Electrical power measurements on my Agattu
- Thread starter 10mph
- Start date
I agree the Pansonic system isn't as easy as some to calculate powers from hill climb performance 10mph, but it's not too difficult. I do actually measure my hill slopes accurately as you'll see here.
Equally my unpowered physical capabulty is easily assessed from hill climbing that way, so these together enable a fair degree of accuracy
Equally my unpowered physical capabulty is easily assessed from hill climbing that way, so these together enable a fair degree of accuracy
Static test - 15 kHz PWM frequency
What I see is 0.8 Volt negative spike followed by damped ringing of about 5 cycles within a time of about a period of about 5 microseconds. Then there is a positive spike with again 5 cycles of ringing. There is then a slight slope in the base line before the next pulse comes 65 microseconds later.
This is measurement was made at the end of a metre or so of wire going from the battery terminals to my handlebars.
My first thought was that this wave form is indicating that the PWM is operating on for 5 microseconds and off for 65 microseconds - a duty cycle of 7.5%. However the odd thing is that as I increase the pedal pressure the duty cycle does not change, but the spike amplitude does. More thought about the circuit is required to explain this.
To eliminate any possibility that these spikes are somehow affecting the readings on my DVM or analogue meter, I shall install some RC filtering.
I don't know if such high frequency pulses, would somehow cause the battery to give a markedly non-linear relationship between DC voltage drop and DC current drawn.
I have carried out the static test. I can see a 15kHz pulse train. The motor is not turning so I assume this is not the motor phase FETs firing but a signal from the PWM running at this frequency.
What I see is 0.8 Volt negative spike followed by damped ringing of about 5 cycles within a time of about a period of about 5 microseconds. Then there is a positive spike with again 5 cycles of ringing. There is then a slight slope in the base line before the next pulse comes 65 microseconds later.
This is measurement was made at the end of a metre or so of wire going from the battery terminals to my handlebars.
My first thought was that this wave form is indicating that the PWM is operating on for 5 microseconds and off for 65 microseconds - a duty cycle of 7.5%. However the odd thing is that as I increase the pedal pressure the duty cycle does not change, but the spike amplitude does. More thought about the circuit is required to explain this.
To eliminate any possibility that these spikes are somehow affecting the readings on my DVM or analogue meter, I shall install some RC filtering.
I don't know if such high frequency pulses, would somehow cause the battery to give a markedly non-linear relationship between DC voltage drop and DC current drawn.
Oscilloscope PWM waveforms
I have got some pictures of the pulses on my oscilloscope screen. This time I have it AC coupled and I have the probe connections reversed so that the +ve pulse on the screen is actually a negative pulse relative to 26.V battery +ve terminal.
It is reassuring to see the pulse width reduces as the pedal pressure is reduced. I am not sure why the pulse amplitude also reduces.
Note: I can only do these measurements on a stationary bike. My oscilloscope is big and heavy and needs plugging into the mains! I would love to see what happens when pedalling with maximum power going into the motor. Where can I get a cheap roller system to run my bike on like the sprinters do to warm up in the velodrome?
I have got some pictures of the pulses on my oscilloscope screen. This time I have it AC coupled and I have the probe connections reversed so that the +ve pulse on the screen is actually a negative pulse relative to 26.V battery +ve terminal.
It is reassuring to see the pulse width reduces as the pedal pressure is reduced. I am not sure why the pulse amplitude also reduces.
Note: I can only do these measurements on a stationary bike. My oscilloscope is big and heavy and needs plugging into the mains! I would love to see what happens when pedalling with maximum power going into the motor. Where can I get a cheap roller system to run my bike on like the sprinters do to warm up in the velodrome?
Attachments
Panasonic Gitane, 3 speed hub, 26" wheels.
As in post # 10, I made up an extension between the underside of the battery plate and the positive wire, looping it up through the frame behind the battery moulding.
