Reply by pedro October 6, 20162016-10-06
On Thu, 06 Oct 2016 15:32:34 -0500, Tim Wescott
<seemywebsite@myfooter.really> wrote:

>On Thu, 06 Oct 2016 12:54:34 -0700, Joerg wrote: >
(snip)
>> A scheme that would work but I haven't seen it implemented is this: >> Apply a current limited 10-12V externally and have the control >> electronics in the supply disconnect it every few seconds. Check open >> circuit voltage of the connected Li-Ion battery, if less than target >> voltage apply another burst. If at target voltage stop and signal a >> green LED for "charge is complete".
>The pack _should_ charge at maximum current until it hits 4.2V/cell, then >stay there with the battery determining the current, then finally the >charge should terminate when the current drops to some fraction of the >cell's maximum current (IIRC 1/10th is the rule of thumb, but don't go >designing a charger based on that).
That is what the charge regime was for the MAX1737 when I used it waaaay back (2002 actually, just checked) in a commercial Li-Ion charger. After a pre-qualification phase (ensuring battery voltage above LVCO and temp within limits), current-limited constant-voltage, terminating when current drops to 10% and a saftey timeout.
Reply by Tim Wescott October 6, 20162016-10-06
On Thu, 06 Oct 2016 14:49:26 -0700, Lasse Langwadt Christensen wrote:

> Den torsdag den 6. oktober 2016 kl. 23.28.41 UTC+2 skrev Joerg: >> On 2016-10-06 13:32, Tim Wescott wrote: >> > On Thu, 06 Oct 2016 12:54:34 -0700, Joerg wrote: >> > >> >> On 2016-10-04 16:05, Dimitrij Klingbeil wrote: >> >>> On 04.10.2016 01:27, Joerg wrote: >> >>>> Hi Folks, >> >>>> >> >>>> Just bought another Li-Ion deal with charger: >> >>>> >> >>>> https://images-na.ssl-images-amazon.com/images/ >> > I/41HL6Xmx1qL._SX466_.jpg >> >>>> >> >>>> The charger is like this little guy, a switcher that can deliver >> >>>> 1 >> >>>> amp: >> >>>> >> >>>> https://images-na.ssl-images-amazon.com/images/ >> > I/41HL6Xmx1qL._SX466_.jpg >> >>>> >> >>>> The typical battery is two cell bundles in series so 8.4V max. >> >>>> To >> >>>> my surprise the charger has 10.2V open loop voltage. It is not a >> >>>> switched deal to overcome wire resistance where it measures the >> >>>> battery voltage every once in a while. Just straight DC. It seems >> >>>> that it relies on the battery's "electric safety vent" to stop at >> >>>> 8.4V. Anybody seen this done? >> >>> >> >>> Hi J&ouml;rg >> >>> >> >>> Not exactly the same, but I've seen a similar configuration in a >> >>> Chinese MP3-playing toy. The configuration looked like: >> >>> >> >>> USB port -> 10 Ohm series resistor -> silicon diode -> LiPo battry. >> >>> >> >>> No other relevant components in the direct power path on the main >> >>> board, >> >>> but the battery did have a small protection board, apparently with >> >>> the classic one chip and 2 transistor solution that seems to be >> >>> common nowadays. >> >>> >> >>> The device in question did have an "interesting" quirk: When fully >> >>> charged, it would fail to operate when the charger was >> >>> disconnected. Switching it off and on with a (mechanical) switch >> >>> would reset it. >> >>> >> >>> >> >> Probably because the battery's OVP had come and needed to be >> >> "unstuck". >> >> >> >> >> >>> Apparently they indeed relied on the protection circuit to work as >> >>> a simple ON/OFF charge controller. It would turn off on overvoltage >> >>> and stay that way, so that when the charger was disconnected, the >> >>> sharp voltage drop to zero (before the the protection has time to >> >>> sense that the overload has gone away) would upset the main ASIC, >> >>> which would get stuck somewhere in the wilderness of an improperly >> >>> timed reset cycle. >> >>> >> >>> >> >> A stuck ASIC would be poor design. >> >> >> >> >> >>> As for safety: There's a good reason to have control and protection >> >>> circuits separate. A single fault condition can't normally kill >> >>> both together (unless the protection is way inadequate). With only >> >>> one power interrupting component (one MOSFET per direction of >> >>> charge / discharge current flow) all it takes is a single fault in >> >>> a semiconductor part to deadlock this thing into an always on >> >>> charging condition, with a third party uncontrolled tolerance >> >>> "nominally 5V" adapter for the power. >> >>> >> >>> >> >> Single-fault safety is why I don't like the scheme where they use a >> >> higher external voltage. I can build a supply with exactly 8.2V but >> >> I tried that and it has one major disadvantage: As the battery >> >> voltage approaches 8.2V the charge current will become very small >> >> due to cable resistance. It'll take all day to recharge a battery to >> >> a reasonable level. That would present a minor inconvencience if I >> >> did a long ride late into the evening and wanted to head out again >> >> in the morning. For safety I ride with bright daytime lights. >> >> >> >> A scheme that would work but I haven't seen it implemented is this: >> >> Apply a current limited 10-12V externally and have the control >> >> electronics in the supply disconnect it every few seconds. Check >> >> open circuit voltage of the connected Li-Ion battery, if less than >> >> target voltage apply another burst. If at target voltage stop and >> >> signal a green LED for "charge is complete". >> >> >> >> If I wanted to do this right it'll be another electronics project >> >> and my hobby time is already occupied with bicycling and beer >> >> brewing. Maybe with a uC but as an analog guys I am not a great >> >> programmer. So possibly I'll do it analog. >> > >> > "proper" charging of a LiPo is 4.2V/cell (or whatever the pack calls >> > for) >> > or some maximum current, whichever is lower. Much like lead-acid, >> > except with more potential for flame. >> > >> > The pack _should_ charge at maximum current until it hits 4.2V/cell, >> > then stay there with the battery determining the current, then >> > finally the charge should terminate when the current drops to some >> > fraction of the cell's maximum current (IIRC 1/10th is the rule of >> > thumb, but don't go designing a charger based on that). >> > >> > >> Ideally yes but for expedient charging this requires sense lines to the >> battery pack which, on a bicycle, you don't have. >> >> > can't just use a four wire cable leaving only the wire from connector to > cells (which I'd expect to be very short) unaccounted for?
He's a user, not a maker in this case. Really, the charging circuitry should be co-located with the battery. -- Tim Wescott Wescott Design Services http://www.wescottdesign.com I'm looking for work -- see my website!
