Even flying wings & inav, i've learned more by watching 5-6 of your videos than from years reading inav docs on PID, Noise, FIlter, etc. You really nail it!
Been trying to tell people this for months. MONTHS! Yelling it from the hilltops.Even reached out to you after your "power it by a separate battery" video. Shout out to Brian White of PID Toolbox. He's been preaching this for a while as well. Great you are getting the word out Mark!
I had some Andlog OSD issues on a bad design FC. With 4s had interference in OSD letters, and with 6s OSD completely gone on throttle high. Tryed any caps but problem were solved by adding good DC/5vor12vDC regulator at vbat/ground out of ESC to FC. I used blheli32 telemetry to get vbat voltage and current data.
Adding a capacitor to VBAT. Where is the best ground to connect minus leg of a capacitor? In the video I can see plus leg of the capacitor connected to BAT pad of FC, but where to connect minus leg? Any ground or the nearest ground to BAT pad or a main minus pad of ESC where battery soldered.
very interesting. since caps on the regulated lower voltages help this suggests a big sudden power draw from the motors is bringing them close to collapsing. which in the first instance would kill their communication over cables due to a low signal level but could even lead to a chip reboot or miscalculations of the ics. you may be able to see esc communication errors without the cap at the moment you spin the motors up like that. i wonder if it would be wise to add more capacity to the lower power lines and maybe even some (10? or 100?)pF to the esc communication lines.
very nice testing. what you have seen should however depend on the board and existing capacity there. also the chosen voltage regulaters play a big role. so it is quite board dependent.
shouldn't have to be this way... issue is there are a lot of manufactures and trying to fix all of their designs is a big job.. also personally as the guy running the MFG program in Betaflight im selective as to who I waste my time with....
I'm keeping this in mind planning on 2100 kv aos ultralight if 4s won't rip enough and if I'll need it. I'll still add caps and see. My current 4s are way heavier than recommended. I'll play around kinda safely
Im having some fc problems it maybe my gyro.... hesitation in my sticks and im getting false arm codes like a heavy load code when my cpu is only at 25 percent and now its kicking a your quad is not level code... im baffled 😢
Hey, thanks for this video. I found your results pretty interesting. Maybe try adding some 0805 or 1206/1210 SMD ceramic capacitors to the FC where you have a Vbat pad next to a ground pad, and a 3.3V pad next to a ground pad. The low impedance and great high frequency performance of a SMD MLCC capacitor may help eek out a bit better result.
I did not test it separately but that exactly what I did with my mamba f722 app that had noise on pitch and now even the logs look like there is no issue at all.
@@rehepeks That's pretty awesome. I sometimes feel like certain FCs are lacking in the decoupling department and I'm sure there's plenty of improvement to be had by a few caps strategically placed here and there. Big electrolytic caps, and a few small SMD MLCC caps. I've been doing this on my last few builds from the get-go so I have no back-to-back comparison, but maybe my next build, I'll do a before-and-after BB log to see.
I would have really going deep in the bench for my goofy SB v2s (I did save them just in case). But it's great that you provide a decent fix! Thank you sensei.
Thanks for the video. I was wondering if it's possible to clear the blackbox log on the speedybee V3 on the app? Perhaps in CLI? Also, I haven't been able to get a readable blackbox log of the v3 FC... I put it in a sealed waterproof build and was hoping to use the app to read blackbox data.
BF has a Mode switch to clear blackbox. See if that works. I have been cleaning through deleting the files in Mass-Storage mode, but I see your setup is all App or Modes (sealed). You should be able to view the BB through the app with a modern smart phone. You have to enable wifi and connect to it. If you have issues, maybe reach out to SpeedyBee. I think it is fiddly. I had to use my wife's Samsung S20. My S7 would not work.
I've got two brand new Tmotor F7HD with the BMI270 that appear to have either some electrical noise or bad gyros, didnt get any issues when it was using the MPU6000 on the first version of the board. I'm gonna try adding some caps to that and see what happens. Its not as bad as your one, but the motor tracers are noisy and spikey on flips/ rolls and the gyro doesnt follow setpoint during the roll/ flip. I have had success by turning the internal filter up to 750hz and using the betaflight gyro filters but its still not as clean as my MPU6000 boards.
Strange. Built a quad with the Tmotor 405 stack with BMI and have horrendous oscillations at idle throttle. With a notch filter at 270 Hz it's fine. Source One v5. th-cam.com/video/ctWYkuLHFCk/w-d-xo.html&feature=share&si=EMSIkaIECMiOmarE6JChQQ Strangely when I changed board rotation in bf the issue disappeared, but of course I couldn't fly that way because the board rotation was 'wrong'. When physically adjusting the board to the bf rotation the issue came back so I stuck to the notch filter for now.
