Transistor Relay Drivers

This is the place for any magazine-related discussions that don't fit in any of the column discussion boards below.
Post Reply
rosborne
Posts: 73
Joined: Mon Jun 23, 2003 1:01 am
Contact:

Transistor Relay Drivers

Post by rosborne »

Are there any hazards associated with using a transistor to drive a 24VDC(coil) relay? Should I put a cap at the base to stop noise from activating my Relay? It seems like I should to me. Any thoughts?
-Rick
Rodney
Posts: 162
Joined: Sat Aug 31, 2002 1:01 am
Location: Titusville, FL USA
Contact:

Re: Transistor Relay Drivers

Post by Rodney »

As long as the transistor has an adequate voltage and current rating (and you put a snubber diode in parallel with the relay coil to kill reverse voltage spikes) there should be no problems. If you are in an area of high noise, a small cap from the base to ground wont hurt if turn on time is no problem.
rshayes
Posts: 1286
Joined: Tue Mar 04, 2003 1:01 am
Contact:

Re: Transistor Relay Drivers

Post by rshayes »

Relays are slow enough that noise is unlikely unless it is extremely high. The inductive kick when they are shut off can be a noise source for other circuits. Diode or capacitor snubbers help this problem.<p>The base drive of the transistor should be high enough to insure transistor saturation. Usually, this is about 10 % of the collector current even if the transistor beta (current gain) is much higher. This overdrive guarentees saturation. If the transistor is not saturated, it dissipates sonsiderably more power, and may overheat. Too much filtering may allow the transistor to stay in a high dissipation condition long enough to burn it out. Turn on times longer than a few milliseconds may be a problem.<p>Placing a filter capacitor directly across the base of the transistor will give unpredictable results, due to the nonlinear characteristics of the transistor. Splitting the base current limiting resistor and connecting a capacitor to the intermediate point will allow using a smaller capacitor and the results will be easier to predict.<p>[ September 10, 2004: Message edited by: stephen ]</p>
User avatar
jwax
Posts: 2234
Joined: Mon Feb 09, 2004 1:01 am
Location: NY
Contact:

Re: Transistor Relay Drivers

Post by jwax »

Have you considered a power MOSFET "transistor"? Low "on" resistance, requires very low drive current, only the threshold voltage to go full on. Some have a built in reverse diode.
WA2RBA
Hello
Posts: 34
Joined: Thu Jul 15, 2004 1:01 am
Contact:

Re: Transistor Relay Drivers

Post by Hello »

What about using a Darlington bipolar transistor?
rosborne
Posts: 73
Joined: Mon Jun 23, 2003 1:01 am
Contact:

Re: Transistor Relay Drivers

Post by rosborne »

In reverse order:<p>Hello, I don't think a Darlington is needed in this case I only need a ß of about 1.5.<p>JWAX, I don't think the Theshold Voltage turns a MOSFET on, it brings it to the threshold of on. In order to use an enhancement MOSFET and to open the channel Vs = 24, Vd = 0. I'm guessing I might need at about 50 V to the gate, I only have 5 V. Furthermore there are handling problems (ESD) with MOSFETs that I would rather not worry about.<p>stephen, I'm concerned about the noise caused by the relay possible momentarily activating the transistor driver of another relay and causing a rapid misfire and making of another relay. I think the odds of this happening are small, but I have a concern. My Ib/Ic = %70 so I could reduce that substantially. I was proposing a cap from base to ground to shunt any incidental voltage spikes to ground. There is a pull down resistor as well hopefully ensuring a good off condition.
What is the problem of the steady on condition? I'm calculating about 10 mW in that condition out of a dissipation of 350 or so. Ah, now I see your worried about the transition, I was going to set the time constant equal to the turn on time of the relay.
I will split the base current limiting resistor as you say. It's only one more resistor. Can you explain in more detail why?<p>Rodney, I'm intrigued by this parallel to coil snubber. I've seen them, but as I sit here trying to think about what they do, I'm struggling and becoming confused. Anyone want to explain . . . ?<p>-Rick
User avatar
jwax
Posts: 2234
Joined: Mon Feb 09, 2004 1:01 am
Location: NY
Contact:

Re: Transistor Relay Drivers

Post by jwax »

Rick- 50 volts threshold to turn a MOSFET "on"?
Yikes, I hope not!
The popular IRF510 has a VGS of 2-4 volts (full "on"):
http://support.radioshack.com/support_s ... /31806.htm
WA2RBA
fsdenis
Posts: 90
Joined: Tue Mar 25, 2003 1:01 am
Location: San Manuel, Arizona
Contact:

Re: Transistor Relay Drivers

Post by fsdenis »

