Battery switching without tears

On Friday, April 10, 2020 at 6:07:46 PM UTC-7, 2G wrote:
On Friday, April 10, 2020 at 4:58:04 PM UTC-7, jfitch wrote:
On Wednesday, April 8, 2020 at 2:51:22 PM UTC-7, 2G wrote:
On Wednesday, April 8, 2020 at 9:04:00 AM UTC-7, jfitch wrote:
On Tuesday, April 7, 2020 at 6:19:36 PM UTC-7, 2G wrote:
On Tuesday, April 7, 2020 at 9:32:07 AM UTC-7, Dan Marotta wrote:
Well then, I must be really confused.Â* If the switch was a BREAK before
MAKE, the equipment would be momentarily without power.Â* Some devices
can survive that, perhaps with internal capacitance, but others will
lose power momentarily.Â* This may or may not be an issue.

Inrush current seems to be a big bugaboo to some people and, in some
cases it is, but what is the time constant associated with the inrush?Â*
How much current are we talking about and for how long? People keep
talking about the high battery "charging" the low battery during the
milliseconds that they are in parallel. Theoretically, yes, practically,
hogwash.Â* The heat required to blow a fuse or burn a wire does not rise
instantaneously; there's counter EMF to reduce the current...Â* So many
details that meant a bunch when you were taking a test back in school
but, practically speaking, don't mean squat in this case.

So, switch your batteries and protect your circuits however you wish.Â*
I'll stick with my make before break switches.Â* I'll even report back
when something fails due to their use, but don't hold your breath.

On 4/6/2020 10:43 PM, 2G wrote:
On Monday, April 6, 2020 at 11:23:21 AM UTC-7, Dan Marotta wrote:
Tom,

Isn't the rotary panel switch in a Schleicher glider a make before break?

OK, I'll stop posting to this thread now.Â* My solution, which I've
relied on for years without ever losing a connection or blowing a fuse
seems to be too simple to be acceptable.

On 4/5/2020 7:45 PM, 2G wrote:
The glider, an ASH 31 Mi, is already wired with a battery selector switch for the avionics, and is the way to go. You definitely don't want to parallel Pb and LFP battery's accidentally.

Tom
--
Dan, 5J
Dan,

Think about it: if it were MAKE before BREAK I wouldn't have had this problem and sure as hell wouldn't post a solution to a non-existent problem.

--
Dan, 5J

If my power switch was a break-before-make (also called a shorting switch) I would not have had a problem and certainly wouldn't have posted a fix to a non-existent problem. But it isn't. It was easier for me to put in a capacitor than replace the switch. It also protects against any switch bounce while switching.

People seem to be overly concerned with inrush current. High inrush current would only happen if there were a substantial amount of power supply capacitance in the LX9000 that needed charging. If that were the case it would handle a millisecond battery switch over without any problem, but it doesn't. Also, there is no inrush current during battery switching as the caps in the LX9000 are already charged.

Tom

The inrush current to the LX9000 is not the issue, nor is switching between batteries while the cap is charged. The issue (if there is one) is the inrush to the discharged cap when you first turn on the mains. It is a pretty complex and loosely speced system. Caps are very loosely spec'd, as are battery internal resistance, switch resistance, and of course the wiring in each glider is unique. That's why I suggested measuring it. When he switch is turned on, it will arc and bounce because that is what switches do.. How much is the question. Quite easy to measure, hard to calculate accurately. It is quite possible that the step impedance in the system limits the inrush to an acceptable value - but you don't know what you don't know.

It would have been easy for LX to spec a switching power supply with enough residual energy to cover, say 20 ms switching interval. Apparently they did not. Or perhaps they are using linear supplies, in which case the garbage can is the appropriate resting place. I'll second Dave's suggestion of using one large battery bank instead of seperate small ones. It is better for battery life too. (In the 31 the second battery may be the starting battery so that isn't practical). Or better still, use a higher capacity LFP - my small instrument battery will power the panel for around 12 hours.

All decent suggestions. I can deduce the series resistance from the time constant of the waveforms, which is about 200 mohm. This includes the battery internal resistance, wiring resistance, and capacitor ESR. So forget about "hundreds of amps" flowing. Any more measurements will have to wait until I get out to the airport again, but including a series resistor of a few ohms, depending upon you total panel current drain, won't hurt. I did look up the NKK switch you mentioned and that is definitely not the one used in my panel.

Tom

200 mohms seems like it would still send a lot of current through the switch, at least briefly. 70A?

I modeled it in Spice, included series inductance, and saw peak currents in the range of 10-20A for a few msec.

Tom

Well, today I did measure the inrush current: the peak current was 9A, very close to what I had simulated with Spice. This current is very brief and totally within the capability of the switch to handle, but a few ohms of series resistance will cut it down to a couple of amps if you are anal about it.

At 03:55 13 April 2020, 2G wrote:
snip

Well, today I did measure the inrush current: the peak current was 9A,
very=
close to what I had simulated with Spice. This current is very brief and
t=
otally within the capability of the switch to handle, but a few ohms of
ser=
ies resistance will cut it down to a couple of amps if you are anal about
i=
t.