Hey, check out Chris Rosser's channel and see his video on the BMI270. He gave some information to the Betaflight devs regarding a lowpass filter setting that could be changed inside the BMI270 with a serial command. I can't remember the name of the gyro filter profile, but it's meant specifically for FCs with the BMI270. Check it out and give that setting a try. Maybe it'll help ya 👍
Thanks for the video. I've messed around with this a few times and found that a 330uf on FC works just as well. It will save you some space to have a smaller one. Still use a 1000uf on the ESC tho. Or just spend $60-$70 and get the speedybee f4 v3 stack.
I'm saying this for years that the huge capacitances (some use over 1000uF) mostly added directly to the xt60 or putting it where the battery inlet cable connects to the PDB/ESC board doesn't do much in terms of filtering. What it does for sure is burning your xt60 plugs with this initial cap charging spark
Well I hadn't thought earlier... but you are right adding a Cap without any resistor sure does a huge current spike... between battery wires... people tend to forget that the placement of cap near the soldering pads is just the routing... there is a complete pi- filter sitting on the board{ pdb/esc}.
you should look up "capacitor derating" or "DC bias" it's when the capacitance is reduce as voltage is increase. A 16V 10uF Multilayer ceramic capacitors (MLCCs) powered at 16V can be actually be 3uF not 10uF. It's crazy. There is a tool call "simsurfing" from murata a semiconductor company where you have graphs showing the effect of voltage and capacitance for each part number they sell. So if for some reason the V3 is better, it could also be a coincidence that they assemble that batch with different brand of capacitors. If those caps are out of stock, it's so easy to say, well change them to something else because well "it's just a capacitor"
I am seeing this exact behavior on a mamba f722. Roll access only completely flipping out in black box. I will try to add a cap on the FC before I buy a new FC.
There are 25v tantalum capacitors under you link in description. Is it okay to use capacitor with lower voltage? 6.3v or even 4v for example? (Add: on 3.3v line)
Caps will pop if the voltage was too high. So definitely don't use 4v caps on a 25V line where the voltage spikes can go up to 35V or even 50V. Use at least 35V rated caps for 6S.
@@TimeFadesMemoryLasts Mark said he using tantalum capacitor on 3.3v, I don't think there are any big spikes. Sure, 35v for vbat is obvious. But what about 3.3v line?
My understanding is Adding a Buck converter on Vbat down converting to 5V , then feeding it to LDO and adding bunch of 10uf & 2.2uf on 3V3 Out should the trick... HOWEVER Adding an ultra LOW drop LDO on 3v3 is a good choice. Typically 200mv drop @1A.
I solved my speedy bee mini 43688 problem on my helicopter after adding filter and adding 100uF on the 3.3 volt. But I already ordered a new expensive t-motor FC with BMI270 on Amazon Prime now being delivered to Amazon Locker. I feel bad for the Amazon driver because I will not pick up the goods. 😢. I will post my helicopter's FPV onboard recording (Walksnail HD) in a few moments. Helis are very sensitive to noise. Solving it is crucial.
Once I ve installed a cap at 3.3v pad of fc. Then it started randomly rebooting itself. Turns out one of the cap legs did not solder to pad and touching once in a while in flight causing fc to reboot because draining the power. Just sharing it if ever something like this happens to someone
Even flying wings & inav, i've learned more by watching 5-6 of your videos than from years reading inav docs on PID, Noise, FIlter, etc. You really nail it!
Thanks!
Been trying to tell people this for months. MONTHS! Yelling it from the hilltops.Even reached out to you after your "power it by a separate battery" video.
Shout out to Brian White of PID Toolbox. He's been preaching this for a while as well. Great you are getting the word out Mark!
As usual, very informative..Believe me you tought me how to configure a fpv drone properly. ..Big thanks to you from Turkey
I had some Andlog OSD issues on a bad design FC. With 4s had interference in OSD letters, and with 6s OSD completely gone on throttle high. Tryed any caps but problem were solved by adding good DC/5vor12vDC regulator at vbat/ground out of ESC to FC. I used blheli32 telemetry to get vbat voltage and current data.
i can't find Vbatt on my FC then where's should i put the capacitor ?
Look for VCC pad.
@@uavtech ah,i see . found it sir. thanks
I always put 2 caps on my quad, one on the battery pads and one on the v bat of the fc I found it helps alot with cheap stacks.
