Amp Sensor for SSG etc.

[Yaro1776]

Been on hiatus since the conference and back at it again after the summer rejuve. Researching amp sensors for the past while that would fit the SSG application and having tried the DVM method, analog meter and clamp on meter over the course of the past two years figured that there must be something better. The something better became a reality after mentioning the issue to a fellow forum participant and expressing the desire to log and display all the SSG parameters on a computer screen using a micro Arduino or RasP.

Being a bit anal about gathering accurate data during any SSG testing can be daunting and a multitasking challenge. Taking voltage reads is straightforward with almost any meter , but the amp reads are a challenge. Normally the amperages in an operating SSG for the Charge and Primary bounce over a range of values forcing the experimenter to estimate an average value and then record correctly. The recording aspect here becomes a challenge when many events are happening simultaneously.

Recently I obtained several Amp sensors from J McDonald james.mcdonald@teslaenergytech.com and put two units through the paces while checking the accuracy against my other trusty meters. The new sensors are externally powered (8-36v) and give a 0-2.5 dc volt output which through a simple calculation yields the amperage. The sensors have 0 to 6 amp DC capability. I found that new sensors are extremely responsive and accurate; when using my older BK DVM the reads on the meter's Bar graph and digital output reflected the normal amp output wave. So these reads were also in close agreement with my clamp on meter. The amp read voltage range of the sensor will work with the analog side of a micro. In addition, when a newer type of DVM with RMS capability (ExTech 430 or similar) is used a steady and stable voltage output is realized without any data bounce. Perfect!

The next steps here are to integrate all into a micro and make the SSG testing and data gathering process more accurate and useful. The ability to present the data on screen and real time is an invaluable tool for tuning and learning the performance characteristics of various SSG variants.

Looking forward to presenting progress in this effort at my usual turtle pace. The micro effort will be of interest since this is the first shot at playing with these little technical gems. I am certain that there will be a number of interesting and engaging speed bumps that will require outside help.

Forward,
Yaro

[Yaro1776]

Have not forgotten this thread - for the past week been having some difficulty in pasting Word and attaching .PDF's.

Yaro

[Yaro1776]

After an extensive period of research into the micro's it became evident that there were too many speed bumps inherent to the Arduino and RasP for my liking and in the end they would not accomplish my goal - Collect real time data,record and display on my computer screen. Yes it is doable with the micro's, but the learning curve with my limited capabilities is too extensive for the purpose.

Having a bit of experience in the process control field, instrumentation and associated software a search was initiated to find a device and software that would do the job without reinventing the wheel. The available selection of hardware and software is mind boggling - but utilizing financial constraint and a bit of research, a DAQ 145 device and software were located at DATAQ.com. The device featured 4 analog voltage input channels with +/- 10 volts along with the oscilloscope 4 channel software and data recording capability. Essentially a starter device with some software limitations for $29.99 + shipping (~$15). A no Brainer for trial purposes.

Upon receiving the device in three days, the unit was temporarily wired to The TET amp sensors and then came the software setup process. Still using XP on my laptop the software installation was relatively straight forward. It took a few hours of play time to figure out and configure the two channels that were to be used for the amp sensors and to create the appropriately scaled EU (engineering units) for the amps. Once this was completed the plain vanilla SSG was fired up in a stable configuration with a wide gap to minimize overall sensitivity to changes.

[Yaro1776]

Continuation of Previous Post

Very pleasantly surprised by the initial results, actually I was blown away by the screen data displays and capabilities. The capabilities allow one to change the data sampling rate from a low value of under a sec to 240 Hz depending on the number of channels in use. For two channels 120Hz is possible. This now became very interesting in that the nominal firing rate for the SSG varies with the RPM - so a 21 magnet wheel @ 257 RPM has a nominal frequency of about 90Hz. Varying the sampling rate (while using the averaging function) shows the amplitude of of the amperage spikes going to the charge battery from the SSG. One could see that the peak draw from the Primary battery was around 4 amps, while the Charge battery was being hit with about 3 amp spikes at the frequency.

