the lie of rapid NiMH self-discharge

[William Sommerwerck]

"Jeff Liebermann" <jeffl@cruzio.com> wrote in message
news:7pcde7pbvf17i2mh332l7mg49e0drms50j@4ax.com...

NiMH cells do NOT totally discharge themselves. They discharge to
about 50% of capacity (by self-discharge) and then just sit there.

Then what is the problem? A 50% loss of capacity is hardly disastrous.

For example, here are some curves I ran for an Energizer 2300ma-hr
cell:
http://802.11junk.com/jeffl/NiMH/Energizer-NiMH-2300.jpg
and for an Duracell 2050ma-hr cell:
http://802.11junk.com/jeffl/NiMH/Duracelll-NiMH-2050.jpg
There were brand new cells used to test the assertion that NiMH cells
need to be charge cycled a few times before they reach full capacity.
Basically, that's true.

Note the voltages (under 1C load). For the Energizer, it starts off
at 1.35v and starts to die at about 1.10v. If you put a load on your
Sanyo cells, I'm fairly sure your 1.23v no load voltage will drop to
something around 1.18v. In other words, your Sanyo batteries
self-discharged until they were almost dead, and stopped.

But they weren't "almost dead".

This morning, I did what I should have done before I posted. Over a period
of a half-hour, I fired the Canon 580EX II over 100 times AT FULL POWER. The
recycling time was 3 to 3.5 seconds (not unreasonable for full-power
recycling), and the final cell voltage was about 1.21V (before the cells had
time to recover).

100 full-power flashes is not "almost dead". Had I fired on automatic, at a
moderate aperture, not using full power, I could easily have gotten 200 or
300 flashes. And had I been willing to tolerate a 5-second recycle, I
probably could have gotten another 100 full-power flashes.

I recently had a similar experience to what you found. I have two old
Norelco rotary shavers. I received both with dead batteries and
replaced them with NiMH cells of dubious origin. I only charge them
when needed, except this time, when a dry squall dropped a few trees
through the power lines.
http://802.11junk.com/jeffl/pics/2011-12-03-Storm/
When I dug out the shavers, the LCD indicator showed a nearly full
charge (about 80%). However, when I tried to use it, both lasted less
than 60 seconds. My guess(tm) is that I hadn't charged it since last
winter (9 months).
That's fairly close to what you observed. There was sufficient
voltage for the LCD indicator to proclaim a nearly full charge, but
not enough stored energy to do much useful work.

That isn't what I observed this morning. The cells, which had sat for almost
two years since being fully charged, were perfectly usable for 100
full-power flashes. I probably have gotten another 100 flashes, but I didn't
want to take the chance of abusing the flash.

The cells are Sanyo Superlattice Alloy EVO, model HR-3U. They were not
shipped pre-charged, and as far as I can tell, they are not
slow-self-discharge ("eneloop") cells. The Sanyo USA site has no information
about them. However, they are available from Batteries America. (Thomas
Distributing doesn't list them.)

 

[William Sommerwerck]

http://store.batteriesamerica.com/hr-3u2700sanyo-brandultra-long-liferechargeableni-mhbatteries-2700mahcapacity.aspx

 

[Ian Field]

"Jeff Liebermann" <jeffl@cruzio.com> wrote in message
news:7pcde7pbvf17i2mh332l7mg49e0drms50j@4ax.com...

On Mon, 12 Dec 2011 04:26:03 -0800, "William Sommerwerck"
grizzledgeezer@comcast.net> wrote:

Four of the cells were 2700mAh Sanyo AAs. They all measured about 1.23V,
rather lower than the 1.4V NiMH cells commonly charge up to, but close to
the "nominal" 1.25V of NiMH and nicad cells. Contrary to Urban Legend,
they
were not completely discharged.

NiMH cells do NOT totally discharge themselves. They discharge to
about 50% of capacity (by self-discharge) and then just sit there.

ISTR NiCd cells frequently became shorted if allowed to self discharge too
often, but I haven't noticed it with NiMh.

 

[Jeff Liebermann]

On Tue, 13 Dec 2011 05:13:46 -0800, "William Sommerwerck"
<grizzledgeezer@comcast.net> wrote:

The cells are Sanyo Superlattice Alloy EVO, model HR-3U. They were not
shipped pre-charged, and as far as I can tell, they are not
slow-self-discharge ("eneloop") cells. The Sanyo USA site has no information
about them. However, they are available from Batteries America. (Thomas
Distributing doesn't list them.)