Adding a capacitor to VBAT. Where is the best ground to connect minus leg of a capacitor? In the video I can see plus leg of the capacitor connected to BAT pad of FC, but where to connect minus leg? Any ground or the nearest ground to BAT pad or a main minus pad of ESC where battery soldered.
You want to capture the spikes before they get to the gyro, so ideal the closes + and - pads to the ESC connection of VBAT that powers the FC.
Fantastic tip, Mark! Thanks a bunch! 😃
Stay safe there with your family! 🖖😊
What type is the small capacitor in the 3.3V?
very interesting. since caps on the regulated lower voltages help this suggests a big sudden power draw from the motors is bringing them close to collapsing. which in the first instance would kill their communication over cables due to a low signal level but could even lead to a chip reboot or miscalculations of the ics.
you may be able to see esc communication errors without the cap at the moment you spin the motors up like that.
i wonder if it would be wise to add more capacity to the lower power lines and maybe even some (10? or 100?)pF to the esc communication lines.
very nice testing. what you have seen should however depend on the board and existing capacity there. also the chosen voltage regulaters play a big role. so it is quite board dependent.
Amazing!!!! What Tantalum cap should I try on a 2S 12A whoop AIO FC?
I use 2 caps, one on the esc and one on the fc vbat
Good idea!
shouldn't have to be this way... issue is there are a lot of manufactures and trying to fix all of their designs is a big job.. also personally as the guy running the MFG program in Betaflight im selective as to who I waste my time with....
I'm keeping this in mind planning on 2100 kv aos ultralight if 4s won't rip enough and if I'll need it. I'll still add caps and see. My current 4s are way heavier than recommended. I'll play around kinda safely
Im having some fc problems it maybe my gyro.... hesitation in my sticks and im getting false arm codes like a heavy load code when my cpu is only at 25 percent and now its kicking a your quad is not level code... im baffled 😢
Thank you. Going to try this. Hopefully it solves the issues I have.
Hey, thanks for this video. I found your results pretty interesting. Maybe try adding some 0805 or 1206/1210 SMD ceramic capacitors to the FC where you have a Vbat pad next to a ground pad, and a 3.3V pad next to a ground pad. The low impedance and great high frequency performance of a SMD MLCC capacitor may help eek out a bit better result.
I did not test it separately but that exactly what I did with my mamba f722 app that had noise on pitch and now even the logs look like there is no issue at all.
@@rehepeks That's pretty awesome. I sometimes feel like certain FCs are lacking in the decoupling department and I'm sure there's plenty of improvement to be had by a few caps strategically placed here and there. Big electrolytic caps, and a few small SMD MLCC caps. I've been doing this on my last few builds from the get-go so I have no back-to-back comparison, but maybe my next build, I'll do a before-and-after BB log to see.
I would have really going deep in the bench for my goofy SB v2s (I did save them just in case). But it's great that you provide a decent fix! Thank you sensei.
Thanks for the video. I was wondering if it's possible to clear the blackbox log on the speedybee V3 on the app? Perhaps in CLI? Also, I haven't been able to get a readable blackbox log of the v3 FC... I put it in a sealed waterproof build and was hoping to use the app to read blackbox data.
BF has a Mode switch to clear blackbox. See if that works. I have been cleaning through deleting the files in Mass-Storage mode, but I see your setup is all App or Modes (sealed).
You should be able to view the BB through the app with a modern smart phone. You have to enable wifi and connect to it.
If you have issues, maybe reach out to SpeedyBee. I think it is fiddly. I had to use my wife's Samsung S20. My S7 would not work.
I've got two brand new Tmotor F7HD with the BMI270 that appear to have either some electrical noise or bad gyros, didnt get any issues when it was using the MPU6000 on the first version of the board. I'm gonna try adding some caps to that and see what happens. Its not as bad as your one, but the motor tracers are noisy and spikey on flips/ rolls and the gyro doesnt follow setpoint during the roll/ flip. I have had success by turning the internal filter up to 750hz and using the betaflight gyro filters but its still not as clean as my MPU6000 boards.
I would be interested to see what you find out.
Strange. Built a quad with the Tmotor 405 stack with BMI and have horrendous oscillations at idle throttle.
With a notch filter at 270 Hz it's fine.
Source One v5. th-cam.com/video/ctWYkuLHFCk/w-d-xo.html&feature=share&si=EMSIkaIECMiOmarE6JChQQ
Strangely when I changed board rotation in bf the issue disappeared, but of course I couldn't fly that way because the board rotation was 'wrong'. When physically adjusting the board to the bf rotation the issue came back so I stuck to the notch filter for now.
