1000W Lipo Charger

[rebelpilot]

Project
I need to build a charger/balancer for 12 packs of Zippy 6S 5Ah 25C Lipo.

Background
I was down in the basement today assembling a parts kit to build a charge controller based on the meanwell charge controller design, and a balancer based on the 12s CellLog CMS-Balancer design. But I got distracted.
I came across four pieces of GPIB equipment I salvaged from old production line testers. I tried selling this stuff online years ago, I had no immediate needs for the equipment and the manuals were very expensive to purchase at that time.
I decided to check if the information was now available. I was able to download user, service, and programming manuals for each piece of equipment.
I have two National Instruments GPIB-PCIIA cards, and an old computer with an ISA bus, but I have been unable to communicate with the cards.

Plan
The current plan is to buy a Prologix GPIB-USB Controller and use EZGPIB to run the system from an old laptop.

http://prologix.biz/

http://www.ulrich-bangert.de/html/downloads.html

The system would consist of two the 20V power supplies connected in series to charge all 12 lipo packs in parallel. The individual cells would be monitored with the 3457A multimeter and resistors would be switched in via the 3488A controller to balance the cells. I have to do some more reading, but I think the 3488A may be able to run the system once it is debugged.

Charge voltage 4.15V/cell
Charge current 50A
Balancing shunt resistor value TBD
Charge termination current TBD

I need help.
Questions:
Is there a better GPIB-PC interface option?
Any alternative software suggestions?
Is this a viable plan?
What balancing shunt resistor value would work, how is it calculated?
What charge termination current would be suitable for 6 packs or 12 packs in parallel?
Any suggestions are welcome.

Equipment Descriptions:

3488A Switch/control unit
View attachment 3




The Agilent 3488A switch/control unit brings versatile, GPIB programmable switching to tests requiring multi-channel measurements. The Agilent 3488A provides signal switching with the integrity and isolation needed for high-performance production test systems. It also offers a flexible, low-cost interconnection for automating experiments on the bench and for developing tests in the lab. The Agilent 3488A is designed to hold up to five of the following switch and control modules:
• DC to 26.5 GHz signal switching
• Up to 50 channels
• Matrix, multiplexer, and general purpose relays
• 40 configuration storage registers
• Digital I/O control and actuation
• 11 switch and control modules
General Purpose Relay (Option 011/HP 44471A) (I have three)
This module consists of 10 SPST independent relays for general signal switching and control of external devices.
Quality connections make this module ideal for switching signals when multiplexing is not required, or for supplying switchable power to the device under test.
Input Characteristics
Maximum Voltage (terminal-terminal or terminal-chassis): 250 V dc, 250 Vac rms, 350 Vac peak
Maximum Current: 2A dc, 2 A ac rms
Maximum Power: 60 W dc, 500 VA ac
Thermal Offset: < 3,μV
DC Isolation (40ºC, 60% RH)
Channel-channel, open channel: >10^11 ohm
AC Isolation/Performance
(50 ohm termination) 100 kHz 1 MHz 10 MHz
Insertion Loss (dB) <0.30 <0.35 <0.90
Crosstalk (dB) <-73 <-53 <-33

Agilent HP 3457A. Multimeter
View attachment 2




Seven functions with 3 1/2 to 6 1/2-digits of resolution extendable to 7 1/2-digits at reading rates from 1 reading per every 2 seconds to 1350 readings per second, and DC volts accuracy as good as 5 ppm.
Ten channels with the 44492A plug-in multiplexer assembly.
Measures DC Volts & Current, Resistance, True RMS AC Volts & Current, Frequency & Period.
Also performs the following math functions: Null, Scale, Offset, RMS Filter, Single Pole Filter, Thermistor Linearization, dB, dBm, % Error, Pass/Fail Limit Testing, and Statistics.

HP6031A Autoranging 1000W GPIB Power Supply
View attachment 1




Output Ratings
Voltage: 0 to 20 V
Current: 0 to 120 A
Autoranging:
20V, 50A
14V, 76A
7V, 120A

Programming resolution
Voltage 5mV
Current 30mA
Programming Accuracy at 25°C ±5°C
Voltage: 0.035%, +15 mV
Current: 0.25%, +250 mA
Ripple & Noise 20 Hz to 20 MHz
Voltage rms: 8 mV
Voltage Peak-to-Peak: 50 mV
Current rms: 120 mA
Readback Accuracy at 25°C ±5°C
Voltage: 0.08%, +7 mV
Current: 0.4%, +100 mA
Load regulation
Voltage 3mV
Current 15mA
AC Input Current
24A @ 120Vac
14A @ 240Vac
Weight
38lb (16.75” x 5.25” x 20.33”)