<http://www.houseofbatteries.com/documents/HR-3U-2500.pdf>
No mention of low self-discharge.

100 flashes does seem like an almost fully charged battery. However,
with such an intermittent load, it would have been nice if you had a
more controlled and more measurable way to test charge level. If what
you say is true, then you may be correct that there's no
self-discharge for a *NEW* NiMH battery. I'm wondering if it might be
a phenomenon that appears later as the battery is used. I'll run my
little discharge test when I have time and get my computer back home.

Incidentally, several people have mentioned that you cannot use the
open circuit voltage as an indication of state of charge. I agree. It
works with batteries that have sloping discharge curves, such as
carbon-zinc and some alkaline. However, batteries that have a flat
discharge curve offer such a small change in terminal voltage, that
the numbers are difficult to distinguish from tolerance and
temperature variations. The only way that seems to work for such
batteries are coulomb counter chips, as found in many laptop
batteries.

More later... bizzeeeeeee
--
# Jeff Liebermann 150 Felker St #D Santa Cruz CA 95060
# 831-336-2558
# http://802.11junk.com jeffl@cruzio.com
# http://www.LearnByDestroying.com AE6KS

 

[Phil Allison]

"Jeff Liebermann"

NiMH cells do NOT totally discharge themselves. They discharge to
about 50% of capacity (by self-discharge) and then just sit there.

** I have several sets of AA NiMH cells - all Sanyo brand and rated at 1700
to 2500 mAH.

Self discharge is a REAL problem, but no worse than with NiCds.

In general, charged cells lose 90% of capacity in about 6 to 8 weeks -
taking a full 6 hour charge at 400mA to recover.

In my Canon A430, the uncharged cells will run the back display for a few
minutes, let me take a few shots and maybe one flash before the camera shuts
down.


.... Phil

 

[Jeff Liebermann]

On Tue, 13 Dec 2011 05:13:46 -0800, "William Sommerwerck"
<grizzledgeezer@comcast.net> wrote:

The cells are Sanyo Superlattice Alloy EVO, model HR-3U. They were not
shipped pre-charged, and as far as I can tell, they are not
slow-self-discharge ("eneloop") cells. The Sanyo USA site has no information
about them. However, they are available from Batteries America. (Thomas
Distributing doesn't list them.)

This might be of interest:
<http://www.stefanv.com/electronics/sanyo_eneloop.html>
It's an independent test of the Sanyo Eneloop NiMH batteries including
some self-discharge tests. Unfortunately, he doesn't compare the
results with the non-LSD batteries. Scroll down to the "Self
Discharge" section heading and note the self-discharge tables.
50% charge loss after 1 year.

There is also this quote from Sanyo:
Storage temperature is of high importance if you measure
self-discharge rate. Higher temperatures substantially
increase self-discharging. It is best to store Eneloops
as cool as possible to keep the charge in the battery.
As a rule-of-thumb, every 10°C increase in storage
temperature is equivalent to doubling the storage time.
Some R/C pilots in Europe put Eneloops in the freezer,
with rather good results.
So, how were your batteries stored?

Presumably, the non-LSD batteries would produce much worse results,
making your miraculous Sanyo HR-3U cells better than Eneloop cells,
which seems rather dubious. Since the Eneloop batteries tested were
brand new, I don't think it's something related to an aging effect.
That leaves your test as an oddity. Are you sure someone didn't
charge your batteries when nobody was looking?



--
# Jeff Liebermann 150 Felker St #D Santa Cruz CA 95060
# 831-336-2558
# http://802.11junk.com jeffl@cruzio.com
# http://www.LearnByDestroying.com AE6KS

 

[Jeff Liebermann]

On Tue, 13 Dec 2011 19:45:07 -0800, Jeff Liebermann <jeffl@cruzio.com>
wrote:

This might be of interest:
http://www.stefanv.com/electronics/sanyo_eneloop.html
It's an independent test of the Sanyo Eneloop NiMH batteries including
some self-discharge tests. Unfortunately, he doesn't compare the
results with the non-LSD batteries.

Ok, I lied[1]. There is a comparison. See:
<http://www.stefanv.com/electronics/sanyo_eneloop.html#compare>
The graph shows the Eneloop battery to be MUCH better at
self-discharge than the conventional NiMH. After 1 year, the Eneloop
retained about 50% of charge, while the conventional NiMH retained
only 2.6%.

