Home   Homebrew Stuff  

A Simple But Complex Relay Driver for Older Power Amps

Driving the Transmit relay in older power amps with modern solid state transceivers often requires some additional circuitry.   Some of these power amps use 120 - 170 vdc relays, yet the SEND signal from solid state rigs can rarely handle more than 16 volts or so.  This dictates that the SEND line be buffered in some way.  The usual way is to put a relay driver between the transceiver and the amp.

As an example, and the immediate problem in front of me, is the Heathkit SB-221 power amp.  The coil of the Transmit relay in this amp is connected to a voltage that is labeled on the schematic as 120 vdc, but in practice is higher.  This voltage measures 175 volts on my particular amp when it is not loaded.  So the transceiver, or itís buffer, must switch this +170 volts to ground.  Connecting your nice new solid state transceiver to the relay input on the SB221 will almost surely zap it.   As a minimum, the amp's relay will always be activated - whether or not the transceiver is transmitting.

There are several ways to solve this problem.  Probably the most straightforward way is to drive a ďbufferĒ relay directly with the SEND line from the transceiver and switch the amplifierís relay coil to ground using the buffer relayís contacts.  Almost any relay with a low voltage coil and relatively high voltage contacts can do this with no additional circuitry, such as shown in Figure 1 below.

schematic of buffer relay circuit

Figure 1

The great thing about this design is that the actual amplifier relay coil voltage (and current) is virtually irrelevant.  Almost any inexpensive relay can handle whatever voltage and current the amp can throw at it.  So this design will work for a wide variety of amplifiers.  Some amps require switching a positive voltage and some require switching a negative voltage.  No matter, this circuit will handle it all without any real attention from the designer.  The only reason to change the design is to accommodate inverse logic from the transceiver Ė that is, if the transceiver SEND line goes to a positive voltage rather than to ground on TX.  This is rather rare, but no doubt exists somewhere in the great world of radiodom.  There are simple ways to accommodate such a change, but they donít interest me at the moment so the solution is left to the student. :-)

However, I have two problems with this simple design (in addition to my dislike of the clickity-clack of relays).

First, the buffer relay adds delay to the overall Transmit/Receive switching time Ė which is usually 5 ms or so.  Even without such a buffer relay, RF energy from the transceiver can reach the amplifier before the amplifierís transmit relay fully switches, resulting in ďhotĒ switching of that relayís contacts.  Adding more delay makes hot switching more likely Ė if not certain.  This, of course, is not good for either the relay or the transceiver (or transmitter, if you are using separates).

My second objection is that this circuit requires an external power supply, which I have labeled V+ 12 vdc in Figure 1.  Iím tired of running 12 volt dc (actually itís 13.8 volts dc) power to every accessory and every solid state rig in the shack.  I have nine power amps in my shack that require a buffer switching circuit of some sort.  So just powering the circuits to switch these power amps adds nine more cables to the accessory power distribution system.  The connection panel for the accessory power is becoming an annoying monster in itself.  So it would be nice if this whole affair could be self-powered by the connection to the power amp.  There would then be no need to use an accessory power source; which means fewer cables, fewer connectors, less to go wrong, less swearing, etc.

The first of my two objections, the T/R delay, is easily addressed by using a couple of high voltage transistors rather than a relay.  Such a circuit reduces the delay added by the buffer circuit to only a few microseconds, which is irrelevant in the greater scheme of things since the ampís relay takes at least 1000 times as long to switch.  The first circuit that might come to mind could look something like the one in figure 2.

schematic of circuit using transistors - step 1

Figure 2

The 756PRO SEND line is ďhighĒ, meaning it is NOT grounded, on Receive.  It is grounded on Transmit.  In reality, the SEND line (in the 756PRO) is pulled up by a resistor to about +6 volts on RX (it can source up to 20 ma) and is pulled down close to ground voltage by a transistor on TX (it can sink up to 200 ma).

