Plug in for EV Prius

Try this

 

I also wanted to build a trickle charger, but was taken aback by one rumor I have heard: the Battery ECU *learns* the characteristics of your battery pack. And charging it, outside the control of the Battery ECU's charging, can screw up those characteristics, and it will not properly charge the battery while driving. Anyone have more info on that?
I just want to top off the SOC to 65-70% before the drive to work in the morning, to avoid the 4-5 miles of charging my car seems to do every day at first. When I park it at night, it usually shows 6 blue bars. But, in the morning, as I start to drive, 1 or 2 of those blue bars are missing, which I find bothersome. I know the NiMH are supposed to lose 1% charge per day, but this is more than that. I wish I had CANVIEW to read the actual battery voltage before and after.

 

You might have to do something fancy to bypass the HV ECU in order to get it to use the extra charge, otherwise the Prius might just dump the extra battery charge if it thinks the SOC is too high

 

<div class='quotetop'>QUOTE(keydiver @ Jun 3 2006, 02:38 AM) [snapback]264729[/snapback]</div>

Saying such is easy, but it is very difficult to maintain the battery level less than 80% SOC.
There is no way to know the direct SOC level of NiMH battery.

Ken@Japan

 

<div class='quotetop'>QUOTE(etyler88 @ Jun 1 2006, 06:52 PM) [snapback]264334[/snapback]</div>

I too am very interested in a plug-in charger for short trips and mileage boost. I think it is easy (conceptually) to do this. First to really get the advantage I need to upgrade my Prius (a 2002) to one that offers EV mode.

Here are my thoughts on the matter:

Use a power-factor-correct (PFC) boost regulator to provide about 100W for charging the battery. This
PFC eval board at digikey would need only minor modifications for this purpose.

Next add a circuit that goes between the battey ECU and the hybrid ECU. This circuit will power the battery ECU when the car is plugged in (and otherwise turned off), and "talk" to the battery ECU just the same as the hybrid controller does. It would basically just ask the battery ECU for it's view of battery SOC, and when it reports 80%, simply turn off (with a relay) the AC input to the boost converter. It wouldn't go on again until the AC power is removed and plugged in again. When there is no AC power, the circuit just re-routes the connections back to the car the way they were to begin with. Of course there would be fail-safes like shutoff if it doesn't get valid messages from the battery ECU, over Voltage on the boost output or high temperatures.

This approach has a number of advantages:
(1) low cost (a few hundred dollars)
(2) minimal modifications to the car
(3) low charge rate ensures no thermal issues with battery (note the battery fan won't have power while charging)
(4) factory battery is still managed by factory ECU algorithm, so the car always "knows" about the battery condition.
(5) automatic "lock-out" of the car when plugged in because the hybrid ECU would be unable to communicate with the battery ECU. You'd get a trouble-code if you tried to start the car before unplugging it.

In principle it is little different to the battery than having a long, not very steep hill to coast down at the end of the day with regenerative "drag" charging the battery. The disadvantage is that it takes about 6 hours to recharge from 40% to 80% SOC with a 100W charger.

Most people would be able to switch to EV mode for the last portion of their trip home which is typically local roads. This would drain the battery just in time to charge overnight for EV mode at the beginning of the next morning commute. For those of us with < 1 mile to work, we could probably be totally EV for those rainy days when walking isn't fun.

I'm guessing the hardest part is going to be finding compatible connectors to allow this to all be installed with no irreverssible change to the car.

 

<div class='quotetop'>QUOTE(haceaton @ Jun 5 2006, 09:28 AM) [snapback]265747[/snapback]</div>

I think the hardest parts are making such circuit and talking to the battery ECU.
Do you know what kind of language is used to talk to the ECU?
I'm sure it isn't Japanese.

Ken@Japan

 

<div class='quotetop'>QUOTE(ken1784 @ Jun 4 2006, 10:39 PM) [snapback]265810[/snapback]</div>

Actually I don't know it yet, but I'm not worried about figuring it out. 5 years ago I figured out how to "talk" to the IE bus in order to add a new stereo with improved features and even designed my own bi-directional driver interface for that bus. Some of the details can be found in the files section of the Yahoo group "prius technical stuff". Given that the messages on the CAN bus for SOC has already been figured out and the information available, I don't think it will be very hard to decipher the direct hybrid<->battery ECU bus traffic since some of it is echoed on the CAN bus.

I was completely serious when I said the hard part would be getting the connectors. I really don't want to cut any wires.

 

Being also interested in a possible 'poor man's plug-in hybrid', I thought it might help to find out a bit more about our stock Prius Nimh battery, and I don't mean the restricted range of operation the Prius insists on.
There seem to be a few misconceptions about what it can or should do.