[1] I'm suppose to be doing my end of year billing and bookkeeping. I
hate doing billing and am therefore easily diverted. If I go broke,
it's all your fault for creating an interesting diversion.

--
# Jeff Liebermann 150 Felker St #D Santa Cruz CA 95060
# 831-336-2558
# http://802.11junk.com jeffl@cruzio.com
# http://www.LearnByDestroying.com AE6KS

 

[William Sommerwerck]

"Jeff Liebermann" <jeffl@cruzio.com> wrote in message
news:isofe7hekgeqpu76oohmlu2qsv82um84t4@4ax.com...

100 flashes does seem like an almost fully charged battery.

But it isn't, of course. The voltage was at the nominal 1.25V when I
started, but this is well below the end-of-charge voltage for NiMH cells.

However, with such an intermittent load, it would have been nice if you
had a more controlled and more measurable way to test charge level.
If what you say is true, then you may be correct that there's no
self-discharge for a *NEW* NiMH battery. I'm wondering if it might be
a phenomenon that appears later as the battery is used. I'll run my
little discharge test when I have time and get my computer back home.

It was "scientific" in the sense of mimicking photographic use. 100+
full-power shots in a half hour, three at a time, is fairly extreme use.

Incidentally, several people mentioned that you cannot use the
open-circuit voltage as an indication of state of charge. I agree.

Of course. That's one of the problems with nicad and NiMH cells.

However, the fact that the cells read about 1.25V showed they WERE NOT fully
charged. But despite having sat for two years, they were able to give 100+
full-power shots. The point is that their were perfectly usable without
having to be recharged. This contradicts belief that NiMH cells rapidly
self-discharge. (Rates of 1% or more per day are stated.)

I might disable the flash's auto-shutoff and let the cells run down to 1.0V,
then see whether it can still fire more than once. (I have no desire to keep
popping the flash.)

PS: Sanyo says their current eneloop cells (rated at 1500 charge cycles)
will hold as much of 75% of their charge for 3 years.

 

[William Sommerwerck]

"Phil Allison" <phil_a@tpg.com.au> wrote in message
news:9kq9n9F8d6U1@mid.individual.net...

"Jeff Liebermann"

NiMH cells do NOT totally discharge themselves. They discharge to
about 50% of capacity (by self-discharge) and then just sit there.

** I have several sets of AA NiMH cells - all Sanyo brand and rated at
1700
to 2500 mAH.

Self discharge is a REAL problem, but no worse than with NiCds.

In general, charged cells lose 90% of capacity in about 6 to 8 weeks -
taking a full 6 hour charge at 400mA to recover.

In my Canon A430, the uncharged cells will run the back display for a few
minutes, let me take a few shots and maybe one flash before the camera
shuts
down.

I wonder if the "spectacular" behavior of my 2700mAh Sanyos has anything to
do with their "superlattice alloy" construction. These appear to be the only
Sanyos using this design. (That's life, I guess.)

http://en.wikipedia.org/wiki/Superlattice

Anyone care to explain this article to me?

 

[William Sommerwerck]

"Jeff Liebermann" <jeffl@cruzio.com> wrote in message
news:j96ge7p5m7rq6jqg7da94igbr666srq504@4ax.com...

There is also this quote from Sanyo:
Storage temperature is of high importance if you measure
self-discharge rate. Higher temperatures substantially
increase self-discharging. It is best to store Eneloops
as cool as possible to keep the charge in the battery.
As a rule-of-thumb, every 10°C increase in storage
temperature is equivalent to doubling the storage time.
Some R/C pilots in Europe put Eneloops in the freezer,
with rather good results.
So, how were your batteries stored?

This is the typical "doubling of chemical reactions with each increase of
10°C" rule.

Right before Christmas 2009, I packed up my camera bag for a visit to Gold
Bar, WA. Two packs of NiMH cells were in the bag. When I came home, I put
the bag on the living room floor, where it has remained for almost two
years. The Pacific Northwest is cooler than the rest of the country. The
cells were exposed to "high" temperatures only intermittently, during the
warmer days of Summer.

Presumably, the non-LSD...

Hmmm... Is there a purple owl on these cells?

...batteries would produce much worse results, making your
miraculous Sanyo HR-3U cells better than Eneloop cells,
which seems rather dubious.

Agreed, but we don't know whether the "superlattice alloy" has magical
powers.

