| Electronics > Projects, Designs, and Technical Stuff |
| How to use the same RELAY on 12V & 24V vehicles |
| << < (4/16) > >> |
| Nusa:
Remembering, of course, that when we say 12V and 24V in automotive systems, we really mean 14.4V and 28.4V when the engine and charging system is running. And whatever you make has to tolerate large spikes, especially during startup, but expect some noise all the time. Also temperature extremes may come into play in some climates. This is understood when you buy automotive relays, but may not be if you source something generic. |
| floobydust:
A linear drop of 24V to 12V is not practical. It's converting the excess voltage into HEAT. OP did not give details about his relay, but high current contacts ~30A need 1.5-2W of coil power. Automotive 40A cube relays i.e. Panasonic CB 12V 117mA 103R coil or 1.4W so you have 1.4W to dissipate (both in the relay and 24-12V dropper) in automotive high ambient temperatures, and a heatsink plus vibration is not a good mix. That's 2.8W of heat on your board for my example. PWM needs no hardware change, as long as the transistor switches good at 20kHz. A software change to go from on/off to PWM on an MCU output pin. I've done it in products and it works great, just as the relay manufacturer suggests. I use a SOT-223 part, BCP56 or BC817 for smaller relays. They are not rated for raw vehicle +ve transients at only ~40V VCE. |
| JDW:
--- Quote from: soldar on August 08, 2019, 03:51:46 pm ---A quick and dirty solution that should work but the main transistor will dissipate some heat at 24 V. Emitter resistor should be calculated for the specific current desired. --- End quote --- That circuit would face more than 24V because when the engine is running, you could get up to 28V on a truck, just like you can see up to 14.4V on a passenger car. Furthermore, I must account for a Load Dump. I will use a TVS with a Vc (voltage cut-off) somewhere between Vc=50v to54v. That Vc must be lower than the maximum voltage all of my regulators can handle in order for the TVS to protect the regulators. So in your schematic, that 12-24V nominal incoming voltage would be AFTER my TVS, which means it would see a maximum voltage of 50 to 54V, depending on which TVS I use. And I am not sure your design is best suited to that. |
| JDW:
--- Quote from: floobydust on August 09, 2019, 12:26:35 am ---A linear drop of 24V to 12V is not practical. It's converting the excess voltage into HEAT. --- End quote --- Not necessarily. Consider the LM9076 datasheet. This 150mA LDO can handle a nominal 40V input (no limit on duration) and can handle 55V for longer period and even 70V for transients of 10ms or less (1% duty). It comes in an SOIC housing version that dissipates heat through the circuit board traces. And because of its currently handling, and the fact it is simpler, less costly, and less noisy than a switcher would be, I am considering the LM9076 to power my 3.3V MCU. Normally, the MCU and all my other circuitry will draw no more than 2mA -- and that's without even sleeping the MCU at all. With all my other circuitry (but NOT including a relay) would draw about 90mA. A 150mA regulator will give me much needed headroom. So when installed in a 24V truck, you are basically dropping 28V (with the engine running) down to 3.3V. So it is possible without resulting in a fireball of heat. What I am considering now is using this Hongfa HFD27/003-H 3V relay (High Sensitivity version with coil resistance of 60Ω). The relay coil can be triggered by as little as 2.4V or as much as 7.0V. I would trigger the GND side of the relay coil using an output pin on my PIC16F1508 MCU via NPN transistor (and yes, I will have a 1N4004 across the relay coil to suppress spikes), and I can use the PWM module of the PIC to send 20kHz or even 40kHz at a 50% or 60% duty cycle to reduce the current to the relay after waiting 0.5s or so for the relay to switch. At a PWM 100% duty, it's Ohm's Law: I = E/R I = 3.3V/60Ω = 55mA (without PWM) And yes, it will indeed draw 55mA because don't forget the voltage is 3.3V. (Again, it's Ohm's Law.) Now with PWM, I could drop that current draw and therefore the heat dissipation down. And technically I do have enough headroom on my LM9076 to accommodate the 3V relay's 55mA even without PWM. But I think PWM is important because this will be a normally open relay used to cut two signal wires coming from the PUSH start switch in a vehicle and the relay will need to be powered the entire time the vehicle is being driven. For trucks, that could be 10 hours straight. I started this thread to see if there were alternatives to using my 3.3V rail to power the relay though. The reason is because even with the 1N4004 diode across the relay coil, I'm concerned about having it on the same 3.3V rail that powers my PIC. Maybe it would be fine. Thoughts? Of course, powering the positive side of the relay coil would eliminate any noise concerns for my MCU. And I could use the 12V of that same relay in such a case. But as you can see from the relay's datasheet, the 12V relay cannot handle anything above 27.7V, and a 24V truck with the engine running will see 28V. So I could not use that for both 12V and 24V. And if I used the 15V relay instead, note that it's lowest voltage is 12.0V, which means anything lower than 12V simply would not switch the relay. Anyway, the current handling of that Hongfa relay is 2A which is really more than I need. I think a max of 500mA or perhaps even 300mA would suffice. But there is no economical solid state relay that would suit my application. They are all pretty outrageously priced. Again, the reason I even need a dual-switch relay is to break 2 signal wires coming from the PUSH Start button (switch) of a vehicle, which will result in starter immobilization. Cutting something else is not optional. I need to cut those two wires. But again, they are signal wires not seeing more than 12V, and not passing more than a couple hundred milliamps. Further thoughts in light of this would be appreciated. THANK YOU! |
| james_s:
If you're dropping 28V down to 3.3V and trying to draw 90mA then the regulator is going to be dissipating more than 2W. I don't know what sort of thermal resistance you can get from a SOIC package offhand but I'd be shocked if it doesn't get searing hot under those conditions. Something a lot of people don't seem to understand about linear regulators is that no matter what type it is, the amount of heat generated depends only on the voltage drop across it and the current being drawn through it and there is no way around this. Just because a regulator is rated to handle 40V on the input and a load of 150mA doesn't mean it can handle both of those limits at the same time, or continuously. 3.3V from a 24V input is a very large drop for a linear regulator, you'll be burning up more than 7 times as much power in the regulator into waste heat as ends up in the load, there is no way around that fact. |
| Navigation |
| Message Index |
| Next page |
| Previous page |