Charging NiMH Bank

[tecnodave]

Particle,

Have you looked into the efficiency of the cells themselves........I'm talking about how much you put in to how much you get out? NiCad and NiMh are not very good at that being less than 60 % at best for the C cell form factor. They have very good power to weight ratio, that is why they are used in power tools. And very fast charge times....45 minutes for a 2.3 a.h. 12 v. NiMh.

  Compare that charge efficiency to 95% or so for a very high quality Flooded Lead Acid such as the Rolls Surette. And maybe even better for Lithium Ferro Phosphate.

I am only using NiCad for my older power tools like the Panasonic for which there are no NiMh available. All my Makita now have NiMh where available. They are environmentally cleaner and have better power to weight ratio but they need to be cycled.......you cannot use only 10% of the charge and pick it up weeks later.......it will be dead and won't charge........you need to run them down and fully charge them to full every week or you will have dead cells months.


td

[Particle]

Charge efficiency for NiMH depends largely on the rate of charge.  At 0.1C it's only around 60% but by 0.5C it's around 95%.  They don't suffer from the memory effect you're describing but they do have relatively high self-discharge.  The discharge isn't terribly concerning though given that they wouldn't be used to store power longer than a day.

Can the Kid be made to stop charging based on battery temperature with a custom set point?  If I could set it for say 50C and install a probe, that'd be an easy way to keep from overcharging regardless of the input current.

[zoneblue]

If the kind of LSD technology involved in the AA eneloops ever makes into bigger cells, that would be kind of cool. Very very high charge efficiencys there.

On balance what i would do is chart out a charge curve for your particular cells, at your typical charge current. Locate the point just prior to peak voltage, and use that as your absorb setpoint. Set absorb to 1 minute, and float to something low enough to get the charge acceptance down to 0.01C.

Repeating my point above, waiting for the cell to heat up is a mistake, never mind that heat reaching a large poorly heat bonded sensor like that. The internal pressure rises exponentially, starting from somewhere around the point where delta v/delta t peaks.

See http://batteryuniversity.com/learn/article/charging_nickel_based_batteries

[Particle]

That's an interesting idea.  If I did charge at say 58V, that should be sufficiently before the peak that it would be safe.  My concern though is that at lower charge currents that peak will occur at lower voltages.

[Particle]

Interesting.  These cells I'm charging are actually cooling with respect to ambient.  I wonder if these cells are actually NiCd with a NiMH wrapper on them.  Crafty Chinese importers.  heh

[Particle]

It has been a while since I last posted on this topic, but I haven't abandoned the project.  I have however abandoned the generic Chinese cells and have instead settled on Tenergy 4/3 AF NiMH cells instead.  I have done some experiments with 40 of the new Tenergy cells configured as 20S2P (20 in series, 2 series in parallel).  I've got another 60 cells coming in a week's time since the pilot has progressed well.  Currently, the reserve capacity is small at about 180 Wh, but that will increase to 450 Wh with the extra cells.

I've learned much about how NiMH chemistry behaves during charge by having manually charged the prototype pack a few times.  The first is that I'm not sure that there is a suitable CV (constant voltage) or CC (constant current) target that can be set if the goal is to do a fast charge.  The reason for this is that charge current doesn't naturally taper off at the end of the charge cycle.  I'll describe what instead happens.  If both a CV (1.425 Vpc) and CC (1C) limit are set (and they should both be), the pack will be limited by the CC limit for the first 60% or so of the charge.  As it charges, the voltage gradually increases until it hits the CV limit.  Once hit, charge current will gradually decrease through about the next 20% of the charge until it hits a minimum of around C/10-C/12.  After this point, the pack will begin to warm for the first time and will do so quickly.  Through the high-current and tapering-off parts of the charge cycle, the pack temperature will only have risen a few degrees.  This increasing temperature combined with the dropping current is indicative of the pack having reached about an 80-85% state of charge.  The tricky, unfortunate bit is that as the temperature rises the pack's internal resistance begins to drop again, and it will start to accept greater charge current even though it has been at the CV limit for quite some time without changing.  Charge current will quickly begin to exceed safe limits if left untended, so it's necessary to reduce the CC limit to C/10 or below once this stage has been hit.

A solar charge controller would seem to need to charge this chemistry somewhere along these lines:

Temperature < (Room + 15C)
- Voltage < 1.400 Vpc : Charge at 1.425 Vpc or 1C max (fast charge)
- Voltage < 1.425 Vpc : Charge at 1.425 Vpc or C/100 max (maint charge)
Temperature > (Room + 15C) & < ((Room + 25C) or 55C)
- Voltage < 1.400 Vpc : Charge at 1.425 Vpc or C/10 max (slow charge)
- Voltage < 1.425 Vpc : Charge at 1.425 Vpc or C/100 max (maint charge)
Temperature > ((Room + 25C) or 55C)
- Do not charge until temperature falls to < (Room + 20C) or 50 C
All
- Do not switch to a higher-current charge mode within 5-10 minutes of having switched to a lower-current charge mode so as to prevent oscillating when at an edge.
- Do not charge when voltage exceeds 1.450 Vpc until pack voltage drops to 1.400 Vpc or below.

Correctly charging the pack with a variable load attached would still be possible I believe, but the temperature probe would be critical.

I've also given thought to what is essentially a bypass system for the battery pack with discrete charge and discharge ports.  The charge side could allow or limit the charge current itself based on the rules above and cut off the charge port when out of the defined operating area.  The discharge side would have no such regulation but have a diode and fuse in line.  The charge controller could then also be attached to the discharge side after the diode, perhaps with an EDLC capacitor in parallel.  This would allow the batteries to charge when it's safe to do so and the charge controller to continue to contribute extra current to the system when more than the pack's charge current is available from solar and the load exceeds that charge current.

A preferred alternative would be to use no bypass system but instead have a charge controller that compensates for a variable load by continuously reading the voltage drop across a standard shunt and adding the instantaneous discharge current to the charge targets.

---

The graph below shows the condition that a charge controller would need to avoid.  When charging at a constant voltage, charge current will eventually start to creep back up.  Charging past this point (current minimum) can be dangerous unless the charge cycle is being actively managed by a person or controller.  Left to themselves the batteries will self-induce failure due to internal pressure and temperatures exceeding safe operating limits.  Prior to I-min, the batteries do not warm appreciably even at high charge currents of 1C.

This graph shows me manually controlling a charge cycle.  The cycle starts with a mostly depleted battery and charge conditions of about 28.5 CV and 7.4 CC.  Once I saw I-min had been passed and temperatures had risen to about 52C, I turned down the CC limit to about 0.8 A before dropping it further to 0.6 A.  This continued for a while as a means to finish the charge and keep the pack in balance.