A far better solution for noise suppression is a big cap and a much smaller cap in paralell, the big cap smooths out the low frequency noise from motors, while a small cap will deal with high frequency switching noise from switching regulators aka onboard voltage regulators vtxs, cams ect ect.. I did this all the time when I used to design synthesiser circuits.. I would never use just one large cap.. A large cap wouldnt be able to sink fast switching currents quick enough, hence why you will always find both in pretty much all electronic circuits these days..
@@nakmax No.. There both very large caps.. A 470uf and somthing much smaller like 100nf or10nf would be better.. Valtage rating doesnt matter as long as its more than your battery puts out at full voltage.. so 35v is ok..
Something i should have added,i make audio equipment using vacuum tube/valves and as most guitarists use 9Vdc power banks i have to run my tubes/valves at proper tube/Valve anode/plate voltages of anything from around 70Vdc to around(for me)350Vdc. I designed an smps boost converter that can be varied from around 60Vdc to about 380Vdc,this is possible due to the series inductor being switched in and out of conduction,the voltage increases due to inductors not being happy to stop conducting,they store energy in the form of a magnetic flux that swells like a balloon it's a field and as long as the inductor core does not saturate it will keep adding to the voltage,the voltage is passed to a high voltage capacitor via a switch,the switch is a either a schottky diode or a super fast diode,voltage dependent as the schottky in general don't have the same voltage range of say a ulf4001 that's ultra fast with the same voltage ratings of a 1N4001,diodes conduct in one direction only and so can act as a switch. So deliberately we can allow a high voltage to be developed,this is with a pwm controller using feedback to skip pulses to control or hold the output voltage,as for uncontrolled Back EMF the field can get quite high very fast. So inductive back EMF can destroy a transistor very quickly,and as such when using inductive loads ie a relay armature,an inductor or any wound inductive component a freewheeling diode should be used to protect against flyback voltages,the diode is actually a current steering diode that in one direction is not conducting and as soon as the reverse voltage appears the diodes junction barrier voltage is overcome allowing the diode to act as a clamp which will hold the back EMF to around 1V a little over the barrier potential of around 0.6v-0.7v generally in that region. Incidentally the old style TVs ,computers and oscilloscopes using crt screens had a flyback transformer inside generating 1000's of volts without breaking a sweat. Again the suppressor cap might stop EMi with the electronics in the same way car radios have/had a suppressor cap to null the EMi noise injected into the power supply via the alternator,or to peoples tv sets as you drove around. The best method of suppression can be with a RLC filter,some shielding,or a simple RC filter using proportionally smaller caps to deal with HF content of the EMi,seeing as a motor rpm is linked to the frequency generated. I did explain the best shielding method to minimise added weight and to give the best shielding,you must realise this shielding will block RF signals ,so external antennas must be used or you aint flying nowhere buddy. Finally Joshua,i messaged about changing out ESC's on a cheapo drone with other mods to use higher current ESC's bigger motors and higher cell count batteries while using a buck or linear regulator to keep original control circuits and cameras etc as per original,to get a lightwieght cheap screamer,sadly no reply. I'd appreciate some input as i'm still a noob at FPV. I think i can give good insight to the electronics side of this,even making cheap shitty gear from china work better than some of the high end ESC's from here in the west(i already did this). Still.i have subscribed and enjoy the vids,i hope i gave good clear information on not getting screwed by bogus info on Back EMF protection circuits for parts that actually cost pennies.
This is why the TVS diodes on ours are so small. Maybe it helps a bit. Doesn't hurt to have it on there. Weighs like 1/3g. So the cap/diode system with the wire attached at the appropriate max length and awg, it's as fool proof and maximally protective as reasonable for a general 5-7" system.
In my experience ESC deaths due to voltage spikes typically occur when plugging in - this is particularly true of FETTEC whom were the first to create a spike absorber, to combat the frequent fires on plug in 🤪
How the ESC dies is mystery #1 in our industry imo. I'm leaning towards physical stress on components resulting in sketchy contacts. I'm hoping to test this board cased in resin soon.....but I'm just one person unfortunately so my results will always be inconclusive.
