Inconsistency between C9000 and BC700 capacity measurements

[espresso]

If I try to measure the remaining capacity of a Sanyo XX that my mouse reports as empty there is a huge difference between what C9000 and BC-700 report at 100 mA discharge rate.
C9000 usually reports that this battery has 0 to few mAh left in it and it reaches 0.9V immediately. But if I try the same battery in BC-700 it reports 200 mAh remaining. The voltage in BC700 starts at 1.16V and then begins to drop gradually during the next 2 hours.

It seems to me that Maha C9000 uses high current discharge pulses that effectively give 100 mA while BC-700 uses much lower current pulses (if at all). Thus I find BC-700 capacity measurement to be more accurate for lower quality and near empty cells as it simulates discharge rates of real life applications more closely. Except for digital cameras and such high tech gadgets.

There are plenty of low drain applications that really don't need super quality Eneloops to work properly and here I can use lower quality cells or older batteries with degraded resistance.
Here is another measurement done with lower quality cells. It shows that C9000 always reports less capacity than BC700.




Notice that even at 3 times higher discharge current, BC-700 measures more capacity from a cell than C9000.
My point is that if I want to do measurements at 100 mA, the charger should really use that current and not some 1A or 2A pulses that simulate 100mA. That's why I find BC-700 to be more accurate for measuring capacity of lower quality cells. After all, the purpose of these chargers is to analyze older and lower quality batteries. I don't need no analyzing of a new Eneloop.

Does anyone know how high discharge pulses are in C9000?

[czAtlantis]

Maha C9000 is really bad charger for small batteries and batteries with high(er) internal resistance. I am in the same situation as you - I also use some batteries in low power devices and they can't be charged successfully in Maha. To be honest I hate it - it is expensive charger and it can't even regulate constant current.

C9000 uses 1A pulses for discharge and 2A for charge.

I also have FK Technics IPC-1 (which is very similar to BC-700) and it uses I think also pulses but they are PWM controlled at high frequency so if you measure it with multimeter it looks like constant current and like you I get much higher readings for older batteries.

[Yamabushi]

@ espresso:

Your test shows that the BC700 reports a higher capacity than the C9000. That doesn't necessarily mean it's more accurate. In the case of the Sanyo XX, it seems that your mouse agrees with the C9000.

Have you performed another referee test, i.e., a third battery testing device or runtime tests in other devices after charging the cells in each charger?

[espresso]

I didn't have time to do more tests as I haven't had C9000 for too long. And cycling/break in tests take so much time to complete. I will test new Varta LSD AAA batteries next week.

In the case of the Sanyo XX, it seems that your mouse agrees with the C9000.

The mouse stops working at around 1.13V. I measured this with a voltmeter under load.
And right after that BC-700 reports 1.16V at 100 mA discharge current. But it manages to draw additional 200 mAh out of that battery until it reaches 0.9V. With Maha, that's not possible, it instantly knocks down the voltage to near 0.9V even though it's discharge current is set to only 100 mA.
That's the reason why I have doubts in accuracy of C9000.

[45/70]

Hi espresso. Your observations and conclusions are correct. The C9000 utilizes a pulsed 1000mA discharge, at whatever discharge rate is chosen and the La Crosse chargers utilize a pulsed 500mA. This does make quite a difference, especially when determining the capacity of small, well used, or weak cells.

As Yamabushi pointed out, the Maha's setup actually seems to work about right, for your mouse cells. It is generally considered that a NiCd or NiMH cell is discharged when the OC voltage reaches ~1.20 Volts. That's not to say that cells in this low state of charge cannot still be used in devices that draw very low current, but the cells are basically discharged.

I too have found that cells that have seen better days, that are now used in indoor/outdoor thermometers, clocks, and such, that don't work well in higher drain devices, require that they be charged in the La Crosse BC-900. The Maha will simply reject them, due to their high IR. These same cells used in my lights, are under performers, for the most part. They either drive the light at an unacceptably low level, perpetual "moon mode", or in some cases, not at all. Again, this is a sign of high internal resistance.

I'll go along with the idea that the C9000 is not the best charger for AAA cells, in some situations. On the other hand, I've found that it weeds out AAA cells quite well, that are used in my pocket AAA lights that draw 1A or more. If the C9000 won't charge an AAA cell, the same cell charged in the La Crosse doesn't perform very well anyway. So, for most devices that I use, the C9000 is still useful for AAA cells.

