15:00 - Looks like quantization effect in low-res ADC. ie. the rpm is so low as to give a conversion function no headroom over integer rounding, thus you lose resolution. As the RPM increases, quantization noise decreases. That's one guess - and if that RPM transducer is designed for "tens of thousands of RPM" - then that would be another supporting factor to my guess; it just isn't designed to be accurate that low - it probably has a "working range" from 1,000-10,000.
+AKorigami Definitely, it's always been a rough block used for holding things down, bashing things, etc, so it was very satisfying to clean it up and turn it into something more precise.
I wanna see something like a washing machine motor spin up that colossal lump of a flywheel :D That'd be a sight! I wonder if the spikes are noise caused by vibrations as the mass spins up, shaking some of the connections on the breadboard? There might not be that much vibration from the mass itself, but that motor wobble might be enough.
+Azayles I have some other motors that would be fun to try, none from washing machine motors but they're of similar size. Encoder issue is sorted out, I'll explain in the next video.
the spikes may be from the wobble combined with the inherent runout of your 3 jaw chuck when you turned the flywheel, producing a harmonic at various rpm. I thought it was too involved in making the flywheel, but then just using what you had laying around is cool. a smaller flywheel, possibly one that weighs only 5-6lb would be more than enough for small RC motors very interesting vid! I'm also looking into making a small dyno, for small 2 stroke bicycle engines and came across this. +1
Many motors are rated by volts/rpm and Newton-meters/amp. Have you considered logging the current into the motor as it spins up the flywheel? That should give you a good way to compare different motors.
+dn I have thought about it, but don't yet have an automated way to log this info. The basic RPM v time graph has been fairly illustrative so far, still experimenting though.
a better way to get that shaft mounted would have been to leave it at interference fit size, shaft in the freezer as cold as it will go, drum in the over as warm as you can get it, press it in then as soon as the temps equalize it should never be coming out.
@@ScottRumschlag They're around 11nm at 6500rpm. However I'm looking at increasing this with a custom efi system based around an Arduino of some sort and some forced induction means.
Just want to have a visible baseline and be able to produce torque graphs to show what changes gain or loose power and as you said commercial systems are quite expensive.
ADD a .01 microfarad capacitor to the pickup lead them to ground... This will clip any erroneous signal, pickup holding capacitance or RF problem ... Try it ... .01 caps are one of my Little secrets.. Along with always having some Op-Amps & thermocouples ... Think about it... It's all bout energy conservation / utilization #SilentWatchDiv Nikola & BuzzBox
When you refer to "we", is that the royal we? I don't see anyone, other than you, doing the things you're describing. Also, what kind of signal does your encoder output?
+MedEighty The encoder puts out a quadrature signal, which is "intercepted" by the 2022 decoder chip, which is then read by the Arduino. I'm looking into simpler systems for later but this works for now. For the "we" thing, I suppose that I mean it in the way a teacher or presenter uses it, so it would be myself and whoever is watching.
Scott Rumschlag Okay then. You may have already looked into these possibilities, but I can think of a couple of possibilities for why you get spikes in your data. - The output of the encoder is not being adequately 'debounced', resulting in erroneous values decoded by the 2022 chip. But the chip's data sheet says that it has built-in noise reduction, so it may be that the clock signal supplied to the 2022 chip is not of a high enough frequency for its sampling and noise reduction circuitry to work properly (for the rate at which the encoder spins). - There's noise in the output signal of the 2022 chip or data output from the 2022 chip is sometimes just about to change, as it is sampled by your code, resulting in an invalid value being read. You may have to 'debounce' the chip's output too. For example, you may have to read the output value at a much faster rate and only treat the most common (within a certain time window) as the correct reading. Or you may just have to be a little quicker with latching the output of the 2022 chip (with signals you send to the OEN, SEL1, and SEL2 pins of the 2022 chip). I'd be very interested to know what is actually causing the spikes. So if you do eventually find out, please let us know.