Secured? the current probe to the frame with bailer twine, bubble wrap and masking tape, the coiled lead reaching the hand held meter. Due to physical constraints the clamp is not centered in the loop.
A nearby steep hill, max about 1 in 3 showed a steady 13.3A over a number of seconds (my max) after a short excursion above 14A.
After the test the clamp zero had drifted to + 0.2A.
The meter claims +/- 0.5% of rdg +/- 2 dgts.
The clamp claims +/-(2% rdg + 2A) over 0 to 600A.
However within the range of normal usage of up to 20A (max of a somewhat better meter) this combination agrees closely with the better meter when it is connected in series with small brushless motors.
Dave.
As in post # 10, I made up an extension between the underside of the battery plate and the positive wire, looping it up through the frame behind the battery moulding.
Secured? the current probe to the frame with bailer twine, bubble wrap and masking tape, the coiled lead reaching the hand held meter. Due to physical constraints the clamp is not centered in the loop.
A nearby steep hill, max about 1 in 3 showed a steady 13.3A over a number of seconds (my max) after a short excursion above 14A.
After the test the clamp zero had drifted to + 0.2A.
The meter claims +/- 0.5% of rdg +/- 2 dgts.
The clamp claims +/-(2% rdg + 2A) over 0 to 600A.
However within the range of normal usage of up to 20A (max of a somewhat better meter) this combination agrees closely with the better meter when it is connected in series with small brushless motors.
Dave.
Attachments
Last edited:
Dave:
I get the idea. Easy to implement and safe. Brilliant. I will try it tomorrow.
Your 13.3A is what I would have expected to measure on mine.
I note your bike goes up 33% hills. The steepest I have tried is 13%, and I am putting in sustained strong effort. I doubt I would manage 20% let alone 33%. My lowest gear is 45 inches. Maybe they really have limited the power on this model in order to get good range.
I get the idea. Easy to implement and safe. Brilliant. I will try it tomorrow.
Your 13.3A is what I would have expected to measure on mine.
I note your bike goes up 33% hills. The steepest I have tried is 13%, and I am putting in sustained strong effort. I doubt I would manage 20% let alone 33%. My lowest gear is 45 inches. Maybe they really have limited the power on this model in order to get good range.
Last edited:
10mph,
The clamp uses a hall sensor that requires to be zeroed before use so I would suggest a best guess of about 13 A because of drift and the compounded inaccuracies due to the specifications of both instruments.
Lacking measurements on this hill I used the gradient function on the link below which may well be optimistic because of the very short test.
The smaller wheels and the Trinitrate might have helped?
Dave
bikehike.co.uk - Course Creator
The clamp uses a hall sensor that requires to be zeroed before use so I would suggest a best guess of about 13 A because of drift and the compounded inaccuracies due to the specifications of both instruments.
Lacking measurements on this hill I used the gradient function on the link below which may well be optimistic because of the very short test.
The smaller wheels and the Trinitrate might have helped?
Dave
bikehike.co.uk - Course Creator
Last edited:
Flecc was right - steady peak power ~400 watts
Following the encouragement from Orraman, I set up to measure the current drawn from the battery in a conventional manner, rather than trying to infer it from my battery voltage drop experiment, reported at the beginning of this thread. NRG is right, there must be some weird non-linearity in the battery's response to pulses.
Orraman advocated a clamp-on ammeter, but the ones in Maplin appear only to measure AC current. Also I would prefer a detachable jaw so I would be able to put the readout on the handle bars rather than down at the crank. The one I found on the web that looked suitable was a Fluke 365 costing some £200. Time I think for a bit of DIY.....
I butchered a shunt from an analogue meter which I had lying around unused. I calibrated it as 3.11 mohm; installed it strapped to the back wheel forks to the rear of the crank, routing the connecting cable through the large gap at the back as Orraman suggested; and inserted it in series with the negative battery lead. The shunt voltage output was then connected to my DVM on the handle bars which was set to the 200mV range. I just have to divide the mV reading by 3.11 to get the current in Amps.
