3.3v GPIO controlled switch for high DC voltage (42V +)

[Ian.M]

Beware: the constant power load still isn't quite working right.  The problem is the transition between resistive and constant power, which has a discontinuity, which causes the sim to hang if you add any series resistance to C1 (the presence of which prevents infinite  dV/dt at the transition, allowing the solver to pass it more easily).

Personally I'd recommend finding a more realistic model for an ESC as a load or using LTspice's native constant power load.

[PhoenixAU]

Thanks Ian - at the moment something simple will do, initially I just want to verify the circuit works as expected. So the native constant load approach sounds good.

I'll worry about trying to simulate Back EMF once i've tackled the basics.

I'd imagine that i'm going to have a hard time finding something that is identical to what im using so near enough will have to be good enough.

[PhoenixAU]

The main issue with your original design is the to high Vgs which will kill the MOSFET, and the needless additional diodes.

I suspected that might have been the case when I initially posted the schematic (thanks).

For protecting against reverse battery connection you can add the same P-MOSFET in series, but with reversed source and drain connection. Also needs a zener to protect against to high Vgs. See here for more info about this.

I can't quite picture that. Is there any chance you could draw a rough circuit diagram showing how that might work, please?

[Ian.M]

You should probably read https://ltwiki.org/index.php?title=B_sources_(complete_reference) for details of LTspice's native constant power load, and also https://ltwiki.org/?title=Undocumented_LTspice

[PhoenixAU]

Why the extra Schottky diodes?

The second series one isn't needed. The one in reverse parallel should be after the first one and it only needs to be big enough to pass the leakage.

Add an emitter resistor so it forms a current sink and thus, the voltage across the collector resistor.

I believe the intention for the second series resistor (D3 in the original post) is to protect against reverse polarity in the instance where the BLDC is manually turned causing the motor acts as a dynamo or perhaps its an attempt to protect against Back EMF when the motor turns off?

It's also my (limited) understanding that if a motor is under heavy load (torque slowing the motor down) then it can also cause some issues due to back EMF.

Although from what I've seen online usually a Schottky diode in reverse polarity across the motor terminals as Flyback/Snubber/Suppressor Diode, I've also seen capacitors across the terminals of brushed DC motors to help filter the noise from the Back EMF

To be honest I don't totally understand why D1 is there although once again I suspect it's also to help tame Back EMF.

[PhoenixAU]

A complete different approach would be to use a Smart High-Side Power Switch  e.g. BTS724G you can put all 4 output in parallel and get a switch current of 8A.
I think there are more beefy versions of this and you can directly switch them with 3.3V

Benno

Thanks for the suggestion Benno, I do remember reading somewhere that they are meant to be more efficient than than using a MOSFET.

Do you know if they are more tolerant to Back EMF than a MOSFET? Apparently if a MOSFET fails due to Back EMF/reverse voltage over current it can cause it to feed voltage through the gate (frying the Microcontroller) or cause the MOSFET to turn on when it's meant to be off

[pcprogrammer]

For protecting against reverse battery connection you can add the same P-MOSFET in series, but with reversed source and drain connection. Also needs a zener to protect against to high Vgs. See here for more info about this.

I can't quite picture that. Is there any chance you could draw a rough circuit diagram showing how that might work, please?

See attached picture. To protect against voltage spikes I added a TVS with 47V rating at the output. The IRF9540 might be to weak for the task with a Id of 19A. Not a lot of head room with a load of 16A. And with Rds on of 0.2 Ohm you still loose 3.2V per MOSFET. There are beter MOSFETS with lower Rds on out there for sure. Power dissipation will also be around 50W per MOSFET. So in this case a schottky diode for the reverse voltage protection might work out better.

The first IRF9540 is to protect against wrong battery connection, and the second IRF9540 is to switch the power on and off. The zener diodes might need a bit more current to work properly, but it gives you an idea of what the circuit looks like.

Edit: I forgot to add a pullup resistor for the second MOSFET, that needs to be parallel to the zener diode.

[radiolistener]

I'd prefer not to use a relay because they can run into issues in the long term with contacts fusing and general reliability.

Usually mosfet lifetime is much shorter than relay, especially when they are working at high voltage. You're needs to have a dozen of replacement mosfets for repair, while relay can work for ten years with no issue

[sparkydog]

So, you say "I'd prefer not to use a relay"... but then explain why you'd prefer to not use a mechanical relay.

