NiMH batteries voltage depression effect

Go To Last Post
30 posts / 0 new
Author
Message
#1
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 1


 

I've been tinkering with the AVR-DB series, and since it has builtin opamps, I decided to make a constant current load, something like this but using the internal opamp and DAC (and the MOSFET is a 2N7000)

 

 

I decided to use it to test some NiMH batteries I have been using for years to power a wireless mouse. So I clobbered together an Arduino program using DxCore, and logged battery voltage data with the AVR-DB ADC.

The results were quite interesting. I knew these batteries were in somewhat poor health, but since powering a mouse is a very undemanding application, I never really cared much.

 

But man, they were in really bad shape. While discharging (at 150 mA) the voltage initialy rose, then peaked in finally dropped as the battery got depleted. Battery capacity had degraded to 1/3 of nominal.

(I discharged to 900 mV, because I didn't want to go to zero, and 900 mV seemed a good number, yeah even though I'm a chemist my knowledge of batteries is pretty low cheeky).

Then I recharged the battery, and did another discharge cycle.

This time the discharge looked "normal" and the battery capacity had increased a lot (still only about 3/4 of nominal, but hey, that's fine by me given how I mistreated these batteries over 10 years or so...)

 

After some research, it turns out this is a know issue with NiMH batteries https://rightbattery.com/tag/voltage-depletion/ but mine really were an extreme case, it seems (probably I never fully discharged them until now - ever).

Always learning smiley

 

 

edit:

Picture with actual circuit (except for serial link, I used realterm and a USB/serial adapter to log data, it's in csv format):

 

 

Also attached program in case someone is interested.

Attachment(s): 

Last Edited: Wed. May 5, 2021 - 01:22 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Interesting.  It would be neat to see V vs time*current-draw for a few constant current values.

This also shows that for this battery voltage is not a great indicator for state of charge.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

MattRW wrote:
This also shows that for this battery voltage is not a great indicator for state of charge.

Indeed.

 

Generally the case for lithium & nickel based batteries; eg, https://www.avrfreaks.net/commen...

Top Tips:

  1. How to properly post source code - see: https://www.avrfreaks.net/comment... - also how to properly include images/pictures
  2. "Garbage" characters on a serial terminal are (almost?) invariably due to wrong baud rate - see: https://learn.sparkfun.com/tutorials/serial-communication
  3. Wrong baud rate is usually due to not running at the speed you thought; check by blinking a LED to see if you get the speed you expected
  4. Difference between a crystal, and a crystal oscillatorhttps://www.avrfreaks.net/comment...
  5. When your question is resolved, mark the solution: https://www.avrfreaks.net/comment...
  6. Beginner's "Getting Started" tips: https://www.avrfreaks.net/comment...
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Yeah, the voltage is pretty much useless to gauge charge, more so because of the memory effects caused by poor usage.

 

MattRW wrote:
It would be neat to see V vs time*current-draw for a few constant current values.

 

True. I'll find a power MOSFET to replace the 2N7000 and try higher currents.

Last Edited: Wed. May 5, 2021 - 01:35 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

You need a power mosfet with a heatsink...aren't you making a power load???

 

One ohm is rather poor, since it greatly limits the possible max current (or maybe it is a bonus)....however, it also limits the gate drive (Vgs) depending on how many amps you want.  0.1 or 0.5 ohm might do you better.

If you want 10 amps, you'd want something more like 0.01 ohm (gives 100mv signal)...notice that is 1W of resistor power (use a 2W to 5W rated part).

When in the dark remember-the future looks brighter than ever.   I look forward to being able to predict the future!

Last Edited: Wed. May 5, 2021 - 04:44 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Do not overlook the forward bias needed by the FET and the output range of the op amp. Can the op amp swing to a high enough voltage to get the 1.5A needed to drive the one Ohm load?  The threshold voltage of a 2N7000 can be as high as 3.0V just to get 1mA. The gate forward bias for 1.5A can be 7V. 

 

Can a DB series op amp, even powered by 5V deliver that?

 

Jim

 

 

 

Until Black Lives Matter, we do not have "All Lives Matter"!

 

 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Can a DB series op amp, even powered by 5V deliver that?