Tom

A while ago, I asked if anyone else has experienced your problem.
Apparently not.
The LX9000 has (presumably) a capacitance of its own which sustains it over
changing batteries in most cases.
If there is a high load from other devices on the Bus then there may not be
sufficient capacitance in the LX9000 and it need some help.

If we isolate the LX9000 from the other devices on the Bus with a schottky
diode, then other devices on the Bus will not drain the LX9000’s
capacitance during switch-over as the diode will be back-biased. If it
still needs some help, a small capacitance can be added on the LX9000 side
of the diode.
This will also lead to a reduced inrush current as the connected
capacitance will be less.

I see this issue with my S80 in my Ventus Ct with two batteries.

The S80 powers my Flarm as designed.

When switching between my main battery and second battery I often, but not always have the Flarm power cycle and sometimes the S80 as well.

I too was thinking about sourcing a make before break switch but I am going to put in a capacitor.

On Monday, April 13, 2020 at 4:21:35 AM UTC-7, Mike N. wrote:
I see this issue with my S80 in my Ventus Ct with two batteries.

The S80 powers my Flarm as designed.

When switching between my main battery and second battery I often, but not always have the Flarm power cycle and sometimes the S80 as well.

I too was thinking about sourcing a make before break switch but I am going to put in a capacitor.

I have an S10 with an internal battery that is no problem, But I also use XCSoar on a OpenVario computer that I built from a kit. That has a big problem, as I switch batteries a couple of times a flight to keep the main battery as charged as possible for starts. So a year ago, I got with an electronics guy at our local supplier and he recommended a properly sized Capacitor and it has worked great. I don't remember the size or model, but it is hidden in the panel.

I researched true make before break switchs and they tended to be large and expensive. Not all switches that say the make before break really do what they advertise.

Bruce

Bruce

Interesting Mike -

I have a Flarm / S8 / Oudie set up. I never have any issues while switching batteries.

I wonder of the Oudie battery somehow supplies the "make" current while the switch is being thrown?

Lou

On Monday, April 13, 2020 at 3:45:04 AM UTC-7, Tim Newport-Peace wrote:
At 03:55 13 April 2020, 2G wrote:
snip

Well, today I did measure the inrush current: the peak current was 9A,
very=
close to what I had simulated with Spice. This current is very brief and
t=
otally within the capability of the switch to handle, but a few ohms of
ser=
ies resistance will cut it down to a couple of amps if you are anal about
i=
t.

Tom

A while ago, I asked if anyone else has experienced your problem.
Apparently not.
The LX9000 has (presumably) a capacitance of its own which sustains it over
changing batteries in most cases.
If there is a high load from other devices on the Bus then there may not be
sufficient capacitance in the LX9000 and it need some help.

If we isolate the LX9000 from the other devices on the Bus with a schottky
diode, then other devices on the Bus will not drain the LX9000’s
capacitance during switch-over as the diode will be back-biased. If it
still needs some help, a small capacitance can be added on the LX9000 side
of the diode.
This will also lead to a reduced inrush current as the connected
capacitance will be less.

Hell of a lot easier just to add a capacitor to the bus line, and it can be significantly smaller than the 39000μ that I used (10000μ would probably be fine).

I did because I am.

Big old power resistors. I think they were 1-2 ohms. However, after considering battery performance implications of deep discharge I normally fly with the switch set to both batteries.

Andy Blackburn
9B


.... a few ohms of series resistance will cut it down to a couple of amps if you are anal about it.

Tom

That could be a significant voltage drop across the resistors. I'd use
a couple power Schottky diodes instead.



On 4/13/20 1:11 PM, Andy Blackburn wrote:
I did because I am.

Big old power resistors. I think they were 1-2 ohms. However, after considering battery performance implications of deep discharge I normally fly with the switch set to both batteries.

Andy Blackburn
9B


... a few ohms of series resistance will cut it down to a couple of amps if you are anal about it.

Tom

On Monday, April 13, 2020 at 1:08:16 PM UTC-7, kinsell wrote:
That could be a significant voltage drop across the resistors. I'd use
a couple power Schottky diodes instead.



On 4/13/20 1:11 PM, Andy Blackburn wrote:
I did because I am.

Big old power resistors. I think they were 1-2 ohms. However, after considering battery performance implications of deep discharge I normally fly with the switch set to both batteries.

Andy Blackburn
9B


... a few ohms of series resistance will cut it down to a couple of amps if you are anal about it.

Tom


Anybody who took EE101 (and most everybody else) can tell you that:
E = I * R
If
I = 1A
R = 1Ω
then
E = 1V

I used Shottky diodes plus power resistors plus capacitors. I'm no EE but I took enough circuits courses to handle this problem. The Shottky diodes keep the batteries from cross-discharging each other, the capacitors keep the instruments powered when the switch is disconnected from battery 1 and before it is connected to battery 2 and the resistors keep the capacitors from drawing too much current when you power them up since they make the circuit (even with the diodes) look like a direct short initially.

Andy

On Monday, April 13, 2020 at 1:08:16 PM UTC-7, kinsell wrote:
That could be a significant voltage drop across the resistors. I'd use
a couple power Schottky diodes instead.