Reply by Joerg October 6, 20162016-10-06
On 2016-10-06 15:37, Lasse Langwadt Christensen wrote:
> Den fredag den 7. oktober 2016 kl. 00.31.15 UTC+2 skrev Joerg: >> On 2016-10-06 15:26, Lasse Langwadt Christensen wrote: >>> Den fredag den 7. oktober 2016 kl. 00.17.59 UTC+2 skrev Joerg: >>>> On 2016-10-06 14:49, Lasse Langwadt Christensen wrote: >>>>> Den torsdag den 6. oktober 2016 kl. 23.28.41 UTC+2 skrev Joerg: >>>>>> On 2016-10-06 13:32, Tim Wescott wrote: >>>>>>> On Thu, 06 Oct 2016 12:54:34 -0700, Joerg wrote: >>>>>>> >>>>>>>> On 2016-10-04 16:05, Dimitrij Klingbeil wrote: >>>>>>>>> On 04.10.2016 01:27, Joerg wrote: >>>>>>>>>> Hi Folks, >>>>>>>>>> >>>>>>>>>> Just bought another Li-Ion deal with charger: >>>>>>>>>> >>>>>>>>>> https://images-na.ssl-images-amazon.com/images/ >>>>>>> I/41HL6Xmx1qL._SX466_.jpg >>>>>>>>>> >>>>>>>>>> The charger is like this little guy, a switcher that >>>>>>>>>> can deliver 1 amp: >>>>>>>>>> >>>>>>>>>> https://images-na.ssl-images-amazon.com/images/ >>>>>>> I/41HL6Xmx1qL._SX466_.jpg >>>>>>>>>> >>>>>>>>>> The typical battery is two cell bundles in series so >>>>>>>>>> 8.4V max. To my surprise the charger has 10.2V open >>>>>>>>>> loop voltage. It is not a switched deal to overcome >>>>>>>>>> wire resistance where it measures the battery voltage >>>>>>>>>> every once in a while. Just straight DC. It seems that >>>>>>>>>> it relies on the battery's "electric safety vent" to >>>>>>>>>> stop at 8.4V. Anybody seen this done? >>>>>>>>> >>>>>>>>> Hi J&ouml;rg >>>>>>>>> >>>>>>>>> Not exactly the same, but I've seen a similar >>>>>>>>> configuration in a Chinese MP3-playing toy. The >>>>>>>>> configuration looked like: >>>>>>>>> >>>>>>>>> USB port -> 10 Ohm series resistor -> silicon diode -> >>>>>>>>> LiPo battry. >>>>>>>>> >>>>>>>>> No other relevant components in the direct power path on >>>>>>>>> the main board, but the battery did have a small >>>>>>>>> protection board, apparently with the classic one chip >>>>>>>>> and 2 transistor solution that seems to be common >>>>>>>>> nowadays. >>>>>>>>> >>>>>>>>> The device in question did have an "interesting" quirk: >>>>>>>>> When fully charged, it would fail to operate when the >>>>>>>>> charger was disconnected. Switching it off and on with a >>>>>>>>> (mechanical) switch would reset it. >>>>>>>>> >>>>>>>>> >>>>>>>> Probably because the battery's OVP had come and needed to >>>>>>>> be "unstuck". >>>>>>>> >>>>>>>> >>>>>>>>> Apparently they indeed relied on the protection circuit >>>>>>>>> to work as a simple ON/OFF charge controller. It would >>>>>>>>> turn off on overvoltage and stay that way, so that when >>>>>>>>> the charger was disconnected, the sharp voltage drop to >>>>>>>>> zero (before the the protection has time to sense that >>>>>>>>> the overload has gone away) would upset the main ASIC, >>>>>>>>> which would get stuck somewhere in the wilderness of an >>>>>>>>> improperly timed reset cycle. >>>>>>>>> >>>>>>>>> >>>>>>>> A stuck ASIC would be poor design. >>>>>>>> >>>>>>>> >>>>>>>>> As for safety: There's a good reason to have control and >>>>>>>>> protection circuits separate. A single fault condition >>>>>>>>> can't normally kill both together (unless the protection >>>>>>>>> is way inadequate). With only one power interrupting >>>>>>>>> component (one MOSFET per direction of charge / >>>>>>>>> discharge current flow) all it takes is a single fault in >>>>>>>>> a semiconductor part to deadlock this thing into an >>>>>>>>> always on charging condition, with a third party >>>>>>>>> uncontrolled tolerance "nominally 5V" adapter for the >>>>>>>>> power. >>>>>>>>> >>>>>>>>> >>>>>>>> Single-fault safety is why I don't like the scheme where >>>>>>>> they use a higher external voltage. I can build a supply >>>>>>>> with exactly 8.2V but I tried that and it has one major >>>>>>>> disadvantage: As the battery voltage approaches 8.2V the >>>>>>>> charge current will become very small due to cable >>>>>>>> resistance. It'll take all day to recharge a battery to a >>>>>>>> reasonable level. That would present a minor inconvencience >>>>>>>> if I did a long ride late into the evening and wanted to >>>>>>>> head out again in the morning. For safety I ride with >>>>>>>> bright daytime lights. >>>>>>>> >>>>>>>> A scheme that would work but I haven't seen it implemented >>>>>>>> is this: Apply a current limited 10-12V externally and have >>>>>>>> the control electronics in the supply disconnect it every >>>>>>>> few seconds. Check open circuit voltage of the connected >>>>>>>> Li-Ion battery, if less than target voltage apply another >>>>>>>> burst. If at target voltage stop and signal a green LED for >>>>>>>> "charge is complete". >>>>>>>> >>>>>>>> If I wanted to do this right it'll be another electronics >>>>>>>> project and my hobby time is already occupied with >>>>>>>> bicycling and beer brewing. Maybe with a uC but as an >>>>>>>> analog guys I am not a great programmer. So possibly I'll >>>>>>>> do it analog. >>>>>>> >>>>>>> "proper" charging of a LiPo is 4.2V/cell (or whatever the >>>>>>> pack calls for) or some maximum current, whichever is lower. >>>>>>> Much like lead-acid, except with more potential for flame. >>>>>>> >>>>>>> The pack _should_ charge at maximum current until it hits >>>>>>> 4.2V/cell, then stay there with the battery determining the >>>>>>> current, then finally the charge should terminate when the >>>>>>> current drops to some fraction of the cell's maximum current >>>>>>> (IIRC 1/10th is the rule of thumb, but don't go designing a >>>>>>> charger based on that). >>>>>>> >>>>>> >>>>>> Ideally yes but for expedient charging this requires sense >>>>>> lines to the battery pack which, on a bicycle, you don't have. >>>>>> >>>>> >>>>> can't just use a four wire cable leaving only the wire from >>>>> connector to cells (which I'd expect to be very short) >>>>> unaccounted for? >>>>> >>>> >>>> It voids any warranty and liability protection if you cut open the >>>> enclosure of the battery pack. There are no screws, it has to be >>>> cut and pried away. >>> >>> but won't the wiring from whatever connector is on the battery pack >>> to the cells inside the pack be very limited? >>> >> >> It is generally a round cable with a standard connector. Looks like this: >> >> https://images-na.ssl-images-amazon.com/images/I/51J4F9sJoTL._SL1000_.jpg >> >> The copper inside isn't very thick, meaning it'll easily drop 100mV. >> Same for the charger cable. During discharge it doesn't matter much >> because good bicycle lights contain a buck converter. In my case that >> keeps the light level the same all the way down to <5V. It does matter >> when charging from a supply which is set to exactly the cut-off voltage >> of 8.2V. >> > > is there a separate cable/connector for charging and "discharging" >
