https://www.ibm.com/docs/en/announcements/archive/ENUSZG91-0116

Bench-Top (B31) and a Rack Mount (R31)

Both models come standard with a 80286/10MHz processor
30 byte (MB) hardfile
one 3.5 Inch 1.44 million byte (MB) Diskette Drive
180W Battery-Backed PSU (requires installation of battery feature #6669)
1MB memory
four (4) customer available I/O expansion slots

MONEY SHOT !

"The IBM 7533 Industrial Computer Model B31 and Model R31 are available
in E/ME/A only. "

No wonder I've never seen or heard of one.

180 Watt Battery Backup * 7541, 7542 6669 57F0842
7561, 7562
180 Watt Power Supply ** 7541, 7542 6738 57F2725
with battery backup 7561, 7562
(Field Installed)
* 7541/42 or 7561/62 above s/n 5502000 order FC6669 PN57F0842.
** 7541/42 or 7561/62 below s/n 5502000 order FC6738 PN57F2725.

I wunder if PN57F2725 is the PSU + battery pack. It would be neat-o
daddy-O to findt out the construction of the battery pack.

"continuous operation through brief main-power interruptions (nominal
default of 10 seconds)"

So the expected normular use of the battery pack is only expeculated to
last 10 seconds. That is not the total capacity. Of course, the run time
on the backup battery depends on the installed options.

There appears to be a PCB that fits over the power header on the system
board, it -MIGHT- convert the 8570-ish output to any required 8530-286
voltages. Dunno. I put our man in Berlin on this. My two semesters of
college German are only gut enough to get my face slapped in a bar...

An internally mounted battery allows continuous operation through brief
main power interruptions (up to 10 (+/- 10%) seconds at full load for a
maximum of one occurrence per hour). The battery is recharged
automatically when the system is connected to normal AC power.

Louis Ohland wrote:
> "continuous operation through brief main-power interruptions (nominal
> default of 10 seconds)"
>
> So the expected normular use of the battery pack is only expeculated to
> last 10 seconds. That is not the total capacity. Of course, the run time
> on the backup battery depends on the installed options.

How big...

"Full Load" is 180W... So we can pull 10 seconds of full load.

How does this equate to Watt Hours?

Louis Ohland wrote:
> An internally mounted battery allows continuous operation through brief
> main power interruptions (up to 10 (+/- 10%) seconds at full load for a
> maximum of one occurrence per hour). The battery is recharged
> automatically when the system is connected to normal AC power.
>
> Louis Ohland wrote:
>> "continuous operation through brief main-power interruptions (nominal
>> default of 10 seconds)"
>>
>> So the expected normular use of the battery pack is only expeculated
>> to last 10 seconds. That is not the total capacity. Of course, the run
>> time on the backup battery depends on the installed options.

How much voltage is used to charge a NiCd battery? The charger is rated
for 26v, but what is the battery voltage?

Louis Ohland wrote:
> An internally mounted battery allows continuous operation through brief
> main power interruptions (up to 10 (+/- 10%) seconds at full load for a
> maximum of one occurrence per hour). The battery is recharged
> automatically when the system is connected to normal AC power.
>
> Louis Ohland wrote:
>> "continuous operation through brief main-power interruptions (nominal
>> default of 10 seconds)"
>>
>> So the expected normular use of the battery pack is only expeculated
>> to last 10 seconds. That is not the total capacity. Of course, the run
>> time on the backup battery depends on the installed options.

More guesstimation.

NiCd charges between 1.4 to 1.5 v per cell. The efficiency of the charge
is @ 83%.

So... 26 / 1.4 is @ 18. 18 * .83 is just shy of 15.

So... Maybe the 180W PSU battery is a 3x5 pack?

Louis Ohland wrote:
> How much voltage is used to charge a NiCd battery? The charger is rated
> for 26v, but what is the battery voltage?
>
> Louis Ohland wrote:
>> An internally mounted battery allows continuous operation through
>> brief main power interruptions (up to 10 (+/- 10%) seconds at full
>> load for a maximum of one occurrence per hour). The battery is
>> recharged automatically when the system is connected to normal AC power.
>>
>> Louis Ohland wrote:
>>> "continuous operation through brief main-power interruptions (nominal
>>> default of 10 seconds)"
>>>
>>> So the expected normular use of the battery pack is only expeculated
>>> to last 10 seconds. That is not the total capacity. Of course, the
>>> run time on the backup battery depends on the installed options.

Louis Ohland wrote:
> More guesstimation.
>
> NiCd charges between 1.4 to 1.5 v per cell. The efficiency of the charge
> is @ 83%.
>
> So... 26 / 1.4 is @ 18. 18 * .83 is just shy of 15.
>
> So... Maybe the 180W PSU battery is a 3x5 pack?