Furthermore, if the acceleration of the wheel is slow one can note a number of areas of instability or transition on the generated screen graph. These transition areas are visible spikes and appear to be frequency related as approximate multiples of about 10 RPM - exception is start up zone of 110 RPM. So there appear to be Major Zones of instability or transition at 167, 227, 257 and 287 RPM. In these areas the displayed graph spikes undergo some changes. These will be more clearly displayed in subsequent posts.

As a proof of concept attached are two screen shot sample .PDFs at a sampling rate of 60s/sec and 3s/sec.

Yaro
SSG1 Amps 2016.pdf

[BobZilla]

Heck yea Yaro, thanks for the tip on that meter. I just ordered one for myself.

[Kamen]

Yaro,

On the top chart what is the violet (light blue) color representing? Are those the current spikes? Where the dark blue is the average?

You gave me a good idea to track the current with my Dataq DI-155 as well. I use that model because it can measure up to 50V so it can track both batteries.

Kamen

[Yaro1776]

DAQ device and software

BobZ and Kamen,

Thanks for the comments. I chose the DATAQ 145 as a trial of the the device and electronics - this worked out well. My first choice was the DATAQ 155 because of the additional inputs and capabilities, but needed to proof this first. I will upgrade to a higher model and software in the near future since I do want to display and record the data real time in Excel.

Anyway, I am using James MacDonald's TET amp sensors (see .pdf attachment) which have performed very well with the SSG and the DAQ parts. The sample Oscope graphs were setup with 2 channels on the device and having the capability of pushing the sampling rate to 120 Hz if desired. This whole combo is excellent for the SSG work - though the upgraded device and price expands the overall capabilities and sampling rate choices. For $30 I can't complain!

The first Oscope graph has the primary amperage on the upper side (0-4 amps) and the charge amps on the lower section (0-4 amps) with an inverted scale for better visualization. The sampling rate here is 60 samples per second with the software averaging function in use. The SSG in this mode produces close to 90 Hz with the amp values bouncing all over the place so the software takes the samples and calculates the amps. The higher number of reads within a specific value range gives the higher density coloration. As you increase the speed of the sampling rate the extreme values are increased. Note the second Oscope graph and how the values are compressed to a thin squiggly line at the 3s/sec rate.

My next post will be a bit more descriptive for the experimenters that are into this kind of thing.

Yaro

TET - Current Senor Data Sheet.pdf

[Yaro1776]

SSG RPM Transition Zone Test 1

I started this research into the SSG transition zones quite awhile back and progressed to the DAQ method along with the TET amp sensors visualizing the inclusion of the battery voltages and rpm - all then exported to Excel for real time data and graphs. Well, the results of the first test of the initial DAQ, software and TET sensor amp package blew me way of course due to the very interesting results.

The SSG configuration used in this test is the standard plain vanilla 8 transistor board without any caps or complexities - 21 magnets on a 23" OD wheel. It has been in use for two years with many hundreds of hours of operation and multiple battery kills.

The first test was configured with a gap of ~0.365" after the various rpm transition zones were identified. The goal here was to investigate the behavior of the Primary and Charge battery amperages in the 257 rpm transition zone by carefully varying the trigger resistance. A lot of play time and patience here to get it right!

So the results are presented in the attached .pdfs and they demonstrate the amp pulsing behavior of this particular SSG configuration over a narrow band of rpm. For reference purposes the DATAQ DI 145 had all four channels active - an important point. Above each Oscope graph are my notes.