The highest current seen was reasonably steady between 14.1 and 14.7 Amps, when climbing a steep (10%-13%) hill in 1st gear at 7-7.5mph. I obtained similar values, between 14 & 15 Amps, when accelerating in 3rd gear on a more modest slope with the speed just below the power ramp down which starts around 14- 15 mph. This corresponds to around 400 watts sustained power from the battery.
The current is also fairly steady at lower demand values around 10 Amps, but at lower currents than this there are big fluctuations as the meter samples roughly every 0.4 seconds - presumably this is where the controller follows the varying pedal torque as each foot is pushed down. I think I may try experimenting with an RC circuit to smooth these fluctuations at lower currents.
Following the encouragement from Orraman, I set up to measure the current drawn from the battery in a conventional manner, rather than trying to infer it from my battery voltage drop experiment, reported at the beginning of this thread. NRG is right, there must be some weird non-linearity in the battery's response to pulses.
Orraman advocated a clamp-on ammeter, but the ones in Maplin appear only to measure AC current. Also I would prefer a detachable jaw so I would be able to put the readout on the handle bars rather than down at the crank. The one I found on the web that looked suitable was a Fluke 365 costing some £200. Time I think for a bit of DIY.....
I butchered a shunt from an analogue meter which I had lying around unused. I calibrated it as 3.11 mohm; installed it strapped to the back wheel forks to the rear of the crank, routing the connecting cable through the large gap at the back as Orraman suggested; and inserted it in series with the negative battery lead. The shunt voltage output was then connected to my DVM on the handle bars which was set to the 200mV range. I just have to divide the mV reading by 3.11 to get the current in Amps.
The highest current seen was reasonably steady between 14.1 and 14.7 Amps, when climbing a steep (10%-13%) hill in 1st gear at 7-7.5mph. I obtained similar values, between 14 & 15 Amps, when accelerating in 3rd gear on a more modest slope with the speed just below the power ramp down which starts around 14- 15 mph. This corresponds to around 400 watts sustained power from the battery.
The current is also fairly steady at lower demand values around 10 Amps, but at lower currents than this there are big fluctuations as the meter samples roughly every 0.4 seconds - presumably this is where the controller follows the varying pedal torque as each foot is pushed down. I think I may try experimenting with an RC circuit to smooth these fluctuations at lower currents.
Here is a reasonably priced clamp meter that can measure dc amps:
TENMA|72-7224|CLAMP METER, WITH FREQUENCY | CPC
TENMA|72-7224|CLAMP METER, WITH FREQUENCY | CPC
10mph,
Your good news is that the newer motors do appear to give more assistance than mine, an advantage that I might well need in the near future.
The clamp probe that I have fastened to the Gitane is much the same as the system you are using. The probe outputs 1mv for each amp in the cable under test within the clamp and this is displayed on the 200mv range on the remote digital multimeter as amps and tenths.
Dave
A neat illustration of the complete system; Fluke 87V/i410 Combo Kit of meter and probe for £470
Fluke 87v-i410 Combo Kit for industrial Applications. Buy online or call
A newer instrument for use with a multimeter, datalogger or 'scope; Current Probe with 3 DC ranges of 0 to 4, 40, and 200A £29
Current Probe with Mini Resolution DC/AC 1mA,23mm Jaw | eBay
Your good news is that the newer motors do appear to give more assistance than mine, an advantage that I might well need in the near future.
The clamp probe that I have fastened to the Gitane is much the same as the system you are using. The probe outputs 1mv for each amp in the cable under test within the clamp and this is displayed on the 200mv range on the remote digital multimeter as amps and tenths.
