Hi all,
I've got some ideas but need a kick in the right direction. I can't find the right direction to take to achieve what I'm trying to.
I want to create an uninterruptable power supply but not one that switches between mains and battery power, but rather one that switches from one cell to another when one becomes depleted. I imagine using two cells and when activated the power would be supplied by the first cell. When the first cell becomes depleted, the circuit detects this and switches to take the supply from the second. The first is then charged whilst the second supplies the power. Then when the second becomes depleted the circuit once again detects this and reverts to the first - which has been charged in the meantime.
If possible I want to avoid any circuitry between the 'active' cell and the load so that the load is getting as much 'pure battery power' as possible and as little noise from any necessary monitoring or switchover circuit.
I had originally imagined that I would just be able to use a bog-standard li-on charger but I now imagine using a wall wart supply and building the charger circuit in the unit with the switchover circuit. My thoughts changed here because I imagine that the switchover circuit will need something to power it separately from the charging power supply.
I'd like to use solid state switching as much as possible but my overiding concern is maintaining the purest of power from the cell itself
It looks like my thinking is not original but this thread from a long time ago didn't go anywhere: Dual battery PSU with low voltage changeover - diyAudio
Anyone have any ideas about how I might start to work through this problem...eg an IC that could form the basis of the switchover circuit? Or, any other thoughts are very welcome.
Cheers,
Crom
I've got some ideas but need a kick in the right direction. I can't find the right direction to take to achieve what I'm trying to.
I want to create an uninterruptable power supply but not one that switches between mains and battery power, but rather one that switches from one cell to another when one becomes depleted. I imagine using two cells and when activated the power would be supplied by the first cell. When the first cell becomes depleted, the circuit detects this and switches to take the supply from the second. The first is then charged whilst the second supplies the power. Then when the second becomes depleted the circuit once again detects this and reverts to the first - which has been charged in the meantime.
If possible I want to avoid any circuitry between the 'active' cell and the load so that the load is getting as much 'pure battery power' as possible and as little noise from any necessary monitoring or switchover circuit.
I had originally imagined that I would just be able to use a bog-standard li-on charger but I now imagine using a wall wart supply and building the charger circuit in the unit with the switchover circuit. My thoughts changed here because I imagine that the switchover circuit will need something to power it separately from the charging power supply.
I'd like to use solid state switching as much as possible but my overiding concern is maintaining the purest of power from the cell itself
It looks like my thinking is not original but this thread from a long time ago didn't go anywhere: Dual battery PSU with low voltage changeover - diyAudio
Anyone have any ideas about how I might start to work through this problem...eg an IC that could form the basis of the switchover circuit? Or, any other thoughts are very welcome.
Cheers,
Crom
Just some quick thoughts... I'm not designing it for you 
, haven't the time
These type of circuits are never as straightforward as they might seem, at least not if they are to be foolproof and reliable.
I would perhaps look at low Rds FET's for directing each cell to the load. With an Rds of below 10 milliohm losses are really minute.
Overlap... you need to ensure one cells output perhaps parallels the other for a few milliseconds to ensure no drop out of supply.
Simple logic circuit to ensure that cell being charged is only active when that cell is not connected to the load and that charging is removed fractionally before that cell is switched into service. That aspect may not be needed depending on whether the circuit fed from the cells can tolerate sudden step changes in applied voltage. Use micropower design techniques to minimise current draw. The logic could possibly be powered from the cells or via the incoming supply and utilise either a "supercap" or NiCad to ensure the logic is always active.
Micropower comparators to set the voltage trip points for each cell.
You have to tackle it all in blocks putting it all together as you go.
, haven't the time These type of circuits are never as straightforward as they might seem, at least not if they are to be foolproof and reliable.
I would perhaps look at low Rds FET's for directing each cell to the load. With an Rds of below 10 milliohm losses are really minute.
Overlap... you need to ensure one cells output perhaps parallels the other for a few milliseconds to ensure no drop out of supply.