Have you considered a SSR?

To be honest I hadn't - but from what I can see anything suitable is going to be around $48USD and I'd require 2 of them (I only had a quick look though).

Do you have any SSRs in mind that are suitable?

Ah, yes, I was forgetting the little $2 ones are only rated for ~0.1A. Unfortunately, Octopart hates me right now, but based on what I could find before it quit working, I would guess you can probably find something for closer to $20. Certainly there are SSRs rated for... hold on, you said you're driving an 18A motor? That's a honking lot of current.

In any case, it's not obvious why you'd need two? (You might need a $2 SSR to control a bigger SSR; the larger ones seem to want higher "coil" voltages.)

On further consideration, I suspect, as others have noted, that there's a way to just "talk" to whatever is driving the BLDC that would make your life ever so much easier. (Even if that means replacing the driver.)

That said, I also suspect you underrate mechanical relays. Unless safety is a critical consideration (in which case you shouldn't be squawking at the idea of paying $50), a suitably rated EMR from a reputable manufacturer shouldn't be terribly expensive and will probably last for years without trouble.

[Siwastaja]

Seeing that VESC apparently is open-source, then the obvious solution is to integrate everything onto that one piece of hardware and software. The motor controller itself already consists of transistors. All you need to do is to not enable the the motor control and put the microcontroller and all relevant circuits in some low-power state. I don't know if that is possible without hardware modifications, but it would be simpler than cutting the high current path and managing high inrush currents to the capacitors without blowing up these extra power switches.

Only if you need bi-directional switching i.e. prevent back-EMF generated by rotating motor going through the diodes of the VESC's existing bridge, only then you would need to add another power switch.

[Zero999]

Why the extra Schottky diodes?

The second series one isn't needed. The one in reverse parallel should be after the first one and it only needs to be big enough to pass the leakage.

Add an emitter resistor so it forms a current sink and thus, the voltage across the collector resistor.

I believe the intention for the second series resistor (D3 in the original post) is to protect against reverse polarity in the instance where the BLDC is manually turned causing the motor acts as a dynamo or perhaps its an attempt to protect against Back EMF when the motor turns off?

It's also my (limited) understanding that if a motor is under heavy load (torque slowing the motor down) then it can also cause some issues due to back EMF.

Although from what I've seen online usually a Schottky diode in reverse polarity across the motor terminals as Flyback/Snubber/Suppressor Diode, I've also seen capacitors across the terminals of brushed DC motors to help filter the noise from the Back EMF

To be honest I don't totally understand why D1 is there although once again I suspect it's also to help tame Back EMF.

There is no need for another diode, because it's impossible for there to be any reverse voltage, with a BLDC and drive.

The motor is a BLDC type and driver. BLDC is a misnomer, it's really a type of three phase AC motor. The driver will have diodes connected between each phase and both of its supply rails, forming a three phase full bridge rectifier. If the motor is manually turned, it'll generate an AC voltage, which is rectified to the same polarity as the driver, hence no further reverse polarity protection is required.

I'll post a schematic with simulation it the above explaination is too difficult to follow.

That would be great

Attached is a simulation. The inductor represents a long cable. The MOSFET is being switched at a high frequency. The yellow plot shows the load current and red the source voltage. The increase in source voltage, due to the long, inductive cable, greatly slows down the switching speed. Connecting an optically isolated MOSFET driver between the gate and as close as possible to the source, would increase the switching speed.

In your case, you're only switching slowly and at a much lower current, so it's unnecessary. The only reason why you might want to use an optically isolated driver, is so you can use an N-channel MOSFET and switch the high side.

Come to think of it, why are you switching it at all? Is it possible to send the BLDC driver a disable command? That would make much more sense, than switching the power.

[Siwastaja]

Come to think of it, why are you switching it at all? Is it possible to send the BLDC driver a disable command? That would make much more sense, than switching the power.

Usual reasons for switching the power:
* Prevention of rectified BEMF voltage (generated by the motor when manually turned un-powered) running into the battery/supply
* Extra safety layer
* Powering off drivers that do not have the enable feature at all, or consume too much power in disabled state, and can't be modified

[PhoenixAU]

Usually mosfet lifetime is much shorter than relay, especially when they are working at high voltage. You're needs to have a dozen of replacement mosfets for repair, while relay can work for ten years with no issue

That said, I also suspect you underrate mechanical relays. Unless safety is a critical consideration (in which case you shouldn't be squawking at the idea of paying $50), a suitably rated EMR from a reputable manufacturer shouldn't be terribly expensive and will probably last for years without trouble.