Probably not NiCd cells, sealed lead acid in a 24V stack (12 cells)
would be more likely.

The back up battery in a 7568 Gearbox is NiCd.

https://www.ardent-tool.com/50_70/Power.html#7561_PSU

"There is a 9.25" (W) x 1.75" (H) hinged cover on the lower outer side
of the PSU. It is fastened at the front by a single captive standard
screw. The cover pivots open on the rear hinges, revealing a three pin
Molex power connector, like a female drive power connector. The battery
cavity is 1.875" deep."

https://i.ebayimg.com/images/g/YiQAAOSwtbViM62W/s-l1600.jpg

So... you think a sealed lead acid battery of 12 cells will fit?

Please give me an example, that will most certainly help...

Rick Ekblaw wrote:
> Louis Ohland wrote:
>> More guesstimation.
>>
>> NiCd charges between 1.4 to 1.5 v per cell. The efficiency of the
>> charge is @ 83%.
>>
>> So... 26 / 1.4 is @ 18. 18 * .83 is just shy of 15.
>>
>> So... Maybe the 180W PSU battery is a 3x5 pack?
>
> Probably not NiCd cells, sealed lead acid in a 24V stack (12 cells)
> would be more likely.
>

Louis Ohland wrote:

> "There is a 9.25" (W) x 1.75" (H) hinged cover on the lower outer side
> of the PSU. It is fastened at the front by a single captive standard
> screw. The cover pivots open on the rear hinges, revealing a three pin
> Molex power connector, like a female drive power connector. The battery
> cavity is 1.875" deep."
>
> https://i.ebayimg.com/images/g/YiQAAOSwtbViM62W/s-l1600.jpg
>
> So... you think a sealed lead acid battery of 12 cells will fit?
>
> Please give me an example, that will most certainly help...

Sure, you can make a SLA battery pack into most any shape you want if
you're not opposed to creating a non-standard design point -- and IBM is
famous for "going their own way", especially in their earlier days. The
key here is that the battery pack needs to put out 180W of power for 10
or 11 seconds, then it gets recharged to be ready for the next power
dip. That's a poor usage case for NiCD, which suffers from the memory
effect if you don't discharge slowly and most of the way towards full
discharge, but lead-acid can dump out power quickly and cycle often,
though doing that lowers the pack lifetime, it would probably need to be
replaced every 3 years or so.

The sort of SLA battery that most people are familiar with today would
be a 1290 pack (12V, 9Ah) which is commonly used in UPS units and
electric scooters for the disabled/elderly. The dimensions on the 1290
are roughly 5 3/4" x 2 1/2" x 3 3/4", so that's not going to fit in the
PSU compartment, but it also has way more energy capacity than is needed
for this application (and you'd need two of them in series to get to
24V). A small 24V stack with 5% of the capacity of a 1290 could be made
as thin as 1 inch, though the thicker you make the plates the more
capacity you'll have available and the pack will last longer.

And IBM made custom SLA pack configurations on many "big iron" products,
like the mainframes and storage boxes. Those packs were huge, heavy and
expensive, but of course they were designed to provide a lot more power
for a longer period of time so that the "big iron" could shut down
gracefully when the power went out. But even the smaller DS5020 storage
boxes had a small battery pack that could make sure that the write cache
was written to DASD when a power outage struck, though that pack used
lithium cells. In today's designs, lithium is going to beat lead-acid
at least 90% of the time. A lot of IBM's ServeRAID adapters also had a
lithium backup battery pack, and all of the crypto adapters had backup
batteries -- all of which were custom IBM designs.

My brain hurts from the charge rates for NiCd.

I must purify myself in the waters of Lake Minnetonka...

"1290 pack (12V, 9Ah)" now here's an actual item to look for, and I >ASS
U ME< there should be plenty of erudite discussions.

Rick Ekblaw wrote:
> The key here is that the battery pack needs to put out 180W of power for
> 10 or 11 seconds, then it gets recharged to be ready for the next power
> dip.  That's a poor usage case for NiCD, which suffers from the memory
> effect if you don't discharge slowly and most of the way towards full
> discharge, but lead-acid can dump out power quickly and cycle often,
> though doing that lowers the pack lifetime, it would probably need to be
> replaced every 3 years or so.

9.25" (W) x 1.75" (H) 1.875" (D)

Battery pack will be shorter, there's a three pin Molex plug and a short
length of cable in there...

