In thinking about Mike Sassnet's suggestion to dump the excess regen
current into a resistive heating element, I came up with what I believe
is a solution to the "Charge Current Limit" problem. I have
successfully tested it yesterday and today. Although my testbed is a
hack and needs some further refinement, it works well and with it the
vehicle has exceptional driveability. It also has some further major
advantages.

What I've done is combine the idea of using a higher voltage for
spoofing the Battery ECU, controlling the spoofing via perceived SOC
feedback, and intermittantly paralleling the new PHEV battery with the
existing OEM battery. The vehicle is run on the OEM battery and the
Battery ECU. No battery tap emulation is required (meaning that the HV
Board is no longer necessary). When perceived SOC (read from the CAN
bus) is below 69%, the PHEV battery pack is placed in parallel with the
OEM pack (currently by a contactor, but in the near future by an IGBT).
Once the perceived SOC reaches 71%, the PHEV battery is removed from the
circuit. I am planning on, but have not yet added, a diode to send
regenerative braking current to the PHEV battery as well as the OEM
battery once the voltage exceeds the current voltage of the PHEV pack,
even when the batteries are otherwise disconnected from each other.

This paralleling charges the OEM battery by holding its voltage higher
than normal, which also raises the ECU's perceived SOC. When the
perceived SOC reaches 71% and the batteries are once again disconnected,
the ECU's perceived SOC will, as it is designed to do, settle on the
true SOC of the OEM battery. Note that for the OEM battery, this
process mimics regenerative braking, nor does it happen when the OEM
battery's SOC is too high.

As the PHEV battery is discharged through this process, less and less
voltage is added when the batteries are paralleled. Eventually, the
PHEV battery is unable to keep the perceived SOC above 69% and remains
in parallel from then on. Its real SOC will equalize at a point where
its open circuit voltage matches that of the OEM battery at the
approximately 60% SOC that the ECU tries to maintain -- and once again,
the perceived SOC will match the OEM battery's real SOC. I therefore
believe that, though lots of voltage shifts and SOC spoofing is going
on, this process will treat the OEM battery well by keeping it within
its normal operating range -- a major requirement of any hybrid battery
scheme.

It will also treat the PHEV battery well by requiring much smaller
current peaks from it and by settling on a minimum SOC predetermined by
the difference in voltages of the two battery packs. And a really low
PHEV battery internal resistance (IR) is not required for proper
operation; increased IR will merely result in reduced EV-only range. I
believe that with this scheme, the lifetime of the PHEV battery should
be at least doubled from that of CalCars' first PbA pack. It's also
likely that battery heating for the PbA pack will no longer be strictly
necessary, though it could still be helpful in preserving PHEV range at
low ambient temperatures.

Yet another potential advantage of this hybrid-battery configuration is
that the overall battery system's IR will at worst match the Prius' OEM
IR, and when the batteries are paralleled -- as will be the case for all
driving beyond the PHEV range -- the combined IR will be around half of
that. This should improve city hybrid-mode mileage, probably fully
making up for the added weight of the PHEV battery pack. It also has so
far eliminated all DTCs due to excessive voltage swings as well as exits
from EV-only mode due to undervoltage (<180V).

The voltage of the PHEV pack necessary for proper spoofing probably
depends on the PHEV battery IR as well as its chemistry. My test
results are as follows:
20 EVP20-12 batteries, 240V nominal: 5 miles EV range until kicked
out of EV-only by perceived SOC hitting 45%
21 EVP20-12 batteries, 252V nominal: 10 miles and 9.7Ah until 45%,
7-8 miles and 7-8Ah until 60% SOC (nominal end-point)
22 EVP20-12 batteries, 264V nominal: may be perfect, but ran out of
time to try until later this week

Note that freeway-speed testing is yet to come, too. There may still be
some further quirks to work out.

I will next continue this testing process and look at the changes
required to both Ryan's Maker Faire and the Electro Energy conversions
to use this configuration in its current unrefined form. There are
specific control, battery voltage, and charger voltage issues for each
conversion. Then I will refine it and come up with a more permanent
solution, using, for example, solid-state instead of wear-prone
contactor switching.

In the meantime I have worked on neither Maker Faire nor Electro Energy
schematics and parts lists, as I knew that they would of necessity be
changing, possibly significantly, to accomodate a necessary solution
(now found) to the CCL conundrum.

/ron

+++++++++++++++++++++++++++++++++++++++++
Ron Gremban, rgremban@...
California Cars Initiative, a nonprofit organization:
http://www.CalCars.org
Moderator & Technical Lead
http://www.priusplus.org
PRIUS+ PHEV Conversion Group: http://groups.yahoo.com/group/priusplus
Newsletter: http://groups.yahoo.com/group/calcars-news
Do-it-yourself PHEVs: http://www.eaa-phev.org
+++++++++++++++++++++++++++++++++++++++++