That's food for thought - thanks very much guys!

I didn't realise that relays had a longer lifespan than mosfets.

(I'd heard lots of bad press about fusing/failing contact due to arcing - which I figured would be an issue with the power i'll be using)

I suppose that most cars utilise relays and mechanical light switches have a fairly long service life.

One pro for relays is that they don't have drain leakage current.

Although I'll have to compare how much power is required to power the coil for the magnet - compared to the alternate solutions.

(If I went with relays I'd be using the Normally Open configuration so the coils would be drawing power while it was turned on.)

[shapirus]

Speaking of mechanical relays, I use a Powercom KIN-1000AP UPS for my desktop PC, and it's been in service since 2005 -- that's almost 20 years!

It uses relays to switch between line power and batteries and also to activate AVR, and these events are far more frequent than I'd wish them to be.

It still works (at least, for an external observer) like new.

[Zero999]

I'd prefer not to use a relay because they can run into issues in the long term with contacts fusing and general reliability.

Usually mosfet lifetime is much shorter than relay, especially when they are working at high voltage. You're needs to have a dozen of replacement mosfets for repair, while relay can work for ten years with no issue

What makes you think that?

There is no reason why a MOSFET will have a shorter life time, compared to a MOSFET. It depends on the application. A relay will have a finite number of on,off cycles, which will depend on the current and voltage. A MOSFET has infinite on, off cycles. A relay will be more robust, when it comes to over voltage and ESD, which will easily kill a MOSFET.

[PhoenixAU]

Attached is a simulation. The inductor represents a long cable. The MOSFET is being switched at a high frequency. The yellow plot shows the load current and red the source voltage. The increase in source voltage, due to the long, inductive cable, greatly slows down the switching speed. Connecting an optically isolated MOSFET driver between the gate and as close as possible to the source, would increase the switching speed.

In your case, you're only switching slowly and at a much lower current, so it's unnecessary. The only reason why you might want to use an optically isolated driver, is so you can use an N-channel MOSFET and switch the high side.

Thanks very much for taking the time to do that Zero999, I really appreciate it!!

In regard to the inductor representing the long cable - have you got a rough idea of how long a cable needs to be before it starts to cause issues?  At this stage I can't foresee any wiring harnesses being much longer than 30cm (12") - I'll use twisted pairs for data lines and keep the higher voltage harnesses separate.

[PhoenixAU]

Come to think of it, why are you switching it at all? Is it possible to send the BLDC driver a disable command? That would make much more sense, than switching the power.

Usual reasons for switching the power:
* Prevention of rectified BEMF voltage (generated by the motor when manually turned un-powered) running into the battery/supply
* Extra safety layer
* Powering off drivers that do not have the enable feature at all, or consume too much power in disabled state, and can't be modified

Exactly

In any case, it's not obvious why you'd need two? (You might need a $2 SSR to control a bigger SSR; the larger ones seem to want higher "coil" voltages.)

Hi guys,
             yep I suppose there has been a little bit of obfuscation.

The circuit in my original post was supplied to keep things easily digestible - the design below essentially uses two of the circuits from the original post and is coupled with an INA169 module for power monitoring, I understand the basics of what was going on, however like many of you I had a few questions to regarding design choices - before my initial post I had tried to simulate it it hope that would answer my questions but the simulation didn't work as I expected - which only led to more questions. I also asked the guys who designed the original circuit, but they didn't seem to know themselves.

The circuit in my original post was derived from the following designs:
   https://github.com/HoverMower/Ardumower_PCBs

   which was derived from:
   https://github.com/Starsurfer78/Ardumower_PCBs/blob/main/Charging_PCB/charge_pcb.pdf

The schematic above performs the following:
            Turns on/off charging to ensure batteries do not get over charged via Q1 + Q2
            Q3 + Q4 controls power to J4, J5, J6, J7, J9 - this could be used to turn off the device to prevent battery undercharge or just to turn off the device when required
            Monitors charge current via an INA169 Module
            Monitors Charge Voltage, Batt Voltage via Voltage dividers
            J8 is a Normally open switch (momentary) used as a start button to turn on the robot

However the original schematic (https://github.com/Starsurfer78/Ardumower_PCBs) above was designed for use with a 29.4V battery.