Xantrex XKW 1 kW programmable DC power supply





High frequency conventional PWM switching supply
Constant voltage with automatic crossover and mode indication
Parallel or series connection
External shutdown, external indicator signals
Remote/local modes
Remote sense, 1 V line loss compensation
Analog programming
LabVIEW® driver
Over voltage protection
Current limit
Over temperature protection
GPIB interface card
Isolated interface card (ISOL)
Note: The XKW Series is not available for purchase in the US.(I think it’s sold as Sorenson in the US)
Output Ratings
Voltage: 0 to 20 V
Current: 0 to 50 A
Programming resolution
Voltage 5mV
Current 30mA
Ripple & Noise 20 Hz to 20 MHz
Voltage rms: 5 mV
Voltage Peak-to-Peak: 60 mV
Load regulation
Voltage 3mV
Current 5mA
Operational AC input voltage
200-250 VAC at 26 A rms 1-phase, or 100-130 VAC at 20 A rms 1-phase, 47-63 Hz (That has to be backwards doesn’t it?)
Weight
18lb (1.7” x 19” x 17.5”)

 

[texaspyro]

I use the same HP equipment (meter/scanner) for my multiple cell battery analyzer. The HP 3457A is my favorite bench meter (except for the 3458A's that cost over $4000 used).

The Prologix is a pretty good GPIB interface. I built my own software compatible controller using a Mega-Donkey graphic LCD touchscreen micro-controller. It speaks RS-232 to the computer and has a HPGL/PCL renderer built in so that it can display instrument dumps (or various status info) on the graphic LCD.

Ulrich's ez-gpib is a great program (except that its scripting language is based upon that suck-nuts bastard demon language from the cesspools of hell a.k.a. Pascal). You should be able to code up the complete charge system using it.

Beware of the input power requirements of running those power supplies at full power. You can probably only drive one of them off a single wall circuit.

At those charge currents you need to make a 4-wire (remote voltage sensing) connection to the packs to compensate for the voltage drop in the charge cables.

Remember to add some pack temperature sensors with maybe an independent shutdown to the power supplies if they over-temp.

 

[rebelpilot]

Thanks for the answers texaspyro, it's good to have some expert advice. I still haven't given up on the NI cards, but my old pc just gave up on me today. Back down to the dungeon in search of another motherboard. One more day of this and I'll bite the bullet and get a new card or the prologix. After doing some research the labview environment looks nice, and I don't think the prologix would be a good choice if I go that route.

 

[rebelpilot]

I spent a long time trying to get a GPIB card up and running, only to have my old PC fail. The gpib cards I have are for an ISA bus computer. ISA motherboards are available still for industrial use, for a premium, but I require a slow computer to operate my old PCB router. I spent three days swapping cards, formatting disks, and cursing. I finally was able to get a motherboard working, so I built a balancer board while I had the opportunity.

View attachment 4


View attachment 3


View attachment 2


View attachment 1





I will build a high voltage alarm board with seven LCD meters to monitor the cells while charging, then I will risk the computer and try again to get the GPIB control up and running. Failing that I can simply use the power supplies and HV alarm to charge. The HP supply has a foldback function that will disable the output when the control switches from constant current to constant voltage mode.

 

[rebelpilot]

I cut and assembled the remaining PCB's today. Here is the meter/hv alarm board being routed.
View attachment 5


Screen shot of ancient DOS router software.

View attachment 4


Board ready for hand drilling. The software has ceased to generate valid drill files. I will be rebuilding the router to use modern software as soon as my next bike is finished.

View attachment 3


Now I need to make the wiring harness, but maybe I should design the cabinet first. (yeah I should have designed everything before I started to build, but I couldn't stop myself)

View attachment 2


Hobby King sent the last shipment of lipo with different ends. PITA!

View attachment 1


Both supplies exposed.

 

[rebelpilot]

I have completed the initial setup and calibration of the power supplies, meters, cell high voltage alarms and balancer circuit.

View attachment 2

View attachment 1





The balancer circuit is based on Gary Goodrum's FAN431 circuit.
Balancer set at 4.15 volts per cell, load is 2.5 ohms connected through polyswitch via D45H8 power transistor.
Cell high voltage alarms set at 4.20 volts per cell.

I measured the intitial voltages of 12 packs of Zippy 25C 6S Lipo.

23.1 3.86 3.86 3.86 3.86 3.86 3.85
23.1 3.86 3.86 3.86 3.85 3.86 3.85
23.1 3.86 3.86 3.86 3.86 3.86 3.85
23.1 3.86 3.86 3.85 3.85 3.86 3.85
23.1 3.86 3.85 3.86 3.86 3.86 3.85
23.1 3.85 3.86 3.86 3.86 3.86 3.84
23.1 3.85 3.85 3.86 3.85 3.85 3.80
23.1 3.86 3.86 3.86 3.86 3.86 3.85
23.1 3.85 3.86 3.85 3.86 3.85 3.82
23.1 3.86 3.86 3.85 3.86 3.86 3.85
23.1 3.85 3.85 3.85 3.86 3.86 3.84
23.1 3.86 3.86 3.86 3.85 3.86 3.85

Can I connect the balance wires and power wires in parallel now and start charging?
I set the foldback voltage of the charging power supply at 24.90 volts. As soon as the output changes from CC to CV the output will shutoff. Is that a sufficient charge?
What current should I charge at?
Can I have the balance board connected when charging all the packs in parallel, or do I have to balance the packs one at a time?

 

[rebelpilot]

.