Most (or many, many)  modern low voltage ( say<60V) power fets will be turned on pretty fully by 5V (or even less)...many can even be driven by 3,3V logic, perhaps to 99% ... You can even get lower fets (like 2V) , but then other negative tradeoffs start to become more evident.

Remember when it used to take 10V to be fully enhanced?

 

The current resistor reduces the effective VGS, so you don't want too much drop there (though it provides a feedback in the desirable direction higher current---> fet driven off)

 

Here, the fet is mostly not wanted to be 100% on, but some regulating amount.   We used tempfets maybe 20 years ago to make a big power load, that only needed to be one for maybe half a minute...so we didn't provide any fan cooling.  We had maybe 7 fets in parallel for several hundred watts.  If we let it run beyond normal time, of course the temperature would climb up...up  ..up.  Eventually one of the tempfets would trip off, then the load was shared by the remaining which would get even hotter, so it was fun to see the rapid collapse of the remaining ones...faster & faster they fell.  The we'd let them cool down & recover.

 

https://rocelec.widen.net/view/p...

The TEMPFET™ switches can be controlled as any discrete FET, but in addition they implement a temperature protection.

 

There are still thermally protected fets around , but not quite as much now...many obsolete!

 

When in the dark remember-the future looks brighter than ever.   I look forward to being able to predict the future!

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

 

avrcandies wrote:
You need a power mosfet with a heatsink...aren't you making a power load???

 

it's not general purpose, the goal is just to test single cell NiMH batteries so the voltage will be limited to 1.5V max and maybe 2A, so 3W power. Yes, I'll need an heatsink but nothing huge. And yeah, reduce the resistor.

 

ka7ehk wrote:

 

Can a DB series op amp, even powered by 5V deliver that?

 

The opamp is rail to rail. But will 5V be enough? I'll have to test I'm not sure.

 

What if I rearrange things like this:

 

 

This way the battery will help bias the MOSFET. Would this be ok or will something blow up because I'm missing something? (quite likely...)

 

edit: 

avrcandies wrote:

The TEMPFET™ switches can be controlled as any discrete FET, but in addition they implement a temperature protection.

 

Thanks, I know there are amazing modern parts around, but this is not for some commercial project, it's just for my own use. I want to use stuff from my junk bin... I have these  https://www.infineon.com/dgdl/irf2807pbf.pdf?fileId=5546d462533600a4015355dea79e18f2  maybe others I'll need to search more carefully...

 

Last Edited: Wed. May 5, 2021 - 07:44 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

So is it "depression" as in the title or depletion? cheeky I can understand batteries getting depressed, they are always drained of their energy.

John Samperi

Ampertronics Pty. Ltd.

https://www.ampertronics.com.au

* Electronic Design * Custom Products * Contract Assembly

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

No , your original circuit is what you want (resistor on low side)...Also, you don't want a high common mode voltage coming into the opamp.

the fet can't turn off (need Vgs very low, like zero)

 

if no flow, Vr1 is zero, like Vbat++ is now at gnd, so Vbat- is negative, same as fet source

You'd need to spit out -1.5V to turn off the fet (to get Vgs zero)!!!!  Trashcan!

When in the dark remember-the future looks brighter than ever.   I look forward to being able to predict the future!

Last Edited: Thu. May 6, 2021 - 12:33 AM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

avrcandies wrote:
You'd need to spit out -1.5V to turn off the fet (to get Vgs zero)!!!!  Trashcan!

 

I know, but the the IRF2807 fet has VGS(th) = 2V minimum, defined as 0.25 mA current. So it will be mostly off cheeky right?

The battery will not be left there, it will only be present during testing.

 

Anyway, I'm aware this is not the "proper" constant load circuit but sometimes I like experimenting, so long as I'm mostly sure stuff won't blow up.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

 

Here's a tip....look for fets with a lower Vgs max  (such as 10V  or 15V)...they will generally have fuller turn on at lower voltages.  Also, look for 60V or less fets, unless you NEED higher voltage.

Once you get above the threshold, the current max capability goes up very rapidly. 

 

In fact you can take a 9v battery & pot to VGS & make a quick pulse current limiter for experiments (it will be very touchy, but useful).

We done that with drivers on the bench, where we didn't want 100amp surges suddenly blowing through while we were fiddling.

 

In any case, you need to use your first circuit with resistor to gnd ...the other will not function so well (at a minimum the opamp would have to provide a negative voltage)

 

so the voltage will be limited to 1.5V ..plan for more!