Hey, check out Chris Rosser's channel and see his video on the BMI270. He gave some information to the Betaflight devs regarding a lowpass filter setting that could be changed inside the BMI270 with a serial command. I can't remember the name of the gyro filter profile, but it's meant specifically for FCs with the BMI270. Check it out and give that setting a try. Maybe it'll help ya 👍
I wonder if you would change board direction and adjust bf accordingly, if another axis would be affected then.
No, it follows that axis if you rotate the board 90-deg.
Thanks for the video. I've messed around with this a few times and found that a 330uf on FC works just as well. It will save you some space to have a smaller one. Still use a 1000uf on the ESC tho. Or just spend $60-$70 and get the speedybee f4 v3 stack.
actually with the added capacity on the lower power lines you can reduce the cap size at the battery and get a similar effect.
Which also has issues with gyro, ~3% having issues with roll or yaw axes
@@OlegShevtsov512 I've heard the reports. Though it doesn't seem as prolific as the v2. Well now we got the v4. Let's see what happens. 😄
I'm saying this for years that the huge capacitances (some use over 1000uF) mostly added directly to the xt60 or putting it where the battery inlet cable connects to the PDB/ESC board doesn't do much in terms of filtering. What it does for sure is burning your xt60 plugs with this initial cap charging spark
Well I hadn't thought earlier... but you are right adding a Cap without any resistor sure does a huge current spike... between battery wires... people tend to forget that the placement of cap near the soldering pads is just the routing... there is a complete pi- filter sitting on the board{ pdb/esc}.
Great, thx for the testing!
you should look up "capacitor derating" or "DC bias" it's when the capacitance is reduce as voltage is increase. A 16V 10uF Multilayer ceramic capacitors (MLCCs) powered at 16V can be actually be 3uF not 10uF. It's crazy. There is a tool call "simsurfing" from murata a semiconductor company where you have graphs showing the effect of voltage and capacitance for each part number they sell. So if for some reason the V3 is better, it could also be a coincidence that they assemble that batch with different brand of capacitors. If those caps are out of stock, it's so easy to say, well change them to something else because well "it's just a capacitor"
Use T-Motor ESCs with a built in 10V BEC. Solves this issue totally.
I am seeing this exact behavior on a mamba f722. Roll access only completely flipping out in black box. I will try to add a cap on the FC before I buy a new FC.
How did it go? I also have a mamba f722 with noisy pitch axis
@@FlaviuFpv I replaced the board. Way too much trouble.
There are 25v tantalum capacitors under you link in description. Is it okay to use capacitor with lower voltage? 6.3v or even 4v for example? (Add: on 3.3v line)
Caps will pop if the voltage was too high. So definitely don't use 4v caps on a 25V line where the voltage spikes can go up to 35V or even 50V. Use at least 35V rated caps for 6S.
@@TimeFadesMemoryLasts Mark said he using tantalum capacitor on 3.3v, I don't think there are any big spikes. Sure, 35v for vbat is obvious. But what about 3.3v line?
@@nikitaak Oh, yeah then 4-6V should be enough. For a 5V line a 4V cap definitely doesn't cut it.
I noted later that was not a great link. I switch so spec a 6.3v cap for the 3.3v.
remember ceramic capacitors failure mode (cracking) is causing a short.
Rushed products, the on-going theme since 2020 👀
Very interesting!
My understanding is Adding a Buck converter on Vbat down converting to 5V , then feeding it to LDO and adding bunch of 10uf & 2.2uf on 3V3 Out should the trick... HOWEVER Adding an ultra LOW drop LDO on 3v3 is a good choice. Typically 200mv drop @1A.
I agree with you，Add another LDO for gyro is a good method.
I solved my speedy bee mini 43688 problem on my helicopter after adding filter and adding 100uF on the 3.3 volt. But I already ordered a new expensive t-motor FC with BMI270 on Amazon Prime now being delivered to Amazon Locker. I feel bad for the Amazon driver because I will not pick up the goods. 😢. I will post my helicopter's FPV onboard recording (Walksnail HD) in a few moments. Helis are very sensitive to noise. Solving it is crucial.
sweet!
It's transients, TVS what will fix it.
Practice practice practice
In soldering and troubleshooting 😂
Once I ve installed a cap at 3.3v pad of fc. Then it started randomly rebooting itself. Turns out one of the cap legs did not solder to pad and touching once in a while in flight causing fc to reboot because draining the power. Just sharing it if ever something like this happens to someone