So not about to risk my own Prius, I have a spare Nimh pack removed from a crashed 2005 Prius. (front crunch, back end where the battery sits unmarked) Before all the cries of caution arise, I will say I am an electrical engineer and strongly recommend anyone without high voltage skills to not-do-this-at-home-folks.

I built an isolated charger out of a modified cheap 1kw Xantrex inverter and charged this spare battery pack by the normal method of a) watching for the slight voltage drop at the end of charge and B) for the rise in temperature that also occurs at this point. I limited charge to 0.3C or ~1.5amps.

Then I discharged the pack into a water-cooled resistor bank to get steady conditions and plotted the result.
initially with just a 4kw load, I accumulated only 4AH before the pack went into end-of-discharge rapid drop.
I though this meant the battery might be bad, but then read up on professional photographers who try lots of different Nimh packs and use the latest smart chargers to get the best performance. They conclude that a pack that isn't used much (or is new) doesn't reach nearly its rated capacity. It has to go through several full charge/discharge cycles before it achieves that. So 3 such cycles later, I can now get 5.5AH. but still not the claimed 6.5AH.

I noticed that when I had fully charged it one evening and didn't do a discharge until a day later, I had somehow 'lost' about 1AH. I've now read that Nimh packs typically self discharge 15-20% in the very first 24 hours and then settle down to ~10% per month (more in a hot climate)

[attachmentid=3759]
You can see from this that from an initial 100% SOC, there is a fairly rapid drop then a much gentler decline for most of the SOC until somewhere about 10% it starts to drop off rapidly. So in the region the Prius confines operation to (80% maximum to 40% minimum) its pretty much all on the straight part of the graph.
This also makes determining the SOC easier as (if the internal resistance is known or can be calculated) the SOC to voltage is actually fairly consistent (at any one temperature) Go outside this range and it gets way more complicated.

So next I decided to measure the internal resistance throughout a discharge cycle. After another full charge (which only raised the battery temperature about 2 degrees C (3F) I then discharged it at a constant 10 amps, with a 20A pulse every 30 seconds, to see the difference that caused (good old R= Voltage difference/current)

[attachmentid=3760]
This shows an internal resistance that is fairly constant at ~0.27 ohms from 90 to 10% SOC but starts to rise rapidly near end of discharge, more than doubling at the point I cut things off at. This gives another clue into what happens when you run-out-of-gas-and-limp-into the-gas-station-on-battery.

As long as you haven't reached this ~90% discharged state, you should be able to restart the engine when re-fuelled. But if you have gone past this, you now have 2 things working against you: the pack voltage is now much reduced and the internal resistance has got much worse so the available current is now so reduced that it may now be impossible to get the ~50amps needed for an engine start. You can see from the graph that trying to take 50A when its down to 170volts means 170-(50*.55) =142volts to the inverter. It probably just refuses to even try.

But is your battery actually dead? Of course not! If it was a plain old Nimh battery bought at the local store, it could perform this near total discharge typically 300 times at least. So a single occurrence should have no effect whatsoever. However......being discharged it might as well be dead because the ability to recharge it is rare indeed. For some reason, Toyota decided for what it ships us not to spend the $5 it would have cost to add a protected charger socket. So to re-charge your 'useless' battery, a tech has to dismantle a good chunk of your car in order to gain access to the bare battery terminals. And then hook up a very expensive and rare charger because its just too much effort to instruct techs safely in high voltage handling and much easier legally to simply say its dead, buy another one.

Conclusion: the pack that is in your car probably has much less than the 6.5AH you thought and so even if you could start at 80% and drop right down to 40%, it would more likely be 40% total of 4AH than of 6.5AH and so more like 1.6AH available (which is good for ~2 miles of very careful driving under 30mph)
2nd conclusion: trickle charging your stock battery requires careful end management to avoid even slight overcharge. It requires monitoring voltage dips and temperature, done manually as you can't (yet) buy a suitable smart charger. But its not impossible. Question is, is ~ 2miles worth all the effort?


to be continued......

 

Thank you for the very nice analysis.

<div class='quotetop'>QUOTE(eflier @ Jun 11 2006, 11:17 AM) [snapback]269369[/snapback]</div>

Your method is detecting 100% SOC level, which means the battery life will be very short.
I believe we had better detect the 80% SOC level to stop charging, but it is more difficult to do so than 100% SOC.

Ken@Japan

 

<div class='quotetop'>QUOTE(eflier @ Jun 10 2006, 07:17 PM) [snapback]269369[/snapback]</div>

I was actually going to call you and see what was all going on with the "spare battery" but you just saved me a long distance call. Interesting!

