DC power for communications site

Greetings,
I am involved with an application where we will have communication cabinets or shelters located in remote locations, but most have
electric service available. However some may not be accessible for days depending on weather. All of the equipment will run on -48VDC os there is no UPS. A rectifier shelf supplied with 240VAC will provide the -48VDC thru a distribution panel (DC circuit breakers) which will power the equipment as well as charge backup batteries. One vendor has suggested that the batteries would be one load on the distribution shelf and the equipment would be the other. Then when the rectifiers go down due to an outage the equipment would be backfed thru the distribution panel. I have not worked with DC power systems but from working with AC power it doesn't seem kosher to backfeed thru a panel. I am hoping that one of you could provide a schematic or system description for this type application.
Secondly, since the sites are remote we would like to have remotely resetable breakers in the distribution panel. We have remote ability to provide contact closures or toggle control voltages as well as read voltage and current.
Looking forward to your comments.
Thanks,
jh
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
It's been "awhile" but in the Old Bell System the offices all ran from 48 volt batteries that were always being charged.
During "normal" periods when AC power was available, there was a compensation cell is series with the 48 volt battery than dropped the actual voltage when being charged back down to the 48 volt range.
When "main" power went down a relay shorted out the compensation cell.
Frankly, that's looks like a good system for your application.
Just have your stuff run from battery all the time and as alternative power is available you charge the battery. The "compensation cell" may not be worth the trouble.
The Old Bell System tended to go first class with the batteries. They were kept well away from the various relay racks and distribution frames so that there would not be any contamination problems. Several hundred sqare feet of floor space was for the batteries.

Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
John Gilmer wrote:

Bullshit. No one in their right mind would 'Short out' a high capacity lead acid battery.
--
Service to my country? Been there, Done that, and I've got my DD214 to
prove it.
  Click to see the full signature.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
wrote:

Good catch. A curiosity now, a break before make double throw would cause a dropout. Seems like the comp cell should be ahead of the load so maybe there was a diode in series between charger and the comp cell to only allow current to flow thru it backwards. Then I think the relay could bypass the comp cell. However it seems this would over charge the comp cell on a regular basis.
jh
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload

Except he's dead wrong.

The "comp cell" holds *no* charge. That is why it can be shorted. It merely provides a voltage drop that is essentially independent of the current.
The original "comp cell" was a CEMF (Counter EMF) cell that used nickel electrodes immersed in an alkaline solution. Because it gassed at a very high rate the electrolyte solution had half an inch or so of mineral oil floating on top.
The drop across a CEMF cell was just over 2 volts, and if provisioned in two arrays of multiple cells in parallel (to handle the current) would allow for just over 4 volts of voltage adjustment in two steps of 2 volts each.
An interesting story... In the late 70's I worked at a troposcatter communications site (which had WECO battery plant equipment) that was about to be replaced with one of those new fangled satellite earth stations (all of which used Lorain power equipment). When this story actually started, we knew it would eventually happen, but weren't sure when.
The battery plants where stacked against a wall, three cells high on steel racks. The top would have been at about 11 feet off the floor. The large 130 volt supply cells were on the upper tier. There was a small window just above the level of the cable racks (at 11 feet off the floor), and in the spring the sun would shine through the windows onto the top tier of batteries for a few hours each day. And it had been doing that for 20 years.
So one day we discovered that a battery on the top tier, directly above the CEMF cells on the bottom tier, had cracked. There was acid all over the wall, all over the cells below it, into the CEMF cells, and about 1/4 inch of acid on the floor for about 12-15 feet in every direction. (This is the worst nightmare you can imagine for someone doing "station checks"! It's a really bad bad-hair day after that.)
Cleaning that up was a huge job. The tiles on the floor had to be replaced. The wall and the steel racks were never the same. The CEMF cells had to be totally rebuilt, after emptying them, scrubbing them out and putting in new electrolyte and mineral oil, they were like new.
And I wish I could tell you we had used our heads, but we did all that by rote and by note, and totally *missed* the important point. Which was that there were two other cells that had had the sun shining on them for 20 years!
So the next spring, *another* cell cracks and we finally do see the light. That window got blocked off!
But by that time we also had an exact date for when everything was going to be turned down, and we just couldn't see all that work going into something that would be shutdown in 14 months. So we did only the minimum. We left those CEMF cells exactly as they were, figuring that if they actually failed we'd do whatever it took.
They didn't fail; the electrolyte turned green. They grew stuff inside the cells, it looked unbelievably ugly. But they kept right on ticking until the day everything was turned off.
--
Floyd L. Davidson <http://www.apaflo.com/floyd_davidson
Ukpeagvik (Barrow, Alaska) snipped-for-privacy@apaflo.com
  Click to see the full signature.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload

