Need Help with Design

BobP122 wrote in news:0a46a588-ade7-48ed-9937- snipped-for-privacy@v29g2000hsf.googlegroups.com:

You could use resistor divider on the base to bias transistor on then connect switch between base and ground (for NPN transistor). Closing the contacts would then ground the base and turn the transistor off...

--
Yeah, make up your mind.

In the one case you want the open switch to turn the transistor on and
the closed switch to turn it off, and in the other you want just the
opposite.

Which is it?

You might recall from his post in the Basics NG that he wants to convert a heating thermostat to a cooling thermostat. So when the tstat closes on low setpoint, he wants to turn his cooling fan off, and when the tstat opens at the high setpoint, he wants to turn his fan on...

Use an EXOR gate.

--
Many thanks,

Don Lancaster                          voice phone: (928)428-4073
Synergetics   3860 West First Street   Box 809 Thatcher, AZ 85552
rss: http://www.tinaja.com/whtnu.xml   email: don@tinaja.com

Please visit my GURU\'s LAIR web site at http://www.tinaja.com

As I recall from your other post, you hooked up a pnp darlington and it worked. But it drains at least a milliamp from your battery while the thermostat contacts remain closed shunting current away from the darlington base, and you asked for a solution that uses less current. Of course, you can't escape having some current going through the contacts, but you can reduce it by an order of magnitude. View in fixed font.

batt | ,-------+-------, | | | 100k | | | / c load +-----| | | \\ e | | | _| | +----||_ t-stat | | | 100k | | | | '-------+-------' | gnd

For the npn transistor, you can choose from any number of small signal transistors. Likewise, there are probably quite a few n-channel mosfets that would serve to drive the load.

I prefer discrete bipolar for this, more rugged and higher parts count. Below is the circuit of the 4's, a standard Darlington with a fast-on slow-off characteristic to overcome tstat contact bounce on close. Q3 and Q4 form an SCR latch activated on tstat opening with Q4 providing base drive to the Darlington. The 4.7M in the base of Q3 provides more than enough current to trigger the latch and draws a mere 3uA during tstat closed state. View in a fixed-width font such as Courier.

. . . Q1 . TIP42 . BATT--+-------------+------- e c ---+---->TO LOAD . (+) | | b | . | | R1 | | . | +--[470]---+ | . | | | | . | | e c -' . | | b . | | R2 | Q2 . | '--[4.7k]----+ 2N4403 . | | . | R3| . | [4.7k] . | | R4 . | D1 +--[4.7k]--. . | TSTAT R5 1N4148| | . +--||-+---+-----[47k]-|>|--+--|>|-----+ . | | | | 1N4148 | . | [4.7M]| | D2 +|C1 . | R7| b R6 c === . '-------e c---[47k]----b 2N4403 |4.7u . | 2N4401 e Q4 | . | Q3 | | . BATT-------+--------------------+----------+---- . (-) | . --- . /// . .

Correction: View in a fixed-width font such as Courier.

. . . Q1 pnp . TIP42 . BATT--+-------------+------- e c ---+---->TO LOAD . (+) | | b | . | | R1 | | . | +--[470]---+ | . | | | | . | | e c -' . | | b . | | R2 | Q2 . | '--[4.7k]----+ 2N4403 . | | pnp . | R3| . | [4.7k] . | | R4 . | D1 +--[4.7k]--. . | TSTAT R5 1N4148| | . +--||-+---+-----[47k]-|>|--+--|>|-----+ . | | | | 1N4148 | . | [4.7M]| | D2 +|C1 . | R7| b R6 c === . '-------e c---[47k]----b 2N4401 |4.7u . | 2N4403 e Q4 npn | . | Q3 pnp | | . BATT-------+--------------------+----------+---- . (-) | . --- . /// . .

Mea Culpa!

Corrected Version Obviously if I can't even "state" the problem right - the solution becomes more difficult. I find it difficult to follow diagrams in 'text mode'. I thought mine would be clearer - and it would have if only I had it right. Thanks

)

Bob, do you have your reader set to show posts in courier or a similar font, with mono-spaced letters like a typewriter? The diagrams are quite legible. If they look like gobbledygook, it's because you have times roman or some such "proportional" font.

If push comes to shove, he can select, copy, and paste the text diagram into Word or something similar, re-select the whole thing and then go into a Format->Font->Courier. Or alternatively, he can open it in Google Groups, and click on the More Options->Original menu item, displaying the diagram in the original Courier format...actually looks like it preserves the original font when it's copied from Google and pasted to Notepad. These are fairly low tech methods.

I don't really care for the really slow turn-off, so Q5 is added to snap it off when C2-R7 detects Q1-Q2 dropping out. This one will be immune to any reasonable tstat contact bounce, handles well over 50ms. View in a fixed-width font such as Courier.

