Many amplifiers use relays in series with the speakers to disconnect the speakers from the amplifiers’ output during power-on and in case of undesireable amplifier states. This includes protection against excessive DC offset on the speaker terminals that may be caused - for example - by a shorted output transistor.
❗ Caution
All information is provided “AS IS”, without warranty of any kind, express or implied.
Electro-Mechanical Relays
The most common type of relays in this application are electro-mechanical relays. Their switch contacts can have a hard time switching off fault conditions like those found in power amplifiers connected to inductive loads like speakers. Reasonably good case: the relay opens quick enough to protect the speaker, but the contacts get pitted by the arc drawn while opening the fault current; this degrades the contacts so that the contact resistance rises - potentially to an extend where the relay has to be replaced. Worse case: the relay only opens too slow to protect the speaker’s voice coil, but gets destroyed in the process.
If the relay was designed to open the contacts during such a fault condition, it would survive without (much of) an issue, but the relays typically used just aren’t. Also during normal operation mechanical contacts may degrade:
- The user may switch on/off the amplifier while a signal is present, so that the relay has to switch the inductive load, over time causing pitting.
- After many years the contacts may oxidize.
Many amplifiers of the 70’s and 80’s share the same/a similar protection relay footprint. Not sure that there is an actual name for it, so I’ll call it the Omron MY2 footprint. (The Omron MY2 is a frequently recommended modern replacement for the relays I’m talking about.) Sometimes the amplifier manufacturers used a 4 pole variant like the one shown in the picture above, paralleling two contacts per channel. Note that similar relays are available from other manufacturers, for example a heavier duty variant from Schurter that is physically larger, but has pretty much the same pin layout/spacing.
Solid State Relays
One solution to avoid the problems of electro-mechanical relays: Solid state relays using low drain-to-source resistance Power MOSFETs. Of course I’m not the only one considering - or implementing - a solid state speaker protection relay. See here for a helpful article by Elliot Sound Products that discusses various implementation options.
The issue with all implementations I’ve found so far: They use a rather large PCB and are therefore a hassle to retrofit. I’d like to have a DPST solid state relay in the form factor and with the footprint of an Omrom MY2 (with only the DPST contacts of course) that I can use in various amplifiers like the Sansui AU-x17 series. It should handle a constant current of 7 .. 10A, with potentially much larger peak currents.
As power MOSFETs I selected Infineon’s IPP026N10NF2S that come in a compact TO-220 case despite their really low max. drain-to-source on resistance of just 2.6mOhm, while also having a high single pulse avalanche energy rating of 430mJ. Even with two of those in series the on-resistance is as low or lower than the contact resistance of typical electro-mechanical relays.
The bigger issue is to supply/control the MOSFETs. However, I found the Texas Instruments TPSI3052 Isolated Switch Driver that seems to be perfectly suited for my design due to the low parts count and the operating modes that allow for up to 60V (absolute maximum) control voltage in the Two-Wire mode. Also, the gate drivers allow for pretty high gate currents. With a total gate charge of 103nC for the selected MOSFETs the TPSI3052 won’t have a hard time to achieve rise/fall times of <1µs. This results in a wide SOA.
The circuit & the PCBs
The typical application of the TPSI3052 couldn’t be much simpler. (If you’re interested just have a look at TI’s datasheet…)
The main challange is the small footprint of the Omron MY2 - no way to fit all components on a single board, so two boards, a main circuit board and a controller circuit board it is. The four TO-220 transistors (two per pole) were placed on the main circuit board in a way that a) minimizes trace length and therefore resistance and b) prevents the metal tabs (drain) from ever touching each other.
The two TPSI3052 Isolated Gate Drivers (one per pole) are placed on the control circuit board that is vertically mounted to the main board. The control circuit board includes a bridge rectifier for the control signal so that the relay can be driven with either polarity. One LED per TPSI3052 indicates the transfer of power to the gate driver side. Under normal operation conditions this indicates that the gates either are already or will soon be driven high.
Both PCBs are 4 layer FR-4 boards to decrease trace resistance on the main circuit board and to make the control circuit board as compact as possible without violating the design rules. The design rules include 0.8mm minimum clearance on the main circuit board, the best I could do…
Measurements
The following measurements are shown for a intentionally current limited 24V control signal (from a lab power supply; the current limit causes the slow rise time) and a 3A load. During the following tests the bridge rectifier was not yet installed.
Rise and fall times
Two crucial characteristics of this circuit are the rise and fall times of the gate-to-source voltage to ensure operation within the MOSFETs’ SOA even for high V_DS/I_C.
As can be seen, with rise and fall times well below 1µs everything is as expected.
Switching behaviour
The relay turns on as soon as the control signal reaches about 5V. (Add about 1.2V for the bridge rectifier in the final application.)
The relay turns off as soon as the voltage falls below about 5V. (Add about 1.2 - 1.2V for the bridge rectifier in the final application.)
Temperature
The following thermal images show the relay with one pole carrying 10A at about 20°C room temperature (see Max temperature!):
The case temperature of the MOSFETs stabilizes at about 65°C (verified with k type thermocouple). Note that loading both poles with 10A at the same time will likely cause the the temperature to rise further.
Relay On-resistance
I measured the on-resistance of the relay to be about 7mOhm not including the pins (under load). This obviously depends on of the MOSFETs and therefore also on the ambient temperature.
Other thoughts
Since there is no audible feedback with a solid state relay, it’s nice to have some form of a protection indicator. This is especially true, if the relay’s LEDs are not visible through the vents of the amplifier’s case. I like Sansui’s implementation of a power light that doubles as a protection indicator by blinking while in protection.
From the protection relay’s perspective, switching the amp on/off with signal applied should not be an issue anymore.
Pratical application
Although this relay is designed as a drop-in replacement for certain applications, it’s important to understand the limiting parameters of the solid state relay as well as both the relay drive circuit and the control signal requirements of the TPSI3052 to evaluate whether this type of a solid state relay may be used in a given application (e.g. retro-fit). There might be cases where this relay can’t be used in the form presented. Those cases may include, but are not limited to:
- high power amplifiers exceeding the electrical (e.g. voltage, current, power, avalanche energy) or thermal contstraints of the MOSFETs
- drive circuits
- that exceed the max. voltage to be applied to the TPSI3052
- that require a certain load/resistance of the relay
- that depend on must-operate/must-release thresholds of electromechanical relays
- higher frequency switching application
- applications where the delay between powering the relay and the relay switching on is less than what the TPSI3052 can provide
Conclusion
Yep, it seems to work just fine, at least in my limited testing in my application. It’s not cheap though, certainly not when compared to an Omron MY2/4 (for example). Is it worth it? For me YES: It was a ton of fun to design and build.