+MedEighty I have often been bothered by this usage of "we" as well, although it is quite harmless. I have noticed a number of other youtubers who do the same thing. I suppose there is some [corporate] stigma with the use of "I" vs "we" as it often means you are not a team player. I will of call BS on that one for sure and will use the appropriate term where applicable and when necessary. Sometimes it makes a single entity appear larger than what they really are in order to gain respect or allow a potential customer to feel more confident about work being performed.
Keith Ward Yes, I agree that if you're a one-man-band, trying to sell something to a customer who prefers working with big organisations, then "we" would be a better term to use than to say "I" (unless you want to get a job with your customer; in which case "I" would definitely be the right term to use). With other things, I think that if you clearly did the work, as is the case here, saying "I" would be more accurate and less pretentious. If you are presenting something you have done as a team, then "we" would be the correct term to use (and, perhaps, identify individuals who did whatever portion, to give credit where credit is due). I once had a colleague who seemed to think documenting his computer code meant telling us a story, taking us on a journey of his code, and explaining and justifying everything that was being done by typing comments like "we now need to open the file and read its contents" and "we must make sure the file has been closed", which was kind of condescending and irritatingly verbose. Or, maybe, he felt that what he was doing was so irrational that he needed to justify it to the reader at every step.
+MedEighty I have seen and read commented code like that as well. Sort of reads like a text book in that the "we" in this case is possibly correct in that you and the instructor or author are doing it together. I agree that this would be annoying to read from a co-worker though.
I love the old drill press left of the window, spent many an hour turning one of those for my Dad when I was a kid...
15:00 - Looks like quantization effect in low-res ADC. ie. the rpm is so low as to give a conversion function no headroom over integer rounding, thus you lose resolution. As the RPM increases, quantization noise decreases. That's one guess - and if that RPM transducer is designed for "tens of thousands of RPM" - then that would be another supporting factor to my guess; it just isn't designed to be accurate that low - it probably has a "working range" from 1,000-10,000.
Love that shiny mass of steel!
+AKorigami Definitely, it's always been a rough block used for holding things down, bashing things, etc, so it was very satisfying to clean it up and turn it into something more precise.
NEVER NEVER NEVER attempt using a late with sleeve sweater !!!!!... for everything else , nice wrok!
vibration from wobbly drive shaft ? can cause spikes
I wanna see something like a washing machine motor spin up that colossal lump of a flywheel :D That'd be a sight!
I wonder if the spikes are noise caused by vibrations as the mass spins up, shaking some of the connections on the breadboard? There might not be that much vibration from the mass itself, but that motor wobble might be enough.
+Azayles I have some other motors that would be fun to try, none from washing machine motors but they're of similar size. Encoder issue is sorted out, I'll explain in the next video.
Scott Rumschlag Top man! Can't wait to see :D
the spikes may be from the wobble combined with the inherent runout of your 3 jaw chuck when you turned the flywheel, producing a harmonic at various rpm. I thought it was too involved in making the flywheel, but then just using what you had laying around is cool.
a smaller flywheel, possibly one that weighs only 5-6lb would be more than enough for small RC motors
very interesting vid! I'm also looking into making a small dyno, for small 2 stroke bicycle engines and came across this.
+1
that cylinder was a better flywheel then I expected.
Many motors are rated by volts/rpm and Newton-meters/amp. Have you considered logging the current into the motor as it spins up the flywheel? That should give you a good way to compare different motors.
+dn I have thought about it, but don't yet have an automated way to log this info. The basic RPM v time graph has been fairly illustrative so far, still experimenting though.
thank you to watch
a better way to get that shaft mounted would have been to leave it at interference fit size, shaft in the freezer as cold as it will go, drum in the over as warm as you can get it, press it in then as soon as the temps equalize it should never be coming out.
Would you be able to use this same style to dyno small petrol engines? Around 125cc
I'm not sure how much torque they produce, but the principles should certainly be the same, just a bigger flywheel.
