Oscilloquartz BVA has been sold. Thank you all who expressed an interest.
Bill Notfaded writes:
Can't I just use a high quality APC backup power system like we use to
power racks of gear in our Telco and compute closets?
Very few UPS's are good at long-run applications, they are typically
built to run a heavy load for minutes, not a tiny load for hours
or even days on end.
That means low efficiency, 75% net efficiency is considered good,
and it goes totally south the further you are from the name-plate load,
because the constant overhead is large.
Some UPSs dont even have a thermal design allowing 24*365 operation.
If you want to power mains kit from batteries, it is usually better
to get a "real" inverter which is built island-grid applications.
But for powering small loads, OCXO's, GPSDO's, Rb's, fire alarms,
emergency lighting etc, the overhead of going from battery voltage
to mains voltage and back is just a unnecessary loss.
--
Poul-Henning Kamp | UNIX since Zilog Zeus 3.20
phk@FreeBSD.ORG | TCP/IP since RFC 956
FreeBSD committer | BSD since 4.3-tahoe
Never attribute to malice what can adequately be explained by incompetence.
We have done exactly that not a week at a time because we have enough high resolution and did it in 1 Volt steps down to 20 V no change. Current stays the same. Bert Kehren In a message dated 9/25/2020 8:44:02 PM Eastern Standard Time, phk@phk.freebsd.dk writes:
Mark Spencer writes:
The bit I am struggling with re using a 24 volt battery system in this
application is what happens when AC power is removed and the terminal
voltage of the battery starts to fall.
Don't overengineer your supply (unless you absolutely want to :-)
It looks like the BVA is specified at 24V ±10%, which is 21.6-26.6V.
Nobody says 24 Volt is optimal in that interval, and I have
never been able to measure any difference on any OCXO, as long as
I stayed a few hundred millivolts above the lower limit.
I would run the BVA for a week at 24V, collecting PPS data against
a GPSDO or other suitable source.
Then I would run it for a week at 22V, collecting the same data.
If I could not tell the data from the two runs apart, I would design
a 22V supply.
The BVA is spec'ed at 3W steady state, at 22V that's ~150mA.
In long-run applications you get the name-plate capacity, so two
12V x 15 Ah VRLA blocks would keep your BVA happy for four days,
24Ah would last a week etc.
"Small" come from a lot of dubious manufacturers and which means
badly recycled and impure lead. For a golfcart that doesn't matter,
but it is a big disadvantage in long-run applications like this.
Saving 5 bucks on lead-acid batteries are usually a bad deal in the
long run, but on the other hand, dont say no if serious battery-people
offer you a couple of pre-owned large batteries cheap, just make
sure you can lift them.
Lead-Acid in long-duration applications should not be discharged much
below 1.85 V, and 12 cells times 1.85V = 22.2 Volt. Subtract a
diode drop, and now you know where the 24V±10% spec comes from in
the first place.
At 27.6 V float-charge, your linear regulator will have to deal with
(27.6-22)V * 150mA = .84W.
A LM317 would do a fine job, but you can reduce the voltage drop
and get longer runtime with a PNP based home-brew supply.
The float-charger should be able to power the BVA and also recharge
the battery in less than a day, so with 15Ah blocks, I would aim
for around 1-1.5 ampere.
Your batteries will live longer if they have individual chargers,
so something like a 2x12VAC 40-50W torroid trafo, two bridges, two
caps, two LM317's plus resistors and trimpots will do fine.
Poul-Henning
PS: practical hints:
Your local car-nut-emporium has fuseholders like these for cheap:
Take two, crimp ring-lugs on one end of them, and bolt them directly
to the battery terminal: one for charger, one for load.
When using them next to batteries, the "contra" diode "backwards
across" the LM317 is mandatory.
"Real" UPS installers often have spare batteries which have been
kept on float-charge but otherwise unused, in case of a failure on
customer site. At end of contract, they dont need them any more.
Telephone people with 2V batteries are interesting too, but their
12V blocks are usually dead by the time the swap them out.
Power-grid people, in particular the nuclear plant people, can also
be a good source very, very good batteries with lots of life in
them still, but be very aware of the weight and the short-circuit
currents.
If you end up with pre-owned liquid electrolyte batteries ("OPzS")
make sure the hydrogen gas can vent safely out of the building.
--
Poul-Henning Kamp | UNIX since Zilog Zeus 3.20
phk@FreeBSD.ORG | TCP/IP since RFC 956
FreeBSD committer | BSD since 4.3-tahoe
Never attribute to malice what can adequately be explained by incompetence.
time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe, go to http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.
Victron Energy has a nice application book on their web site the have super products, I used them on my 64 foot GOFAST. Besides products they also discuss battery technologies and applications. My hard copy is 15 years old. Here is the link in English https://www.victronenergy.com/upload/documents/Book-Energy-Unlimited-EN.pdf Bert Kehren In a message dated 9/25/2020 9:02:32 PM Eastern Standard Time, phk@phk.freebsd.dk writes:
Bob kb8tq writes:
Two basic types of SLA / VRLA batteries out there: AGM and Gel. For the AGM
variety, you want a bit more to your charger.
Yes, if you buy batteries for N * 100K $money, spending more money
on your charger is a good investment.
But if you have two batteries doing float-charge/long-run/tiny-load,
in a benign environment in a corner of your lab, the internal
variance between batteries in that same production lot, will
overshadow any imaginable benefit a fancy charger could bring.
I haven't worked with Gel cells in a while so they may indeed be easier.
They are for the same reasons they suffer cyclic use better.
But for a float/long-run application, it makes absolutely no difference.
--
Poul-Henning Kamp | UNIX since Zilog Zeus 3.20
phk@FreeBSD.ORG | TCP/IP since RFC 956
FreeBSD committer | BSD since 4.3-tahoe
Never attribute to malice what can adequately be explained by incompetence.
time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe, go to http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.
On 9/25/20 11:07 PM, Poul-Henning Kamp wrote:
Bill Notfaded writes:
Can't I just use a high quality APC backup power system like we use to
power racks of gear in our Telco and compute closets?
Very few UPS's are good at long-run applications, they are typically
built to run a heavy load for minutes, not a tiny load for hours
or even days on end.
That means low efficiency, 75% net efficiency is considered good,
and it goes totally south the further you are from the name-plate load,
because the constant overhead is large.
Some UPSs dont even have a thermal design allowing 24*365 operation.
As people have found when modifying those inexpensive $100 UPSes to use
an external battery. They're cost sensitive so the electronics just
gets to maximum temperature as the designed in battery runs out. It
might be cheaper to add mass than a (noisy) fan, for instance.
If you want to power mains kit from batteries, it is usually better
to get a "real" inverter which is built island-grid applications.
They do make UPSes for long term running but not as a static inverter -
typically you can identify them because they come with options for
external batteries. Or,you choose which battery pack to add.
But for powering small loads, OCXO's, GPSDO's, Rb's, fire alarms,
emergency lighting etc, the overhead of going from battery voltage
to mains voltage and back is just a unnecessary loss.
There is, also, the Tesla PowerWall approach..
13.5 kWh, 7kW peak, 5 kW avg, $6500
You can keep a lot of oscillators humming for quite a while with 13.5
kWh, even at reduced efficiency.