Hello Bob, I do electronics design and ESC repairs for living. I've been trying to analyze this issue for ages and this is what I've been theorizing from my own experience. Different kind of failures tell us what actually happened. If the ESC dies non-destructively, most of the times it will be the microcontrollers failed to get power or failing themselves. For the first issue, many times the ceramic capacitors failed shorted on one of the power stages. This can be due to poor placement where the capacitor experience physical stress. As for the second issue, I suspect induced transient on the signal line connected to each of the microcontroller like the ADC pins for EMF detection or other pins because this is how you kill a microcontroller other than by physical damage. The way ESC dies destructively also tell us what could be happening. If the ESC failed without bringing the motor with it (single phase short), the FETs are the problem. Either this could be underrated parts or poor quality. If the ESC failed while bringing the motor with it, then the short happened across phase. Could be caused by the same issue as before, failure to detect the correct EMF to the point of coil saturation (overcurrent), or other external trigger that excites the FETs abnormally. How would I come up with a reliable ESC design? Full-size 5x6mm FETs for typical 5 inch setup. Even a dual 5x6mm FETs is better spec-wise than 2x 3x3mm with smaller overall footprint. 2 capacitors in series configuration in case of one failing, it won't make a dead short. Larger resistors on gate inrush limiter because the smaller one can't really handle the peak drive current and often fail high resistance.
@@Asu01 I just had this happen where one ESC failed in a 4in1. but the motor works still. I thought it was the motor at first because it was just twitching but after swapping it I found that wasn't the case. No components get hot on the esc. Usually something gets hot on a dead esc from my experience. Oddly enough, that's not the case here.
I'd say the ESCs die by overcurrent. Those voltage dips are probably caused by an overcurrent happening in the motors. When you have a "short-circuit like" situation, the voltage will decrease in function of the current getting increased, that could explain why people burn their esc when having a jammed motor.
@@Asu01 I just randomly came back to these and saw your comment. Design an ESC. We'll get it made and see how far it gets. The industry would absolutely love to have a highly reliable board.
Isolate and the decouple the crap of the microcontroller and whatever other gate driving components there are, then increase all the ceilings on the mosfets as far as absolute ratings. Whatever is serving as the freewheeling, beef that up too. If anything, just to try avoiding ground issues I would actually just try to design the setup so that the low side (gate drive) of the ESC was completely isolated and using its own battery supply or a good isolated DC to DC converter. It's either that or take all of the filtering and decoupling real seriously by doing more than just tossing a couple electrolytics on that are supposed to be low-esr. A common practice for decoupling and snubbing when just winging it would be to stack values for various frequency ranges such as a 1nF/100nF/1uF set of decent ceramics. Then you have to actually tidy up everything else that is sharing the same system or close enough for magnetic interference. All the wiring is just inductance laying around. Good low esr caps on top is also good but shouldn't be the end-all. I would imagine any level of improper switching of the fets can result in a shoot-through that will kill them relatively fast. Larger and more balanced lower inductance PCB layouts would help but then you're increasing the size and weight. Factors outside of the ESC design itself can seem to be causes of their failure even if they are being operated within reasonable limits. If they are not conformal coated I would imagine just moisture and debris could cause issues. Or the wiring and grounding. When I think about ESC's I think about VFD's, except VFD's are typically running 3ph AC induction motors. While they are eventually all going to suffer the same deaths for the same reasons no matter their filtering or protection.. what typically keeps them from dying prematurely is just the fault protections.. They are using switches (usually IGBTs) which are rated for very high short circuit currents for a certain duration of time, and that duration is longer than it takes for the built in overcurrent protection to activate. It just kind of seems like given the size limitations of what most people need, it's not really feasible to build an ESC that is going to never die because it has every perceivable protection.
I think what blows the ESC is the heating up of the fete either melts the solder it touches next to it not creating a voltage spike but creating a short.
It is exactly as you assume. If the motor is blocked it will generate its highest torque and therefore the highest current. So either the motor winding will overheat first or the esc will blow up.
A low impedance cap will always filter while other protection devices like TVS diodes are meant for much larger voltages. We luckily dont really have those, the back emf spikes are of very limited energy and if the cap is capable of sucking them up youre good to go. A much smaller cap would probably be fine too. And I would definitely stick with 50V caps from experience unrelated to drones.
It is over current because, without some way to limit it, the energy being supplied by the battery, to the motor (through the esc), is going to be converted into heat instead of mechanical energy. The heat changes the properties of the various layers inside the FET, causing it to behave more like a variable resistor with positive feedback until it burns out the internal trace to the pad - or worse.