Dave
A neat illustration of the complete system; Fluke 87V/i410 Combo Kit of meter and probe for £470
Fluke 87v-i410 Combo Kit for industrial Applications. Buy online or call
A newer instrument for use with a multimeter, datalogger or 'scope; Current Probe with 3 DC ranges of 0 to 4, 40, and 200A £29
Current Probe with Mini Resolution DC/AC 1mA,23mm Jaw | eBay
NRG: I am generally a 10 mph rider. Well perhaps 10 to 14 to be exact. I only ride faster for test or experimental purposes. The bike has more than enough power for me since I generally use it on minimum assist, except for the occasional steep hill, strong headwind, or extreme sloth/tiredness.
I am curious as to what happens to the maximum power on very long hills. Does it stay at 400 watts from the battery, which presumably is significantly more than 250 watts into the road? Or does the controller reduce power as the motor gets too hot? I think my longest hill climb so far at something like maximum power has been about 2 minutes.
Orraman: Thanks for that link to the current probe. That is rather closer to my price. However, I think I am now completely happy with my shunt which cost me nothing and is smaller.
I am curious as to what happens to the maximum power on very long hills. Does it stay at 400 watts from the battery, which presumably is significantly more than 250 watts into the road? Or does the controller reduce power as the motor gets too hot? I think my longest hill climb so far at something like maximum power has been about 2 minutes.
Orraman: Thanks for that link to the current probe. That is rather closer to my price. However, I think I am now completely happy with my shunt which cost me nothing and is smaller.
Attachments
Long ago I devised a near continuous moderate climb in the North Downs of almost 9 miles, it being in two parts with a short but very steep fast downhill that linked the two climbs within a few seconds. On the Agattu, my subjective observation is that the power remained roughly constant over the 45 minutes of the climb.
Static torque and power
This post has turned to be rather long I am afraid. In summary, I have measured the stalled torque provided by the Panasonic motor when pressing the pedals with the bike held stationary. When I am in first gear the force at the wheel trying to drive the bike forward increases from around 9 lb with just my force on the pedal of 32 lb and the motor off to somewhere around 24 to 30 lb with maximum motor assist. This helps nicely on hill starts, or for very rapid acceleration from traffic lights. The power taken from the battery is 164 watts, and after about a minute stalled in this way the overheating warning starts flashing.
My bike has a first gear of 44 inches (distance moved forward for 2 radians revolution of the pedals, ie for 1 pedal rev divided by pi). So the gearing on my bike is equivalent to a rear wheel of 22 inches radius coupled to the pedal crank with an overall 1:1 gearing ratio. My pedal operates on a radius of 6.9 inches. Ignoring the frictional losses in the chain, gears and bearings, this ratio means that a force of 32 lbs on the pedal will yield a drive force from the rear wheel onto the road of 10 lbs.
I was able to get a rough experimental confirmation of this calculated drive force by setting up the bike with a rope attached to the rear tyre pulling a spring scale:
I could read the spring scale to an accuracy of about +/- 2 lb. this is nicely close to the value of 10lb which I calculated above from the gearing ratio.
With the bag weighing 32 lb on one pedal crank I measured 9 +/- 2 lb at the rear wheel.
46 lb on the pedal gave 10.5 +/- 2 lb at the rear wheel.
It is interesting to note that with the pedal force of 46 lbs I would just be able to hold a bike + rider weight of 210 lb stationary on a hill with gradient 10.5/210 = 5% ( 1 in 20) I would need more pedal pressure to actually start riding up such a hill. By standing on one pedal I would be able to increase the force by about 4 times and so start riding up the hill (alternatively just hold stationary on a 20% hill)
Next, I removed the pedal weight and turned the motor power on so that the torque sensor would calibrate. I then replaced the 32 lb weight on the pedal and obtained the following readings on the spring scale:
Assist level 1 (minimum): 17-19 lb (less 9 lb from the pedal weight = 8-10 lb from the motor) current drawn by motor 1.4 to 1.6 Amp = 38 Watt
Assist level 2: 23-25 lb (less 9 lb from the pedal weight = 14-16 lb from the motor) current drawn by motor 2.8 to 3.4 Amp = 78 Watt
Assist level 3 (maximum): 24-30 lb (less 9 lb from the pedal weight = 15-21 lb from the motor) current drawn by motor 6.5 Amp = 164 Watt
I am not sure the measurement was reliable when in assist level 3, for I noticed that after less than a minute stalled at 164 watts from the battery the mode 3 light started flashing, according to the Kalkhoff manual this indicates motor over heating. I therefore discontinued the measurements. It is be be expected that a motor rated at a continuous OUTPUT of 250 watts might overheat if operated stalled. In this situation all the 164 watts from the battery would be being dissipated in the motor windings and the controller.
