BQ35100: BQ35100: ScaledR measuredZ

tao zhang19,

1) Ra table does correspond to SOH 0-100%. However it is not evenly distributed. There are more point towards the empty to give a better accuracy near empty when impedance is increasing exponentially. 

2) MeasuredZ is the raw impedance measured by the gauge during the pulse you are applying while the gauge is active. 

3) ScaledR is the measuredZ * the scale factor which is learned on the "Learning Pulse". on the learning pulse the gauge takes Ra0 vs. MeasuredZ to establish the scale factor. This is used to account for parallel cells or cell to cell variation. Since these are primary cells there is no way to "learn" the actual cell attached.

The data you should here it seem like something is wrong with your pulse and the gauge is not accurately calculating measuredZ correctly. i would expect the values to be pretty stable. Some things to check:

1) During the load are you getting 100mV drop in the voltage

2) Is the pulse load long enough (32mSec +)

3) Is the battery as rested as it can be and near OCV when GE is put high. 

first thing you need to do is make sure you are getting consistent measuredZ values. 

Hope this helps.

Hope this helps

zhang19,

A Value of 32767 is max int. This does not look like a correct value. If you impedance was truly 32767. a current of 100mA would have an IR drop of about 3V and make the cell unusable. 

for a very simple test on what you should be getting around would be scope your Dsg pulse and current. then basic ohms law to calculate the impedance (MaxV - MinV)/ current. 

Zhang19,

1) For that particular post and chemID that was being used 8.8K was the Ra14 point which corresponds to empty. Each battery and ChemID will have a different value.

2) Impedance is easy. it is simply (OCV - loadedV)/Current. The 100mV requirement come from the math itself. Without a big enough voltage delta there is not enough resolution to accurately calculate the impedance. The drop i am talking about is on the voltage of the battery itself. When you apply the load does you battery voltage lower by at least 100mV. If it doesn't i would increase the load until you do see 100mV to see if that improves you measuredZ value. There is no requirement for current amplitude only enough to cause 100mV IR drop on the battery. I expect the load to be in the 100's of mA range. 

3) I expect you are not getting enough voltage drop or you load current is not high enough. MeasuredZ is the Raw value the gauge is calculating based on the IVT data collected after the gauge start command is issued. Due to cell to cell variation the impedance of each cell will be slightly different. To account for this on the "Learning pulse count" the gauge will do a comparison of measuredZ to Ra0 (full battery). This ratio computes a scale factor. From then on the measuredZ is multiplied by the scale factor to produce the ScaledR. The ScaledR is what is used to compare against the Ra table to establish the batteries SOH.

4) Each pulse should produce a different measuredZ, but i expect them to be very similar until the end when we should see a large increase in the value. The procedure you need to do is simple

a. Start with battery relaxed at OCV

b. Set GE pin high

c Send Gauge start command

d Apply load pulse of high enough amplitude to cause battery voltage to drop by 100mV

e remove load

f send gauge stop command

g wait for G_DONE bit (15sec)

h Read measured Z

Your first step should not be worrying about SOH at all until you get reliable measuredZ values. I suggest you do this process on our EVM and benchtop equipment as opposed to in system. Having a well rested battery to start with will be critical.