Also keep in mind that cell capacity varies, depending on the rate of discharge. For NiCd and NiMH cells the "standard discharge" rate is 0.2C. This is quite low, compared to the current draw of many of our lights. For higher drain devices useable capacity is more important. This requires a higher discharge rate, similar to the rate at which the device will discharge the cell(s), to determine the cell's useful capacity. For these applications, the C9000 works out quite well.

Dave

[SilverFox]

Hello Espresso,

Dave covered the difference in discharging methods as far as the discharge current is concerned.

Another difference is when the discharge begins.

The BC700 immediately starts the discharge as soon as the charge is completed. The cells are "hot off the charger." This is something that has been noted by the RC people for years. If your battery pack is hot off the charger, it will perform better in comparison to one that has been sitting for awhile. Sitting allows the surface charge to dissipate.

The C9000 in Break In mode charges, then lets the cells rest for 30 minutes before starting the discharge. This allows the surface charge to bleed off and not influence the actual capacity.

Since most battery use conditions involve using a batter over a longer period of time, the capacity minutes the surface charge capacity is more representative.

My position on cell condition is well known. I believe that when a cell drops below 80% of its initial capacity, it is time to recycle it. CRAP cells have higher internal resistance, higher self discharge rates, and tend to miss charge termination when charging. This results in more damage being done to the cell from overcharging.

Sanyo 2500 mAh cells usually start out at around 2200 mAh. 80% of that comes in at 1800 mAh. Your cells come in below that on the Break In test.

There is a grey area between 80% and 60%. 60% of 2200 mAh is 1320 mAh, so if voltage holds under load and there are no missed terminations during charging, you may venture into the grey area and take your chances.

By the way, a Sanyo 2500 mAh cell that is in good condition can easily test out to over 2600 mAh if you charge at 3000 mA and discharge at 1000 mA.

Tom

[Smells_Familiar]

As a layperson, is there any advantage in the C-9000 charging with 2000mA pulses and discharging at 1000mA pulses besides it being able to determine if a cell is appropriate for higher drain devices?

At this point, I think I'd like the C-9000 to charge and discharge with more frequent and less intense charge pulses, and would petition Maha to do this, I just don't know what the downside would be.

[espresso]

I don't think there is any advantage for our batteries. I think Maha engineers chose this design only to make their own work easier.

Sanyo 2500 mAh cells usually start out at around 2200 mAh. 80% of that comes in at 1800 mAh. Your cells come in below that on the Break In test.

The results in the table are not from XX but from lower quality AA batteries labeled as Camplus 2800 mAh. They came with one of my chargers and I haven't used them much. Anyway, I will be comparing some Varta AAA soon and we'll see how they perform in both chargers.

[Yamabushi]

The C9000 in Break In mode charges, then lets the cells rest for 30 minutes before starting the discharge.

My C9000 (production code 0K0GA) rests the cell 1 hour, which IIRC is consistent with the IEC test cycle. Were earlier versions different?

[SilverFox]

Hello Yamabushi,

Oops, I was going from memory and what is called for in the standard for cycle life testing. When I do a Break-In I just set it up and come back a day or two later and I don't pay much attention to how long it is resting. I should have gone through the Maha thread to find this, but I just went from memory.

Sorry about that. I will go with your experience of having a 1 hour wait.

Tom

[Mr Happy]

The C9000 is not perfect. A rigorous scrutiny will find many design decisions that may be questioned or considered less than optimal. However, we have to be realistic and realize that few things indeed are perfect. The imperfections are things that sometimes are thought to give a device "character".

However, regardless of the imperfections, the C9000 has many things to commend it, not least because where are you going to find something else with the same or better positive attributes for the same money? Sure, you can criticize it, but it remains a very good charger with many excellent aspects to its design and a manufacturer that stands solidly behind it.

I am aware of its flaws, but I will still buy another one without hesitation, and recommend the same to others. No, it is not perfect, but it is still a very good charger and it weeds out badly performing cells with an uncompromising rigor.

[Smells_Familiar]

I would buy another too, but as of now I hope Maha would reduce the pulses to 1000mA charge and 500mA discharge. I'm going to contact Maha and request they do this and I encourage anyone else who agrees to do the same.

But before I contact them I want to make sure I know what I'm talking about. I know the advantages and disadvantages of the current c-9000 setup. I also know the advantages of having lower intensity charge and discharge pulses. What I'd like to learn first before contacting Maha is 1) Are there any disadvantages to a lower intensity and more frequent charge and discharge pulse? 2) What are the advantages of a less frequent higher intensity pulse?