 

<div class='quotetop'>QUOTE(eflier @ Jun 10 2006, 10:17 PM) [snapback]269369[/snapback]</div>

Obtaining a spare battery is cool, but hardly sounds like a "poor man's" approach.

One critical issue that nobody seems to talk about is the cell equalization in the battery. The battery ECU does connect to many taps along the battery, and I don't think we really know exactly what it does. Some have theorized that it only monitors the equalization so it can throw a trouble-code if the cells are too unequal. Personally I highly doubt that is all it does. I suspect it actually equalizes them from time to time, presumably with a loading scheme. It is not really believable that so many cells in series could all be so perfectly matched that no damage is done over the thousands of charge-discharge cycles with no active equalization.

This is another reason that I want to keep the battery ECU alive during plug-in charging. Not only does it monitor voltage, temperature and do coloumb-counting to determine SOC, I believe that it may equalize the string during charging too.

Since you had access to a "spare" hybrid battery, perhaps you could also obtain the battery ECU in order to see what's inside, and look for evidence of high-power loads(s) that could be used for equalization.

I also am not sure I accept that the 40-80% range is on a lesser A-h rating. I belive that Wayne actually measured the A-h of charge/discharge and concluded that the 40% delta actually corresponded to 2.6 A-h.

Finally, it seems that the battery ECU literally talks to the HV ECU on the CAN bus, thus the battery ECU "language" is already figured out and published.

 

<div class='quotetop'>QUOTE(haceaton @ Jun 12 2006, 01:20 AM) [snapback]269543[/snapback]</div>

You're right.
2004 Prius has 28 of 7.2V modules and there are wires among two modules.
The THHS scanner can check the SOC every two modules and the Toyota service spec is to replace the whole HV battery unit if the delta-SOC was more than 20%.

How is your project going?

Ken@Japan

 

<div class='quotetop'>QUOTE(ken1784 @ Jun 11 2006, 03:05 PM) [snapback]269591[/snapback]</div>

I have completed step 1 I have a 2006 Prius now.

 

<div class='quotetop'>QUOTE(haceaton @ Jun 11 2006, 09:20 AM) [snapback]269543[/snapback]</div>

The battery ECU contains no high power components, just optically isolated relays to all the battery sense taps and the usual micro-chip stuff. So where does that leave equalisation?

What I was trying to explain was how the stock battery performs outside the parameters Toyota sets, because I couldn't find documentation on that anywhere. (If anybody knows of published data anywhere, please post a link.) Also possibly to dispel the myth that a single discharge to zero will destroy the battery somehow: its only effectively dead because you can't conveniently charge it.

BTW, my idea of a 'poor' battery is still lead-acid.

 

The Batt ECU monitors every other cell for charge/discharge, impedance of the cells, overall temperature of the pack to activate the cooling blower for safety, voltage differential of the cells and CAN architecture to the rest of the prius ECU's for SOC control but not any active equalization to the cells. Also, the Hybrid ECU plays a major rol on the Synergy drive control of the HVbattery. Thi's a very intelligent combination of electrical's for the efficiency and durabillity ( under Toyota spec's) of the Hybrid Synergy Drive design that we are trying to crack.

 

Norm have you figured out why the pack is split 56V and 144v instead of 100-100v?

 

<div class='quotetop'>QUOTE(Frank Hudon @ Jun 11 2006, 04:07 PM) [snapback]269685[/snapback]</div>

The battery is made up of 28 sets of 7.2v packs, arranged with their + and - ends alternating. Each pack is connected to the next one with a short flat jumper plate. The only difference is that between pack 19 and 20, counting from the -ve end, instead of the usual flat jumper there are 2 orange cables going off ~12" to the orange service plug, which also has a built-in 120A fuse. Pulling out that service plug is just the same as removing any of the flat jumpers, except much easier and safer.

With it removed, the battery is effectively safe as both -ve and +ve ends are left floating. I don't think it matters too much which pair of packs this 'removable' jumper is installed between. So while the battery is currently split into 2 different sections by this jumper, nominally 137 volts and 65 volts, thats only true with the plug removed, so I wouldn't read too much into it. Maybe someone made up 100,000 service sockets with 12" of cable on them so thats where the connection ended up being....

 

<div class='quotetop'>QUOTE(ken1784 @ Jun 10 2006, 10:42 PM) [snapback]269422[/snapback]</div>

Besides the fact that the Prius will not allow 100% SOC, what other evidence do we have that the battery life will be very short? As we've discussed before, the batteries in my Rav are the same chemistry, and they have been charged to 100% almost every single day for the past four years. No loss of capacity has yet been detected.