 

 

 

When in the dark remember-the future looks brighter than ever.   I look forward to being able to predict the future!

Last Edited: Thu. May 6, 2021 - 11:12 AM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

avrcandies wrote:

Here's a tip....look for fets with a lower Vgs max  (such as 10V  or 15V)...they will generally have fuller turn on at lower voltages.  Also, look for 60V or less fets, unless you NEED higher voltage.

Once you get above the threshold, the current max capability goes up very rapidly. 

 

Thanks for the tips, I'll see if I can find some fets matching those specs around.

 

avrcandies wrote:

In fact you can take a 9v battery & pot to VGS & make a quick pulse current limiter for experiments (it will be very touchy, but useful).

 

Yeah good idea I'll use it to find a suitable fet from my parts bin.

 

Also I noticed the opamp oscillates under certain conditions. Maybe because of gate capacitance? Any advice on how to address that problem?

I found this  https://itectec.com/electrical/electronic-constant-current-source-with-mosfet-opamp-and-instrument-amp/ seems to contain some ideas.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 1

I was going to mention you should add maybe 100 or 1k ohms between the opamp and gate (smaller,  for really fast response , like MHz)

i usually tap off the sense pickup with some resistance (to the opamp) and add a very small cap between the opamp out & - terminal.

 

Make sure your sense resistor is non-inductive & its gnd is tight with the opamp gnd...so the sensed voltage is not wobbled getting to the opamp*---will cause bouncing troubles

 

*that's an advantage of something like the INA180  ..since it has two input leads (differential) that can go to the sense resistor & resistor can wobble around on gnd & still be seen by the AVR accurately**.

But you generally need to be dealing with AMPS, not mA to care about that.

 

** even better if you use a 4-terminal (Kelvin) sense resistor, not needed at low currents & higher sense drop voltages

When in the dark remember-the future looks brighter than ever.   I look forward to being able to predict the future!

Last Edited: Thu. May 6, 2021 - 06:10 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

avrcandies wrote:

I was going to mention you should add maybe 100 or 1k ohms between the opamp and gate (smaller,  for really fast response , like MHz)

i usually tap off the sense pickup with some resistance (to the opamp) and add a very small cap between the opamp out & - terminal.

 

Thanks yes. I had to do all of the above to stabilize the opamp, I used 1k from opamp output to gate, 1k from sense point to (-) terminal and 100nF cap between (-) and output (maybe the cap is too large? You said "very small").

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

 and 100nF cap between (-) and output (maybe the cap is too large? You said "very small").

I'd usually add a 100-220 pf, perhaps up to 1000pf....more than that & you need to get out your calculator.  You should always do a stability analysis, or at a minimum do a step response on the bench.

When in the dark remember-the future looks brighter than ever.   I look forward to being able to predict the future!

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Ok, after some step response experiments I found better values for the resistors (increased the sense pickup resistor to 100k) and now I can use a 1000pF cap (or maybe less) without ringing or overshoot.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

El Tangas wrote:
... and 100nF cap between (-) and output (maybe the cap is too large?
Per a simulator, stable with and without the capacitor; capacitor reduces bandwidth by three orders of magnitude (DAC's frequency)

Measuring stability -

A NanoVNA would be a fit for a 330pF capacitor (plus or minus); lower frequency VNA for larger capacitor though at an order of magnitude greater price (a power supply may have an an instability in the ones of KHz range ... any control system down to a Hz)

 


NanoVNA | Very tiny handheld Vector Network Analyzer

Specifications | NanoVNA V2 Official Site

NanoVNA V3

A $50 Vector Network Analyzer | The Embedded Muse 382 by Jack Ganssle

Using the Network Analyzer - Digilent Reference

 

OPAMP Specifications | AVR® DB Family

Constant-Current Driver Using OPAMP Peripheral | Constant-Current Driver Using the Analog Signal Conditioning (OPAMP) Peripheral

for attached LTspice schematics, Bode plot is V(sense)/V(X1:gen)

Easy Control Loop Measurement and Verification (ON Semiconductor)

LTspice | Design Center | Analog Devices

 

Attachment(s): 

"Dare to be naïve." - Buckminster Fuller

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Yes with the 2N7000 there are no problems, but with a power mos the opamp has oscillation problems, I suppose because of gate capacitance being much larger for a power mos than for a 2N7000.