The results are as follows:
1) As the rpm is slowly increased to the transition zone the characteristics of the amp pulses are modified to very discernible narrow pulses that slowly increase in width.
2) The width of the pulses slowly increase to the point where they take up most of the graphical display.
3) Of particular interest are Oscope graphs 5 and 6 which show the pulse becoming a wide band with unusual amp reads.
4) As the transition zone is exited the pulses then start to narrow in width.
5) The behavior in Oscopes 5 and 6 was duplicated several times on different occasions - highly skeptical of the data values here and attribute this anomaly to wiring and config of the DI 145.

So the above is a short description of why the test work was diverted into another direction - again being highly skeptical of the above results the entire wiring system was revamped and the two channels not in use closed down. I did also use my clamp on amp meter to verify the values for O 5and 6, but without success at this time. I progressed with other testing and did not obsess with these unusual and highly improbable results since the entire config was still in preliminary test mode.

At this time I have no real explanation as to the behavior in the transition zones - the other zones and speeds are still being investigated. The plain vanilla SSG still has a lot interesting aspects to its performance. The right instrumentation and software helps to unlock these hidden gems.

Yaro
SSG Amps 2016a.pdf

[James McDonald]

{At this time I have no real explanation as to the behavior in the transition zones - the other zones and speeds are still being investigated. The plain vanilla SSG still has a lot interesting aspects to its performance. The right instrumentation and software helps to unlock these hidden gems.}

The recorded current of the primary and charge circuits really shows the classic theory of a coil building up its magnetic field.
The primary current versus the charge current shows the coils current lagging by 90 degrees. During the 90 degree period I see
the graph showing that 25% of that time period the SSG was in unity or slightly over unity. I noted that this time period was equal
to 5.61 seconds in the unity or slightly over unity. These currents sensors combined with the DAQ device and software is the only
way to show in a graph the data that the SSG really does achieve over unity. Be that it is only for a several seconds but the graph
does show what John Bedini had been saying about the SSG achieving unity or over unity but there was not a way of measuring that.
Once a capacitor dump circuit is added the 25% time of the graph in either unity or over unity may be able to be increased by another
25%. This is very good data showing the peak performance of the experimenter's SSG. This way of measuring the performance of the
SSG can be used to compare one experimenter's SSG to another experimenter's SSG. Still not all DMM's are created equal and with RMS
type measurements being done by these DMM's thus further reduces the ability to compare SSG setups. And just when we thought there
was nothing more to be learned from the mechanical SSG machine!!!

Thanks Yaro for all your experimental work!!!!


-- James

[Yaro1776]

Hey James,

Thank you for taking the time to examine the SSG amp charts. Your explanation makes sense, though it is beyond my current level of expertise. I do have two other sets of Test data that I will share that do show the lag you are pointing out. Interestingly these results are repeatable, to a certain degree, at other RPMs and wheels.

The Amp testing demonstrates that it is possible to have a real time value of the amp ratio for rapid comparison of various gap, RPM and trigger value combinations - that is the valuable, and also fun aspect, of this particular methodology.

Yaro

[RS_]

Hi Yaro,

I missed this Thread..... I have been testing the Arduino Micro, and the Spark Fun Hall Effect 5A current sensor..... works great for measuring 1+ Amps...

But did not work so good on the really small Joule Thief size SSG's, that only pull 20-50mA, to much noise.....

I will have to check on the DAQ and TET current sensor, and see about getting one, and see if i can measure these small mA currents with it......

also, You seam to be measuring the out put of the SSG after the diode...? is that correct....

I always measure the Primary Battery input Current x Voltage = W x t = Jin to a SG/SSG, then measure the Secondary Battery discharge output Current x Voltage = W x t = Jout to a load, to get Jin / Jout = COP....

You seam to be doing something else for a different reason......

[James McDonald]

Hi Yaro,

I missed this Thread..... I have been testing the Arduino Micro, and the Spark Fun Hall Effect 5A current sensor..... works great for measuring 1+ Amps...

But did not work so good on the really small Joule Thief size SSG's, that only pull 20-50mA, to much noise.....