Since the Eneloop batteries tested were
brand new, I don't think it's something related to an aging effect.
That leaves your test as an oddity. Are you sure someone didn't
charge your batteries when nobody was looking?

Well, it was around Christmas... Elves, perhaps? Brownies?

Speaking of which... One of Stan Freberg's less-well-known recordings is
"Yulenet", with Joe Friday trying to convince a doubter named Grudge that
there really is a Santa Claus. When they visit the North Pole, they're
greeted by a brownie from the South Pole (Daws Butler doing a silly
mock-Southern voice) who's helping out while Santa is away. I can imagine
the flap such a joke would cause today...

 

[William Sommerwerck]

"Jeff Liebermann" <jeffl@cruzio.com> wrote in message
news:907ge75n0maqfu0l93mevdpr2o6bdmgun7@4ax.com...

On Tue, 13 Dec 2011 19:45:07 -0800, Jeff Liebermann <jeffl@cruzio.com
wrote:

This might be of interest:
http://www.stefanv.com/electronics/sanyo_eneloop.html
It's an independent test of the Sanyo Eneloop NiMH batteries including
some self-discharge tests. Unfortunately, he doesn't compare the
results with the non-LSD batteries.

Ok, I lied[1]. There is a comparison. See:
http://www.stefanv.com/electronics/sanyo_eneloop.html#compare
The graph shows the Eneloop battery to be MUCH better at
self-discharge than the conventional NiMH. After 1 year, the Eneloop
retained about 50% of charge, while the conventional NiMH retained
only 2.6%.

[1] I'm suppose to be doing my end of year billing and bookkeeping. I
hate doing billing and am therefore easily diverted. If I go broke,
it's all your fault for creating an interesting diversion.

Extremely interesting.

I'll contact Sanyo and ask them about the self-discharge of my specific
cells.

 

[josephkk]

On Tue, 13 Dec 2011 19:45:07 -0800, Jeff Liebermann <jeffl@cruzio.com>
wrote:

On Tue, 13 Dec 2011 05:13:46 -0800, "William Sommerwerck"
grizzledgeezer@comcast.net> wrote:
The cells are Sanyo Superlattice Alloy EVO, model HR-3U. They were not
shipped pre-charged, and as far as I can tell, they are not
slow-self-discharge ("eneloop") cells. The Sanyo USA site has no information
about them. However, they are available from Batteries America. (Thomas
Distributing doesn't list them.)

This might be of interest:
http://www.stefanv.com/electronics/sanyo_eneloop.html
It's an independent test of the Sanyo Eneloop NiMH batteries including
some self-discharge tests. Unfortunately, he doesn't compare the
results with the non-LSD batteries. Scroll down to the "Self
Discharge" section heading and note the self-discharge tables.
50% charge loss after 1 year.

There is also this quote from Sanyo:
Storage temperature is of high importance if you measure
self-discharge rate. Higher temperatures substantially
increase self-discharging. It is best to store Eneloops
as cool as possible to keep the charge in the battery.
As a rule-of-thumb, every 10°C increase in storage
temperature is equivalent to doubling the storage time.
Some R/C pilots in Europe put Eneloops in the freezer,
with rather good results.
So, how were your batteries stored?

Presumably, the non-LSD batteries would produce much worse results,
making your miraculous Sanyo HR-3U cells better than Eneloop cells,
which seems rather dubious. Since the Eneloop batteries tested were
brand new, I don't think it's something related to an aging effect.
That leaves your test as an oddity. Are you sure someone didn't
charge your batteries when nobody was looking?

In for a penny's worth; i add that this might be some freaky low self
discharge cells for the given process, perhaps at some process corner for
the set.

?-)

 

[William Sommerwerck]

I'll try to make this quick...

Canon's specs for the 580EX II flash appear to be based on nicad or NiMH
cells starting at 1.25 volts. As I explained, when the unloaded voltage was
at 1.21V, I had no trouble getting more than 100 full-power flashes, which
meets the 100 - 700 flash spec in the book. After letting the flash sit,
running, for several hours, the unloaded voltage was about 1.18V. After
removing and replacing the cells, the flash charged up once, taking more
than 7 seconds. After firing it, it would not recycle.

Several points... The cells had sat for two years, but delivered at least
the spec'd number of flashes. Some NiMH cells might self-discharge quickly,
but these Sanyos did not. (I was surprised that, throughout the discharge,
the cells' voltages were virtually identical, never differing by more than
about 10mV. This suggests very tight manufacturing tolerances.)