@@ScottRumschlag They're around 11nm at 6500rpm. However I'm looking at increasing this with a custom efi system based around an Arduino of some sort and some forced induction means.
Just want to have a visible baseline and be able to produce torque graphs to show what changes gain or loose power and as you said commercial systems are quite expensive.
How do you measure angular acceleration??
Change in speed over a given time period
ADD a .01 microfarad capacitor to the pickup lead them to ground... This will clip any erroneous signal, pickup holding capacitance or RF problem ...
Try it ... .01 caps are one of my Little secrets.. Along with always having some Op-Amps & thermocouples ... Think about it... It's all bout energy conservation / utilization
#SilentWatchDiv
Nikola
&
BuzzBox
When you refer to "we", is that the royal we? I don't see anyone, other than you, doing the things you're describing. Also, what kind of signal does your encoder output?
+MedEighty The encoder puts out a quadrature signal, which is "intercepted" by the 2022 decoder chip, which is then read by the Arduino. I'm looking into simpler systems for later but this works for now.
For the "we" thing, I suppose that I mean it in the way a teacher or presenter uses it, so it would be myself and whoever is watching.
Scott Rumschlag Okay then.
You may have already looked into these possibilities, but I can think of a couple of possibilities for why you get spikes in your data.
- The output of the encoder is not being adequately 'debounced', resulting in erroneous values decoded by the 2022 chip. But the chip's data sheet says that it has built-in noise reduction, so it may be that the clock signal supplied to the 2022 chip is not of a high enough frequency for its sampling and noise reduction circuitry to work properly (for the rate at which the encoder spins).
- There's noise in the output signal of the 2022 chip or data output from the 2022 chip is sometimes just about to change, as it is sampled by your code, resulting in an invalid value being read. You may have to 'debounce' the chip's output too. For example, you may have to read the output value at a much faster rate and only treat the most common (within a certain time window) as the correct reading. Or you may just have to be a little quicker with latching the output of the 2022 chip (with signals you send to the OEN, SEL1, and SEL2 pins of the 2022 chip).
I'd be very interested to know what is actually causing the spikes. So if you do eventually find out, please let us know.
+MedEighty I have often been bothered by this usage of "we" as well, although it is quite harmless. I have noticed a number of other youtubers who do the same thing. I suppose there is some [corporate] stigma with the use of "I" vs "we" as it often means you are not a team player. I will of call BS on that one for sure and will use the appropriate term where applicable and when necessary. Sometimes it makes a single entity appear larger than what they really are in order to gain respect or allow a potential customer to feel more confident about work being performed.
Keith Ward Yes, I agree that if you're a one-man-band, trying to sell something to a customer who prefers working with big organisations, then "we" would be a better term to use than to say "I" (unless you want to get a job with your customer; in which case "I" would definitely be the right term to use).
With other things, I think that if you clearly did the work, as is the case here, saying "I" would be more accurate and less pretentious. If you are presenting something you have done as a team, then "we" would be the correct term to use (and, perhaps, identify individuals who did whatever portion, to give credit where credit is due).
I once had a colleague who seemed to think documenting his computer code meant telling us a story, taking us on a journey of his code, and explaining and justifying everything that was being done by typing comments like "we now need to open the file and read its contents" and "we must make sure the file has been closed", which was kind of condescending and irritatingly verbose. Or, maybe, he felt that what he was doing was so irrational that he needed to justify it to the reader at every step.
+MedEighty I have seen and read commented code like that as well. Sort of reads like a text book in that the "we" in this case is possibly correct in that you and the instructor or author are doing it together. I agree that this would be annoying to read from a co-worker though.
3:52 Turn down for what!
its late but those spikes are probably cause by inductive spikes in the motors, try googling a flyback diode to solve that one
Where iz wiring and code?
Take the weight to a scrap yard and have it checked for radioactivity. Kid you not.
+D irk3672 Bad experience? I think we found this one on the side of the road about 20 years ago. It's been our go to hold-down weight ever since.