What does a diode and/or capacitor in parallel with a coil (inductor) do to snub (reduce) inductive "kick" noise?<p>There is usually no problem with noise associated with turning a coil on. The current in the coil gradually increases as inductive reactance decreases.<p>The coil "charges" as the electric current comes up to maximum over time. It stores energy in the magnetic field associated with the coil.<p>If we suddenly disconnect the coil from its drive circuit as we do when we turn off the transistor, then the coil will naturally attempt to force current through itself at the same maximum level as when it was on, and in the same direction. <p>In other words the coil becomes a source trying to maintain current flow amount and direction even though it has been open-circuited.<p>To become a source, it reverses its polarity. If it had positive on top and negative on bottom, it will have negative on top and positive on bottom after being open circuited. <p>And the killer is that the voltage across it will rise to the level necessary to continue current flow at the level it had before being open-circuited.<p>This means that the coil may suddenly be a "battery" with 5000V across it; whatever is necessary to CREATE a circuit that will permit continuing the max current flow at the instant of disconnection. <p>Insulation between coil windings may break down and permit conduction with a short. Or other parts in the neighborhood may arc-over to provide a path for current flow.<p>The trick is to provide a path for current to continue flowing at its maximum at transistor turnoff and gradually decrease to zero as the coil discharges stored magnetic energy.<p>This is called snubbing.<p>A capacitor in parallel will do it by absorbing the energy as charge. <p>A resistor in parallel will do it, but will heat while the transistor is on.<p>A diode will do it nicely. It won't conduct while the transistor is on. Only when the transistor is off and the coil polarity reversal turns the diode on. Additionally, the diode will have some capacitance and slow the rate of change of current flow that way while it is turning on.<p>The slower the rate of change of current flow as the coil discharges, the lower the voltage it generates to discharge itself.
rosborne
Posts: 73
Joined: Mon Jun 23, 2003 1:01 am
Contact:

Re: Transistor Relay Drivers

Post by rosborne »

jwax,
I see Vgs = 4 volts on figure 5 on this spec as 0 Amps regardless of Vds.MOSFETspec
I don't use MOSFET regularly so I could be wrong, but the threshold voltage doesn't look like full on to me. I guess I'm missing something. :/
-Rick
rshayes
Posts: 1286
Joined: Tue Mar 04, 2003 1:01 am
Contact:

Re: Transistor Relay Drivers

Post by rshayes »

Most power MOSFETs are specified with 10 volts of gate drive. These have a threshold voltage that is not tightly controlled, but is probably in the 3 to 6 volt range. However, the drain current is specified to be a minimum value at 10 volts drive no matter what the threshold voltage is.<p>Some power MOSFETs are called logic level devices. These are specified for minimum drain currents with drive voltages of 5 volts. This allows them to be driven directly from a 5 volt logic family. This requires a lower threshold voltage, probably around 2 volts or so. These devices are a little more difficult to make, and usually cost more.<p>The drain current at the specified drive voltage is the most significant characteristic. The threshold voltage is mainly significant when you need to guarentee that the device is turned off.<p>The base of a saturated transistor is a fairly low impedance, on the order of 10 to 100 ohms. From a small signal point of view, a shunt capacitor would have to be quite large to have a significant filtering action in the millisecond range. Usually the drive resistor is in the 1 to 10 K range for small signal transistors. If this resistor is split, the impedance at the midpoint will be in the 250 ohm to 2.5 K range, and the capacitor can be smaller.<p>The change in base voltage required to change a transistor from off to on is about 50 millivolts or so for a collector current chance of about 10 to 1. The capacitor across the base must keep voltage changes lower than this. Again, this requires a large capacitor.<p>The corresponding voltage change at the center of the drive resistor may be a volt or so. This further reduces the value of the capacitor.<p>Another way of slowing down the rise time of the collector voltage is to connect the capacitor from collector to base. This creates a feedback loop that makes the collector voltage ramp up and down rather than changing abruptly. These slower edges are less likely to couple to other circuits.
rosborne
Posts: 73
Joined: Mon Jun 23, 2003 1:01 am
Contact:

Re: Transistor Relay Drivers

Post by rosborne »

Thanks Stephen,
I'm still a little confused by the Vth, but I think I get the split resistor thing now. I think your saying that the low impedance the Capacitor would see makes for a very small time constant, so we need to use the split resistance to slow down the discharge of the cap. Right? Looking at it from that perspective seems to make the cap a bad idea, cuz if it did recieve an impulse of sufficient energy, the discharge might fire the transistor.<p>It still seems to me looking at the family of curves for different values of Vgd, that when Vgd = Vth, Id = 0. I can see how a logic level power MOSFET could fit my need, but I only need about 37.5 mA to drive the relay coil. I guess I could find a smaller MOSFET.
-Rick
User avatar
jwax
Posts: 2234
Joined: Mon Feb 09, 2004 1:01 am
Location: NY
Contact:

Re: Transistor Relay Drivers

Post by jwax »

Thanks for the clarification, Stephen.
The Fig. 5 referred to is plotted with 80 microsecond pulses, with a max duty cycle of .5%.
5vdc on the gate will turn it on.
WA2RBA
rshayes
Posts: 1286
Joined: Tue Mar 04, 2003 1:01 am
Contact:

Re: Transistor Relay Drivers

Post by rshayes »

Your relay coil appears to be 750 ohms. The maximum power dissiaption in the relay driver occurs at half the power supply voltage. In this case, 14 volts squared divided by 750 results in about 260 milliwatts. The dissipation is lower for all other conditions. A metal can package like the TO-5 (2N2218) should be able to handle this much power. In theory, these transistors can dissipate about 600 milliwatts, but the case is slightly above boiling water when they do. Some of the plastic types with a metal tab, but lighter duty than the TO-220 might also be adequate.<p>The curves on a data sheet don't mean very much. Figure 5 is represented as a typical part, but as you noted, the threshold voltage is about 4 volts. The specified range for threshold voltage is 2 to 4 volts, so this represents a transistor at the high end of the specified range. At 5 volts gate drive, the drain current will probably be greater than 1 amp, but could be as much as 9 or 10 amps if you had a device with a 2 volt threshold. This part would easily drive your 750 ohm relay, but a cheaper and smaller device would also be adequate.
Post Reply

Who is online

Users browsing this forum: Alexa [Bot] and 48 guests