It was also interesting that the flash "conked out" well before the cells
reached 1.0V. This suggests that this flash is /not/ designed to work down
to 1V per cell, the traditional "standard" of battery-operated designs.

 

[Phil Allison]

"William Sommerwanker"

Canon's specs for the 580EX II flash appear to be based on nicad or NiMH
cells starting at 1.25 volts.

** The manual clearly states that it is based on Alkaline cells.

This implies that the unit will operate with much lower cell voltages and
currents than modern digital cameras require.

As I explained, when the unloaded voltage was
at 1.21V,

** It is totally meaningless to quote unloaded voltages of MiNH, NiCd or
alkaline cells.

Only when loaded as in the application does terminal voltage become
meaningful.



.... Phil

 

[who where]

On Fri, 16 Dec 2011 05:11:33 -0800, "William Sommerwerck"
<grizzledgeezer@comcast.net> wrote:

I'll try to make this quick...

Canon's specs for the 580EX II flash appear to be based on nicad or NiMH
cells starting at 1.25 volts. As I explained, when the unloaded voltage was
at 1.21V, I had no trouble getting more than 100 full-power flashes, which
meets the 100 - 700 flash spec in the book. After letting the flash sit,
running, for several hours, the unloaded voltage was about 1.18V. After
removing and replacing the cells, the flash charged up once, taking more
than 7 seconds. After firing it, it would not recycle.

Several points... The cells had sat for two years, but delivered at least
the spec'd number of flashes.

" at least the MINIMUM spec'd number of flashes" - which you would
expect given (a) the expanse of the range specified (100-700) and (b)
the proximity of your cells' voltage to the Canon staring point.

Some NiMH cells might self-discharge quickly,
but these Sanyos did not. (I was surprised that, throughout the discharge,
the cells' voltages were virtually identical, never differing by more than
about 10mV. This suggests very tight manufacturing tolerances.)

Sanyo is probably the most highly regarded name in both NiCd and NiMH
manufacture. You shouldn't be surprised.

It was also interesting that the flash "conked out" well before the cells
reached 1.0V. This suggests that this flash is /not/ designed to work down
to 1V per cell, the traditional "standard" of battery-operated designs.

 

[William Sommerwerck]

Several points... The cells had sat for two years, but delivered at least
the spec'd number of flashes.

" at least the MINIMUM spec'd number of flashes" - which you would
expect given (a) the expanse of the range specified (100-700) and (b)
the proximity of your cells' voltage to the Canon staring point.

100 is for a full-power flash. The 700 refers to a partial-power flash,
under auto-exposure.

 

[Phil Allison]

"William Sommerwanker is Full of Shit "

100 is for a full-power flash.

** The flash energy input is probably about 20 Joules, ie 330uF and 350V.

Allowing 5 Joules for losses, the energy required for 100 flashes is 2500
Joules.

A fully charged, 2500mAH NiMH has a capacity of 11,000 Joules ( 1.2 x 2.5 x
3600 )

Four of them have a capacity of 44,000 Joules

So, your "magic" Sanyo cells had under 6% of normal capacity.



.... Phil

 

[mike]

who where wrote:

On Fri, 16 Dec 2011 05:11:33 -0800, "William Sommerwerck"
grizzledgeezer@comcast.net> wrote:

I'll try to make this quick...

Canon's specs for the 580EX II flash appear to be based on nicad or NiMH
cells starting at 1.25 volts. As I explained, when the unloaded voltage was
at 1.21V, I had no trouble getting more than 100 full-power flashes, which
meets the 100 - 700 flash spec in the book. After letting the flash sit,
running, for several hours, the unloaded voltage was about 1.18V. After
removing and replacing the cells, the flash charged up once, taking more
than 7 seconds. After firing it, it would not recycle.

Several points... The cells had sat for two years, but delivered at least
the spec'd number of flashes.

" at least the MINIMUM spec'd number of flashes" - which you would
expect given (a) the expanse of the range specified (100-700) and (b)
the proximity of your cells' voltage to the Canon staring point.

Some NiMH cells might self-discharge quickly,
but these Sanyos did not. (I was surprised that, throughout the discharge,
the cells' voltages were virtually identical, never differing by more than
about 10mV. This suggests very tight manufacturing tolerances.)

Sanyo is probably the most highly regarded name in both NiCd and NiMH
manufacture. You shouldn't be surprised.