When the 756PRO is in RX mode, Q1 is turned on, which turns off Q2.  So no current flows through the ampís relay coil through Q2.  However, some current will flow through the relay coil and through Q1, but it is too small to cause the relay to activate.  In this circuit, that current will be about 7 ma.  This current will cause the ampís relay coil to dissipate about 400 mw whenever the amp is powered up and in receive mode, which is clearly suboptimal.

When the 756PRO is transmitting, the SEND line is grounded, which turns Q1 off.  This allows R2 to pull the base of Q2 up, which turns it on and pulls the amp relay coil to ground through the 1k resistor, R1.  This turns the coil on and puts the amp in TX mode.

Of course, the real problem with this circuit is that the amp goes into TX mode whenever the transceiver is turned off or disconnected!  I know one ham who actually built and used this circuit for a year or so.  He just made sure he turned the amp off first.  But that is no way to live, so we need a bit smarter driver circuit.

Since the SEND signal from most transceivers goes to ground when transmitting, it is reasonable to think in terms of using a PNP transistor.  In general, the emitter of a PNP transistor that is being used as a switch will be connected to a positive voltage.  To turn the transistor on, one pulls the base down to a lower voltage through a current limiting resistor.  This leads us to a circuit like the one in Figure 3.

simple pnp transistor circuit

Figure 3

This circuit operates in the following way:  When the transceiver goes into TX mode, the SEND line is pulled down toward ground.  This pulls the base of Q1 to a lower voltage than the emitter, which is initially at 175 volts, and turns Q1 On.  The base current is limited to about 8 ma by R3.  When Q1 is On, it sources current to the base of Q2 - through R2.  This turns Q2 On and pulls R1 to ground, causing about 19 ma to flow through the amplifierís relay coil.  This activates the relay and puts the amp in TX mode.  It also causes V2 to drop to about 19 volts.  This means that the emitter of Q1 will then be at 19 volts and the current through R3 (which is Q1's base current) will drop to about 700 ua. Since Q1's beta is at least 25, Q1 is capable of sourcing at least 17 ma, which is way more than Q2's base needs to keep Q2 saturated. In fact, the collector current of Q1 will be limited by R2 to less than 1 ma - which is still way more than the base of Q2 needs to keep Q2 saturated.

So everything seems fine with this circuit.  And it is - as long as the transceiver SEND line is driven by a relay or an open collector transistor that can handle 170 volts.  In reality, most transceivers will clamp the SEND voltage at a level equal to or less than the transceiverís supply voltage, which is usually 13.8 volts.  If this is the case, Q1 will still be turned ON when the transceiver is in RX mode.  So the amp will always be in TX mode.  This is a major limitation, since most modern transceivers do not use a relay or a high voltage transistor to drive the SEND line. To fix this problem, we must make the voltage on the emitter of Q1 less than the minimum voltage the SEND line will go to when it is in RX mode.

We could do this easily by supplying Q1 with its own power supply of some voltage, say 5 volts, that is less than the ďhighĒ, or RX, state of the SEND line.  But remember, our goal is to make this driver self-powered.  So no additional power supply is allowed.  As a result, we will have to make our own 5 volt supply using only the voltage available to us from the amplifierís relay coil.  Which is just what we have done below in Figure 4.

simple pnp driver with zener

 Figure 4

In this design we have added a 5 volt zener diode to provide 5 volts on the emitter of Q1.  Now this circuit works properly Ė putting the amplifier in TX mode only when the SEND line is in TX mode.  Disconnecting the SEND line or turning the transceiver off will not cause the amp to go into TX mode.  Therefore, all would seem to be well.  However, it is difficult to choose exactly the right component values to make this circuit work very well, and it is not very versatile Ė it is quite sensitive to the ampís relay voltage and coil resistance.  The value of Rz must be high enough so that the current through it is not so high in RX mode that the relay is activated yet low enough in TX mode (when V2 is only 19 volts) to keep the zener in its regulated region (plus supply the current required to turn Q2 on).  In theory this circuit solves our problem, but it is not terribly practical or robust.