Initially I'll be using a 42V 4.4AH battery (10S2P) from a hoverboard - the battery has a built in BMS but it's pretty basic.
However there's a high chance I'll upgrade the battery at some point in the future to get longer run times (anywhere up to 55V 20Ah) although that all depends on how it performs with the battery I have.

Connected to this "Power PCB" via J4, J5, J6, J7, J9
   Meanwell DDR-30L-5 30W 5V DC/DC converter  (to power all 5V devices incl. Arduino Nano and Nvidia Jetson)
   hacked hoverboard motor driver + BLDC motors  (for driving robot)
   Additional motor driver + BLDC motor  (for cutting grass)

Arduino Nano @3.3v (controls Power PCB and some other peripherals)
Nvidia Jetson Nano 4GB B01 (main MCU - controls Arduino nano via USB cable)
   Stereo depth camera
   IMU
   RTK GPS

I'll also have include an emergency stop so that it kills power to all motors but not the computers - that way it can resume what it was doing if the "emergency" has passed (like a pause button).

So yeah one of my main concerns I suppose was if the mosfet fails, I don't want it surging through the gate pin and frying everything via the connection to the Arduino GPIO however obviously that can be prevented via isolation such as by using an optocoupler, capacitive isolated gate driver, relay etc...

[sparkydog]

I'd heard lots of bad press about fusing/failing contact due to arcing - which I figured would be an issue with the power I'll be using.

It's true that contact welding can happen, but so far as I know it's more of an issue with relays being abused beyond their rated capacity. There are also various ways to manage arcing.

Elsewhere you said "BLDC is a misnomer, it's really a type of three phase AC motor", but we also seem to be talking about disconnecting a battery. What's the actual load current that the relay needs to interrupt? My understanding is that AC is usually a lot less problematic than DC as far as contact welding. (Basically, it's hard to sustain an arc with AC because the voltage drops to zero 120 times per second.)

"Battery" also implies a portable, possibly hand-held application. Is that the case? I ask because, if you are highly space-constrained, I'm told you can get some serious industrial contactors (e.g. the kind used to control air conditioners) for $20 or less. These are not small components, but they can handle significant loads with high reliability.

Usual reasons for switching the power:
* Prevention of rectified BEMF voltage (generated by the motor when manually turned un-powered) running into the battery/supply
* Extra safety layer
* Powering off drivers that do not have the enable feature at all, or consume too much power in disabled state, and can't be modified

Exactly

This suggests that your goal is more electrical isolation than needing to switch a very large current. Specifically, if you can "soft stop" the motor before opening the relay, such that the relay needs to carry a lot of current but never needs to interrupt a lot of current... then you can definitely be a lot less paranoid about contact welding. Arcing happens when you try to interrupt large currents. If you're using the relay for physical isolation (in which capacity an EMR is far superior to a FET), but you can arrange to only open/close it when the load current is low, that will greatly improve reliability.

[PhoenixAU]

Elsewhere you said "BLDC is a misnomer, it's really a type of three phase AC motor", but we also seem to be talking about disconnecting a battery.

It was Zero999 that was talking about AC motor.  It's a battery powered project - I'll be switching DC (either on or off).

However, it might be possible that there could be some AC current introduced into the system if the BLDC motor was manually turned while unpowered or via Back EMF or if the motor was turned off/stopped suddenly or is under heavy torque loads ( https://www.portescap.com/en/newsroom/whitepapers/2021/12/running-a-brushed-dc-motor-as-a-generator ) - I'm not sure what safegaurds (if any) have been included in the design of the Hoverboard driver PCB or the the other BLDC motor controller - if they have sufficient safegards in place it might not even be something I need to consider.

In saying that if those circuits were isolated properly that would be less of an issue - Only the motor drivers, the Meanwell DC/DC stepdown and the battery BMS would be on the 42V circuit (And the INA169 power monitor). I'm sure the Meanwell stepdown would shield the low power circuits.

If you can "soft stop" the motor before opening the relay, such that the relay needs to carry a lot of current but never needs to interrupt a lot of current... then you can definitely be a lot less paranoid about contact welding. Arcing happens when you try to interrupt large currents. If you're using the relay for physical isolation (in which capacity an EMR is far superior to a FET), but you can arrange to only open/close it when the load current is low, that will greatly improve reliability.

That could be done - it could easily be handeld by software for shutdown sequences.

