Current sinking in HP 6632A

[jkrbasu]

Hey guys!

I have an HP 6632A DC Power Supply. I'd like to use it to cycle Lithium Ion Batteries.

Setting the positive current limit is a piece of cake.

However, when I try to use the unit in current sinking mode, I always get this stinking offset of ~ -250 mA.

I know, that's what the manual says the offset is in current sinking mode. But does anyone know how to change or turn off the offset? I'd really like to be able to sink currents less than 250 mA!

Muchas gracias-

Roger

[pigrew]

As far as I know, there isn't a way to sink less than 250mA. You probably would need to rewire the circuit.
Also, when dealing with batteries, you should disable the crowbar circuit, by moving an internal jumper.

Addendum: also note that the accuracy of the negative limit is 125mA+0.8%.

It looks like this limit is due to a positive offset of 1.25V (Vos) added to the current limit. You may be able to reduce this voltage (maybe for the current limit circuit only, not the voltage circuit, and recalibrate). Halving the voltage should half the negative offset. You may also be able to play with adding an offset in the negative comparator circuit.

[jkrbasu]

Hey thanks for the heads up with regards to the accuracy. I think I'm going to put a current sense in line to accurately measure the current.

So based on what you're saying, I could also just put a digital potentiometer in line to reduce that offset, correct? Then I wouldn't have to monkey with the internals.

[pigrew]

Hey thanks for the heads up with regards to the accuracy. I think I'm going to put a current sense in line to accurately measure the current.

So based on what you're saying, I could also just put a digital potentiometer in line to reduce that offset, correct? Then I wouldn't have to monkey with the internals.

The current displayed is fairly accurate, so there is little need for the current sense resistor. The accuracy I mentioned is the accuracy of the negative current limit offset.

The potentiometer would need to be installed within the unit., but yes, it could work. Thinking about the circuit more, it may make more sense to add a physical potentiometer in the unit to adjust the negative offset by adding an additional input to the summing op amp, and tweak the offset to be closer to 50mA (or so )

Regardless, you'll need to open the unit to disable the crowbar circuit.

[jkrbasu]

Just opened up the unit and am looking around for the jumpers to take out the crowbar, but can't seem to find them! Here's a photo of the inside.

Do I really need to take the board out and potentially remove that heatsink to find the jumpers? Have you done this modification?

Thank you!

[pigrew]

The jumpers are wires soldered onto the board. See the user manual, section 4-15 for details. Both W3 and W4 should be removed. They are located next to the heat sink fan.

I've not removed them on my unit.

[jkrbasu]

By the way- since we're on the topic...Any recommendations for affordable DC power supplies that source/sink ~0-5A that do not have this current offset issue in sink mode?

[alm]

The only reason these power supplies are affordable is that they get sold on eBay cheap because they were bought in large numbers for ATE applications. Newer power supplies from the same series, like the 6632B and 66332A, have mostly eliminated this offset (lower limit 5 mA). Affordable two quadrant power supplies are not that common.

Other options that are sometimes available for cheap are Kepco BOP units. The BOP 36-6M and 50-4M might suit your requirements. Keep in mind that the units with the single op-amp symbol have a severely reduced current sinking limit compared to sourcing (the 36-5M can only sink 1A). Very similar (both specs and look) are the EMI/Lambda BOSS units. Like the HP power supplies, list prices are very high (and the BOSS is discontinued), so you will need to be lucky finding a good deal. I think I paid well under $100 for them in untested/broken condition.

The third option would be to split the work and have either a separate electronic load, or have a separate power supply for low power levels. For example, the Keithley 2302/3/4/6 communication/high speed power supplies can also sink with I believe a very good low current resolution down to a few mA. But power is quite limited (see datasheet). Can sometimes be found for around $100.

[Dr. Frank]

What about the 6632B?
It's redesigned from the A version, and quite cheap on the 2nd hand market.
And it would be quite familiar for you.

Its lowest sink current is -5mA, and its setting resolution is 2 or 3mA, readback accuracy being much more precise, of course, see pictures.
Additionally, it has a low current range of +/- 20mA, but is precise for low currents in the upper current range (+/- 20mA .. 6.66A) also.

I have chosen this B version, as I also want to do precise low current charge / discharge measurements on rechargeable batteries.

Frank

[jkrbasu]

Hey thanks very much for the recommendation! I actually have two 6632B units, so I think I will move over to those for testing. The 6632A unfortunately I think will have to go back to eBay.... I just don't have the time to figure out a way to limit that offset.

Interesting and encouraging to hear that others are using these power supplies for testing batteries! Any comments on how accurate your measurements are relative to industry standards like ARBIN, MACCOR, etc.? I know Jeff Dahn's group in Canada abandoned the industry standards and made their own testing infrastructure using Keithleys, which is partly what inspired me to do the same. And now there's a spinoff out of Dahn's group called NOVONIX that are selling their own testers.

[jkrbasu]

Hmm. still having the same problem! Setting current to 0.05A, getting almost 0.3A sink current! Am I doing something wrong...?

[alm]

It is not set to 6632A compatibility mode, is it? Then it will add the offset to stay compatible with software developed for the 6632A .

[jkrbasu]

Ah I see. Any idea how to change it from compatibility mode to...SCPI mode?

[alm]

Via GPIB/RS-232: SYST:LANG SCPI. For front panel, see attachment.

[jkrbasu]

Yes! Thank you a million times! It is working now!