During these static measurements I noticed that every 18 seconds the battery current dropped momentarily and the forced was removed for a fraction of a second. The current and force was then resumed at a slightly different value. This brief interruption every 18 seconds is readily observed on my bike when standing stationary with weight on the pedal, eg at lights with the bike held by the brakes ready to release and accelerate when the light changes. It causes a brief slight jerk on the pedal.
I also tried different weights on the pedal:
7 lb or 14 lb were not sufficient to cause the motor's torque sensor to cause the motor to draw current. However, I noticed that with the 14 lb weight hanging from the pedal, if I added a further force by hand, the motor would come on, and remain on at constant current with just the 14 lb weight after my hand pressure was removed.
The accuracy of reading the wheel force with the spring scale was limited in these experiments. One could set up with a better scale to make up more accurate measurements, but I am not sure, that this would be particularly useful. One can already see from the data that in maximum assist the effect of a pedal force of 32 lbs is roughly tripled from about 9 lb to 24 - 30 lb, so one should be able to easily start this bike on hills steeper 10%, as I have observed in practice.
Measurements when moving
I am more interested in making measurements to confirm the assistance provided when moving. This is not so easy. Flecc has described the the use of hills of known slopes, and that is what I want to try with better instrumentation to measure the battery current.
I have managed to smooth the fluctuating current readings one gets during pedalling at modest force by using a 5 MOhm resistor in series with the 10 MOhm input impedance my voltmeter in conjunction with a 0.5 microF capacitor across the meter. Unfortunately this setup causes varying offsets up to 1 mV. I have not yet tracked down the cause - in the resin dipped ceramic capacitor or the carbon series resistor?
I would like to have a data logger. There are nice little USB loggers available but these only take 1 sample per second. I want at least 10 samples per second, so I can start to record and analyse the pedal stroke. I could get a Cycle Analyst with its logger for some £200 which would do the job (and a lot more which I don't absolutely need). The Eagle Tree data logger device sold to RC enthusiasts is quite a bit cheaper but I am not sure whether the 100 Amp current measurement capability will have enough resolution for accurate measurement of my currents of just a few Amps. The Eagle Tree has to be inserted in series between the battery and the motor, and it will not quite as simple as the current shunt which I have already manged to install.
Do any of the people on here, who have electronics experience, have any advice on a suitable (cheap) datalogger to record up to say 25 amps at 10 samples/sec.?
This post has turned to be rather long I am afraid. In summary, I have measured the stalled torque provided by the Panasonic motor when pressing the pedals with the bike held stationary. When I am in first gear the force at the wheel trying to drive the bike forward increases from around 9 lb with just my force on the pedal of 32 lb and the motor off to somewhere around 24 to 30 lb with maximum motor assist. This helps nicely on hill starts, or for very rapid acceleration from traffic lights. The power taken from the battery is 164 watts, and after about a minute stalled in this way the overheating warning starts flashing.
My bike has a first gear of 44 inches (distance moved forward for 2 radians revolution of the pedals, ie for 1 pedal rev divided by pi). So the gearing on my bike is equivalent to a rear wheel of 22 inches radius coupled to the pedal crank with an overall 1:1 gearing ratio. My pedal operates on a radius of 6.9 inches. Ignoring the frictional losses in the chain, gears and bearings, this ratio means that a force of 32 lbs on the pedal will yield a drive force from the rear wheel onto the road of 10 lbs.
