Do you even know why MLCC are required next to switching stage? and for e.g. do you know why they need to be freaking 2x voltage rated? Look at the pictures do they look like 100V capable? Regarding ESR that whole product will not have not even a single component related to low ESR at all. Do you think 600uF capacitors does even have low ESR? They are cheap Chinese no names yes they do have capacity and something looking similar to original values to make it work but it will not have anything even close related to any other capacitors you can get at distributors as Digikey or etc. Low ESR donât make me laugh clearly you donât know a shit about Chinese productsâŚ
Do you even know what tolerance values mean? That at higher temperatures and other conditions the resistor will be still within 1%-5% or etc if you use it on a breadboard without any much hard conditions you will still read pretty close resistance even with 5% tolerance rated resistors so it does not matter much, but if you at normal conditions get tolerance deviations when its something wrong with them.
I am not saying VESC is good, it lacks protection, filters, and etc stuff on every piece of the schematic block⌠Those DRV work perfectly fine in consumer products and etcâŚ
Your argument points would be valid if the parts used in their products wouldnât use no name componentsâŚ
Itâs because of transients, at fully charged 12S, the DRV8302 dies to positive transients,
At really high currents, the DRV8302 (I think) dies to a negative transient.
There are countless application notes from companies such as IR, Infineon, TI, etc. that address this problem.
It is as easy as adding a tuned resistor and capacitor between the node and ground, (RC snubber) that literally costs pennies extra and maybe only an hour of testing on a workbench with an oscilloscope.
Another device, called a high side node clamp, might be required at high currents, especially considering the godawful reverse recovery time on the IR MOSFETâs body diode, to protect the pins of the gate driver from negative transients.
I actually do know why bypass capacitors are needed next to anything that switches, I donât think you do; I also know that you donât use ceramic caps to bypass power stages, (their power density is insufficient). Ceramic capacitors need to be 2x rated because they lose most of their capacitance at their rated voltage, you learn something new everyday.
Fun fact, did you know the original VESC 4.12 schematic calls for 50V rated ceramic caps on the supply rails?
Another trap people fall into is using a ridiculous amount of bypass capacitance on power stages, you donât really need that much.
Is it really a âChinese no-nameâ? is the capacitor unmarked and unbranded? You can always ask them where they bought the capacitor from. Even then, itâs not a problem if it falls within specification.
Do you know what âESRâ stands for and why it is important?
Do you also know why it is such a horrible idea to put your electrolytic capacitors on a separate board from the power stage, separated by several inches of wire? If you do, then youâd understand why it was somewhat necessary to have a giant 10uf 100V MLCC on the poorly designed PCBs of the 4.12 and not necessary on a PCB that actually HAS THE ELECTROLYTIC CAPACITORS ON THEM.
You are just using the lack of a ridiculously overpriced (and unnecessary) capacitor as an excuse to devalue the ESC.
And btw, nowadays, most of your Integrated Circuits and passive components from big name companies, like ON Semi and ST Microelectronics, are actually fabricated in plants located in China. I donât understand your obsession with marking products as âChineseâ and then immediately devaluing them for no reason.
I donât think you know what tolerance means. TEMPERATURE COEFFICIENT is a completely separate parameter used to qualify resistors and other passive components. It is normally expressed as parts per million per degree Celsius. Vishay has a really cool (and expensive) resistor that completely mitigates temperature effects.
Tolerance is the initial deviation from the nominal value when it leaves the factory.
Were the 10k resistors I used 10k Ohms? More like 10.2k and 9.7k when I measured with a multi-meter. Was that a 5% deviation? No, of course not. Was it greater than 1% deviation? Yes. What does this mean? It means 1% tolerance is already more than good enough for most products, including professional U.S. products.