No, same cable. But on my mountain bike I mounted this fairly heavy battery in a sturdy box, screwed down via a bracket but cushioned with rubber. Added a fuse and a switch. Two switched cables come out for front and back light plus a non-switched one for charging or connecting miscellaneous loads. I left the original cable on there, coiled up and strapped in. This has withstood hard rock hits. -- Regards, Joerg http://www.analogconsultants.com/
Reply by Lasse Langwadt Christensen October 6, 20162016-10-06
Den fredag den 7. oktober 2016 kl. 00.31.15 UTC+2 skrev Joerg:
> On 2016-10-06 15:26, Lasse Langwadt Christensen wrote: > > Den fredag den 7. oktober 2016 kl. 00.17.59 UTC+2 skrev Joerg: > >> On 2016-10-06 14:49, Lasse Langwadt Christensen wrote: > >>> Den torsdag den 6. oktober 2016 kl. 23.28.41 UTC+2 skrev Joerg: > >>>> On 2016-10-06 13:32, Tim Wescott wrote: > >>>>> On Thu, 06 Oct 2016 12:54:34 -0700, Joerg wrote: > >>>>> > >>>>>> On 2016-10-04 16:05, Dimitrij Klingbeil wrote: > >>>>>>> On 04.10.2016 01:27, Joerg wrote: > >>>>>>>> Hi Folks, > >>>>>>>> > >>>>>>>> Just bought another Li-Ion deal with charger: > >>>>>>>> > >>>>>>>> https://images-na.ssl-images-amazon.com/images/ > >>>>> I/41HL6Xmx1qL._SX466_.jpg > >>>>>>>> > >>>>>>>> The charger is like this little guy, a switcher that > >>>>>>>> can deliver 1 amp: > >>>>>>>> > >>>>>>>> https://images-na.ssl-images-amazon.com/images/ > >>>>> I/41HL6Xmx1qL._SX466_.jpg > >>>>>>>> > >>>>>>>> The typical battery is two cell bundles in series so > >>>>>>>> 8.4V max. To my surprise the charger has 10.2V open > >>>>>>>> loop voltage. It is not a switched deal to overcome > >>>>>>>> wire resistance where it measures the battery voltage > >>>>>>>> every once in a while. Just straight DC. It seems that > >>>>>>>> it relies on the battery's "electric safety vent" to > >>>>>>>> stop at 8.4V. Anybody seen this done? > >>>>>>> > >>>>>>> Hi J&ouml;rg > >>>>>>> > >>>>>>> Not exactly the same, but I've seen a similar > >>>>>>> configuration in a Chinese MP3-playing toy. The > >>>>>>> configuration looked like: > >>>>>>> > >>>>>>> USB port -> 10 Ohm series resistor -> silicon diode -> > >>>>>>> LiPo battry. > >>>>>>> > >>>>>>> No other relevant components in the direct power path on > >>>>>>> the main board, but the battery did have a small > >>>>>>> protection board, apparently with the classic one chip > >>>>>>> and 2 transistor solution that seems to be common > >>>>>>> nowadays. > >>>>>>> > >>>>>>> The device in question did have an "interesting" quirk: > >>>>>>> When fully charged, it would fail to operate when the > >>>>>>> charger was disconnected. Switching it off and on with a > >>>>>>> (mechanical) switch would reset it. > >>>>>>> > >>>>>>> > >>>>>> Probably because the battery's OVP had come and needed to > >>>>>> be "unstuck". > >>>>>> > >>>>>> > >>>>>>> Apparently they indeed relied on the protection circuit > >>>>>>> to work as a simple ON/OFF charge controller. It would > >>>>>>> turn off on overvoltage and stay that way, so that when > >>>>>>> the charger was disconnected, the sharp voltage drop to > >>>>>>> zero (before the the protection has time to sense that > >>>>>>> the overload has gone away) would upset the main ASIC, > >>>>>>> which would get stuck somewhere in the wilderness of an > >>>>>>> improperly timed reset cycle. > >>>>>>> > >>>>>>> > >>>>>> A stuck ASIC would be poor design. > >>>>>> > >>>>>> > >>>>>>> As for safety: There's a good reason to have control and > >>>>>>> protection circuits separate. A single fault condition > >>>>>>> can't normally kill both together (unless the protection > >>>>>>> is way inadequate). With only one power interrupting > >>>>>>> component (one MOSFET per direction of charge / > >>>>>>> discharge current flow) all it takes is a single fault in > >>>>>>> a semiconductor part to deadlock this thing into an > >>>>>>> always on charging condition, with a third party > >>>>>>> uncontrolled tolerance "nominally 5V" adapter for the > >>>>>>> power. > >>>>>>> > >>>>>>> > >>>>>> Single-fault safety is why I don't like the scheme where > >>>>>> they use a higher external voltage. I can build a supply > >>>>>> with exactly 8.2V but I tried that and it has one major > >>>>>> disadvantage: As the battery voltage approaches 8.2V the > >>>>>> charge current will become very small due to cable > >>>>>> resistance. It'll take all day to recharge a battery to a > >>>>>> reasonable level. That would present a minor inconvencience > >>>>>> if I did a long ride late into the evening and wanted to > >>>>>> head out again in the morning. For safety I ride with > >>>>>> bright daytime lights. > >>>>>> > >>>>>> A scheme that would work but I haven't seen it implemented > >>>>>> is this: Apply a current limited 10-12V externally and have > >>>>>> the control electronics in the supply disconnect it every > >>>>>> few seconds. Check open circuit voltage of the connected > >>>>>> Li-Ion battery, if less than target voltage apply another > >>>>>> burst. If at target voltage stop and signal a green LED for > >>>>>> "charge is complete". > >>>>>> > >>>>>> If I wanted to do this right it'll be another electronics > >>>>>> project and my hobby time is already occupied with > >>>>>> bicycling and beer brewing. Maybe with a uC but as an > >>>>>> analog guys I am not a great programmer. So possibly I'll > >>>>>> do it analog. > >>>>> > >>>>> "proper" charging of a LiPo is 4.2V/cell (or whatever the > >>>>> pack calls for) or some maximum current, whichever is lower. > >>>>> Much like lead-acid, except with more potential for flame. > >>>>> > >>>>> The pack _should_ charge at maximum current until it hits > >>>>> 4.2V/cell, then stay there with the battery determining the > >>>>> current, then finally the charge should terminate when the > >>>>> current drops to some fraction of the cell's maximum current > >>>>> (IIRC 1/10th is the rule of thumb, but don't go designing a > >>>>> charger based on that). > >>>>> > >>>> > >>>> Ideally yes but for expedient charging this requires sense > >>>> lines to the battery pack which, on a bicycle, you don't have. > >>>> > >>> > >>> can't