The new GRINSPECTOR

Batteries, Chargers, and Battery Management Systems.
Post Reply
User avatar
Matador
10 kW
10 kW
Posts: 593
Joined: Jun 29, 2016 10:00 pm

The new GRINSPECTOR

Post by Matador » Apr 15, 2017 8:21 am



I could'nt find any specification of this R&D product that Grin is deveopping on their website yet.
It can discharge you battery at different discharge rate and give capacity at different loads.
You can datalog the [CA-Grinspector] unit complex to a computer and plot all the data.

AND !!!! Calculate DC internal resistance.
Now that's very interesting.... DCIR measurement can be complex too. Hence i'm intrigued how they do it.

Can you measure DC-IR in function of State of Charge and get a cruve of the DC-IR evolution through whole the discharge cycle ??? (DCIR on y-axis versus SOC on x-axis) ?

Very intrigued... And enthousiast!!!

I'd like to do discharge cycles with different resistor values, and Justin say you can interchange the big resistors (or parrallel or serie them) to have different loads.
If there is many different values of resistors, you can plot a graph that's :

V = - r I + E in the y = mx + b form (the slope m = - r)
where
V = voltage (V)
r = DC internal resistance (Ohms)
I = current (A)
E = electomotive force (aka float voltage of battery with zero load)

The more points you have on the graph, the more refined the DCIR measurement becomes.

A graph with 5 or more points and you can get some very accurate DCIR measurement.
I did it on a smaller scale here (on one 18650 cell) : https://endless-sphere.com/forums/viewt ... O#p1276187
Image if for each measurement of the discharge curve you get a set of 5 measures like this to get the precise "r" value. Than plot "r" against the state of chare/voltage.

As you may notice, I took measurement precise given voltage (state of charge). DCIR can vary with SOC... And my method has a flaw : you have to take into considaration that voltage drop is actually voltage drop + a bit of voltage loss from normal discharge.
Now if the Grinspector does a multipoints (multiresistance values) testing not just at one discrete SOC but for the whole SOC range/curve. This tool is what I have been looking for a long time now and did not find before.
I'd like to understand this product in more details...

I imagin myself running a battery for 10 cycles, with 10 different loads(10 different external resistors). I coud obtain 10 dicharge curves that sag differently.
Then for that graph with 10 curves, I get 10 datapoint for every discrete SOC value plotted on x-axis.
I then get a very precise DC-IR value at any given SOC....

markz
10 GW
10 GW
Posts: 4753
Joined: Jan 09, 2014 11:38 pm
Location: Alberta Canada

Re: The new GRINSPECTOR

Post by markz » Apr 15, 2017 4:39 pm

I'd like to know the price of the device when it is released.

Justin said in a month, and to dealers. So interesting.
Wonder if us enthusiests can buy it.

User avatar
Matador
10 kW
10 kW
Posts: 593
Joined: Jun 29, 2016 10:00 pm

Re: The new GRINSPECTOR

Post by Matador » Apr 15, 2017 7:01 pm

markz wrote:I'd like to know the price of the device when it is released.

Justin said in a month, and to dealers. So interesting.
Wonder if us enthusiests can buy it.
I'm pretty sure they'd sell it to us enthousiast also. Why loose such a part of the market. As you can see from my post above, I'm already looking for improvised ways to do it using resistors. The need exists. DIY enthousiast do such things. Why not let them buy the Grinspector rather than let them build their own improvised (and possibly unreliable or even dangerous) device. If Grin doesnt do it the chinese will probably come up with their own device. I'd rather buy stuff from a reputable company or reputable seller... Like Agilent, Siemens, Variant, or Grin for that matter.

I'm the kind of guy who would buy a DIY solder Geiger counter kit and build it... BUT not much spare times these days....

markz
10 GW
10 GW
Posts: 4753
Joined: Jan 09, 2014 11:38 pm
Location: Alberta Canada

Re: The new GRINSPECTOR

Post by markz » Apr 15, 2017 7:14 pm

I'd buy it if the price is right just to support Justin and his work, because I'd only use it once every so often, its not like I buy new battery packs, or make packs often, but I'd like to know the health of my packs and track them.

markz
10 GW
10 GW
Posts: 4753
Joined: Jan 09, 2014 11:38 pm
Location: Alberta Canada

Re: The new GRINSPECTOR

Post by markz » Nov 04, 2017 10:49 pm

Available now for 2 bills! loons tho ;)
http://www.ebikes.ca/shop/electric-bicy ... ation.html

also, just did a quick user search for Justin_le with keyword being grinspector and nothing came up. The video sums it up nicely though!

User avatar
amberwolf
100 GW
100 GW
Posts: 23948
Joined: Aug 17, 2009 6:43 am
Location: Phoenix, AZ, USA, Earth, Sol, Local Bubble, Orion Arm, Milky Way, Local Group
Contact:

Re: The new GRINSPECTOR

Post by amberwolf » Nov 06, 2017 1:48 am

Looks like it only does up to 60V full charge, and 35A discharge rates.

Voltage can be upped by changing a FET to a 100V type for 72V packs.