How do you know they are âno-nameâ components? Do you ever see a passive component (that isnât a giant electrolytic) branded with a part name or manufacturer name? Even then does it really matter? The reason why passive components are COLOR-CODED rather than branded is because their production is standard and no secret to anyone. The manufacturer doesnât really matter unless the components were consistently out of specification, in which case if they were, problems would arise very quickly and wouldâve been reported on this thread. No, a 0.5V difference is not significant unless you are trickle charging the batteries (which this device does not), and if youâre really worried about overcharging your batteries during regen, then just charge your battery pack to 90 or 95%, or just change the voltage cutoffs in VESC tool to compensate, sheesh.
Youâd be pretty hard pressed to find the DRV8302 used in a consumer product other than the VESC (which honestly really should use separate gate driver ICs anyway, thatâs why ebike controllers do, and they donât break unless you crash into a building). The reason why people still buy the chip from TI is because they donât subject it to (or beyond) its absolute maximum ratings on a regular basis, most engineers are rational enough to use the chip in 24V maybe 36V systems, but they donât push it further, and if they do, they take precautions to ensure it doesnât explode, transient suppression being one of those precautions.
Iâm not sure what was going through Benjamin Vedderâs head when he decided to use inferior International Rectifier MOSFETs and completely leave out transient suppression when designing the VESC, maybe it was so he could sell more VESCs as many would inevitably break, or he completely derped out and forgot that switching transients exist.
You just mentioned one word bypass capacitors and not a single theory why you donât need them.
Haha haha
Look at the BOM list, capacitors resistors are usually from well known names like panasonic, vishay and other good brands.
I know at least 10 manufacturers who make electronic modules based on DRV for biggest brand tools and etc (Those brands never do it themselves they buy already created modules for this kind of stuff) You can take some better brand brushless battery-powered electric tools and you will find DRV inside there.
Again lack of knowledge⌠He sold just maybe several of those vesc by himself everything else was other people who actually make them and sell it.
tl;dr: Long paragraphs of useless information, definitely visible lack of knowledge just trying to show off⌠Any real EE would laugh of thisâŚ
@Gamer43
You may well be very competent in electronics but you are very misinformed on the history of the VESC. It is only recently that Ben began to make a substantial amount of money with his design. Before that, he sold a couple of hand soldered VESC (I have one of these early birds) and published the design fully open source.
You donât need to make a contest of offensive statement with Kug3list trying to prove whoâs got the biggest !
If you have better to share as a fully open source and applicable design, please donât restrain yourself and do it ! We appreciate very much your knowledge (well at least I do), there is no use to humiliate those who have other arguments (or less knowledge I cannot judge).
I apologize for that, Ben mustâve been using it in lower voltage designs and felt transient supression was unnecessary (because it is in that situation, and would be a waste of time figuring out the right values to insert for a snubber of any sort). It just really bugs me that sellers of his design misled customers into thinking it could handle 12S at 60A, only to leave them with a pound and a half of electronic waste because they stressed the controller beyond its limits.
I am working on a controller for a college thesis, and I will most definitely make its files open-source once I complete it.
So, for no-electro engineers guys like me, but huge VESC fan, user and rider , what will you suggest guys ? I ask to all of you Kug3, Gamer, HyperIon etc
We should use vesc at alf what said the seller ?
What could we do to upgrade it with what we could, more than buy heatsink case ? Change Caps ? air forced etc ?
But most important good parameter set !
I saw that Enertion engineers recommended 30amps batt max? (Right?)
I ask cause I used VESC at 12s like alf of us here, without any problem, but I never set them at the limit too (except V6 that handle a lot) .
What will be the ârealâ/âmore safetyâ parameter and setup for the 4.20hw ?
Thanks for your help
12S is fine for BLDC mode since there is a lot less switching going on than in FOC, and problems normally arise on a fully charged battery, since the voltage is so high (50.4V).
If you have no heatsinking or forced air cooling, I wouldnât run more than 30A continuous through the motor, I have a VESC from TB and one I hand soldered (with different MOSFETS, IPB014N06), the power stage is almost too hot to touch running only 20A (around 10 continuous) motor current on the ESC from TB.
With proper heatsinking, you could probably get away with 40-50A continuous motor current, but the TO-263 package is very difficult to heatsink properly with an external heatsink, especially since the high side and low side MOSFETs are on opposite sides of the board.