If anyone knows please speak up. I don't want to request they change something that would actually make the charger worse.

[Yamabushi]

I suspect that the intensity of the pulses is based on the maximum charge and discharge rates. If they reduce the pulse intensity as you suggest, it probably means you lose the ability to charge at 2000 mA and discharge at 1000 mA.

If that's the case, I would oppose the change. I have devices that draw at 900 to 1000 mA so I test my batteries at those rates. Getting a higher test number is pointless if it does not reflect real world performance.

Note that the BC700 can only charge and discharge at about 1/3 the rate of the C9000 (maximum charge 700 mA and maximum discharge 350 mA). It produces higher test results because it can't test using the IEC method.

[Mr Happy]

As it stands, the C9000 uses a 2 second cycle and it applies charge or discharge pulses for up to 90% (approximately) of that time.

When charging it measures the voltage in the gap when the charging current is not applied. This is why the displayed voltage on the LCD is much lower than you would expect for the given charging current. For example if charging at 2 A the peak voltage on the cell should exceed 1.6 V at the end of charge but you never see that. The reason is that the cell voltage drops substantially when the charging current is removed and this no current voltage is what the charger measures. Unfortunately, what is wrong with this is that the charger is trying to do minus delta-V termination when charging--and you can only properly observe and measure the minus delta-V voltage drop if you measure the voltage while the charging current is applied. Since the C9000 does not measure the minus delta-V signal in the correct manner, it has trouble detecting end of charge and often misses the signal. To correct that shortcoming, the designers applied a "fix" where they impose a maximum cell voltage of 1.47 V and terminate when this is reached. For most good cells, most of the time, this will cause termination to occur before the peak voltage has been reached and gives the illusion of reliable termination. It's an ugly fudge, but it works OK with eneloops and other modern cells.

When discharging it measures the voltage in the gap when the discharging current is applied. What is wrong here is that heavy discharge pulses applied to cells with a higher internal resistance will cause the voltage to plummet and will lead to a termination of discharge before the cell is fully discharged. A better design would measure the voltage in the gaps between discharge pulses and set the termination voltage at a slightly higher value such as 1.0 V to compensate.

So it turns out that both charging and discharging methods have problems. It is certainly possible that these problems could be reduced if the cycle time was made shorter, for example 1 second or 0.5 seconds. This would more closely approximate a continuous charge and discharge current.

One may ask why the C9000 does not measure the charging voltage while current is applied, as would be the normal correct approach? A possible answer is the noise imposed on the voltage from the switching power supply. When viewed on an oscilloscope this is quite significant. It would need some decent filtering to smooth out this noise and recover the underlying cell voltage. Perhaps there was some reason they didn't make the design this way (cost or complexity)?

[Russel]

Here are some oscilloscope captures to help illustrate:


Above is the waveform of a MH-C9000 charging an Eneloop AA cell set to 1000mA. You can see the 2 second 2000mA charge pulse at about 45% duty cycle, averaging about 1000mA. (keep in mind that charging at 2000mA with at 2000mA pulse is still only 90% on and 10% off so that the charger can measure voltage during the off time.)



Above you can see the difference in voltage between the charge pulse and off time with the C9000 set to 1000mA charge current.



Above is the C9000 charging at 50mA in breakin mode. A very short charge pulse at 2000mA. It really shows the "noise" on the charge current.

I know you are talking about discharging, but when I read MrHappy's post I figured I would post these captures to help illustrate part of what he is talking about. I'll try and post some of the C9000 discharging when I get a chance.

[Viking]

This guy has tested both the BC-700 and Maha C9000 , and some other chargers as well.


It seems he is getting most out of the batteries in the BC-700

http://danaco.net/

[apagogeas]

This guy has tested both the BC-700 and Maha C9000 , and some other chargers as well.


It seems he is getting most out of the batteries in the BC-700

http://danaco.net/

Actually not really. He gets the capacity reading out of each charger, so basically BC-700 will always show a higher capacity compared to MC-9000 because of the way it measures the capacity; this doesn't mean BC-700 charges more fully the cells or in a better way somehow or the opposite. These data say absolutely nothing about the charging performance of each charger. The correct way to do this test would be to decide on only one charger to perform the discharge test (which one is irrelevant) and perform the charging individually on each with the recommended approach.

[Yamabushi]

I agree with apagogeas up to a point but there are a few more variables ...