This suggests that your goal is more electrical isolation than needing to switch a very large current.

Yes, I'd like to be able to turn it on or off via an Microcontroller (on 3.3v GPIO OUT) - for system start up, system shut down and battery charge management.

I have no need for high speed switching for these switches, I wont be using PWM - they will either be on or off.

However, I'd prefer for it to be in a normally open (off) state if it looses power.

The charger that I have at the moment is a 42V 2A charger. The charge current would at most be 7A at 50.4V - that's if I upgraded the battery and charger.

The only time that I'd need to actually switch a very large current while the motors are running would be via the the emergency kill switch however that would be handled by a mechanical switch - I could use a DPST switch there for extra safety. Tripping the emergency kill switch would not effect the current state of the Relay/Contactor/Mosfet etc..

Note: Drain leakage current from a mosfet probably wouldn't be an issue due to the frequency that the robot would be in use - grass grows year round where I am. However, knowing that the batteries are always fully charged would be helpful.

"Battery" also implies a portable, possibly hand-held application. Is that the case? I ask because, if you are highly space-constrained, I'm told you can get some serious industrial contactors (e.g. the kind used to control air conditioners) for $20 or less. These are not small components, but they can handle significant loads with high reliability.

Yep it's battery powered - a robot mower.

Size, weight and electrical efficiency are all considerations.

Weight and electrical efficiency will eat into run time.

Obviously I don't want it to be too bulky - but I'm flexible and I'm open to ideas! 

How big are these industrial contactors? Have you got any examples?

[Zero999]

It was Zero999 that was talking about AC motor.  It's a battery powered project - I'll be switching DC (either on or off).

However, it might be possible that there could be some AC current introduced into the system if the BLDC motor was manually turned while unpowered or via Back EMF or if the motor was turned off/stopped suddenly or is under heavy torque loads ( https://www.portescap.com/en/newsroom/whitepapers/2021/12/running-a-brushed-dc-motor-as-a-generator ) - I'm not sure what safegaurds (if any) have been included in the design of the Hoverboard driver PCB or the the other BLDC motor controller - if they have sufficient safegards in place it might not even be something I need to consider.

In saying that if those circuits were isolated properly that would be less of an issue - Only the motor drivers, the Meanwell DC/DC stepdown and the battery BMS would be on the 42V circuit (And the INA169 power monitor). I'm sure the Meanwell stepdown would shield the low power circuits.

As I said before, any AC generated by the motor, will be rectified by the driver.

The output stage has six transistors, each containing a body diode, forming a three phase bridge rectifier. Any AC or back EMF will be rectified and fed into the power supply.

https://www.digikey.com/en/articles/how-to-power-and-control-brushless-dc-motors



https://www.electronics-tutorials.ws/power/three-phase-rectification.html

[sparkydog]

How big are these industrial contactors? Have you got any examples?

Mmm... I'm not actually seeing much that isn't listed as a contactor (and $50+) that's rated for DC voltage above 24V-30V. Interrupting DC is hard. 🙂 (Finding $2 relays rated for 250V AC is a piece of cake.) That said, after much digging, AZDC105 might work but it's still ~$25.

It looks like you can find some automotive contactors that are not much bigger than a "D" battery, but most general-purpose ones are a few inches on a side. Maybe a little larger than a typical Rubik's Cube. Some DIN-mount models are about the size of a household circuit breaker. Not necessarily a deal-breaker on a mower, but bigger than you'd want in most hand-held tools (at least until you start talking about something more like a chainsaw or leaf blower).

[Siwastaja]

What makes you think that?

It's radiolistener. He has usually pretty good advice, you just need to run absolutely everything he says through an inversion filter. This transforms his 0% correctness rate to near 100%. For example, when he says that relays have longer lifetimes than MOSFETs, you obviously negate that --> relay has shorter lifetime --> correct!

[BennoG]

What makes you think that?

It's radiolistener. He has usually pretty good advice, you just need to run absolutely everything he says through an inversion filter. This transforms his 0% correctness rate to near 100%. For example, when he says that relays have longer lifetimes than MOSFETs, you obviously negate that --> relay has shorter lifetime --> correct!

Yes I also find that mosfet has a far longer lifetime than mechanical relais.
I have burned mechanical reilas controlling 3 phase AC motors in 1 month switching over 100 times a day.
Replaced it with an inverter MOSFET based and it is running for 8 years now without problems.

Benno