Amp Sensor for SSG etc.

[BobZilla]

Hey Yaro,
I just got my meter and I am just goofing around with it. Pretty neat but I didn't realize I didn't get any gain control on the low end model, that kind of sucks lol.

So I'm actually not sure I follow what your doing here but I did notice something when playing with my meter. These shifts you see in the pattern are a result of the sample rate and the rise/fall of the wave. It is sort of like a phase shift which then stretches the chart while transitioning.

I used to see a similar effect with my digital multi meter when running charts, it only sampled 1 per second so depending on where the pulse was at you could get higher or lower peaks, it is relation of the sample rate. Over time you would see the pattern develop

[Yaro1776]

Hey BobZ,

Appreciate your reply and yes the $30 meter and software does have its limitations. Overall it is a good and useful tool kit for the experimenter - what I like about it is the fact that for low bucks you can track battery charge and load discharge. Simple to reduce the sample rate down to the lowest levels and set the data function to average - the y axis can be scaled to whatever you need. The data sampling in the average mode algorithm takes all the data within the period and produces a screen data point. If one wants to discharge a battery under load, setting the sampling rate to 0.05 (two channels) will yield close to 10 hours of averaged voltages producing a nice on screen graph curve. Same would apply to the charge cycle. Of course, you would need a simple voltage divider for this since +/- 10 volts DC are the limits of the DI 145.

Combining the DI 145 or similar with easy software gives a great tool set for the beginning experimenter. So voltage and amp measurement along with RPM and you are good to go with the SSG.

With respect to the direction of this thread - well a bit uncertain on that, but I will finish up one more post and take it from there. When the RPM and Magnet combo are combined they yield board firing HZ - then the transition zones become interesting.

Stay warm, balmy -21F this morning,
Yaro

[BobZilla]

I totally agree Yaro, it's still a great meter to have on hand. I had watched the video's on the site and saw the gain control where they were scoping down the single wave and thought wow, then found out I won't be able to do it so a little disappointed is all. It can still be used for a lot of useful analysis though.

There are very interesting things that are real at the transitions, no argument there. All I was trying to say was be careful in interpreting what that meter shows. You can simultae the graph behavior of the real shift prctically anywhere just by touching the wheel for a second though because the sample rate is fixed but if you change the rpm then you get a phase shift. Your doing great work so please don't think I am trying to muck things up.

Another way you might try looking into the shift is what I used to do a lot when I first started noticing it. If you take a fully charged front and tune just barley into one spike, i mean right on the edge of the transition it will naturally fall back to two spikes after a short time because the volts drop into normal ranges. You could look at it from reverse. When a machine drifts back from 1 to 2 or 2 to 3 there is usually a huge boost for a few seconds while the wheel growls and settles in. I know you know how a machine behaves already but I mention it in case you had not though about looking in reverse at the transition.

[Yaro1776]

SSG DAQ Test 3

Presented in the following is the last series of graphs for the SSG transition zone tests. Test 3 used the the high friction wheel at the same rpm of 257, but it required a much smaller gap, down to 0.125". The max RPM in this config was about 261 RPM with a 12 ohm resistor. By tweaking the resistor value lower the rpm was stabilized in the 256 to 257 RPM area. Very sensitive!

The graphs for this config are nearly identical in pulse behavior to the two prior tests. The physical amp values were confirmed with a clamp on meter and direct voltage read form the TET amp sensor - so the data is valid.

Graph 1 depicts the 257 RPM transition zone behavior at a sampling rate of 3s/sec - note how the pulse width and shape changes slowly.

Graph 2 and the other following graphs depict a much closer look at the amp pulses with a sampling rate of 60s/sec. It is interesting to watch the morphing of the pulses into a relatively steady state.

Graph 5 depicts the slow progression into a steady state of amp values for approximately 11.5 sec. During this progression from graph 1-6 nothing on the SSG is altered. The transition happens without any outside push or interference and will oscillate back and forth in pulse behavior and width.