It was also interesting that the flash "conked out" well before the cells
reached 1.0V. This suggests that this flash is /not/ designed to work down
to 1V per cell, the traditional "standard" of battery-operated designs.

Make sure you're comparing apples with apples.
The unloaded voltage of a cell is irrelevant.
I've found it very difficult to get the unloaded voltage of NiMH below
1.2V. Discharge it down to .8V, remove the load and let it sit and
it will creep back up to 1.2V. But it's still dead and can't supply
much current.
A flash is a VERY high current device. Once the LOADED voltage gets
much below 1V, it's too weak for a flash. The ONLY useful voltage
measurement is with the intended load.

A useful measurement is internal resistance. Use a square-wave load from
1/2A to 1A. Measure the P-P amplitude of the cell voltage and use that
to calculate a resistance dV/dI. Try it at different states of charge.

Calculate the voltage drop from your load current and the ISR.
Multiply that by the number of series cells and it's easy to see
why high-current loads quit working long before the open-circuit
voltage gets below 1.2V.

 

[William Sommerwerck]

"Phil Allison" <phil_a@tpg.com.au> wrote in message
news:9l2gmuFjfeU1@mid.individual.net...

"William Sommerwanker is Full of Shit "

Phil Allison is a foul-mouthed shmuck.

By the way, in German "shmuck" means "jewelry" or "adornment". Billy Wilder
gets a funny gag out of this in "One, Two, Three".

100 is for a full-power flash.

The flash energy input is probably about 20 joules, ie 330uF
and 350V. Allowing 5 joules for losses, the energy required for
100 flashes is 2500 joules.

The losses are probably greater than that. When I left the flash running,
without firing it, I was surprised that it conked out after about two hours.

I suspect this flash provides more than 20Ws output. But the specs are
silent on this.

A fully charged, 2500mAH NiMH has a capacity of 11,000 joules
(1.2 x 2.5 x 3600). Four of them have a capacity of 44,000 joules.
So, your "magic" Sanyo cells had under 6% of normal capacity.

I can't argue with plausibly-chosen numbers, and I won't. However...

You're still missing the point. We have been told that NiMH cells lose
several percent of their capacity every week. (Let's say 3%, and assume it's
a linear loss, rather than exponential.) After 102 weeks, the cells should
have been dead, dead, dead. They were not. They had no trouble powering the
flash to its spec'd number of full-power flashes. What do you want, for
heaven's sake?

The point about "voltage creep" of nicads & NiMH cells was well-taken.
However, I measured the cells' voltages within about 15 seconds of shutting
off the flash. Furthermore, the under-load voltage is arguably not that
important if the DUT works as it's spec'd. Which this flash did.

Please note that I actually performed an experiment! I tested cells that
should have been useless. They were not. The statement that conventional
NiMH cells are generically incapable of holding a charge for extended
periods is simply NOT TRUE. Not because "I say so", but because I have
empirical evidence.

Yesterday I pulled out my Sunpak 622 Super, a "professional" potato-masher
flash. It has four 5500mAh NiMH C cells made by CTA, whoever that is. (They
came from Overstock.) The last time I charged them was about six months ago.
I flicked the power switch, and unit came to full power in 6 seconds.
Full-power recycling was 4 seconds -- not great, but not bad, either,
especially for "dead" cells..

 

[William Sommerwerck]

"mike" <spamme9@gmail.com> wrote in message
news:jch2n5$n5e$1@dont-email.me...

Make sure you're comparing apples with apples.
The unloaded voltage of a cell is irrelevant.
I've found it very difficult to get the unloaded voltage of NiMH below
1.2V. Discharge it down to .8V, remove the load and let it sit and
it will creep back up to 1.2V. But it's still dead and can't supply
much current.
A flash is a VERY high current device. Once the LOADED voltage gets
much below 1V, it's too weak for a flash. The ONLY useful voltage
measurement is with the intended load.

A useful measurement is internal resistance. Use a square-wave load from
1/2A to 1A. Measure the P-P amplitude of the cell voltage and use that
to calculate a resistance dV/dI. Try it at different states of charge.

Calculate the voltage drop from your load current and the ISR.
Multiply that by the number of series cells and it's easy to see
why high-current loads quit working long before the open-circuit
voltage gets below 1.2V.

I shouldn't have said anything about the voltage.

The point is that the cells "should" have been dead, but weren't. After
nearly two years, they powered the flash to its spec's number of flashes.