Max battery current is whatever your battery can handle, since it is the motor current that is run through each phase in the power stage.
For the Flipsky 4.20 hardware, I would be very comfortable running 40 A continuous through it and donât doubt the 50A rating, especially since it comes with its own heatsink, and doesnât suffer from high switching losses that the 4.12 hardware does (in fact, it has minimal switching losses and is not even worth calculating at this point since it is so low with the NTMFS5628NL, it never amounts to more than a watt within the usable current and voltage range). With forced air, you could probably pull 70A out of the Flipsky 4.20, but take that with a grain of salt.
Always be careful with fully charged 12S, one stray transient and you have to heatgun or hot air a new DRV8302, but I would bet that the failure rate is lower since the parasitics on the PCB will most definitely be better. Again, should be reasonably ok with BLDC mode, but with all the extra switching in FOC, I would not run FOC at fully charged 12, I would undercharge the battery to 48V or even lower. My friendâs ESC from DIY ElectricSkateboard died when he tested FOC at fully charged 12S (on one of those huge 6374 motors for that matter).
Thanks for your very helpfull feedback !
I charge 4.16v/cell so Iâm a bit less than 50v and no direct downhill around my house (I wiil set it at 4.15v/cell or even less).
So a 10s setup I more appropriate/appreciate by the vesc ?
When do you say 40A for the 4.20 its motor or batt max ?
I ask cause I used 12s on all my setup now single or dual all set at motor max > max amps from reseller characteristics, batt around the max too but a bit less (5 / 10Amps less than) whitout any problem. But I prefer to ask Im waiting my 4.20 dual and mini.
Once again thanks you
10S works great on the VESC, and I am guessing that is the application it was originally designed for.
40A max for batt and 60A motor should be fine, since you wonât be running those currents all the time.
Were you running BLDC mode at 12S? Since you undercharge your batteries thatâs most likely why your VESCs havenât exploded, which is smart, improves battery life and keeps the controller safe.
The VESC killer is fully charged 12S running Field Oriented Control, BLDC mode (slow decay) seems to skirt by because it switches only one phase instead of all three.
thx haha it wasnât voluntary, just, even if I got no downhill immediately starting from home, I wonât "overcurrent " (?!..not sure about what Iâm saying in this English word but hope you get it ^^) cause after an uphill theyâre always a downhill !.. And as seller sail them and used them at 12s ⌠I used them too without any doubt until now ⌠with the new design and the lack of feedback on them.
Yes bldc on 4.10, Focbox and v6 at 12s so around 49.92v when charged.
Thanks again for your highlight !!! much appreciated
Thanks, I am trying to figure out a way to do an on-the-fly resistance measurement so that my controller will be able to estimate motor temperature without a sensor and compensate accordingly.
My idea right now is that every once in a while, when a zero cross in the BEMF is expected on one of the phases, the controllers pulls the phase to high impedance before the zero cross, and measures the BEMF. Really close to the zero cross, it will then inject current into the phase and then measure the rate of current decay. Using the previously measured BEMF, it could then extrapolate the BEMF into the current decay period (since the BEMF is linear with respect to time for the most part in this region) and then do a resistance calculation (accuracy may vary greatly). Problems I could see with this is whether itâll end up interfering with the Sliding Mode Observer, but there should be a way to compensate for it considering it directly measured the BEMF on one of the phases.
Or I could be approaching this totally wrong and there is actually a way to measure resistance using d-axis current injection, but I have yet to see an implementation or paper on this.
No problem. If your controller has been working at 12S for a while, there shouldnât be anything to worry about as long as you keep your set limits where they are. Some DRV8302s are more rugged than others, and I guess my friend just got unlucky with his.
47uF 100V is pretty freaking big. There are only available in the through-hole package and still are obsolete. Biggest available capacitance in 1210 package and 100V is 10uF
It was most likely a typo and they meant 4.7uf, which is already more than enough. Application notes from Texas Instruments themselves only recommend 1uf of ceramic bypass.