So from the results of this test one can conclude that for a given 8 transistor board the SSG will produce a number of definite transition zones that can be correlated to rpm. For the 23" wheels and 21 magnets this will relate to a frequency dictated by the various rpm's as demonstrated by tests 1 through 3. The low and high friction wheels will produce the same transition behavior at the same RPM or frequency.

One can only speculate as to how this impacts the overall performance of the SSG - further tests are required to gain this information. Certainly the frequency related behavior is very interesting and it does raise a number of questions.

Comments and opinions related to the above testing are always welcomed! Thank you for your attention.

So for the time being the initial transition zones and behavior have been identified - What is Next? Hmmm...
Yaro
SSG3 Amps 2016.pdf

[Yaro1776]

Another way you might try looking into the shift is what I used to do a lot when I first started noticing it. If you take a fully charged front and tune just barley into one spike, i mean right on the edge of the transition it will naturally fall back to two spikes after a short time because the volts drop into normal ranges. You could look at it from reverse. When a machine drifts back from 1 to 2 or 2 to 3 there is usually a huge boost for a few seconds while the wheel growls and settles in. I know you know how a machine behaves already but I mention it in case you had not though about looking in reverse at the transition.

BobZ,

Appreciate your thoughtful reply and insights with respect to the transition zone behavior. Your LED method really introduced me to the spike transitions awhile ago and it has intrigued me since then. Apologize for not mentioning this fact sooner and giving credit where it is due - I did not connect the dots on this until your post.

I use 0.025" plastic shim stock recycled from a vendor's plastic calendar for setting up the wheel gaps - found this perfect for slowing down the the wheel by letting one flap against the magnets...

Yaro

[BobZilla]

It's all good Yaro,

I was going to refrain from further comment because I did not want to start pushing my opinions all over your post, I respect your work.

With that said I have something for you that you may not know about in respect to the transitions as we are calling them. Did you know that the back end impedance plays a very significant role? It is not always tied to RPM as it appears. I will tell a brief story of how I discovered what I am saying.

I have gone through many phases of experimentation and for a long time I was really into cap dumping. I discovered that if I set my resistance just right and set my dump voltage/time period just right I could trigger a transition intentionally over and over. It has been quite a long time since I ran this way so I have to go from what I remember on some of it but basically I had a 60k uf cap and I would charge up to about 35v. I would tune the machine just a little past the transition from 2 to 1 spikes. After a dump it would fall into 2 spikes and as the impedance grew with the voltage in the cap Iit would shift into one spike. I looked at it as a clever way to get a boost into the cap each dump because when it down shifts a lot of power comes through. Now along with this yes the RPM changes slightly because just as when we get to the top of a battery charge we speed up, the same will happen on a cap if your tuned right. Anyway it is something for you to think about as your investigating the transitions, impedance is a culprit.

Happy hunting

[James McDonald]

So how much is the TET sensor, i will order 1 or 2.

can i adjust for the 1.5A, or is that a factory pre set...?

Hi RS_ --

Sorry for the delay on answering your post but I had to confirm with the factory technical support if the
current sensor would measure down low enough to cover the 20 mA up to 100 ma range. As it stands now the current
sensor has a low end limitation even if I reconfigure it from 6 Amps maximum current down to 1.5 Amps maximum
current. Below is the statement from the technical applications support personal. The cost of the current sensor I
have already emailed that to you from my email address james.mcdonald@teslaenergytech.com . I am going to
ask them if they have a current sensor that will work that low. If they do not sell one I am sure no one else has
one since I went through two other current sensors in my R & D effort to decide on the current sensor I am using
now.

Technical Application Support Email to me is below.

The Sparkfun board is using an Allegro ACS712. The ACS712 is a Hall cell based, open-loop current sensor (5A).

Our current sensor is a magnetic probe based and is a closed-loop sensor (6A).

They are technically both current sensors measuring similar magnitude currents. In reality they are an apples to oranges comparisons.

I do not typically criticize a competitors product, so I will not go into details of the ACS712.

Our current sensors are more accurate and have better bandwidth and response time (much, much, much better).

A clamp-on current probe is not accurate enough to check our sensor with, however it is accurate enough to
check an ACS712 current sensor. The ACS712 current sensor is an open-loop device which has more sources of
uncertainty than a closed-loop.