Louis Ohland wrote:
> My brain hurts from the charge rates for NiCd.
>
> I must purify myself in the waters of Lake Minnetonka...
>
> "1290 pack (12V, 9Ah)" now here's an actual item to look for, and I >ASS
> U ME< there should be plenty of erudite discussions.
>
> Rick Ekblaw wrote:
>> The key here is that the battery pack needs to put out 180W of power
>> for 10 or 11 seconds, then it gets recharged to be ready for the next
>> power dip.  That's a poor usage case for NiCD, which suffers from the
>> memory effect if you don't discharge slowly and most of the way
>> towards full discharge, but lead-acid can dump out power quickly and
>> cycle often, though doing that lowers the pack lifetime, it would
>> probably need to be replaced every 3 years or so.

9.25" (L) x 1.75" (W) 1.875" (D)

This sounds better...

Louis Ohland wrote:
> My brain hurts from the charge rates for NiCd.
>
> I must purify myself in the waters of Lake Minnetonka...
>
> "1290 pack (12V, 9Ah)" now here's an actual item to look for, and I >ASS
> U ME< there should be plenty of erudite discussions.
>
> Rick Ekblaw wrote:
>> The key here is that the battery pack needs to put out 180W of power
>> for 10 or 11 seconds, then it gets recharged to be ready for the next
>> power dip.  That's a poor usage case for NiCD, which suffers from the
>> memory effect if you don't discharge slowly and most of the way
>> towards full discharge, but lead-acid can dump out power quickly and
>> cycle often, though doing that lowers the pack lifetime, it would
>> probably need to be replaced every 3 years or so.

Rick, what would be the minimum rated Watt-Hour for the battery pack?
180W over 10 seconds?

Assuming a 12v battery pack...

Louis Ohland wrote:
> 9.25" (L) x 1.75" (W) 1.875" (D)
>
> This sounds better...
>
> Louis Ohland wrote:
>> My brain hurts from the charge rates for NiCd.
>>
>> I must purify myself in the waters of Lake Minnetonka...
>>
>> "1290 pack (12V, 9Ah)" now here's an actual item to look for, and I
>> >ASS U ME< there should be plenty of erudite discussions.
>>
>> Rick Ekblaw wrote:
>>> The key here is that the battery pack needs to put out 180W of power
>>> for 10 or 11 seconds, then it gets recharged to be ready for the next
>>> power dip.  That's a poor usage case for NiCD, which suffers from the
>>> memory effect if you don't discharge slowly and most of the way
>>> towards full discharge, but lead-acid can dump out power quickly and
>>> cycle often, though doing that lowers the pack lifetime, it would
>>> probably need to be replaced every 3 years or so.

Louis Ohland wrote:
> Rick, what would be the minimum rated Watt-Hour for the battery pack?
> 180W over 10 seconds?
>
> Assuming a 12v battery pack...
Louis, your battery engineer is just going to shake his head sadly at
you, as the watt-hour rating for this battery pack is going to be
irrelevant because you're not looking for it to provide continuous power
for even a minute, much less an hour. Since you want to assume a 12V
pack instead of my suggested 24V, to get to 180W you need 15A of current
coming out of that pack for the 10 or 11 seconds required -- actually a
bit more because you will have some losses involved inverting the DC
power output into AC to back-feed the PSU -- unless the battery backup
is just used to feed the DC section of the PSU, which might be the way
IBM did it.

As I think about it, feeding the DC section of the PSU is more likely
than doing the AC inversion. Since the highest output voltage is 12V, a
12V battery pack would be just fine. But you said the charger is rated
for 26V, so that would suggest a 24V battery pack. It's all getting
complicated, for sure. But if you pressed your battery engineer for a
watt-hour rating for this hypothetical battery pack, it would probably
be in the high single digits, maybe as much as 10 -- certainly nowhere
close to the 100+ rating of the 1290 12V SLA battery pack.

Louis Ohland wrote:
> Rick, what would be the minimum rated Watt-Hour for the battery pack?
> 180W over 10 seconds?
>
> Assuming a 12v battery pack...

So after all of that thinking and postulating, I went back and looked at
the Announcement Letter and found:

Battery, Nicad (Internal) 6669 57F0842

so it's a NiCD pack after all. I'll bet that using it just a few times
to supply power during brief power outages killed the pack, those cells
would get HOT.

Rick, the simplest [and slowest!] rate is c/1, a lot SIMMpler control
circuit and it makes the least heat. I've seen c/5 mentioned, much
shorter time, but more heat and it slaps the battery pack like the
naughty boy it is.

So... if you got to SEE the battery charging circuit, could you
guesstimate capacity?

The pittances we have of 26v and 40mA could mean whatever you wandt.
Measuring the output connector in the battery well -MIGHT- tell us the
actual charging voltage, which could be less than 26v. Maybe the charger
input is 26v at 40mA?