 

And thanks for the links yes that's an interesting device.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Reduce your gate resistor to 100 ohms.  If you have external noise causing severe problems (including inductance of the leadwires of a remote heat-sunk fet), a ferrite bead around the Fet's gate pin (right at the fet) can be a real help. 

Make sure your opamp has a solid power pin cap.   Forget about the 2N7000  ...it can burn to a crisp in no time.

 

Show your schematic & a photo of your build!

 

That VNA sure beats the boat anchor I have!!

When in the dark remember-the future looks brighter than ever.   I look forward to being able to predict the future!

Last Edited: Sun. May 16, 2021 - 09:09 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0


You're welcome.

El Tangas wrote:
I suppose because of gate capacitance being much larger for a power mos than for a 2N7000.
Indeed

AVR DB op amp drive is current-limited and typical for a jellybean CMOS op amp; can current amplify to drive power transistors.

 

OPAMP Specifications | AVR® DB Family (ISC = 30mA)

MCP6481/MCP6482/MCP6484 Data Sheet (ISC is approx 30mA at 5V)

TND6093 - Low VCE(sat) BJT's in Automotive Applications (ON Semiconductor)

[top of page 4]

 

"Dare to be naïve." - Buckminster Fuller

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

avrcandies wrote:
That VNA sure beats the boat anchor I have!!
Boat anchor must be the referencewink

"Old" T&M instruments can be of excellent value; some can still be calibrated.

 

"Dare to be naïve." - Buckminster Fuller

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

You do not need MOSFET Gate Driving when you are in linear region. This schema is for very fast switching use only. And the misspelled R LAOD is the inductor or transformer usually.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

You do not need MOSFET Gate Driving when you are in linear region.

That's partly true, if you are not moving around too much or don't need a super-fast pulse response (or have a fairly steady loading).  For general purpose low freq current source, just an opamp works very good. Of course if the load is appearing suddenly or the setting changes suddenly, then the gate needs a good kick.

When in the dark remember-the future looks brighter than ever.   I look forward to being able to predict the future!

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Having the negative feedback and all about is the science many fathoms deep.

 

How fast the MOSFET will react is also interesting, and it may be disputed. For example- if the load is changed that fast- is it better then to use Power Darlington type, it may follow the signal most quickly, guess.

 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0


Certainly, any time you cause the FET the need to react quickly & in a large amount, the gate drive requirement can be large.  Note that the high gain of the fet (gm, transconductance) may still mean the gate does not need to change much for a large output increase (or decrease)--that's what amplification is all about.

 

The gate input capacitance has a rather complex model, that can be categorized into 3 semi-overlapping "modes", for a good overall gate drive approximation. If you want to switch the gate of/off at MHZ, a surprising amount of current is needed (and power ued).

 

 

Don't you hate those annoying terms?

When in the dark remember-the future looks brighter than ever.   I look forward to being able to predict the future!

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

 

Not at all, and you just scratched the surface.

All AVR fans should know that the world is not only digital 0,1- the real analogue word is the science on its own and such heavy formulae are common.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

As my electronics tutors used to beat into us: there's no such thing as digital electronics. Everything is analog. Digital is just a subset where you agree on the analogue limits.

 

Apropos of which: the board I'm currently testing uses an LM358 to level convert a serial signal from 3.3v 'ttl' to 12v... works fine at the design frequency, but when you forget and leave the uart at 115,200 instead of 4,800 or 7,812... nope, it don't work then :)

 

Neil

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

LM358 is extremely suitable for LF anything, for example the simple I/U convertor. Cheap, too. What about LF358, please.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Earth Ground by Dr. Howard Johnson

[near top]

In my opinion the most important point to make with regard to grounding is that the input to every digital logic gate is a DIFFERENTIAL amplifier.

...

barnacle wrote:
Apropos of which: the board I'm currently testing uses an LM358 to level convert a serial signal from 3.3v 'ttl' to 12v
tops out 18V though minimum is 3V; IIRC, also in hex and octal :

CD40109B CMOS Quad Low-to-High Voltage Level Shifter (20V Rating) | TI.com

 

"Dare to be naïve." - Buckminster Fuller