just use a four wire cable leaving only the wire from > >>> connector to cells (which I'd expect to be very short) > >>> unaccounted for? > >>> > >> > >> It voids any warranty and liability protection if you cut open the > >> enclosure of the battery pack. There are no screws, it has to be > >> cut and pried away. > > > > but won't the wiring from whatever connector is on the battery pack > > to the cells inside the pack be very limited? > > > > It is generally a round cable with a standard connector. Looks like this: > > https://images-na.ssl-images-amazon.com/images/I/51J4F9sJoTL._SL1000_.jpg > > The copper inside isn't very thick, meaning it'll easily drop 100mV. > Same for the charger cable. During discharge it doesn't matter much > because good bicycle lights contain a buck converter. In my case that > keeps the light level the same all the way down to <5V. It does matter > when charging from a supply which is set to exactly the cut-off voltage > of 8.2V. >
is there a separate cable/connector for charging and "discharging" -Lasse
Reply by Joerg October 6, 20162016-10-06
On 2016-10-06 15:26, Lasse Langwadt Christensen wrote:
> Den fredag den 7. oktober 2016 kl. 00.17.59 UTC+2 skrev Joerg: >> On 2016-10-06 14:49, Lasse Langwadt Christensen wrote: >>> Den torsdag den 6. oktober 2016 kl. 23.28.41 UTC+2 skrev Joerg: >>>> On 2016-10-06 13:32, Tim Wescott wrote: >>>>> On Thu, 06 Oct 2016 12:54:34 -0700, Joerg wrote: >>>>> >>>>>> On 2016-10-04 16:05, Dimitrij Klingbeil wrote: >>>>>>> On 04.10.2016 01:27, Joerg wrote: >>>>>>>> Hi Folks, >>>>>>>> >>>>>>>> Just bought another Li-Ion deal with charger: >>>>>>>> >>>>>>>> https://images-na.ssl-images-amazon.com/images/ >>>>> I/41HL6Xmx1qL._SX466_.jpg >>>>>>>> >>>>>>>> The charger is like this little guy, a switcher that >>>>>>>> can deliver 1 amp: >>>>>>>> >>>>>>>> https://images-na.ssl-images-amazon.com/images/ >>>>> I/41HL6Xmx1qL._SX466_.jpg >>>>>>>> >>>>>>>> The typical battery is two cell bundles in series so >>>>>>>> 8.4V max. To my surprise the charger has 10.2V open >>>>>>>> loop voltage. It is not a switched deal to overcome >>>>>>>> wire resistance where it measures the battery voltage >>>>>>>> every once in a while. Just straight DC. It seems that >>>>>>>> it relies on the battery's "electric safety vent" to >>>>>>>> stop at 8.4V. Anybody seen this done? >>>>>>> >>>>>>> Hi J&ouml;rg >>>>>>> >>>>>>> Not exactly the same, but I've seen a similar >>>>>>> configuration in a Chinese MP3-playing toy. The >>>>>>> configuration looked like: >>>>>>> >>>>>>> USB port -> 10 Ohm series resistor -> silicon diode -> >>>>>>> LiPo battry. >>>>>>> >>>>>>> No other relevant components in the direct power path on >>>>>>> the main board, but the battery did have a small >>>>>>> protection board, apparently with the classic one chip >>>>>>> and 2 transistor solution that seems to be common >>>>>>> nowadays. >>>>>>> >>>>>>> The device in question did have an "interesting" quirk: >>>>>>> When fully charged, it would fail to operate when the >>>>>>> charger was disconnected. Switching it off and on with a >>>>>>> (mechanical) switch would reset it. >>>>>>> >>>>>>> >>>>>> Probably because the battery's OVP had come and needed to >>>>>> be "unstuck". >>>>>> >>>>>> >>>>>>> Apparently they indeed relied on the protection circuit >>>>>>> to work as a simple ON/OFF charge controller. It would >>>>>>> turn off on overvoltage and stay that way, so that when >>>>>>> the charger was disconnected, the sharp voltage drop to >>>>>>> zero (before the the protection has time to sense that >>>>>>> the overload has gone away) would upset the main ASIC, >>>>>>> which would get stuck somewhere in the wilderness of an >>>>>>> improperly timed reset cycle. >>>>>>> >>>>>>> >>>>>> A stuck ASIC would be poor design. >>>>>> >>>>>> >>>>>>> As for safety: There's a good reason to have control and >>>>>>> protection circuits separate. A single fault condition >>>>>>> can't normally kill both together (unless the protection >>>>>>> is way inadequate). With only one power interrupting >>>>>>> component (one MOSFET per direction of charge / >>>>>>> discharge current flow) all it takes is a single fault in >>>>>>> a semiconductor part to deadlock this thing into an >>>>>>> always on charging condition, with a third party >>>>>>> uncontrolled tolerance "nominally 5V" adapter for the >>>>>>> power. >>>>>>> >>>>>>> >>>>>> Single-fault safety is why I don't like the scheme where >>>>>> they use a higher external voltage. I can build a supply >>>>>> with exactly 8.2V but I tried that and it has one major >>>>>> disadvantage: As the battery voltage approaches 8.2V the >>>>>> charge current will become very small due to cable >>>>>> resistance. It'll take all day to recharge a battery to a >>>>>> reasonable level. That would present a minor inconvencience >>>>>> if I did a long ride late into the evening and wanted to >>>>>> head out again in the morning. For safety I ride with >>>>>> bright daytime lights. >>>>>> >>>>>> A scheme that would work but I haven't seen it implemented >>>>>> is this: Apply a current limited 10-12V externally and have >>>>>> the control electronics in the supply disconnect it every >>>>>> few seconds. Check open circuit voltage of the connected >>>>>> Li-Ion battery, if less than target voltage apply another >>>>>> burst. If at target voltage stop and signal a green LED for >>>>>> "charge is complete". >>>>>> >>>>>> If I wanted to do this right it'll be another electronics >>>>>> project and my hobby time is already occupied with >>>>>> bicycling and beer brewing. Maybe with a uC but as an >>>>>> analog guys I am not a great programmer. So possibly I'll >>>>>> do it analog. >>>>> >>>>> "proper" charging of a LiPo is 4.2V/cell (or whatever the >>>>> pack calls for) or some maximum current, whichever is lower. >>>>> Much like lead-acid, except with more potential for flame. >>>>> >>>>> The pack _should_ charge at maximum current until it hits >>>>> 4.2V/cell, then stay there with the battery determining the >>>>> current, then finally the charge should terminate when the >>>>> current drops to some fraction of the cell's maximum current >>>>> (IIRC 1/10th is the rule of thumb, but don't go designing a >>>>> charger based on that). >>>>> >>>> >>>> Ideally yes but for expedient charging this requires sense >>>> lines to the battery pack which, on a bicycle, you don't have. >>>> >>> >>> can't just use a four wire cable leaving only the wire from >>> connector to cells (which I'd expect to be very short) >>> unaccounted for? >>> >> >> It voids any warranty and liability protection if you cut open the >> enclosure of the battery pack. There are no screws, it has to be >> cut and pried away. > > but won't the wiring from whatever connector is on the battery pack > to the cells inside the pack be very limited? >