Our current sensors are used where accuracy and speed are needed. Grid connect solar inverters and high performance
motor drives are good examples.

Your design can be modified to measure 1.5A by routing the four sensor bus bars in series to loop the current four times
in the same direction through the sensor. This magnetic multiplication makes 1.5 Amps look like 6 Amps to the sensor.

When measuring below 0.5% of the nominal rating of the sensor, in this case perhaps less than 30 mA, the sensors Class
B amplifier output and the coercivity of the sensors magnetic core, can cause zero crossing non-linearities to manifest.
The exact point is not precisely defined and magnitude of the effects are also not defined. The magnitude of the non-linearities
are smaller than the defined accuracy of the sensor (0.7% or 42mA).

Hope this helps in your decision on what to look for in a current sensor.

Thanks,-- James

[RS_]

Hi James,

sorry i took so long to respond, as we were out of town last week....

So it seams that measuring things in the 30ma range is tough to do.....

I have been using some cheap china 2A current meters, that can measure the 10 to 40ma range very well. If i could get the out put from that meter into a ardunio, so that i can do the V x A =W on these small energizers, that would be great......

other wise, some other better sensor out there would be nice for these small measurements

[James McDonald]

Hi James,

sorry i took so long to respond, as we were out of town last week....

So it seams that measuring things in the 30ma range is tough to do.....

I have been using some cheap china 2A current meters, that can measure the 10 to 40ma range very well. If i could get the out put from that meter into a ardunio, so that i can do the V x A =W on these small energizers, that would be great......

other wise, some other better sensor out there would be nice for these small measurements

Hi RS_ --

This problem in measuring very low currents would require a special shunt resistor with an OpAmp circuit to amplify the voltage up
high enough so that the analog input on an Ardunio Microcontroller would be able to measure a voltage up above the noise. Then you
would have a special formula in the Ardunio program to convert it to a current measurement. The formula would subtract out the noise
and the tolerances of all the parts being used to have a left over pure current value.

-- James

[Yaro1776]

SSG Transition Zones Charge Tests

To complete the cycle of this testing in the SSG transition zones of 227 and 257 RPM a set of tests were run to document the charging behavior in these zones. The intent here is to evaluate if there is any noticeable charging benefit form operating the SSG in these areas compared to wheel rpm's outside of the Zone.

The tests were run over a five day period using the standard 8 transistor board and the high friction wheel. This wheel is less sensitive to the usual SSG amp fluctuations and speed changes than low friction wheel.

The attached Excel .pdf presents the data from 6 individual tests. The amperage data was taken from a DAQ graph during the charge cycle (TET amp sensors) and the voltages (BK 389) were noted in written form during the duration of each test. The RPM's were taken from the digital read of the Bell 100 bicycle speedometer. The Gap measurement for each test was taken from a marked location on the wheel for consistency. Plastic shims of a given thickness were used to raise or lower the coil to the specified gap. The discharge load was a NAPA 9 watt rated auto marker bulb 0.67 Ah as measured by the BK and the DAQ graphs.

The Charge battery was charged for a given interval (1.4 hrs) and then rested for about 15 minutes and then loaded with the 9 watt bulb until the voltage dropped to 12.16 volts. Test 1 is the exception with a 12.20 volt discharge stop point. The Primary battery was charged with 10A12 charger.

The results of the test do not show any real improvement in charging efficiency in the Transition zones. One should take note of the D/C ratio (see test sheet) for the Charge battery that was used for these test runs. This ratio depicts the Ah taken from the Charge battery vs the input from the SSG. Note that this process is fairly consistent and that the conversion of battery potential for this particular battery is about 87 percent. This information will be useful when the Cap Dump circuit is tested at a later future date.

The data from test 6 will be rerun - everything during this test was perfect, with the exception of the final results which are lower than anticipated.

Done with this testing for the time being. Very comfortable with the instrumentation and will upgrade to the next level DI 155 from DATAQ for the better features.

Thanks for all the attention on this project and the very relevant suggestions.

Yaro
SSG Transition Zone2016 Test Data.pdf