Can lab power supplies sink current?

[IntegraR0064]

I'm wondering if you can help me understand how lab power supplies treat sinking current.

Some background:  I do a lot of lithium ion battery testing.  When I'm done with a battery, I want to set it to 3.6-3.7V for storage.  As I'm sure you're aware fully charged li-ion batteries are 4.2V, and fully discharged are 2.5-3V (they don't like being at these extremes for a long time, which is why I want to set them to 3.6ish).  Sometimes I charge the battery before deciding I'm actually done with it, whereas sometimes I don't.

So I was trying to think of a simple way to do this and ideally have one device that goes both ways.  I was thinking I could just get a few battery holders and wire each positive contact to the positive of a power supply with a power resistor in between in series (we'll say around 1 ohm).  Wire the negative of the power supply to the negative of the battery.  Set the power supply to 3.65V.

Then:
Scenario A)  If the battery is discharged, it will be 2.6-3V, the power supply will be 3.65V, the power supply will charge the battery.  If I have a 1 ohm resistor inline that will mean it'll be about 1A at first then slowly dwindle down as it gets closer to 3.65V and it will asymptotically approach 3.65V.  This part seems clear cut to me and everything will work great as far as I know up to the max current of the supply (5A in this case).  I'm sure there are more elegant ways to do it but this should work.
Scenario B)  If the battery is charged, it will be 4.2V and the power supply will be 3.65V, so the power supply will see negative current provided by the battery.  Is this a problem?  Is there a limit to how much current I can do, from the power supply's perspective?  Any input on how much that limit probably is?

The power supply I'm using is this one:  http://www.sra-solder.com/korad-kd3005d-precision-variable-adjustable-30v-5a-dc-linear-power-supply-digital-regulated-lab-grade

Any thoughts would be appreciated.

[IntegraR0064]

Also, if this sounds like a bad idea, I'd be curious to hear input on another way to accomplish what I'm trying to do.

[0xdeadbeef]

There are some supplies that can also act as current sinks (like Hameg/RS HM8143 or Gossen Metrawatt SSP/SLP) but typical power supply can't.
For a supply that can also sink current, you need another output stage design (push/pull).

[IanB]

Some background:  I do a lot of lithium ion battery testing.  When I'm done with a battery, I want to set it to 3.6-3.7V for storage.  As I'm sure you're aware fully charged li-ion batteries are 4.2V, and fully discharged are 2.5-3V (they don't like being at these extremes for a long time, which is why I want to set them to 3.6ish).  Sometimes I charge the battery before deciding I'm actually done with it, whereas sometimes I don't.

Just so you know, a fully discharged lithium ion cell should be not much below 3.6 V open circuit when fully discharged (maybe down to about 3.3 V if deeply drained). They really should not be below 3 V.

[rx8pilot]

Why not a DC load connected to a PC to control and monitor the discharge?

[dom0]

Some supplies do this (2 quadrant and 4 quadrant supplies, some are "operational power supplies" - basically a power op-amp).

Some (decent) supply designs have bias current sinks with wildly varying currents on the output. Something between a few mA and up to ~200 mA. You can easily measure this with another lab supply connected in parallel in CC mode, ramp up the current until the regulation of the DUT fails. That's the output stage bias current, and the maximum current it can sink.

Oh, and some other supplies have something sometimes called dynamic down-programming, essentially a class B output stage. Those can sink current until the shunt transistor blows (if unprotected).

[IntegraR0064]

Thanks everyone!  Really appreciate the input.  That's what I figured.  My circuit will work for the charging side but for the discharging I think I need a separate circuit that doesn't sink into the power supply.  I'll have to put some thought into how to do that, probably using some kind of comparator circuit instead of sinking direct into the PS.

I have ways to do this with computer controlled loads that I have, but that requires too much manual input.  Really hoping for something simple where I can just put the battery in and that's it.  My electronics is sooo rusty but I think it's time to get out the wire wheel and take the rust off .

Actually I wonder if I could just put the battery in series with a resistor and 6 diodes?  That might be the ultra simple way to do it.

Some background:  I do a lot of lithium ion battery testing.  When I'm done with a battery, I want to set it to 3.6-3.7V for storage.  As I'm sure you're aware fully charged li-ion batteries are 4.2V, and fully discharged are 2.5-3V (they don't like being at these extremes for a long time, which is why I want to set them to 3.6ish).  Sometimes I charge the battery before deciding I'm actually done with it, whereas sometimes I don't.

Just so you know, a fully discharged lithium ion cell should be not much below 3.6 V open circuit when fully discharged (maybe down to about 3.3 V if deeply drained). They really should not be below 3 V.

Thanks but that's actually not true.  If you look at the spec sheets for li-ion batteries generally they're discharged to 2.5 or sometimes even 2.0V depending on the particular battery and its rated minimum voltage.  Keeping it higher does lengthen the lifespan but for testing like I'm doing against the spec sheets they need to be discharged that low.

[LaurentR]

Buy an RC (Remote Control) Lithium charger. The good ones have a storage mode to automatically charge/discharge the battery to put it at the appropriate voltage. I have a dual-channel one that can drain 2A/20W. Some can do more.

Laurent

[joesixpack]

The Keithley 2281s power supply / battery simulator will do what you want.  You can even load profiles for various batteries. 

[IanB]

Thanks but that's actually not true.  If you look at the spec sheets for li-ion batteries generally they're discharged to 2.5 or sometimes even 2.0V depending on the particular battery and its rated minimum voltage.  Keeping it higher does lengthen the lifespan but for testing like I'm doing against the spec sheets they need to be discharged that low.