It is generally a round cable with a standard connector. Looks like this: https://images-na.ssl-images-amazon.com/images/I/51J4F9sJoTL._SL1000_.jpg The copper inside isn't very thick, meaning it'll easily drop 100mV. Same for the charger cable. During discharge it doesn't matter much because good bicycle lights contain a buck converter. In my case that keeps the light level the same all the way down to <5V. It does matter when charging from a supply which is set to exactly the cut-off voltage of 8.2V. -- Regards, Joerg http://www.analogconsultants.com/
Reply by Lasse Langwadt Christensen October 6, 20162016-10-06
Den fredag den 7. oktober 2016 kl. 00.17.59 UTC+2 skrev Joerg:
> On 2016-10-06 14:49, Lasse Langwadt Christensen wrote: > > Den torsdag den 6. oktober 2016 kl. 23.28.41 UTC+2 skrev Joerg: > >> On 2016-10-06 13:32, Tim Wescott wrote: > >>> On Thu, 06 Oct 2016 12:54:34 -0700, Joerg wrote: > >>> > >>>> On 2016-10-04 16:05, Dimitrij Klingbeil wrote: > >>>>> On 04.10.2016 01:27, Joerg wrote: > >>>>>> Hi Folks, > >>>>>> > >>>>>> Just bought another Li-Ion deal with charger: > >>>>>> > >>>>>> https://images-na.ssl-images-amazon.com/images/ > >>> I/41HL6Xmx1qL._SX466_.jpg > >>>>>> > >>>>>> The charger is like this little guy, a switcher that can deliver 1 > >>>>>> amp: > >>>>>> > >>>>>> https://images-na.ssl-images-amazon.com/images/ > >>> I/41HL6Xmx1qL._SX466_.jpg > >>>>>> > >>>>>> The typical battery is two cell bundles in series so 8.4V max. To > >>>>>> my surprise the charger has 10.2V open loop voltage. It is not a > >>>>>> switched deal to overcome wire resistance where it measures the > >>>>>> battery voltage every once in a while. Just straight DC. It seems that > >>>>>> it relies on the battery's "electric safety vent" to stop at 8.4V. > >>>>>> Anybody seen this done? > >>>>> > >>>>> Hi J&ouml;rg > >>>>> > >>>>> Not exactly the same, but I've seen a similar configuration in a > >>>>> Chinese MP3-playing toy. The configuration looked like: > >>>>> > >>>>> USB port -> 10 Ohm series resistor -> silicon diode -> LiPo battry. > >>>>> > >>>>> No other relevant components in the direct power path on the main > >>>>> board, > >>>>> but the battery did have a small protection board, apparently with the > >>>>> classic one chip and 2 transistor solution that seems to be common > >>>>> nowadays. > >>>>> > >>>>> The device in question did have an "interesting" quirk: When fully > >>>>> charged, it would fail to operate when the charger was disconnected. > >>>>> Switching it off and on with a (mechanical) switch would reset it. > >>>>> > >>>>> > >>>> Probably because the battery's OVP had come and needed to be "unstuck". > >>>> > >>>> > >>>>> Apparently they indeed relied on the protection circuit to work as a > >>>>> simple ON/OFF charge controller. It would turn off on overvoltage and > >>>>> stay that way, so that when the charger was disconnected, the sharp > >>>>> voltage drop to zero (before the the protection has time to sense that > >>>>> the overload has gone away) would upset the main ASIC, which would get > >>>>> stuck somewhere in the wilderness of an improperly timed reset cycle. > >>>>> > >>>>> > >>>> A stuck ASIC would be poor design. > >>>> > >>>> > >>>>> As for safety: There's a good reason to have control and protection > >>>>> circuits separate. A single fault condition can't normally kill both > >>>>> together (unless the protection is way inadequate). With only one power > >>>>> interrupting component (one MOSFET per direction of charge / discharge > >>>>> current flow) all it takes is a single fault in a semiconductor part to > >>>>> deadlock this thing into an always on charging condition, with a third > >>>>> party uncontrolled tolerance "nominally 5V" adapter for the power. > >>>>> > >>>>> > >>>> Single-fault safety is why I don't like the scheme where they use a > >>>> higher external voltage. I can build a supply with exactly 8.2V but I > >>>> tried that and it has one major disadvantage: As the battery voltage > >>>> approaches 8.2V the charge current will become very small due to cable > >>>> resistance. It'll take all day to recharge a battery to a reasonable > >>>> level. That would present a minor inconvencience if I did a long ride > >>>> late into the evening and wanted to head out again in the morning. For > >>>> safety I ride with bright daytime lights. > >>>> > >>>> A scheme that would work but I haven't seen it implemented is this: > >>>> Apply a current limited 10-12V externally and have the control > >>>> electronics in the supply disconnect it every few seconds. Check open > >>>> circuit voltage of the connected Li-Ion battery, if less than target > >>>> voltage apply another burst. If at target voltage stop and signal a > >>>> green LED for "charge is complete". > >>>> > >>>> If I wanted to do this right it'll be another electronics project and my > >>>> hobby time is already occupied with bicycling and beer brewing. Maybe > >>>> with a uC but as an analog guys I am not a great programmer. So possibly > >>>> I'll do it analog. > >>> > >>> "proper" charging of a LiPo is 4.2V/cell (or whatever the pack calls for) > >>> or some maximum current, whichever is lower. Much like lead-acid, except > >>> with more potential for flame. > >>> > >>> The pack _should_ charge at maximum current until it hits 4.2V/cell, then > >>> stay there with the battery determining the current, then finally the > >>> charge should terminate when the current drops to some fraction of the > >>> cell's maximum current (IIRC 1/10th is the rule of thumb, but don't go > >>> designing a charger based on that). > >>> > >> > >> Ideally yes but for expedient charging this requires sense lines to the > >> battery pack which, on a bicycle, you don't have. > >> > > > > can't just use a four wire cable leaving only the wire from connector > > to cells (which I'd expect to be very short) unaccounted for? > > > > It voids any warranty and liability protection if you cut open the > enclosure of the battery pack. There are no screws, it has to be cut and > pried away.