I always pictured the prop strike killing an esc by essentially increasing prop resistance in the same way trying to run a 12inch prop on a 5inch setup. Your asking it to spin the world in this case. You learn about over propping burning up your esc. If the prop can't move its the same ish as having it over propped right?
I should have added you should add some caps in series ie first cap of say 1000uF,then 100nF then 100pF to eliminate HF noise.The values are proportional.
not to say that i actually know the cause of esc blowing, but the inductance on the circuit can cause voltage spikes from suden changes in current, tho an increase in current can cause voltage drop or sag
Is there anything that reduces the spark when connecting a battery? I recently moved up to 6S and noticed a huge difference in the spark created when connecting the higher voltage battery. Is it due to the remaining voltage in the capacitor?
Looks like prop strikes result in Amperage spikes. I guess to protect the ESC you would need a fuse. That said, I do not suggest putting a fuse the ESC.
That voltage spike from a motor being stopped explains why a row of 5 capacitors literally exploded off my esc yesterday.... I was wondering how that happened when it wasn't even that hard of a crash.
In a 4s 3.5" quad, should i use 25v or 35v? And what uf, 1000 or less? Should, I use 1 or more ? My vista is connected to my directly on my battery, in the connection the drone had it self a small capacitor 16v 100uf, should i change it? Please help!!!
Very interesting and fun clip. Learned something new along with Joshua. I actually smoked a motor trying to turtle out of a tree. 5 inch race quad, aikon 20x20 esc vs a 2207 2450kv motor.
external TVS won't save the internal diode in the n-mosfet... so that thing will die regardless if it wants to. When a 3 phase motor abruptly stops the fight is with the coil in the motor and the mosfet on that channel and nobody else. What I am wondering is ... where exactly are we measuring the noise / voltage ?
I believe a 4in 1 ESC is under heavier stress when connected to 4 motors doing their own thing (some speeding up while other ones are braking). I have a couple of quads in which I've burned one ESC on a 4-in-1 board, and kept running with one separate ESC, both these kwads have been incredibly reliable since. A lot of pilots including me, believe in the effectiveness of TVS diodes+50V cap on 4in1 escs... I've myself never burned an ESC equipped like that... But I know there is nothing scientific here ;-)
I took a high speed impact with a wooden pole with my racing drone trying to hit gaps, the frame was perfectly fine, but the 4 in 1 esc was slightly bent and 1 esc stopped working :D
The largest current draw will always happen under "locked rotor" conditions. They can be massively higher than the running amps. Back EMF can also wipe out the processor, so it depends on the failure mode. These are all vastly more likely when the manufacture cuts corners, or uses poor engineering practices, which is not always reflected in the retail price.
Nobody seems to mention active breaking / damped mode in BLheli... Wouldn't that tend to generate higher voltage more than a prop strike?... Especially in a high air flow situation like an abrupt throttle cut at high speed?... I'm curious about people's thoughts on this as I don't know much about it...
What the test shows is that there is a current spike happening,which results in a voltage drop because the current limit gets cought. Does anyone know if this "maxing out" of the current has been checked as a reason for the ESC death? So basically the only current safety would be using a battery with lower current rating than the ESC, "abusing" the battery as limiter.
Forgive me for I'm not 100% fluent in electrical engineering. But what I assume is happening is t hat when a motor get bound by something the coils just perceive it as a substantial amount of drag. When the temp of the motor heats up resistance increases substantially and the ESC tries to over compensate and burns out to over current. If this were true a test could be done with a high amp ESC and whatever the equivalent a lower amp motor would be. If the motor smokes before the ESC and you could measure amps from the ESC it would confirm this theory.
Its a lot to drop in such a short time for sure. Im in 30 months now and just topped 20k. I did the math just out of curiosity and I've averaged over $700/month. Ive kept a record from day one of every penny spent. Maybe I shouldn't have lol
The problem is that voltage spikes are generated FROM the motor and propagate to the INPUT. So putting TVS on the input is the LAST place where you should place TVS diode.
I think the actual scenario when TVS diode is useful was not tested. When you have a big capacitor and connect a battery there is a lot of inrush current. And If your wires have enough inductance, high inrush current can cause the voltage to "overshoot" and probably kill MOSFETs. I don't know if that ever happens in reality (killing ESC while plugging in the battery) but having just one small TVS with the cap seems like a reasonable thing. I honestly can not imagine ANY other scenario when this diode will be useful. But anyway these tests are great, thanks for sharing.