I was able to get a rough experimental confirmation of this calculated drive force by setting up the bike with a rope attached to the rear tyre pulling a spring scale:
I could read the spring scale to an accuracy of about +/- 2 lb. this is nicely close to the value of 10lb which I calculated above from the gearing ratio.
With the bag weighing 32 lb on one pedal crank I measured 9 +/- 2 lb at the rear wheel.
46 lb on the pedal gave 10.5 +/- 2 lb at the rear wheel.
It is interesting to note that with the pedal force of 46 lbs I would just be able to hold a bike + rider weight of 210 lb stationary on a hill with gradient 10.5/210 = 5% ( 1 in 20) I would need more pedal pressure to actually start riding up such a hill. By standing on one pedal I would be able to increase the force by about 4 times and so start riding up the hill (alternatively just hold stationary on a 20% hill)
Next, I removed the pedal weight and turned the motor power on so that the torque sensor would calibrate. I then replaced the 32 lb weight on the pedal and obtained the following readings on the spring scale:
Assist level 1 (minimum): 17-19 lb (less 9 lb from the pedal weight = 8-10 lb from the motor) current drawn by motor 1.4 to 1.6 Amp = 38 Watt
Assist level 2: 23-25 lb (less 9 lb from the pedal weight = 14-16 lb from the motor) current drawn by motor 2.8 to 3.4 Amp = 78 Watt
Assist level 3 (maximum): 24-30 lb (less 9 lb from the pedal weight = 15-21 lb from the motor) current drawn by motor 6.5 Amp = 164 Watt
I am not sure the measurement was reliable when in assist level 3, for I noticed that after less than a minute stalled at 164 watts from the battery the mode 3 light started flashing, according to the Kalkhoff manual this indicates motor over heating. I therefore discontinued the measurements. It is be be expected that a motor rated at a continuous OUTPUT of 250 watts might overheat if operated stalled. In this situation all the 164 watts from the battery would be being dissipated in the motor windings and the controller.
During these static measurements I noticed that every 18 seconds the battery current dropped momentarily and the forced was removed for a fraction of a second. The current and force was then resumed at a slightly different value. This brief interruption every 18 seconds is readily observed on my bike when standing stationary with weight on the pedal, eg at lights with the bike held by the brakes ready to release and accelerate when the light changes. It causes a brief slight jerk on the pedal.
I also tried different weights on the pedal:
7 lb or 14 lb were not sufficient to cause the motor's torque sensor to cause the motor to draw current. However, I noticed that with the 14 lb weight hanging from the pedal, if I added a further force by hand, the motor would come on, and remain on at constant current with just the 14 lb weight after my hand pressure was removed.
The accuracy of reading the wheel force with the spring scale was limited in these experiments. One could set up with a better scale to make up more accurate measurements, but I am not sure, that this would be particularly useful. One can already see from the data that in maximum assist the effect of a pedal force of 32 lbs is roughly tripled from about 9 lb to 24 - 30 lb, so one should be able to easily start this bike on hills steeper 10%, as I have observed in practice.
Measurements when moving
I am more interested in making measurements to confirm the assistance provided when moving. This is not so easy. Flecc has described the the use of hills of known slopes, and that is what I want to try with better instrumentation to measure the battery current.
I have managed to smooth the fluctuating current readings one gets during pedalling at modest force by using a 5 MOhm resistor in series with the 10 MOhm input impedance my voltmeter in conjunction with a 0.5 microF capacitor across the meter. Unfortunately this setup causes varying offsets up to 1 mV. I have not yet tracked down the cause - in the resin dipped ceramic capacitor or the carbon series resistor?
I would like to have a data logger. There are nice little USB loggers available but these only take 1 sample per second. I want at least 10 samples per second, so I can start to record and analyse the pedal stroke. I could get a Cycle Analyst with its logger for some £200 which would do the job (and a lot more which I don't absolutely need). The Eagle Tree data logger device sold to RC enthusiasts is quite a bit cheaper but I am not sure whether the 100 Amp current measurement capability will have enough resolution for accurate measurement of my currents of just a few Amps. The Eagle Tree has to be inserted in series between the battery and the motor, and it will not quite as simple as the current shunt which I have already manged to install.
