ARISSat-1 battery eclipse voltage decreasing
The ARISSat-1 Battery voltage is decreasing each eclipse period. It therefore is taking longer for the Battery to charge up to 32.5V to allow the switch from Low Power to High Power when the satellite enters an illumination period.
Kenneth Ransom, N5VHO has plotted the battery min/max for the last 8 days. We see that the battery voltage is decreasing at a faster rate than expected. Kenneth's graph can be found on the arissat1.org site under FAQ http://www.arissat1.org/v3/index.php?option=com_content&view=section&...
Take advantage of the High Power mode as much as possible over the next few days.
73, Gould, WA4SXM for the ARISSat-1/RadioSkaf-b team
On 8/11/11 8:46 PM, Gould Smith wrote:
Kenneth Ransom, N5VHO has plotted the battery min/max for the last 8 days. We see that the battery voltage is decreasing at a faster rate than expected. Kenneth's graph can be found on the arissat1.org site under FAQ http://www.arissat1.org/v3/index.php?option=com_content&view=section&...
Disclaimer: what follows is entirely my own personal opinion, not necessarily that of the ARISSat-1 team or anyone else. I might be missing information that would affect my analysis.
Ken's plot is extremely valuable. It strongly suggests battery failure, not a negative power budget as Tony, AA2TX suggests. Note the sharp and consistent *increase* in daylight battery voltage that occurred late on Aug 11, roughly coincident with sharp *decreases* of voltage during eclipse and reports of computer resets and extended low power operation. This is inconsistent with a negative orbit-average power budget. If that were the case, the battery would never reach full charge and the voltage would never rise so high.
This plot suggests that one or more cells in the battery have lost significant capacity. The cell with the lessened capacity reaches full charge very quickly and then goes into overcharge, allowing the overall voltage of the string to rise. Then of course the voltage falls very quickly in eclipse as the impaired cell(s) rapidly discharge.
As Luc Leblanc points out, silver-zinc (Ag/Zn) batteries have long been the battery of choice in the aerospace industry where their high energy density (almost comparable to li-ion), long shelf life and reliability outweighs the obviously high cost.
The Apollo/Saturn program apparently used Ag/Zn batteries exclusively, powering everything from the guidance system in the Saturn V to the lunar module to the portable life support systems the astronauts used on the lunar surface.
But these features are associated with Ag/Zn as a *primary* battery, one that is never recharged. Only the three entry batteries in the command module were ever recharged, and then only a few times during a mission after moderate discharges. Those Ag/Zn batteries had rated cycle lives measured in the single digits; Luc's figures show only modest improvements since then.
Another unusual feature of Ag/Zn batteries is a two-stage discharge process. When the battery is fully charged, the positive plate contains AgO, silver oxide with the silver in the +2 valence. As the battery discharges, the silver is first reduced to Ag2O, with silver in the +1 valence, and then eventually to metallic silver, Ag, with valence 0.
This two-stage conversion of silver means that fairly high voltages are needed to fully charge the battery. During the Apollo 7 mission, the shakedown flight of the Apollo command module in 1968, the entry batteries could not be fully recharged because the battery charger could not lift the battery terminal voltage high enough through the line resistance. Over 39V is needed to fully charge a nominal 28V battery. According to Ken's graph the voltage has never been this high, and we don't know its initial state of charge. This suggests another possible failure mechanism besides exceeding the battery's cycle life: despite a positive power budget, the battery was never fully recharged, and overdischarge of one or more cells caused cell reversal and damage.
-Phil
Disclaimer: what follows is entirely my own personal opinion, not necessarily that of the ARISSat-1 team or anyone else. I might be missing information that would affect my analysis.
Ken's plot is extremely valuable. It strongly suggests battery failure, not a negative power budget as Tony, AA2TX suggests. Note the sharp and consistent *increase* in daylight battery voltage that occurred late on Aug 11, roughly coincident with sharp *decreases* of voltage during eclipse and reports of computer resets and extended low power operation. This is inconsistent with a negative orbit-average power budget. If that were the case, the battery would never reach full charge and the voltage would never rise so high.
Phil,
How about you designing circuitry/program to charge "cells" individually rather than the "battery". Or has that been tried??
73 Dave
73, Dave, WB6LLO [email protected]
Disagree: I learn....
Pulling for P3E...
On 8/13/11 6:11 PM, Dave Guimont wrote:
How about you designing circuitry/program to charge "cells" individually rather than the "battery". Or has that been tried??
That's a great idea; I wish I thought of it first, but it was Lou McFadden W5DID who published a paper at the AMSAT Symposium a few years ago on a new modular power system.