Nope, not bullshit. He's *precisely* correct.
A "compensation cell" is in *no* *way* anything that would be described as "a high capacity lead acid battery". CEMF cells (an alkaline cell with nickle electrodes), resistors, diodes, and probably 3 devices that I can't remember, have all been used as the "compensation cell", and virtually any of them can be directly shorted.
--
Floyd L. Davidson <http://www.apaflo.com/floyd_davidson
Ukpeagvik (Barrow, Alaska) snipped-for-privacy@apaflo.com
  Click to see the full signature.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
The "compensation cell" is NOT a lead acid battery, typically it was a selenium rectifier rated for the CHARGE current of the battery string.
Jim http://telco-power.com
also a DD214 holder.....
wrote:

Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload

In effect it is a UPS, just not the kind you are used to seeing.
If I interpret your post correctly all the equipment will run on an appropriatly sized DC power supply.
A backup battery bank is then being charged by an appropriatly sized maintance charger.
The usual practice is to 'float' the backup batteries by connecting then to the load through a diode which prevents the main supply from overcharging them.

Remote breaker reseting should be done with a great deal of consideration. How likely witll the result be a fire? How much damage will result by forcing a reset to a shorted power supply? An anaogy would be: the guy that repeatedly resets the CB to the garage with out fixing the fault until the garage burns down.

Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
Thanks for the posts gentlemen.
Yes we are in some ways simulating a UPS and in fact life would be easier if all the equipment ran on 120VAC. But I didn't get a choice. However, there may be a chance to kill two rats with one rock here (used to say 'kill 2 birds...' until we started keeping pet birds). The diode could protect the battery bank from overcharge and at the same time provide a little voltage drop. We don't have a final spec on the tolerance allowed for the -48VDC but four fully charged batteries in series would provide 52.8 VDC.
The remote reset will be a remote 'manual' operation if I may use such a term. We have monitoring equipment that can read current and voltage and the readings are recorded on a periodic (period TBD) basis. When the breaker is reset we will take readings in a tight loop for a period of time and if the load remains normal will look for signs of transients. I am a engineer by degree but have been a software weenie for 20+ years but I will be reluctant to go along with automating this process. Perhaps a line voltage monitor would be appropriate so we could detect incoming transients. However there is pressure to keep the instrument count down. One for cost purposes, two, more components reduce MTBF.
Any recommendations for DC power shelf vendors?
Best Regards,
John
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
I should add that we have found some remotely controlled PDUs, one source is specpower.com. Still looking for others as well as power shelves.
jh
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload

The choice of 48 VDC is a good one.
I worked in telecommunications equipment rooms, all of which were powered by 48 volt battery plants, so I admit to being habituated. That included hands on experience with dozens of installations that match what you have described, ranging in size from the Fairbanks Toll Center to small satellite stations that fit into one rack.
Many of these sites handle either FAA circuits or DoD circuits that require immediate response, and the cost of putting a craftperson on site exceeded $2000 just for the aircraft. Reliability is the top priority.

There is no diode, as such.