Do any of the people on here, who have electronics experience, have any advice on a suitable (cheap) datalogger to record up to say 25 amps at 10 samples/sec.?
Attachments
Does your data logger have to be portable or are you just talking about static tests?
There are various ways of using the sound card in record mode on a pc/laptop as a simple data logger. I have used this sort of scheme as an event logger in the past and it works fine.
Here is on of many google hits
http://hardandsoftware.mvps.org/sound_card.htm
There are various ways of using the sound card in record mode on a pc/laptop as a simple data logger. I have used this sort of scheme as an event logger in the past and it works fine.
Here is on of many google hits
http://hardandsoftware.mvps.org/sound_card.htm
Last edited:
mike_j : Thanks for a very interesting suggestion. I want to do mobile datalogging, and I have an Acer One note book which I could put in the pannier bag. Also I see it has an input connection for an external mic.
Unfortunately your linked article rather puts a damper on it,
Unfortunately your linked article rather puts a damper on it,
I need DC up to, say, 10Hz.
Breadboarding 10mph's Ammeter
Thank you very much for the PM and various tips in followup messages. I am going to summarise how far I have got in building a datalogging ammeter for use with my Agattu.
My ammeter uses the PICAXE microcontroller, suggested by Orraman, to convert the output from a Hall effect current sensor to a digital signal which can be displayed on an LCD to be mounted on the handlebars, and can also recorded at regular intervals for later download to my PC via a plug-in USB cable.
Here is a photo of my bread board, When complete it will fit in the black box, which will be attached to my handlebars, and the Hall sensor board will be mounted at the bike's battery terminal to sense the current drawn by the bike's motor.
The Picaxe microcontoller is a versatile microprocessor, and is easy to programme on my PC. The program is then downloaded to the chip through a USB cable. The USB cable can then be disconnected until it is required again to readout any data which has been collected, and processed by the program and then stored within the Picaxe chip.
It will be possible to store current readings at intervals of about 100ms and so record the varying current drawn as each stroke on the Agattu pedals is sensed by its torque sensor.
Alternatively, it will be possible to write a program which will average the current over a few seconds, and display it on the LCD on the handle bars, as well as storing the varying currents during a long ride for later download.
STATUS
The Picaxe chip and the LCD work very well. For the picture I wrote a program to display a simple message - there is ample room on the 2-line to display in real-time various data readouts. As well as the input from the Hall ammeter, the Picaxe has several other input channels with ADCs which could in principle be used to collect data from other sensors.
I have not got the Hall Effect sensor working yet. I bought an Allegro ACS712ELECTR-20A. I had not realised quite how small the Surface Mount Package would be. Its pin pitch is 1.27mm - just half the pitch of the Vero-type strip board which I have. I devised a way of mounting it at an angle on the strip board, shown in the enlarged section. My soldering tools and technique need improving for such small chips - this is the first device of this size which I have ever tried soldering.
I got all the connections soldered, but the device is not working - the output should be sitting at half way between the 5v and 0V for no current flowing through the heavy current connectors - but it seems to be sitting at 5V, furthermore the Hall device is drawing no current from the 5V supply - it should be drawing about 10 mA.
Perhaps I have somehow wrecked the device. I have now ordered a tiny adapter board which will properly mount the SOIC style package and convert to an 8 Pin DIL which will match the holes on the strip board. I may have to order some more Hall sensors at £4.50 each.
Dave,
Thank you very much for the PM and various tips in followup messages. I am going to summarise how far I have got in building a datalogging ammeter for use with my Agattu.
