 - which reduces the base current to TR4. Note that we never use the 100 volt figure. The equation would be identical regardless of voltage.
Maybe looking at it another way would help: TR3 will do its damndest to keep the voltage drop across the 180 equal to its own Vbe. The only way it can do that is by telling TR4 "you're applying too much current to the 180".
But in any event - you're getting a 1800% increase in voltage and only a 50% increase in current.
I mean in my circuit, without the current limiter.
How did the circuit designer come to pick the 100K resistor value?
Ok, thanks. That makes sense, so I'm wondering why SwitcherCAD shows such a variation in the current. I'll have to play around with the simulator some more.
seems
Vce
bit
is
used
probably
N-Channel
I have 50 of them as of this afternoon when the UPS guy came. ;-) I also ordered 10 Zetex ZVN2106A parts (lower Rdson/more current but otherwise similar) to keep on hand.
Never mind, I understand how it works now. The variation is due to the current flowing thru the 100K resistor. After doing some tinkering, simulating and reading I guess it's not such a bad little circuit after all. :-)
BTW, I had to increase R1 from 5.6M to 10M to reliably work on my second phone line. It seems the threshold for the see-saw effect of TR1 turning off. The threshold before was at about 10V, but I have a cheap phone that only loads the line down to just under 12V. Rasing the value to 10M moved the threshold to about 20V. This value should work on just about any phone line.
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.