The "correct" float voltage is specified by the manufacturer of the batteries.
The voltages of interest are the "high voltage cutoff" (and you want a separate circuit that will shutdown the plant if that voltage is exceeded), the "charge voltage", the "float voltage", and the "low voltage cutoff" (which like the maximum voltage, should have a separate detection circuit to disconnect the batteries from any load when this voltage is reached).
Normal operation, when AC is available, will be at the float voltage (roughly 52-54 volts). If you place the plant on "charge", to equalize the cells (for example that should always be done if water is added to any cell), the regulator should adjust the voltage to the "charge voltage" (about 56 volts, which like the float voltage depends on the specific batteries used).
If AC power is lost, the battery begins to discharge. There *absolutely* *must* be a low voltage disconnect (which will operate at about 42 volts), or your batteries will be destroyed by an extended power failure.
As has been mentioned, in "the good old day" there was also usually some sort of a device to reduce the operating bus voltage when the plant was floating. CEMF cells and other devices were common. That is much less common today than it once was.
For most relatively small installations today, gel cells rather than wet cells, are preferred.

I've never seen any of that sort of equipment used... Take that as a recommendation to assume it isn't needed!
On the other hand, a single telephone line provided by the closest thing to "diversity" that is possible, connected to equipment that can power cycle and/or do software resets on as many other equipments as possible is *definitely* a money making asset. Ideally that would be included as part of the alarm system.
Also, the typical configuration for a small installation such as you've described, is to have a redundant pair of regulators, each of which is large enough to handle the load without the other. Typically they are operated in parallel, each with about half the normal load.
With larger installations the number of charging units operated in parallel is greater than two, even if the capacity of each is also increased. The design target then would be that the maximum charge rate (to recover from a discharge down to the low voltage cutoff, for example) would require all chargers to be active.

I can only remember two manufacturers... WECo (AT&T in the old days) and Lorain Power Systems.
Certainly there must be others, but Lorain comes *very* highly recommended.
Oh, incidentally... a little DC-AC inverter unit can provide 120VAC for the few things that require protected AC power.
--
Floyd L. Davidson <http://www.apaflo.com/floyd_davidson
Ukpeagvik (Barrow, Alaska) snipped-for-privacy@apaflo.com
  Click to see the full signature.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
Great post Floyd, a lot of good info which generate more questions.

Pardon my density, I wish I had a schematic of your connections. It sounds like the batteries were being charged by the same source as feeds the load. If that is the case we will be drawing about 30 amps on an continuous basis. If the battery and the equipment load are in parallel how are the batteries maintained at float?

Correct my PDU has an LVD and an OVD.

Not familiar with off the shelf CEMF devices. Will have to surf.

I have heard this b4. Is it because they don't vent and de-hydrate?

Actually will have customer edge routers on VPNs on redundant lines provided by a major telco. Attached to the routers will have an SNMP device that has analog inputs, relay contacts, as well as digital I/O. Monitoring is dictated so I have no choice tho I agree that we are adding complexity and reducing reliability at the same time.

Pretty much all the equipment is redundant, including battery racks.

Thanks,
jh
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload

The charger is rated according to the load you are supplying. I have several in by shed but I cannot get at them at the moment to comfirm the amperage rating but it is quite high. They are switch mode devices.
--
Stuart Winsor

From is valid but subject to change without notice if it gets spammed.
  Click to see the full signature.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload

When AC power is present, the charge circuit regulates the voltage at the "float voltage" specified for the batteries. At that voltage the batteries are taking a charge, though it is relatively small. Hence the battery is just part of the load as far as the AC charging equipment is concerned.
If the AC fails, then the battery supplies power to the load, but it will be at a lower voltage. Typically the float voltage would be about 53 volts, and the battery itself would be 48 volts when fully charge, dropping down to 42 volts before a low voltage cutout prevents a total discharge (which would destroy the battery cells).
The battery is always connected. It has, in addition to stored energy, the attribute of being a wonderful filter too.
In fact, you can visualize the whole thing as if the battery itself is just a really big capacitor with fairly high leakage. If the AC power is disconnected the load is supplied from the charge in the capacitor until it is drained. The only difference is that a battery has a much larger stored charge, so it takes much longer to discharge it.
...

Horrible things, that you don't want to use!