I will have to check on the DAQ and TET current sensor, and see about getting one, and see if i can measure these small mA currents with it......

also, You seam to be measuring the out put of the SSG after the diode...? is that correct....

I always measure the Primary Battery input Current x Voltage = W x t = Jin to a SG/SSG, then measure the Secondary Battery discharge output Current x Voltage = W x t = Jout to a load, to get Jin / Jout = COP....

You seam to be doing something else for a different reason......

Hi RS_

I am using the same TET current sensors on my two SG Machines. They are hooked in series with the batteries.
Both the Primary and Charge batteries. So the input current draw is measured and the output current draw is measured.
It looks like the ACS712 5 Amp current sensor's frequency bandwidth is DC to 80 kHz. The TET current sensor is from
DC to 200 kHz. The main reason for using a TET current sensor is it did not degrade the RPM of the SG wheel when
put in series with the batteries. Right now the TET current sensor upper end is setup for a maximum of 6 Amps. This
can be modified to be 1.5 Amps very easily.

TET = Tesla Energy Technologies.

The current is measured using a different method. The chip doing the work is the DRV401 chip made by Texas
Instruments. (Fluxgate External Type Current Sensor) The internal measurement is done with a current probe which
is similar to the clamp meters.

http://www.ti.com/product/DRV401

-- James

[RS_]

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...?

[Yaro1776]

RS,

Thanks for your input! Your calc method for COP is the correct way to approach this measurement, however for the time being during the setup process the amps are a very useful indicator since the battery voltages during the testing are not that dissimilar - 12.4 volts vs. 12.8 volts (big batteries with lots of cushion are used during setup).

Since this a plain Jane SSG the amp measurements are on the cables to the batteries (- Primary) and + Charge). Over the many hours of testing this equipment, the amp method is a reasonable indicator of performance. It may not be precise, but it is a quick assessment of capability.

When all the pieces and parts are together with Excel in the DAQ mix, there will be a real time display of instantaneous Amp and Energy ratio. Just getting all this together slowly when the initial testing DAQ graphs distracted my attention. Us old dogs are easily distracted when we catch the scent of something interesting.

Keep on sniffing,
Yaro

[Yaro1776]

SSG DAQ Test 2

After the completion of the first test and a review of the results the SSG wiring configuration was modified and the DI 145 was re-positioned along with better input data cables. The software was configured to have the displayed graph scale of 0 - 4 amps with only two channels live - the other two available data input channels were disabled.

The purpose of Test 2 was twofold: the first aspect was to see if the same transition characteristics observed in the first test could be duplicated with the modified SSG configuration and the second goal was to change wheels and attempt to observe the transition zone at a different RPM of 227 - another identified transition zone.

Graphs 1 through 4 depict the startup behavior of the SSG and the further tweaking of the trigger resistance to dial into the transition zone. The subsequent graphs show the development of the transition zone pulses and also the characteristic of the pulses outside of this zone. The use of the sampling rate is a very neat tool in that it allows the researcher to zoom in on the area in question - essentially enlarging the specific view area and amp spikes.

So the first goal in this test was achieved with a better resolution to the amp values and pulse behavior duplication.

Now the low friction wheel 1 was replaced by the higher friction wheel 2 and tuned by varying the gap and resistance to obtain the approximate 227 RPM. Graph 5 depicts the transition zone with the "W" shape. This shape seems to identify the transition zone... Further slow tweaking of the trigger resistance created the pulse waves seen in Graph 6 similar to the patterns observed in Test 1. There are more graphs available at the higher sampling rate but information overload is not necessary - there is a similar progression of the pulse pattern!

The next post will deal in detail with Wheel 2 operating at 257 RPM (nearly max possible speed for this wheel) and demonstrate the amp pulse behavior with a different configuration.

Any errors or omissions in the above are on my shoulders.

The distraction continues,
Yaro
257 RPM Unstable Zone ESF.pdf