Interesting... looks like I won't be worrying/rushing to update my fleet with one of the 15 or so tvs diodes I invested in! I've also read that having a smallish cap on the 5v or 12v rail (whichever is being used for the cam/vtx) is beneficial -I think Pidtoolboxguy did the testing... JB, on a separate note, for analogue builds do you advise powering the camera from the vtx directly (i.e. using the vtx's filtered voltage output) or from the FC (using perhaps a common ground pad if both vtx and cam are being supplied the same voltage)? Do you prefer using 9v or 5v for the cam if it can handle both? -I ask as I noticed you using the FC in your recent build video (Thank you)
Electronics usually have built in TVS, if they fail the whole board is more likely to fail, running a supplementary TVS adds an extra layer of protection, that is all. I stopped burning stuff out after I started adding TVS diodes, so I will continue.
There's 1 other reply to this message (denoted by the blue "1 reply" text, which will now say "2 replies"), but for some strange reason I can't see it?! Can anyone else read it?
A far better solution for noise suppression is a big cap and a much smaller cap in paralell, the big cap smooths out the low frequency noise from motors, while a small cap will deal with high frequency switching noise from switching regulators aka onboard voltage regulators vtxs, cams ect ect.. I did this all the time when I used to design synthesiser circuits.. I would never use just one large cap.. A large cap wouldnt be able to sink fast switching currents quick enough, hence why you will always find both in pretty much all electronic circuits these days..
so, on a 5 inch quad for example, one cap 35V 470μF and one 50V 220μF?
@@nakmax No.. There both very large caps.. A 470uf and somthing much smaller like 100nf or10nf would be better.. Valtage rating doesnt matter as long as its more than your battery puts out at full voltage.. so 35v is ok..
@@petrokemikal thnx mate
Something i should have added,i make audio equipment using vacuum tube/valves and as most guitarists use 9Vdc power banks i have to run my tubes/valves at proper tube/Valve anode/plate voltages of anything from around 70Vdc to around(for me)350Vdc.
I designed an smps boost converter that can be varied from around 60Vdc to about 380Vdc,this is possible due to the series inductor being switched in and out of conduction,the voltage increases due to inductors not being happy to stop conducting,they store energy in the form of a magnetic flux that swells like a balloon it's a field and as long as the inductor core does not saturate it will keep adding to the voltage,the voltage is passed to a high voltage capacitor via a switch,the switch is a either a schottky diode or a super fast diode,voltage dependent as the schottky in general don't have the same voltage range of say a ulf4001 that's ultra fast with the same voltage ratings of a 1N4001,diodes conduct in one direction only and so can act as a switch.
So deliberately we can allow a high voltage to be developed,this is with a pwm controller using feedback to skip pulses to control or hold the output voltage,as for uncontrolled Back EMF the field can get quite high very fast.
So inductive back EMF can destroy a transistor very quickly,and as such when using inductive loads ie a relay armature,an inductor or any wound inductive component a freewheeling diode should be used to protect against flyback voltages,the diode is actually a current steering diode that in one direction is not conducting and as soon as the reverse voltage appears the diodes junction barrier voltage is overcome allowing the diode to act as a clamp which will hold the back EMF to around 1V a little over the barrier potential of around 0.6v-0.7v generally in that region.
Incidentally the old style TVs ,computers and oscilloscopes using crt screens had a flyback transformer inside generating 1000's of volts without breaking a sweat.
Again the suppressor cap might stop EMi with the electronics in the same way car radios have/had a suppressor cap to null the EMi noise injected into the power supply via the alternator,or to peoples tv sets as you drove around.
The best method of suppression can be with a RLC filter,some shielding,or a simple RC filter using proportionally smaller caps to deal with HF content of the EMi,seeing as a motor rpm is linked to the frequency generated.
I did explain the best shielding method to minimise added weight and to give the best shielding,you must realise this shielding will block RF signals ,so external antennas must be used or you aint flying nowhere buddy.
Finally Joshua,i messaged about changing out ESC's on a cheapo drone with other mods to use higher current ESC's bigger motors and higher cell count batteries while using a buck or linear regulator to keep original control circuits and cameras etc as per original,to get a lightwieght cheap screamer,sadly no reply.
I'd appreciate some input as i'm still a noob at FPV.
I think i can give good insight to the electronics side of this,even making cheap shitty gear from china work better than some of the high end ESC's from here in the west(i already did this).