I know that, but those discharge curves show the voltage under load, during the discharge. That is not the same thing as the open circuit voltage when not loaded.

[IntegraR0064]

Buy an RC (Remote Control) Lithium charger. The good ones have a storage mode to automatically charge/discharge the battery to put it at the appropriate voltage. I have a dual-channel one that can drain 2A/20W. Some can do more.

Laurent

Good suggestion, I have considered that.  I do have one of those (only single channel though).   Only reason it's not ideal is a)  I'm hoping to have like 4-8 bays and b)  ideally I'd just plop the battery in there without having to press buttons so I think I'm going to keep thinking about a good way to make something up that's totally passive.  I know on mine it's many button presses to get to a mode like that.

The Keithley 2281s power supply / battery simulator will do what you want.  You can even load profiles for various batteries. 

.

That would be pretty awesome but it's a little rich for a small project like this .

Thanks but that's actually not true.  If you look at the spec sheets for li-ion batteries generally they're discharged to 2.5 or sometimes even 2.0V depending on the particular battery and its rated minimum voltage.  Keeping it higher does lengthen the lifespan but for testing like I'm doing against the spec sheets they need to be discharged that low.

I know that, but those discharge curves show the voltage under load, during the discharge. That is not the same thing as the open circuit voltage when not loaded.

This is true and most of the time they do end up a bit over 3V but if for example you're doing a low current test then the under load voltage isn't that far off from loaded voltage.  My company generally specializes in really high drain low internal resistance batteries so if you test one of those at low current the voltage sag is negligible or close to it.

[KL27x]

I know that, but those discharge curves show the voltage under load, during the discharge. That is not the same thing as the open circuit voltage when not loaded.

I get you.

http://i18.photobucket.com/albums/b103/klee27x/DSCF60331_zpsdzbxapzk.jpg

I used a microcontroller. When the comparator flops, it turns the load on for a minimum set interval (~20 seconds), ignoring the comparator in the interim. So when it approaches target voltage and the comparator starts to flip, basically it will only listen to the comparator every 20 seconds. And the length of the pause starts out as near nothing, but gets longer and longer as the float voltage takes longer to rise past the set point. When the comparator stops flopping for at least 10 seconds, it turns on an LED and stops. You could also use a 555 timer if you don't do micros. The one-shot pulse would be used to make the load stay on for a minimum time period.

This was the simplest algorithm I could come up with where it stops according to Vfloat (without having to account for ESR), and which wouldn't leave the circuit in oscillation that takes hours/days to reach target Vfloat.

***all that said, I haven't used this circuit in about 10 years. Not worth the effort for me.

[IntegraR0064]

I know that, but those discharge curves show the voltage under load, during the discharge. That is not the same thing as the open circuit voltage when not loaded.

I get you.

http://i18.photobucket.com/albums/b103/klee27x/DSCF60331_zpsdzbxapzk.jpg

I used a microcontroller. When the comparator flops, it turns the load on for a minimum set interval (~20 seconds), ignoring the comparator in the interim. So when it approaches target voltage and the comparator starts to flip, basically it will only listen to the comparator every 20 seconds. And the length of the pause starts out as near nothing, but gets longer and longer as the float voltage takes longer to rise past the set point. When the comparator stops flopping for at least 10 seconds, it turns on an LED and stops. You could also use a 555 timer if you don't do micros. The one-shot pulse would be used to make the load stay on for a minimum time period.

With just a comparator, you will have to either know exact ESR for your battery to know what voltage it needs to stop under the load. You would need to adjust this for different batteries. This was the simplest algorithm I could come up with that wouldn't leave the circuit in oscillation that would take hours to reach target voltage.

***all that said, I haven't used this circuit in about 10 years. Not worth the effort for me.

This is great, now we're moving on the right track.

Actually I far prefer micros to anything else and have been known to make circuits with microcontrollers when I could have just put a couple components together .  So I'm trying to resist the urge here but if it's needed it's needed.

This is one of those things where as long as it ends up somewhere between 3.5-3.8, ideally 3.6-3.7, I'll be happy.  So I'm definitely trying to keep things simple.  I'll try to digest exactly how I would implement what you're saying tomorrow.

[KL27x]

I never discharged any other type of battery other than 18650 cells. So when I say it was accurate to within a hundredth of a volt, take that into consideration. A 20 second draw over a 9.3 ohm power resistor is not as fine a resolution to a smaller battery. To an 18650 cell, it is pretty small. The circuit went through probably dozens of rounds of oscillation, IIRC. Maybe closer to a hundred than not.


 

[bson]

Some, like my HP 66312A do.  This particular model is called a "Dynamic Measurement DC Supply".  But it has another neat trick up its sleeve: it can float the outputs.

[cncjerry]

I have a keithley 228 four quadrant that I picked up pretty cheaply.  It is programmable for sink or source,etc and I use it on lipos all the time.  It is only 100w but it is about double your typical charger for a 4.2 cell that does only 5amps as this unit will do 10A at 10V.  You can do charge/discharge cycles limited pulse charges,etc.  I think I paid $100 for it and I don't think the seller knew what he had.

[Psi]

You can build your own source /sink psu out of a high current opamp a pot and a few resistors
You can get them with 8Amp outpus, not cheap though.$20 etc

[AndyC_772]

HP 6632B can either source or sink current. Expensive new, but fairly cheap and readily available on Ebay.

[timofonic]

You can build your own source /sink psu out of a high current opamp a pot and a few resistors
You can get them with 8Amp outpus, not cheap though.$20 etc

Can you provide some schematic? Please

[dom0]

https://www.google.de/search?q=op+amp+current+source