but won't the wiring from whatever connector is on the battery pack to the cells inside the pack be very limited? -Lasse
Reply by Joerg October 6, 20162016-10-06
On 2016-10-06 14:49, Lasse Langwadt Christensen wrote:
> Den torsdag den 6. oktober 2016 kl. 23.28.41 UTC+2 skrev Joerg: >> On 2016-10-06 13:32, Tim Wescott wrote: >>> On Thu, 06 Oct 2016 12:54:34 -0700, Joerg wrote: >>> >>>> On 2016-10-04 16:05, Dimitrij Klingbeil wrote: >>>>> On 04.10.2016 01:27, Joerg wrote: >>>>>> Hi Folks, >>>>>> >>>>>> Just bought another Li-Ion deal with charger: >>>>>> >>>>>> https://images-na.ssl-images-amazon.com/images/ >>> I/41HL6Xmx1qL._SX466_.jpg >>>>>> >>>>>> The charger is like this little guy, a switcher that can deliver 1 >>>>>> amp: >>>>>> >>>>>> https://images-na.ssl-images-amazon.com/images/ >>> I/41HL6Xmx1qL._SX466_.jpg >>>>>> >>>>>> The typical battery is two cell bundles in series so 8.4V max. To >>>>>> my surprise the charger has 10.2V open loop voltage. It is not a >>>>>> switched deal to overcome wire resistance where it measures the >>>>>> battery voltage every once in a while. Just straight DC. It seems that >>>>>> it relies on the battery's "electric safety vent" to stop at 8.4V. >>>>>> Anybody seen this done? >>>>> >>>>> Hi J&ouml;rg >>>>> >>>>> Not exactly the same, but I've seen a similar configuration in a >>>>> Chinese MP3-playing toy. The configuration looked like: >>>>> >>>>> USB port -> 10 Ohm series resistor -> silicon diode -> LiPo battry. >>>>> >>>>> No other relevant components in the direct power path on the main >>>>> board, >>>>> but the battery did have a small protection board, apparently with the >>>>> classic one chip and 2 transistor solution that seems to be common >>>>> nowadays. >>>>> >>>>> The device in question did have an "interesting" quirk: When fully >>>>> charged, it would fail to operate when the charger was disconnected. >>>>> Switching it off and on with a (mechanical) switch would reset it. >>>>> >>>>> >>>> Probably because the battery's OVP had come and needed to be "unstuck". >>>> >>>> >>>>> Apparently they indeed relied on the protection circuit to work as a >>>>> simple ON/OFF charge controller. It would turn off on overvoltage and >>>>> stay that way, so that when the charger was disconnected, the sharp >>>>> voltage drop to zero (before the the protection has time to sense that >>>>> the overload has gone away) would upset the main ASIC, which would get >>>>> stuck somewhere in the wilderness of an improperly timed reset cycle. >>>>> >>>>> >>>> A stuck ASIC would be poor design. >>>> >>>> >>>>> As for safety: There's a good reason to have control and protection >>>>> circuits separate. A single fault condition can't normally kill both >>>>> together (unless the protection is way inadequate). With only one power >>>>> interrupting component (one MOSFET per direction of charge / discharge >>>>> current flow) all it takes is a single fault in a semiconductor part to >>>>> deadlock this thing into an always on charging condition, with a third >>>>> party uncontrolled tolerance "nominally 5V" adapter for the power. >>>>> >>>>> >>>> Single-fault safety is why I don't like the scheme where they use a >>>> higher external voltage. I can build a supply with exactly 8.2V but I >>>> tried that and it has one major disadvantage: As the battery voltage >>>> approaches 8.2V the charge current will become very small due to cable >>>> resistance. It'll take all day to recharge a battery to a reasonable >>>> level. That would present a minor inconvencience if I did a long ride >>>> late into the evening and wanted to head out again in the morning. For >>>> safety I ride with bright daytime lights. >>>> >>>> A scheme that would work but I haven't seen it implemented is this: >>>> Apply a current limited 10-12V externally and have the control >>>> electronics in the supply disconnect it every few seconds. Check open >>>> circuit voltage of the connected Li-Ion battery, if less than target >>>> voltage apply another burst. If at target voltage stop and signal a >>>> green LED for "charge is complete". >>>> >>>> If I wanted to do this right it'll be another electronics project and my >>>> hobby time is already occupied with bicycling and beer brewing. Maybe >>>> with a uC but as an analog guys I am not a great programmer. So possibly >>>> I'll do it analog. >>> >>> "proper" charging of a LiPo is 4.2V/cell (or whatever the pack calls for) >>> or some maximum current, whichever is lower. Much like lead-acid, except >>> with more potential for flame. >>> >>> The pack _should_ charge at maximum current until it hits 4.2V/cell, then >>> stay there with the battery determining the current, then finally the >>> charge should terminate when the current drops to some fraction of the >>> cell's maximum current (IIRC 1/10th is the rule of thumb, but don't go >>> designing a charger based on that). >>> >> >> Ideally yes but for expedient charging this requires sense lines to the >> battery pack which, on a bicycle, you don't have. >> > > can't just use a four wire cable leaving only the wire from connector > to cells (which I'd expect to be very short) unaccounted for? >
It voids any warranty and liability protection if you cut open the enclosure of the battery pack. There are no screws, it has to be cut and pried away. -- Regards, Joerg http://www.analogconsultants.com/
Reply by Lasse Langwadt Christensen October 6, 20162016-10-06
Den torsdag den 6. oktober 2016 kl. 23.28.41 UTC+2 skrev Joerg:
> On 2016-10-06 13:32, Tim Wescott wrote: > > On Thu, 06 Oct 2016 12:54:34 -0700, Joerg wrote: > > > >> On 2016-10-04 16:05, Dimitrij Klingbeil wrote: > >>> On 04.10.2016 01:27, Joerg wrote: > >>>> Hi Folks, > >>>> > >>>> Just bought another Li-Ion deal with charger: > >>>> > >>>> https://images-na.ssl-images-amazon.com/images/ > > I/41HL6Xmx1qL._SX466_.jpg > >>>> > >>>> The charger is like this little guy, a switcher that can deliver 1 > >>>> amp: > >>>> > >>>> https://images-na.ssl-images-amazon.com/images/ > > I/41HL6Xmx1qL._SX466_.jpg > >>>> > >>>> The typical battery is two cell bundles in series so 8.4V max. To > >>>> my surprise the charger has 10.2V open loop voltage. It is not a > >>>> switched deal to overcome wire resistance where it measures the > >>>> battery voltage every once in a while. Just straight DC. It seems that > >>>> it relies on the battery's "electric safety vent" to stop at 8.4V. > >>>> Anybody seen this done? > >>> > >>> Hi J&ouml;rg > >>> > >>> Not exactly the same, but I've seen a similar configuration in a > >>> Chinese MP3-playing toy. The configuration looked like: > >>> > >>> USB port -> 10 Ohm series resistor -> silicon diode -> LiPo battry. > >>> > >>> No other relevant components in the direct power path on the main > >>> board, > >>> but the battery did have a small protection board, apparently with the > >>> classic one chip and 2 transistor solution that seems to be common > >>> nowadays. > >>> > >>> The device in question did have an "interesting" quirk: When fully > >>> charged, it would fail to operate when the charger was disconnected. > >>> Switching it off and on with a (mechanical) switch would reset it. > >>> > >>> > >> Probably because the battery's OVP had come and needed to be "unstuck". > >> > >> > >>> Apparently they indeed relied on the protection circuit to work as a > >>> simple ON/OFF charge controller. It would turn off on overvoltage and > >>> stay that way, so that when the charger was disconnected, the sharp > >>> voltage drop to zero (before the the protection has time to sense that > >>> the overload has gone away) would upset the main ASIC, which would get > >>> stuck somewhere in the wilderness of an improperly timed reset cycle. > >>> > >>> > >> A stuck ASIC would be poor design. > >> > >> > >>> As for safety: There's a good reason to have control and protection > >>> circuits separate. A single fault condition can't normally kill both > >>> together (unless the protection is way inadequate). With only one power > >>> interrupting component (one MOSFET per direction of charge / discharge > >>> current flow) all it takes is a single fault in a semiconductor part to > >>> deadlock this thing into an always on charging condition, with a third > >>> party uncontrolled tolerance "nominally 5V" adapter for the power. > >>> > >>> > >> Single-fault safety is why I don't like the scheme where they use a > >> higher external voltage. I can build a supply with exactly 8.2V but I > >> tried that and it has one major disadvantage: As the battery voltage > >> approaches 8.2V the charge current will become very small due to cable > >> resistance. It'll take all day to recharge a battery to a reasonable > >> level. That would present a minor inconvencience if I did a long ride > >> late into the evening and wanted to head out again in the morning. For > >> safety I ride with bright daytime lights. > >> > >> A scheme that would work but I haven't seen it implemented is this: > >> Apply a current limited 10-12V externally and have the control > >> electronics in the supply disconnect it every few seconds. Check open > >> circuit voltage of the connected Li-Ion battery, if less than target > >> voltage apply another burst. If at target voltage stop and signal a > >> green LED for "charge is complete". > >> > >> If I wanted to do this right it'll be another electronics project and my > >> hobby time is already occupied with bicycling and beer brewing. Maybe > >> with a uC but as an analog guys I am not a great programmer. So possibly > >> I'll do it analog. > > > > "proper" charging of a LiPo is 4.2V/cell (or whatever the pack calls for) > > or some maximum current, whichever is lower. Much like lead-acid, except > > with more potential for flame. > > > > The pack _should_ charge at maximum current until it hits 4.2V/cell, then > > stay there with the battery determining the current, then finally the > > charge should terminate when the current drops to some fraction of the > > cell's maximum current (IIRC 1/10th is the rule of thumb, but don't go > > designing a charger based on that). > > > > Ideally yes but for expedient charging this requires sense lines to the > battery pack which, on a bicycle, you don't have. >
can't just use a four wire cable leaving only the wire from connector to cells (which I'd expect to be very short) unaccounted for? -Lasse
Reply by Joerg October 6, 20162016-10-06
On 2016-10-06 13:32, Tim Wescott wrote:
> On Thu, 06 Oct 2016 12:54:34 -0700, Joerg wrote: > >> On 2016-10-04 16:05, Dimitrij Klingbeil wrote: >>> On 04.10.2016 01:27, Joerg wrote: >>>> Hi Folks, >>>> >>>> Just bought another Li-Ion deal with charger: >>>> >>>> https://images-na.ssl-images-amazon.com/images/ > I/41HL6Xmx1qL._SX466_.jpg >>>> >>>> The charger is like this little guy, a switcher that can deliver 1 >>>> amp: >>>> >>>> https://images-na.ssl-images-amazon.com/images/ > I/41HL6Xmx1qL._SX466_.jpg >>>> >>>> The typical battery is two cell bundles in series so 8.4V max. To >>>> my surprise the charger has 10.2V open loop voltage. It is not a >>>> switched deal to overcome wire resistance where it measures the >>>> battery voltage every once in a while. Just straight DC. It seems that >>>> it relies on the battery's "electric safety vent" to stop at 8.4V. >>>> Anybody seen this done? >>> >>> Hi J&ouml;rg >>> >>> Not exactly the same, but I've seen a similar configuration in a >>> Chinese MP3-playing toy. The configuration looked like: >>> >>> USB port -> 10 Ohm series resistor -> silicon diode -> LiPo battry. >>> >>> No other relevant components in the direct power path on the main >>> board, >>> but the battery did have a small protection board, apparently with the >>> classic one chip and 2 transistor solution that seems to be common >>> nowadays. >>> >>> The device in question did have an "interesting" quirk: When fully >>> charged, it would fail to operate when the charger was disconnected. >>> Switching it off and on with a (mechanical) switch would reset it. >>> >>> >> Probably because the battery's OVP had come and needed to be "unstuck". >> >> >>> Apparently they indeed relied on the protection circuit to work as a >>> simple ON/OFF charge controller. It would turn off on overvoltage and >>> stay that way, so that when the charger was disconnected, the sharp >>> voltage drop to zero (before the the protection has time to sense that >>> the overload has gone away) would upset the main ASIC, which would get >>> stuck somewhere in the wilderness of an improperly timed reset cycle. >>> >>> >> A stuck ASIC would be poor design. >> >> >>> As for safety: There's a good reason to have control and protection >>> circuits separate. A single fault condition can't normally kill both >>> together (unless the protection is way inadequate). With only one power >>> interrupting component (one MOSFET per direction of charge / discharge >>> current flow) all it takes is a single fault in a semiconductor