Still.i have subscribed and enjoy the vids,i hope i gave good clear information on not getting screwed by bogus info on Back EMF protection circuits for parts that actually cost pennies.
This is why the TVS diodes on ours are so small. Maybe it helps a bit. Doesn't hurt to have it on there. Weighs like 1/3g. So the cap/diode system with the wire attached at the appropriate max length and awg, it's as fool proof and maximally protective as reasonable for a general 5-7" system.
Hmm it could help with voltage spikes on plug in maybe. Only use I can see for it.
In my experience ESC deaths due to voltage spikes typically occur when plugging in - this is particularly true of FETTEC whom were the first to create a spike absorber, to combat the frequent fires on plug in 🤪
How the ESC dies is mystery #1 in our industry imo. I'm leaning towards physical stress on components resulting in sketchy contacts. I'm hoping to test this board cased in resin soon.....but I'm just one person unfortunately so my results will always be inconclusive.
Hello Bob, I do electronics design and ESC repairs for living. I've been trying to analyze this issue for ages and this is what I've been theorizing from my own experience.
Different kind of failures tell us what actually happened. If the ESC dies non-destructively, most of the times it will be the microcontrollers failed to get power or failing themselves. For the first issue, many times the ceramic capacitors failed shorted on one of the power stages. This can be due to poor placement where the capacitor experience physical stress. As for the second issue, I suspect induced transient on the signal line connected to each of the microcontroller like the ADC pins for EMF detection or other pins because this is how you kill a microcontroller other than by physical damage.
The way ESC dies destructively also tell us what could be happening. If the ESC failed without bringing the motor with it (single phase short), the FETs are the problem. Either this could be underrated parts or poor quality. If the ESC failed while bringing the motor with it, then the short happened across phase. Could be caused by the same issue as before, failure to detect the correct EMF to the point of coil saturation (overcurrent), or other external trigger that excites the FETs abnormally.
How would I come up with a reliable ESC design? Full-size 5x6mm FETs for typical 5 inch setup. Even a dual 5x6mm FETs is better spec-wise than 2x 3x3mm with smaller overall footprint. 2 capacitors in series configuration in case of one failing, it won't make a dead short. Larger resistors on gate inrush limiter because the smaller one can't really handle the peak drive current and often fail high resistance.
@@Asu01 I just had this happen where one ESC failed in a 4in1. but the motor works still. I thought it was the motor at first because it was just twitching but after swapping it I found that wasn't the case. No components get hot on the esc. Usually something gets hot on a dead esc from my experience. Oddly enough, that's not the case here.
I'd say the ESCs die by overcurrent. Those voltage dips are probably caused by an overcurrent happening in the motors. When you have a "short-circuit like" situation, the voltage will decrease in function of the current getting increased, that could explain why people burn their esc when having a jammed motor.
@@Asu01 I just randomly came back to these and saw your comment. Design an ESC. We'll get it made and see how far it gets. The industry would absolutely love to have a highly reliable board.
Isolate and the decouple the crap of the microcontroller and whatever other gate driving components there are, then increase all the ceilings on the mosfets as far as absolute ratings. Whatever is serving as the freewheeling, beef that up too. If anything, just to try avoiding ground issues I would actually just try to design the setup so that the low side (gate drive) of the ESC was completely isolated and using its own battery supply or a good isolated DC to DC converter. It's either that or take all of the filtering and decoupling real seriously by doing more than just tossing a couple electrolytics on that are supposed to be low-esr. A common practice for decoupling and snubbing when just winging it would be to stack values for various frequency ranges such as a 1nF/100nF/1uF set of decent ceramics. Then you have to actually tidy up everything else that is sharing the same system or close enough for magnetic interference. All the wiring is just inductance laying around. Good low esr caps on top is also good but shouldn't be the end-all. I would imagine any level of improper switching of the fets can result in a shoot-through that will kill them relatively fast. Larger and more balanced lower inductance PCB layouts would help but then you're increasing the size and weight.
