0
\$\begingroup\$

I would like to build a DIY solid state relay for my car to enable some circuits (less than 1A current absorption) only when the accessories are powered up (independently from engine on or off).

I have some spare IRF4905 P-MOSFET, some optocouplers, some TVS diodes 18V and various Zeners.

My idea was this one (I found no optocoupler in Circuitlab, I made it myself more or less). I omit from the schematic the diode protecting from inverse voltage.

schematic

simulate this circuit – Schematic created using CircuitLab

IRF4905 accepts at most Vgs=20V. In a car I could get short spikes higher than that. Hoe can I protect it using a 18V TVS diode or various Zener I have?

An idea could be to do this, since the IRF4905 requires only 4V of Vgs for complete turn-on.

schematic

simulate this circuit

Would that work? other ideas? Or should I really use a traditional mechanical relay?

\$\endgroup\$
5
  • \$\begingroup\$ IMo the first circuit with zener is OK. One other concern is that you can get high voltage spikes on output, when some other inductive load kicks back, this can be solved usin avalanche rated MOSFET, that starts conducting even in opposite if the voltage is higher than avalnache breakdown voltage, so the MOSFET doesn't suffer the breakdown voltage. \$\endgroup\$ – Marko Buršič May 26 '16 at 16:31
  • \$\begingroup\$ Well the electrical circuit of the car has already somewhere else a TVS diode at 24V (I put it myself as part of electronics.stackexchange.com/questions/225315/… ), so I know I shouldn't get more than 24V as Vbatt. And against reverse I have a diode (I forgot to mention it, I added it now to the explanation). \$\endgroup\$ – FarO May 26 '16 at 16:36
  • 1
    \$\begingroup\$ Since you already have a 24V TVS, you could omit the Zener because the resistor divider won't reduce the voltage below 12V. But better safe than sorry. \$\endgroup\$ – Barleyman May 26 '16 at 17:00
  • \$\begingroup\$ IMO, want to build a switch + diode circuit. NMOSFET, since having lower Rds are more suitable that PMOSFET + diode, but requires a charge pump, not easy to DIY. In case you don't want spend time to diy, I have found a circuit, didn't bougt it yet (same problematics, I want to turn off loads) here the LTC4359 is ideal diode + switch circuit: aliexpress.com/store/product/… \$\endgroup\$ – Marko Buršič May 27 '16 at 8:19
  • \$\begingroup\$ Datasheet:cds.linear.com/docs/en/datasheet/4359fb.pdf, you will also find various gate protection circuits, maybe it will help you for your DIY. \$\endgroup\$ – Marko Buršič May 27 '16 at 8:25
0
\$\begingroup\$

The zener idea is a good one. If you can sacrifice turn on time, and your only dealing with short transient pulses on the supply side, then maybe adding a capacitor on the gate might be another way (it will slow things down) to stop transients, especially short ones. An inductor in series with R2 could accomplish a simmilar thing assuming the pules are short and you can deal with a slower turn on time.

An Solid State Relay is essentially the same thing, there are some constraints that come with them (resistance, some have a voltage drop) but they also have isolation and some have robust inputs with protection.

\$\endgroup\$
1
  • \$\begingroup\$ Inductors are difficult to come by by chance in my toolbox, and even if I have some of hem and also a Victor VC9805A+ with inductance meter (a Chinese one), they are still difficult to measure. I'll probably go for the capacitor. \$\endgroup\$ – FarO May 27 '16 at 8:47
1
\$\begingroup\$

Would that work? other ideas?

In the 2nd circuit I would make R3 much lower than R2 or you will be naturally reducing the drive voltage to the gate under normal conditions and this will result in higher on resistance in the MOSFET.

I say this because the CTR of the opto may be 50% (for example) and this would mean only about 1 mA flowing into the collector. Peak loads may far exceed average load currents as well so getting a few more volts driving the MOSFET isn't a bad idea.

\$\endgroup\$
6
  • \$\begingroup\$ How do you get that exactly? -4V gth P-mos gate driven at -6V source voltage? \$\endgroup\$ – Barleyman May 26 '16 at 16:56
  • 1
    \$\begingroup\$ @Barleyman optimum on-resistance occurs at Vgs = 15V although 10V aint too bad but 6V may not be sufficient for some loads. I know the op says 1A but who knows what the peak value might be. Also given the CTR of the opto every bit of drive voltage that can be maximized should be. \$\endgroup\$ – Andy aka May 26 '16 at 17:07
  • 1
    \$\begingroup\$ I agree. R3 could be as low as 100 ohms and still no chance of parasitic oscillation. M1 is ok if no severe voltage spikes on the 12 volt line, if M1 is in an OFF state. I would put a .1uF cap across the Vkey input to dampen contact noise. \$\endgroup\$ – user105652 May 26 '16 at 17:34
  • \$\begingroup\$ Ah the fool me, taking OP word for it without checking datasheet. However with ~50A capacity at 6V, I think it can drive 1A load without too much effort.. \$\endgroup\$ – Barleyman May 26 '16 at 19:19
  • \$\begingroup\$ @Sparky256 100 ohm on R3 means that for a 24V spike (6V on it) I would get 0.36W dissipation, I would need 1/2W resistor. Better use >470 ohm to use the common 1/8W or 1/4W resistors. \$\endgroup\$ – FarO May 27 '16 at 8:38

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.