My ammeter uses the PICAXE microcontroller, suggested by Orraman, to convert the output from a Hall effect current sensor to a digital signal which can be displayed on an LCD to be mounted on the handlebars, and can also recorded at regular intervals for later download to my PC via a plug-in USB cable.
Here is a photo of my bread board, When complete it will fit in the black box, which will be attached to my handlebars, and the Hall sensor board will be mounted at the bike's battery terminal to sense the current drawn by the bike's motor.
The Picaxe microcontoller is a versatile microprocessor, and is easy to programme on my PC. The program is then downloaded to the chip through a USB cable. The USB cable can then be disconnected until it is required again to readout any data which has been collected, and processed by the program and then stored within the Picaxe chip.
It will be possible to store current readings at intervals of about 100ms and so record the varying current drawn as each stroke on the Agattu pedals is sensed by its torque sensor.
Alternatively, it will be possible to write a program which will average the current over a few seconds, and display it on the LCD on the handle bars, as well as storing the varying currents during a long ride for later download.
STATUS
The Picaxe chip and the LCD work very well. For the picture I wrote a program to display a simple message - there is ample room on the 2-line to display in real-time various data readouts. As well as the input from the Hall ammeter, the Picaxe has several other input channels with ADCs which could in principle be used to collect data from other sensors.
I have not got the Hall Effect sensor working yet. I bought an Allegro ACS712ELECTR-20A. I had not realised quite how small the Surface Mount Package would be. Its pin pitch is 1.27mm - just half the pitch of the Vero-type strip board which I have. I devised a way of mounting it at an angle on the strip board, shown in the enlarged section. My soldering tools and technique need improving for such small chips - this is the first device of this size which I have ever tried soldering.
I got all the connections soldered, but the device is not working - the output should be sitting at half way between the 5v and 0V for no current flowing through the heavy current connectors - but it seems to be sitting at 5V, furthermore the Hall device is drawing no current from the 5V supply - it should be drawing about 10 mA.
Perhaps I have somehow wrecked the device. I have now ordered a tiny adapter board which will properly mount the SOIC style package and convert to an 8 Pin DIL which will match the holes on the strip board. I may have to order some more Hall sensors at £4.50 each.
Attachments
Hi 10mph, I have been reading your posts regarding the Panasonic unit with interest. The one thing I am curious about which is whether you have established up to date measurements for when the various power down steps kick in and at what motor speed the power cuts off completely?
A timely question, since I have just recovered from a run to my nearby hill, testing my breadboard of 10MPH's AMMETER V0.1!
I will report later on how the version V0.1 breadboard worked on the bike. But for now just two pics. The first showing the LCD readout and the Picaxe board location on my handlebars:
and the scond shows the location of the hall sensor in the extended battery lead:
While riding back home into slight wind on the flat or slight downhill gradient (<1%), I watched the current consumption (averaged from 100 samples taken over 2 seconds and displayed on the LCD on my handlebars). I noticed how the current cut off as my speed varied from around 15 mph to 18 mph.
Max assist - top gear - gear 3rd gear on my 3 speed Agattu.
15 mph max current of 8 A available
17mph about 2 to 3 A available
17.5 mph some current available
17.9 mph no current available
I put the cut off at just above 17.5 mph which is 28.1 kph. This is 12% above the Euro limit of 25kph.
This agrees with my earlier observations where I saw the ramp down start somewhere above 14 mph. It is also in line with the typical type of ramp down shown in the latest Kalkhoff manual for an 8 speed model:
This is all with my 3 speed Agattu in top gear; the gearing is shown here:
The tyres are as supplied by 50 cycles fairly fat Continental Townrides.
Attachments
Related Articles
-
MTF Enterprises announces acquisition of EMU Electric Bikes- Started by: Pedelecs
-
Wisper 806T folding bike wins Which? ‘Best Buy’- Started by: Pedelecs
-
Sustrans calls for protected cycle lanes- Started by: Pedelecs
-
Amazon launch their first UK e-cargo micromobility hub- Started by: Pedelecs