Factors outside of the ESC design itself can seem to be causes of their failure even if they are being operated within reasonable limits. If they are not conformal coated I would imagine just moisture and debris could cause issues. Or the wiring and grounding. When I think about ESC's I think about VFD's, except VFD's are typically running 3ph AC induction motors. While they are eventually all going to suffer the same deaths for the same reasons no matter their filtering or protection.. what typically keeps them from dying prematurely is just the fault protections.. They are using switches (usually IGBTs) which are rated for very high short circuit currents for a certain duration of time, and that duration is longer than it takes for the built in overcurrent protection to activate. It just kind of seems like given the size limitations of what most people need, it's not really feasible to build an ESC that is going to never die because it has every perceivable protection.
Voltage dips occur on current surges - you are right, thats the answer. And that current cooks the ESC (and/or solder).
I think what blows the ESC is the heating up of the fete either melts the solder it touches next to it not creating a voltage spike but creating a short.
It is exactly as you assume. If the motor is blocked it will generate its highest torque and therefore the highest current. So either the motor winding will overheat first or the esc will blow up.
Ideally you’d want a high-volume cap on the input side of a motor ESC. Easy to do with discreet ESC’s, but not so much with the X-in-1’s.
A low impedance cap will always filter while other protection devices like TVS diodes are meant for much larger voltages. We luckily dont really have those, the back emf spikes are of very limited energy and if the cap is capable of sucking them up youre good to go.
A much smaller cap would probably be fine too. And I would definitely stick with 50V caps from experience unrelated to drones.
Then isn't the issue is not if a TVS diode is good or not, but should we install a voltage limiter diode or a current limiter diode?
It is over current because, without some way to limit it, the energy being supplied by the battery, to the motor (through the esc), is going to be converted into heat instead of mechanical energy. The heat changes the properties of the various layers inside the FET, causing it to behave more like a variable resistor with positive feedback until it burns out the internal trace to the pad - or worse.
Why wouldn't you have provided the link to the document your showing?
I always pictured the prop strike killing an esc by essentially increasing prop resistance in the same way trying to run a 12inch prop on a 5inch setup. Your asking it to spin the world in this case. You learn about over propping burning up your esc. If the prop can't move its the same ish as having it over propped right?
I should have added you should add some caps in series ie first cap of say 1000uF,then 100nF then 100pF to eliminate HF noise.The values are proportional.
not to say that i actually know the cause of esc blowing, but the inductance on the circuit can cause voltage spikes from suden changes in current, tho an increase in current can cause voltage drop or sag
Is there anything that reduces the spark when connecting a battery? I recently moved up to 6S and noticed a huge difference in the spark created when connecting the higher voltage battery. Is it due to the remaining voltage in the capacitor?
Looks like prop strikes result in Amperage spikes. I guess to protect the ESC you would need a fuse. That said, I do not suggest putting a fuse the ESC.
That voltage spike from a motor being stopped explains why a row of 5 capacitors literally exploded off my esc yesterday.... I was wondering how that happened when it wasn't even that hard of a crash.
In a 4s 3.5" quad, should i use 25v or 35v? And what uf, 1000 or less? Should, I use 1 or more ?
My vista is connected to my directly on my battery, in the connection the drone had it self a small capacitor 16v 100uf, should i change it?
Please help!!!
wait, my battery becomes quickly depleted after dropping below 3.5V.. it drops down to 2,8 almost instantly. Could it be because of a capacitor?
Very interesting and fun clip. Learned something new along with Joshua. I actually smoked a motor trying to turtle out of a tree. 5 inch race quad, aikon 20x20 esc vs a 2207 2450kv motor.
But what about to protect from an emp component
12:55 Is there any kind of electrical component that prevents current spikes? Or maybe a setting in betaflight to limit current draw or something
Fuse.
external TVS won't save the internal diode in the n-mosfet... so that thing will die regardless if it wants to. When a 3 phase motor abruptly stops the fight is with the coil in the motor and the mosfet on that channel and nobody else. What I am wondering is ... where exactly are we measuring the noise / voltage ?
Where's the link for that chart?
How does he keep it attached to the side of the quad like that? 00:30
12:13 do jammed motors also pull more current without bidirectional Dshot, where the ESC doesn't know how the motor is spinning?
the esc always knows how the motor is spinning, bdir dshot gives the fc that information.
I believe a 4in 1 ESC is under heavier stress when connected to 4 motors doing their own thing (some speeding up while other ones are braking).
I have a couple of quads in which I've burned one ESC on a 4-in-1 board, and kept running with one separate ESC, both these kwads have been incredibly reliable since.