. . . Q1 pnp . TIP42 . BATT-----+----------+-------- e c ----------+---------TO LOAD . (+) | | b | . | | R1 | | . | +-[470]-+---+- e c -----+ . | | | b | . | | | | Q2 | . | | | R2 | 2N4403 | . | | '-[4.7k]-+ pnp | . | | | |0.47u . | | | |C2 . | | Q5 pnp | === . |+ | TIP42 | | . === +------- e c -------------------. . | | b | | | . | | | | R7 | | . | .-------+--|

--
Use a non-proportional font like Courier to read and write ASCII
schematics and the process becomes much easier.

ISTM that if you use a SPDT (or a SPSTNC) relay to run the fan the
solution to your problem becomes trivial:

BAT+>--+--------+        
       |  TSTAT |
       | /      |C
     | O        O 
     | O   NO   |   NC
       |   O--> |--+------------+
  
That is, with the thermostat open the relay will be de-energized and
the fan will be running because the normally closed and common
contacts will be closed,  but when the thermostat is closed the relay
will be energized, breaking the contact between the NC and C contacts,
turning the fan off.

JF

--
Hmm... I must have missed it.

ISTM this would be the easiest way: (View in Courier)

.BAT+>--+--------+        
.       |  TSTAT |
.       | /      |C
.     | O        O 
.     | O   NO   |   NC
.       |   O--> |--+------------+

JF

Hi, I'll have open some pretty dust covered books - for the transistor part of the ciruit at leat. Am I following? t-stat OPEN: Transistor OFF & Mosfet ON (Load Grounded ON) t-stat CLOSED: Transistor ON & Mosfet OFF (Load Floating OFF )

Really nice simple solution with negligible power consumption in "Load OFF" state. Thanks BobP

you're very close -- with the thermostat open circuit, both the transistor and the mosfet are ON, and so is the load. with the thermostat contacts closed, the transistor and mosfet are both OFF, and so is the load.

With the thermostat open circuit, the upper 100k resistor feeds current into the base of the npn transistor, turning it on. In turn, this provides a connection between the battery voltage and the mosfet gate, which then turns on.

When the thermostat contacts close, they shunt current away from the transistor base causing the transistor to turn off. That takes the voltage away from the mosfet gate. The lower 100k resistor then bleeds the charge off the mosfet gate, allowing it to turn off.

I forgot how much current you said the fan draws, but you might need a mosfet in a TO-220 case. Look for Rds in the datasheet to figure out how much power it will dissipate according to current squared times Rds. The TO-220 can deal with a watt or two and not need a heatsink. Radio Shack sells the IRF510, which has about half an ohm resistance. You can get mosfets with much lower resistance than that, though. Something I left out of the circuit is a freewheeling diode across the fan. Connect the cathode (stripe) to the battery positive and the anode to the other end of the fan, where it connects to the mosfet drain. You take this precaution because when the mosfet turns off, the inductance of the fan creates a pulse. With a diode across the fan, the pulse travels harmlessly through the diode instead of hitting the mosfet. If you have some 1N400x diodes on hand, you could use one of them.

Mike

e

Hi Mike (aka gearhead), Boy, those books are dustier than I thought. Thanks again

ile

.

P.S. You might be able to use more than 100k. Try megohm resistors and see how it works.

Hi again Mike, If you could indulge me one more 'elucidation'? I'm dusting off some old books - analyzing your circuit. Fets were an academic subject only for me - never to be used (exception cmos chips). They still intimidate me. Wouldn't the FET part of your circuit when ON have B+ right at its gate?

B+ to gate B+ thru resistor to gate

| \\ e | | \\ e | | | _| t-stat | | | +----||_ | 100k | t-stat | | | | _| | 100k | | +----||_ | | | | | '--------+-------' '-------+---------' | | gnd gnd

Shouldn't it (perhaps) be: B+to gate thru Resistor? Thanks again BobP

and

y while

the

rent.

e

w in

=A0 =A0|

=A0 |

=A0_|

=A0 =A0 =A0 =A0|

For a bipolar transistor, you would need a resistor in the gate to limit current. In contrast, mosfets are voltage-controlled devices. No dc current flows into the gate, so you don't need a current-limiting resistor. Just make sure the voltage appied to the gate doesn't exceed the Vgs (gate-source voltage) listed on the datasheet, typically 20 volts. The mosfet model includes a capacitor from gate to source, a very

*tiny* capacitor. You can find the actual capacitance on datasheets. It measures in nanofarads or picofarads. A mosfet contains a body diode from drain to source. For a n-mos, the cathode corresponds to the drain and the anode corresponds to the source. When you apply a voltage at the gate, it opens the channel through the body of the mosfet and allows it to conduct like a resistor. If you remove the voltage from the gate and drain the charge from the capacitor, the channel shuts off and the mosfet becomes a diode again, blocking current. You only have a significant amount of current flowing in the gate when switching the mosfet at high frequency and the gate capacitance conducts ac. In an application like that you would probably put a small resistor on the order of 10 ohms goes in the gate to prevent ringing. I don't think ringing or oscillation will be a problem in your application, but if you want to bung a 10 ohm resistor in there it won't do any harm.