Simple logic circuit to ensure that cell being charged is only active when that cell is not connected to the load and that charging is removed fractionally before that cell is switched into service. That aspect may not be needed depending on whether the circuit fed from the cells can tolerate sudden step changes in applied voltage. Use micropower design techniques to minimise current draw. The logic could possibly be powered from the cells or via the incoming supply and utilise either a "supercap" or NiCad to ensure the logic is always active.
Micropower comparators to set the voltage trip points for each cell.
You have to tackle it all in blocks putting it all together as you go.
Hey no worries, I want the challenge ;-)
I'm not fussed about making sure they overlap. If a UPS can switchover in less than 100mS and not crash a server then that's good enough for me...unless I'm missing something.
In my simple mind I'd imagined that I could use the solid state equivalent of relays to switch between cells. I hadn't come across micropower comparators - thank you. From a brief read, when the voltage triggers the micropower comparator (MC), the MC triggers the 'relay' to switch to the other cell and also switch the charger into the circuit for the offline battery.
The battery is on charge whenever it's offline (the charging circuit taking care of whether it should actually be being charged or not).
Yes, blocks...break it down...good idea.
I'm not fussed about making sure they overlap. If a UPS can switchover in less than 100mS and not crash a server then that's good enough for me...unless I'm missing something.
In my simple mind I'd imagined that I could use the solid state equivalent of relays to switch between cells. I hadn't come across micropower comparators - thank you. From a brief read, when the voltage triggers the micropower comparator (MC), the MC triggers the 'relay' to switch to the other cell and also switch the charger into the circuit for the offline battery.
The battery is on charge whenever it's offline (the charging circuit taking care of whether it should actually be being charged or not).
Yes, blocks...break it down...good idea.
Always think outside the box when designing a one off circuit.
You could use the comparator to drive a FET directly (if you have say 5 volts or more available) or you could use photo voltaic couplers (think solar cell illuminated by an LED to give a true floating drive voltage of up to 16 volts if two couplers in the package are series connected). They make brilliant solid state relays. Depending on voltages, you could look at switching the cell negative rather than the positive which might make the drive easier (low voltage ground referenced).
Solid state relays using FET's are detailed here. These using two FET's are bidirectional, a refinement you probably don't need but you will see the idea.
This thread will tell you all you need to know on solid state relays,
http://www.diyaudio.com/forums/solid-state/191449-output-relays-13.html#post2659578
You could use the comparator to drive a FET directly (if you have say 5 volts or more available) or you could use photo voltaic couplers (think solar cell illuminated by an LED to give a true floating drive voltage of up to 16 volts if two couplers in the package are series connected). They make brilliant solid state relays. Depending on voltages, you could look at switching the cell negative rather than the positive which might make the drive easier (low voltage ground referenced).
Solid state relays using FET's are detailed here. These using two FET's are bidirectional, a refinement you probably don't need but you will see the idea.
This thread will tell you all you need to know on solid state relays,
http://www.diyaudio.com/forums/solid-state/191449-output-relays-13.html#post2659578
This looks like a simple solution LT1579 - 300mA Dual Input Smart Battery Backup Regulator - Linear Technology but I'd rather that the battery supply wasn't regulated.
This too might be worth looking at: LTC4417 - Prioritized PowerPath Controller - Linear Technology but rather than prioritising a supply we want to flip flop between batteries only when one becomes exhausted and not, as I guess would happen using this, the first battery would take over as soon as the voltage became high enough to trigger (but possibly not fully charged).
This too might be worth looking at: LTC4417 - Prioritized PowerPath Controller - Linear Technology but rather than prioritising a supply we want to flip flop between batteries only when one becomes exhausted and not, as I guess would happen using this, the first battery would take over as soon as the voltage became high enough to trigger (but possibly not fully charged).
Another interesting take on SS v mech: noise - What kind of effects does a relay have on signals? - Electrical Engineering Stack Exchange
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.