I wonder what it would take to up the discharge rate to something useful for large / high-current packs? (like what I use on the trike, which could do 200A continuous and 400A for 10sec per the spec sheets)

User avatar
justin_le
1 MW
1 MW
Posts: 1994
Joined: May 28, 2007 3:27 am
Location: Vancouver
Contact:

Re: The new GRINSPECTOR

Post by justin_le » Nov 09, 2017 5:07 pm

amberwolf wrote:Looks like it only does up to 60V full charge, and 35A discharge rates.
Yeah. I really was hoping we'd be able to pull of the standard unit to work with up to 72V batteries OK but in the end the component selection for the mosfets for charge (p-channel) and discharge control made this difficult. Our main target for this is for bike shops that deal with ebikes where 99% of what they encounter are 36V / 48V / 52V batteries and where a 10-20A characterization is more than sufficient.
I wonder what it would take to up the discharge rate to something useful for large / high-current packs?
Basically if you don't _need_ the test to terminate the discharge and start the charge cycle automatically, and are fine running it just as a discharge test until the BMS trips, then you can skip the Grinspector base station completely. Take an existing V2 CA device, load on the Grinspector firmware, and hook it up inline with your battery using a shunt and load that is appropriate for your discharge currents. You'll get to use the software that makes the discharge plots and will be able to test batteries right up to the full 150V rating of the CA device and at whatever current you want. If you did want the automatic charging functionality on higher power and voltage packs, then you could take the CA's throttle output signal and have it turn a relay board on and off as shown in this image here, basically replicating the base station functionality with appropriately rated parts:
Image
Matador wrote:AND !!!! Calculate DC internal resistance.
Now that's very interesting.... DCIR measurement can be complex too. Hence i'm intrigued how they do it.
Can you measure DC-IR in function of State of Charge and get a cruve of the DC-IR evolution through whole the discharge cycle ??? (DCIR on y-axis versus SOC on x-axis) ?
Basically in the initial implementation we have it so that IF you run two more more discharges at different current levels, then the software will compare the terminal voltages between the two plots at a handful of SOC levels and compute an average/effective DC internal resistance value. However, for that to work it means that the user needs a load bank that lets them switch the discharge current and do two or more discharge tests.

We'll be modifying the software fairly soon so that it has an option to do a periodic pause during the discharge cycling to get an internal resistance value on just a single discharge test at a fixed current. This is what we've been doing during the QC of our LiGo battery modules, and it allows for some pretty nice graphs of the DCIR versus the SOC level. Here is a recent test on a pack that failed QC. We have it so that every 0.1 Ah the discharge current stops, the voltage is measured at 0 amps, and compared to the voltage when current was flowing.
LiGo QC Discharge Example.jpg
LiGo QC Discharge Example.jpg (92.61 KiB) Viewed 365 times
Since we're communicating directly with the BMS in this case we can see all the individual cell voltages and do a cell by cell resistance comparison. In this case, cell #10 (purple) has a much higher resistance than the others, averaging 40.7 mOhm compared to ~35mOhm, and you can see that during the discharge this was reflected in a lower voltage too since there is more terminal voltage drop. So the Grinspector software will be doing something similar, but it would just be at the pack level voltage and resistance, not the cell level.
Big Dummy Frame (yes This One, thanks ES!) with prototype 26" Grin all-axle front hub, Phaserunner controller, and 52V 11Ah Cellman triangle pack
My website: http://www.ebikes.ca
Please contact via email, info@ebikes.ca, rather than PMs, which are disabled

markz
10 GW
10 GW
Posts: 4753
Joined: Jan 09, 2014 11:38 pm
Location: Alberta Canada

Re: The new GRINSPECTOR

Post by markz » Nov 09, 2017 5:55 pm

Impressive!

User avatar
amberwolf
100 GW
100 GW
Posts: 23948
Joined: Aug 17, 2009 6:43 am
Location: Phoenix, AZ, USA, Earth, Sol, Local Bubble, Orion Arm, Milky Way, Local Group
Contact:

Re: The new GRINSPECTOR

Post by amberwolf » Nov 10, 2017 1:41 am

justin_le wrote: Our main target for this is for bike shops that deal with ebikes where 99% of what they encounter are 36V / 48V / 52V batteries and where a 10-20A characterization is more than sufficient.
I figured that was probably why--no reason to spend more on it for the big market, for the few that might need more. :)

Basically if you don't _need_ the test to terminate the discharge and start the charge cycle automatically, and are fine running it just as a discharge test until the BMS trips, then you can skip the Grinspector base station completely. Take an existing V2 CA device, load on the Grinspector firmware, and hook it up inline with your battery using a shunt and load that is appropriate for your discharge currents. You'll get to use the software that makes the discharge plots and will be able to test batteries right up to the full 150V rating of the CA device and at whatever current you want. If you did want the automatic charging functionality on higher power and voltage packs, then you could take the CA's throttle output signal and have it turn a relay board on and off as shown in this image here, basically replicating the base station functionality with appropriately rated parts:
Wow--thanks! That, I can do. :)

I have a bunch of EIG 20Ah cells that it'd be nice to test out; the voltage isn't as big a deal but the current is, since they're 5C cells (100A), and I'd like to characterize them at the high end of that, since my trike already pulls more than that (I'm running 2P) and will be even harder on them once I build the better controllers.

It'll only be that hard on them during acceleration (a couple of seconds at a time at that higher rate, just under four seconds now), or while pulling heavy loads like 600lbs of dog food, or both dogs plus cargo, etc., or riding up hills. Not something I'll do a lot, but....

markz
10 GW
10 GW
Posts: 4753
Joined: Jan 09, 2014 11:38 pm
Location: Alberta Canada

Re: The new GRINSPECTOR

Post by markz » Nov 10, 2017 5:02 pm

Amberwolf - Keep us updated in a thread with lots of pictures if you decide to try it out.

User avatar
tomjasz
1.21 GW
1.21 GW
Posts: 3565
Joined: Mar 29, 2014 1:45 pm
Location: Out riding, MN USA
Contact:

Re: The new GRINSPECTOR

Post by tomjasz » Nov 29, 2017 12:36 am

Anyone else using the Grinspector?

Justin, any future for OS X?
Thanks Justin_le we're here thanks to you. All the best to the mods for their tireless work keeping it on an even keel.

Post Reply