For a fair comparison, you would also have to charge at the same 700 mA rate on the other 3 chargers.

Also, if you test the same batteries on each charger, you condition the batteries and, especially if they are new, their performance will improve in the subsequent test. You would have to test several cycles, e.g., C9000, BC-700, IQ328, BC-9009, C9000, BC-700, IQ328, BC-9009, .... until the conditioning effect is minimized or you would need to test different sets (statistically valid sample size) of batteries on the four chargers.

[Viking]

Actually not really. He gets the capacity reading out of each charger

True enough , i didn't get that.
It doesn't really say much then.

[LetThereBeLite]

True enough , i didn't get that.
It doesn't really say much then.

It seems to me his spreadsheet is actually quite useful. You just happened to interpret the data in a certain way. His data clearly shows all data--it's not hiding anything. The data allows one to compare one cell against another on all 4 chargers. I think the proper way to use his data is to compare one battery against another tested on the same charger, i.e. compare

Tenergy (Blue) AA 2600 tested on C9000 against

PowerEx AA 2700 tested on C9000 or

Tenergy (Blue) AA 2600 tested on BC700 and

PowerEx AA 2700 tested on BC700. The data speaks for itself but the one making the conclusion/interpolation could make some bad conclusions/interpretations. :-)

[apagogeas]

His test assumes that the real capacity of a battery will be somewhere between the lowest and highest capacity found, so he takes the average and assumes this would be more realistic than just using the capacity returned by any particular charger. Although it makes some sense, still it is of course a very rough estimate of the actual capacity, no better or worse than just picking the result of any charger. Still these data can't be used to conclude which charger does a better charging job.

[espresso]

As I promised, here is a comparison with new Varta Longlife Accu 800 mAh LSD AAA cells.



This time I chose higher discharge current on BC700 just to show the obvious. The batteries are new (cycled less than 10 times and they didn't measure much increase in capacity from the first charge - couple of mAh).

What I don't like is the big difference between two cells from the same pack. Is this ok even after several cycles?

[czAtlantis]

Did you run the discharge in the same charger's slot? Or you just simultaneously discharged batteries in different slots? I am asking because there might be inconsistency between charger's slots and slot 1 and slot 4 can give different readings even if those battereis are exactly same.

[espresso]

They were in different slots. But even before this test, one of the cells always had less capacity than the other, no matter what slot it was in. Just asking if this is acceptable. I mean, the capacity is more than 10% less than advertised on the pack.

[SilverFox]

Hello Espresso,

You are asking is it Black, or is it White.

It is Grey...

Normally you would look at the manufacturers specification sheet for the cell and see what it is supposed to be. Unfortunately, the specification sheets for Varta cells are very difficult to find.

The industrial AAA Varta cells are rated at 730 mAh if that helps, but I was unable to find a discharge graph for that to see how close they come to their labeled capacity.

Tom

[Yamabushi]

What I don't like is the big difference between two cells from the same pack. Is this ok even after several cycles?

I wouldn't worry about them if you are using them individually but if you're using them as a pair, they seem pretty mismatched (7 - 8% spread).

For comparison, a relatively new set of Duracell (Eneloop) AA 800 mAh tested on a C9000: 789, 785, 784, 779. Less than 1% spread when used in pairs in my Preon 2.

[Bones]

From William Chueu of MahaEnergy:

...

6) Capacity difference between the MH-C9000 and BC-900

BC-900, according to our measurement, tends to overestimate the capacity quite a bit (for instance, the Powerex 2700mAh capacity (1A charge/discharge) typically gets 2750mAh on the BC-900, when our laboratory-grade instrument consistently give about 2550mAh). This has to do with two things: 1) accuracy of the discharge current measurement, and 2) accuracy of the time base. The MH-C9000 has a 1% tolerance in the discharge current measurement. Furthermore, the MH-C9000 uses a quartz oscilattor time base (you can see them if you open the charger, I think there are some photos here on CPF) whereas the BC-900 uses a resistor-capacitor (RC) network as a time base which is less accurate.

I trust the BREAK-IN capacity on the MH-C9000. It typically is within 2% of the capacity measured on a calibrated battery analyzer.

I hope this answers some of the questions raised.

William

[espresso]

That's interesting. I guess he is right on this. However, I still doubt it's more accurate on cells with increased resistance. And some of them it won't test at all. Those might be basically dead, I know... but still.

@Yamabushi
Unfortunately, they're stuck together in a blood pressure measuring device