His idea was to turn each cell into an intelligent energy storage module and to connect those modules in parallel to a power bus. There'd be a DC-DC converter between the cell and the bus so they could operate at different voltages. If the cell in one module failed, it would disconnect itself from the bus instead of dragging it down.
He found it a challenge to achieve high efficiency in the DC-DC converter with low voltage batteries. For the lower voltage chemistries (e.g., NiMH at 1.2 V) it might be necessary to compromise by using two or maybe three cells in series per module. A lithium ion module would need only one cell since they operate at a much higher voltage of 3.6V.
The beauty of his scheme is that not only would this be far more robust against individual cell failures than a single series string, you could fly several kinds of batteries to see which functions best.
With the proper command to a module controller you could perform a controlled discharge of its cell for a capacity test. That's kinda neat.
Some modules could use supercaps. They have very high cycle lives (~500,000) but low energy density (0.35 watt-hour for a D-cell sized cap). You'd always discharge them first, or perhaps keep one in reserve to keep a computer going. Since they're capacitors you'd need the DC-DC regulator to produce a constant voltage as they discharge.
Thinking more about his scheme, you could program each module with a command like "Keep the power bus at +12V by pumping up to 2A into it until you're 50% depleted" or "charge at 1A max unless the bus voltage falls to 11.5V". The computer could issue a new command at any time, such as one to stop all charging when the satellite enters eclipse.
One module might contain just a load resistor to act as a shunt regulator to keep the bus voltage from going too high.
And one or two modules might contain high-density (e.g., lithium) primary batteries as emergency fallbacks to keep things going long enough for the command stations to figure out what's going on.
-Phil
ARRISat-1 was mainly in eclipse over North America since the launch but starting August 15 the first orbits over NA will be out of eclipse and if the batteries can hold and the inside temp does not cooked the electronics we will have a chance to listen and work it at a more suitable time slot instead having to set the alarm clock in the middle of the night...Plus we will have the high power.
We all hope for the best but as projected we already exceed the +- 8 days life expectancy lets say we lived on borrowed time.
"-"
Luc Leblanc VE2DWE Skype VE2DWE www.qsl.net/ve2dwe DSTAR urcall VE2DWE WAC BASIC CW PHONE SATELLITE
serves me right for trying at midnight my time... :( not a soul on the satellite, EXCELLENT overhead pass.... Long...
Sigh...
Lee W5LMM
Dave, A simple switching inverter connected to the solar cells with multiple windings would prevent over charging as the voltage output would be regulated to the voltage of a single cell and the windings providing isolation between cells. This would insure an equal charge to each cell preventing early battery failure. Using FET switches for rectification can boost efficiencies to over 90% even using input voltages as low as 2 Volts D.C.
Art, KC6UQH
-----Original Message----- From: [email protected] [mailto:[email protected]] On Behalf Of Dave Guimont Sent: Saturday, August 13, 2011 6:11 PM To: Phil Karn Cc: [email protected] Subject: [amsat-bb] Re: ARISSat-1 battery eclipse voltage decreasing
Disclaimer: what follows is entirely my own personal opinion, not necessarily that of the ARISSat-1 team or anyone else. I might be missing information that would affect my analysis.
Ken's plot is extremely valuable. It strongly suggests battery failure, not a negative power budget as Tony, AA2TX suggests. Note the sharp and consistent *increase* in daylight battery voltage that occurred late on Aug 11, roughly coincident with sharp *decreases* of voltage during eclipse and reports of computer resets and extended low power operation. This is inconsistent with a negative orbit-average power budget. If that were the case, the battery would never reach full charge and the voltage would never rise so high.
Phil,
How about you designing circuitry/program to charge "cells" individually rather than the "battery". Or has that been tried??
73 Dave
73, Dave, WB6LLO [email protected]
Disagree: I learn....
Pulling for P3E... _______________________________________________ Sent via [email protected]. Opinions expressed are those of the author. Not an AMSAT-NA member? Join now to support the amateur satellite program! Subscription settings: http://amsat.org/mailman/listinfo/amsat-bb
__________ Information from ESET NOD32 Antivirus, version of virus signature database 6374 (20110813) __________
The message was checked by ESET NOD32 Antivirus.
__________ Information from ESET NOD32 Antivirus, version of virus signature database 6375 (20110813) __________
The message was checked by ESET NOD32 Antivirus.
participants (6)
-
Art McBride -
Dave Guimont -
Gould Smith -
Lee Maisel -
Luc Leblanc -
Phil Karn