A lot of pilots including me, believe in the effectiveness of TVS diodes+50V cap on 4in1 escs... I've myself never burned an ESC equipped like that... But I know there is nothing scientific here ;-)
So it’s current spikes instead of voltage spikes that kill ESCs. So… how do you protect against current spikes?
I took a high speed impact with a wooden pole with my racing drone trying to hit gaps, the frame was perfectly fine, but the 4 in 1 esc was slightly bent and 1 esc stopped working :D
Maybe run the same tests but log the current through the fets.
The largest current draw will always happen under "locked rotor" conditions. They can be massively higher than the running amps. Back EMF can also wipe out the processor, so it depends on the failure mode. These are all vastly more likely when the manufacture cuts corners, or uses poor engineering practices, which is not always reflected in the retail price.
Great Video as always.
Nobody seems to mention active breaking / damped mode in BLheli... Wouldn't that tend to generate higher voltage more than a prop strike?... Especially in a high air flow situation like an abrupt throttle cut at high speed?... I'm curious about people's thoughts on this as I don't know much about it...
What the test shows is that there is a current spike happening,which results in a voltage drop because the current limit gets cought. Does anyone know if this "maxing out" of the current has been checked as a reason for the ESC death? So basically the only current safety would be using a battery with lower current rating than the ESC, "abusing" the battery as limiter.
Forgive me for I'm not 100% fluent in electrical engineering. But what I assume is happening is t hat when a motor get bound by something the coils just perceive it as a substantial amount of drag. When the temp of the motor heats up resistance increases substantially and the ESC tries to over compensate and burns out to over current. If this were true a test could be done with a high amp ESC and whatever the equivalent a lower amp motor would be. If the motor smokes before the ESC and you could measure amps from the ESC it would confirm this theory.
My hypothesis is that the jamming of the motor causes a major spike in ohms and that causes a BUILD up of current that gets built up on the ESC
I just got into FPV exactly 21 days ago and I already spent over $1,500. What the f*** I went crazy over this
Its a lot to drop in such a short time for sure. Im in 30 months now and just topped 20k. I did the math just out of curiosity and I've averaged over $700/month. Ive kept a record from day one of every penny spent. Maybe I shouldn't have lol
потратить нужно столько , сколько есть. И немного оставить на поесть
The problem is that voltage spikes are generated FROM the motor and propagate to the INPUT. So putting TVS on the input is the LAST place where you should place TVS diode.
Perhaps it works as a short circuit.
Should have bent a prop and compared the noise compared to no damage
I think the actual scenario when TVS diode is useful was not tested. When you have a big capacitor and connect a battery there is a lot of inrush current. And If your wires have enough inductance, high inrush current can cause the voltage to "overshoot" and probably kill MOSFETs. I don't know if that ever happens in reality (killing ESC while plugging in the battery) but having just one small TVS with the cap seems like a reasonable thing. I honestly can not imagine ANY other scenario when this diode will be useful.
But anyway these tests are great, thanks for sharing.
They blow because the drivers that control the fets loose power and shoot through happens between high and low sides.
Interesting... looks like I won't be worrying/rushing to update my fleet with one of the 15 or so tvs diodes I invested in! I've also read that having a smallish cap on the 5v or 12v rail (whichever is being used for the cam/vtx) is beneficial -I think Pidtoolboxguy did the testing...
JB, on a separate note, for analogue builds do you advise powering the camera from the vtx directly (i.e. using the vtx's filtered voltage output) or from the FC (using perhaps a common ground pad if both vtx and cam are being supplied the same voltage)? Do you prefer using 9v or 5v for the cam if it can handle both? -I ask as I noticed you using the FC in your recent build video (Thank you)
Electronics usually have built in TVS, if they fail the whole board is more likely to fail, running a supplementary TVS adds an extra layer of protection, that is all.
I stopped burning stuff out after I started adding TVS diodes, so I will continue.
There's 1 other reply to this message (denoted by the blue "1 reply" text, which will now say "2 replies"), but for some strange reason I can't see it?! Can anyone else read it?
I’ve got the best analog video by using a Matek BEC for my VTX and the camera from the VTX.
Dear ELRS Devs, Joshua wants to take you out to drinks and dinner. Are you available? ; 0
The gear are getting “futuristic” like those power-saving devices plugged into the wall
what is that tshirt lol
It dies due to high amperage draw.
TVS diode may not have fast enough response to be useful.
Ok, not gonna bring up what I saw in the graphs. Ink blots and shit...😵💫