We can greatly improve this circuit by using a perfect zener diode.  This perfect zener diode would require no minimum current through it to ensure regulation yet could still handle relatively large current levels.  This theoretical circuit could be further improved if Q2 was a perfect transistor that required virtually zero base current.  Perfection is reportedly not possible, so we must be content with simple excellence.  The good news is that a nearly perfect zener diode actually exists:  the LM4040.  It is actually an integrated circuit but it functions just like a zener.  We can keep it happy with only 60 ua of current running through it.  This means we wonít have to draw much current through Rz, certainly not enough to even tempt the ampís relay to activate.  In addition, we can make Q2 more nearly perfect by adding another transistor and turning it into a Darlington pair.  This will greatly reduce the base current required to turn on Q2, which reduces the range of currents the zener must handle.  All very good stuff.   Having made these changes, we end up with the circuit in Figure 5.

driver with LM4040

Figure 5

This circuit works quite well.  You may notice that we have also added a few additional parts to soup it up a bit.  We added D1 to make sure that Q1ís reverse voltage limits cannot be exceeded in the event that the SEND line is pulled to a voltage substantially higher than 5 volts.  We added R3, R6, and R7 to ensure that the transistors switch off quickly. And, of course, Q1 can now be a low voltage (read cheap) transistor.

When in RX mode, this circuit draws only half a milliamp through the relay coil.  This will cause the coil to dissipate only about 2 milliwatts, which is not a concern. 

One more item to take care of is RFI.  We certainly donít want RF to get into the circuit and cause erratic operation.  Unfortunately, the high resistances used (especially R1 and R4) provide fertile ground for RFI problems.   

So we added a few chokes and caps and wind up with the circuit in Figure 6.  The RF chokes are comprised of ferrite beads.  D2 was also added to take care of any inductive kickback we might get from the relay coil.   We also added some LED's, which make it easy to see what is going on.  LED1 illuminates when the SEND line is in RX mode.  LED2 indicates that the amp is on and connected - ready to go, if you will.  And LED3 illuminates when the relay coil is energized, indicating that we have successfully keyed the amp.

driver with rfi protection and LEDs

Figure 6

This circuit acts as it should, even when the transceiver is turned off or when the SEND line is unplugged, and it requires no separate power supply.  The only connections are to the transceiverís SEND line output and the ampís RELAY input.   So it meets our basic design goals.

This circuitís operation is straightforward:  When the transceiver is in RX mode, the SEND line will be pulled up to a positive voltage.  As long as that voltage is higher than about 4.4 volts (or the SEND line output is an open collector) Q1 will be turned off and R4 will pull the base of Q3 down to ground, keeping Q3 and Q2 turned off.  As a result, the relay in the SB221 will not be energized and the amp will be in RX mode.  All is well:  whether the transceiver is in RX mode, turned off, or disconnected - the amp will be in the correct state.

When the transceiver is in TX mode, the SEND line is essentially grounded.  This will turn Q1 on and cause it to source current from its collector into the base of the Darlington comprised of Q3 and Q2.  This turns on Q2 and Q3 and connects the SB221 relay coil to ground through R5.  This places the amp in TX mode.  Again, all is well.

Conclusion:   So we have found the perfect amp relay driving circuit, right?  Well, not exactly.  A better solution would also provide electrical isolation between the transceiver and the amp to minimize the potential for ground loops.  However, this is impossible (well, at least impractical) as long as all the power for the circuit comes from the amp.  If we were willing to take some power from the transceiver side we could power the input of an opto-isolator - and this would isolate the amp from the transceiver.  But I want this unit to be self-powered, so I chose to forgo isolation.

It seems amazingly complicated in view of its rather simple function.  But, then again, maybe the mission we have given it wasn't really all that simple:  invert logic, switch very quickly, shift voltage level, drive an inductive load, work without a power supply, operate with a variety of transceivers, operate with a variety of amplifier relay voltages, be impervious to RFI, etc. 

It does all this and works well.  It is fast, and it requires no power supply.  So itís a keeper.

In its own way, itís elegant.  Complex, but simple.

Home   Homebrew Stuff