That is probably the main attraction. Maintenance free is a good policy for unmanned sites... :-)

As long as the "SNMP device" can be accessed via either of the redundant lines, that appears to fit the bill.
Monitoring shouldn't actually reduce reliability. Alarm systems are a bit of a black art though, and it's really difficult to see the potentials until there is some actual operating experience. Six months before installation the poor design engineer just isn't going to know all the "interesting" things that operations will be able to tell him he did wrong, after they've used it for a couple years!
--
Floyd L. Davidson <http://www.apaflo.com/floyd_davidson
Ukpeagvik (Barrow, Alaska) snipped-for-privacy@apaflo.com
  Click to see the full signature.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload

This is normal and allowed for in the spec for 48V DC operated comms equipment.
--
Stuart Winsor

From is valid but subject to change without notice if it gets spammed.
  Click to see the full signature.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
Http://telco-power.com

Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload

Not entirely sure what you mean by this.
Modern chargers designed for this application are constant voltage and the load and the charger are connected to the battery at the same point
--
Stuart Winsor

From is valid but subject to change without notice if it gets spammed.
  Click to see the full signature.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
wrote:

Some method of limiting charging current to a depleted battery bank while simultaneously operating equipment when main power is restored. There are probably several schemes to accomplish this.
My speculation is that load will have a high peak current at a low duty cycle, something like a digipeater. I admit its just a guess but it would probably need a high current regulated supply.
Some method of maintaining normal operation when the battery fails should be included. (I have seen melted shorted gel cells)
Schemes to extend battery life by deep cycling them abound and could be included as desired (and budget alows).

It's simple easy and reliable but you have to assume supply current will not exceed a value that will harm the batterys under max charge conditions and you also have to assume the batterys will never short.

Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload

It's fairly simple in practice. Each charging unit (commonly called a "rectifier") has a current limiter, and can only supply so much current. The voltage regulation has to be common though, and in some way shared by all of the charging units. That can be as simple as the units sequentially come online and if a given unit cannot regulate the voltage without exceeding maximum current, it will go into current limiting and pass voltage regulation to the next unit in the sequence.
After an extended AC power outage it might well be that for a short time all of the charging units will be in current limiting. The voltage will slowly rise until it reaches whatever is set for voltage regulation.
Generally the voltage will always be set to the float voltage, and in an automatic mode the only variation from that would be when all charging units are in current limiting.
As the battery becomes charged the current it can take is reduced. When it goes down sufficiently, the plant will go into voltage regulation rather than current regulation.
However, it _is_ an interesting experience to watch a large plant go through that cycle.

The instantaneous peak current requirements need only be within the limits of the /batteries/, not the charging mechanism. The charging mechanism need only be large enough to handle the average load plus sufficient to eventually recharge a discharged battery.
Of course generally the peak capacity of the charging system is going to be high enough to recharge the battery at the manufacturer's specified normal rate, plus handle the highest "normal" load likely to be encountered at the same time.

That's provided by the Armstrong with Wrench method!
Unbolt any part of the series connections for the battery, and everything runs on the charging units instead of the battery plant. It works, but is not without some consequence. The battery provides filtering and voltage regulation, which ceases to exist. Typically a higher level of power line hum will be experienced without the battery.

Beware of trying to homebrew such schemes!
The manufacturer of the batteries can provide detailed information on a specific type of battery, and that is what should be followed. There are distinctly *different* instructions for virtually every battery type, and just picking one off the wall is likely to shorten battery life significantly.

Of course. But assuming the supply current is correct is merely saying that the system is properly designed.
Batteries shorting is not exactly a common occurrence, nor is it particularly exciting. Most telephone company battery plants last for a couple of decades without having any significant problems. Between 20 and 30 years they take a lot of babying, and how far they go depends greatly on how well they were maintained in those first 20 years.
--
Floyd L. Davidson <http://www.apaflo.com/floyd_davidson
Ukpeagvik (Barrow, Alaska) snipped-for-privacy@apaflo.com
  Click to see the full signature.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
my brother works on these type of setups..... the chargers and the batterys and the phone system are all in parellel if the phone system draws 100 amps at 48 volts dc then they will have 3 of 4 chargers in parallel so that the batterys can charge if they run completly dead and so they won't trip the charger offline..... and redundinacy if one of the chargers fail the rest of the units will not trip offline...

Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload

Polytechforum.com is a website by engineers for engineers. It is not affiliated with any of manufacturers or vendors discussed here. All logos and trade names are the property of their respective owners.