part to >>> deadlock this thing into an always on charging condition, with a third >>> party uncontrolled tolerance "nominally 5V" adapter for the power. >>> >>> >> Single-fault safety is why I don't like the scheme where they use a >> higher external voltage. I can build a supply with exactly 8.2V but I >> tried that and it has one major disadvantage: As the battery voltage >> approaches 8.2V the charge current will become very small due to cable >> resistance. It'll take all day to recharge a battery to a reasonable >> level. That would present a minor inconvencience if I did a long ride >> late into the evening and wanted to head out again in the morning. For >> safety I ride with bright daytime lights. >> >> A scheme that would work but I haven't seen it implemented is this: >> Apply a current limited 10-12V externally and have the control >> electronics in the supply disconnect it every few seconds. Check open >> circuit voltage of the connected Li-Ion battery, if less than target >> voltage apply another burst. If at target voltage stop and signal a >> green LED for "charge is complete". >> >> If I wanted to do this right it'll be another electronics project and my >> hobby time is already occupied with bicycling and beer brewing. Maybe >> with a uC but as an analog guys I am not a great programmer. So possibly >> I'll do it analog. > > "proper" charging of a LiPo is 4.2V/cell (or whatever the pack calls for) > or some maximum current, whichever is lower. Much like lead-acid, except > with more potential for flame. > > The pack _should_ charge at maximum current until it hits 4.2V/cell, then > stay there with the battery determining the current, then finally the > charge should terminate when the current drops to some fraction of the > cell's maximum current (IIRC 1/10th is the rule of thumb, but don't go > designing a charger based on that). >
Ideally yes but for expedient charging this requires sense lines to the battery pack which, on a bicycle, you don't have. -- Regards, Joerg http://www.analogconsultants.com/
Reply by Tim Wescott October 6, 20162016-10-06
On Thu, 06 Oct 2016 12:54:34 -0700, Joerg wrote:

> On 2016-10-04 16:05, Dimitrij Klingbeil wrote: >> On 04.10.2016 01:27, Joerg wrote: >>> Hi Folks, >>> >>> Just bought another Li-Ion deal with charger: >>> >>> https://images-na.ssl-images-amazon.com/images/
I/41HL6Xmx1qL._SX466_.jpg
>>> >>> The charger is like this little guy, a switcher that can deliver 1 >>> amp: >>> >>> https://images-na.ssl-images-amazon.com/images/
I/41HL6Xmx1qL._SX466_.jpg
>>> >>> The typical battery is two cell bundles in series so 8.4V max. To >>> my surprise the charger has 10.2V open loop voltage. It is not a >>> switched deal to overcome wire resistance where it measures the >>> battery voltage every once in a while. Just straight DC. It seems that >>> it relies on the battery's "electric safety vent" to stop at 8.4V. >>> Anybody seen this done? >> >> Hi J&ouml;rg >> >> Not exactly the same, but I've seen a similar configuration in a >> Chinese MP3-playing toy. The configuration looked like: >> >> USB port -> 10 Ohm series resistor -> silicon diode -> LiPo battry. >> >> No other relevant components in the direct power path on the main >> board, >> but the battery did have a small protection board, apparently with the >> classic one chip and 2 transistor solution that seems to be common >> nowadays. >> >> The device in question did have an "interesting" quirk: When fully >> charged, it would fail to operate when the charger was disconnected. >> Switching it off and on with a (mechanical) switch would reset it. >> >> > Probably because the battery's OVP had come and needed to be "unstuck". > > >> Apparently they indeed relied on the protection circuit to work as a >> simple ON/OFF charge controller. It would turn off on overvoltage and >> stay that way, so that when the charger was disconnected, the sharp >> voltage drop to zero (before the the protection has time to sense that >> the overload has gone away) would upset the main ASIC, which would get >> stuck somewhere in the wilderness of an improperly timed reset cycle. >> >> > A stuck ASIC would be poor design. > > >> As for safety: There's a good reason to have control and protection >> circuits separate. A single fault condition can't normally kill both >> together (unless the protection is way inadequate). With only one power >> interrupting component (one MOSFET per direction of charge / discharge >> current flow) all it takes is a single fault in a semiconductor part to >> deadlock this thing into an always on charging condition, with a third >> party uncontrolled tolerance "nominally 5V" adapter for the power. >> >> > Single-fault safety is why I don't like the scheme where they use a > higher external voltage. I can build a supply with exactly 8.2V but I > tried that and it has one major disadvantage: As the battery voltage > approaches 8.2V the charge current will become very small due to cable > resistance. It'll take all day to recharge a battery to a reasonable > level. That would present a minor inconvencience if I did a long ride > late into the evening and wanted to head out again in the morning. For > safety I ride with bright daytime lights. > > A scheme that would work but I haven't seen it implemented is this: > Apply a current limited 10-12V externally and have the control > electronics in the supply disconnect it every few seconds. Check open > circuit voltage of the connected Li-Ion battery, if less than target > voltage apply another burst. If at target voltage stop and signal a > green LED for "charge is complete". > > If I wanted to do this right it'll be another electronics project and my > hobby time is already occupied with bicycling and beer brewing. Maybe > with a uC but as an analog guys I am not a great programmer. So possibly > I'll do it analog.
"proper" charging of a LiPo is 4.2V/cell (or whatever the pack calls for) or some maximum current, whichever is lower. Much like lead-acid, except with more potential for flame. The pack _should_ charge at maximum current until it hits 4.2V/cell, then stay there with the battery determining the current, then finally the charge should terminate when the current drops to some fraction of the cell's maximum current (IIRC 1/10th is the rule of thumb, but don't go designing a charger based on that). -- Tim Wescott Wescott Design Services http://www.wescottdesign.com I'm looking for work -- see my website!