Rick Ekblaw wrote:
> Louis Ohland wrote:
>> Rick, what would be the minimum rated Watt-Hour for the battery pack?
>> 180W over 10 seconds?
>>
>> Assuming a 12v battery pack...
>
> So after all of that thinking and postulating, I went back and looked at
> the Announcement Letter and found:
>
> Battery, Nicad (Internal)                        6669  57F0842
>
> so it's a NiCD pack after all.  I'll bet that using it just a few times
> to supply power during brief power outages killed the pack, those cells
> would get HOT.
>

Louis Ohland wrote:
> Rick, the simplest [and slowest!] rate is c/1, a lot SIMMpler control
> circuit and it makes the least heat. I've seen c/5 mentioned, much
> shorter time, but more heat and it slaps the battery pack like the
> naughty boy it is.
>
> So... if you got to SEE the battery charging circuit, could you
> guesstimate capacity?
>
> The pittances we have of 26v and 40mA could mean whatever you wandt.
> Measuring the output connector in the battery well -MIGHT- tell us the
> actual charging voltage, which could be less than 26v. Maybe the charger
> input is 26v at 40mA?

OK, let's go back for another round of the guessing game. If IBM did
not go for custom-built NiCD cells, you have a few common cell sizes to
choose from: C size, sub-C size, AA size, AAA size, VARTA size pucks (3
cells stacked that are roughly a dime diameter, a bit less than an inch
long), the individual pucks used in the VARTA batteries that can stacked
to whatever length you desire in series, and then a handful of odd sizes
that largely never caught on. Given that the battery pack must fit
within a 9.25" x 1.75" x 1.875" space, if the AA size was used, they
could be in a 3x3x4 configuration and the wiring would likely be 3x4 in
series, 3 stacks of 3x4 in parallel, giving you a nominal pack voltage
of 14.4V. Each cell would likely be 600mAh meaning the pack would be
21.6 Wh.

If sub-C cells were used instead of AA, the configuration would likely
be 2x2x4 and the pack voltage would be either 9.6V or 19.2V depending on
whether you wired it as 2 parallel 2x4 or every cell serial. Neither of
those choices seems likely, so I suspect the pack used AA sized cells.

Knowing that the pack has a 3 wire connection suggests positive,
negative and sense wires (most likely). It could also mean one negative
and 2 positive if the sub-C cells were used and the pack was providing
two 9.6V outputs.

You could also get into some wonky possibilities if the battery backup
was ONLY providing 5V DC power, letting the hard drive "slide" for that
10 or 11 second period of backup power, but thinking about that makes my
brain hurt.

Rick, though I love posing possumbilities in obtuse waze, this battery
thing is past being deduced.

Let's put this on the back burner until we either have a battery pack
-OR- an actual 180W PSU so we could measure voltages and examine the
charging circuit.

Rick Ekblaw wrote:
> Louis Ohland wrote:
>> Rick, the simplest [and slowest!] rate is c/1, a lot SIMMpler control
>> circuit and it makes the least heat. I've seen c/5 mentioned, much
>> shorter time, but more heat and it slaps the battery pack like the
>> naughty boy it is.
>>
>> So... if you got to SEE the battery charging circuit, could you
>> guesstimate capacity?
>>
>> The pittances we have of 26v and 40mA could mean whatever you wandt.
>> Measuring the output connector in the battery well -MIGHT- tell us the
>> actual charging voltage, which could be less than 26v. Maybe the
>> charger input is 26v at 40mA?
>
> OK, let's go back for another round of the guessing game.  If IBM did
> not go for custom-built NiCD cells, you have a few common cell sizes to
> choose from:  C size, sub-C size, AA size, AAA size, VARTA size pucks (3
> cells stacked that are roughly a dime diameter, a bit less than an inch
> long), the individual pucks used in the VARTA batteries that can stacked
> to whatever length you desire in series, and then a handful of odd sizes
> that largely never caught on.  Given that the battery pack must fit
> within a 9.25" x 1.75" x 1.875" space, if the AA size was used, they
> could be in a 3x3x4 configuration and the wiring would likely be 3x4 in
> series, 3 stacks of 3x4 in parallel, giving you a nominal pack voltage
> of 14.4V.  Each cell would likely be 600mAh meaning the pack would be
> 21.6 Wh.
>
> If sub-C cells were used instead of AA, the configuration would likely
> be 2x2x4 and the pack voltage would be either 9.6V or 19.2V depending on
> whether you wired it as 2 parallel 2x4 or every cell serial.  Neither of
> those choices seems likely, so I suspect the pack used AA sized cells.
>
> Knowing that the pack has a 3 wire connection suggests positive,
> negative and sense wires (most likely).  It could also mean one negative
> and 2 positive if the sub-C cells were used and the pack was providing
> two 9.6V outputs.
>
> You could also get into some wonky possibilities if the battery backup
> was ONLY providing 5V DC power, letting the hard drive "slide" for that
> 10 or 11 second period of backup power, but thinking about that makes my
> brain hurt.