New codes for old 2003 still won't start.

In the past I would have said this is the OP's best bet, as the going price for 2G salvage battery assemblies had been $400-$500 each plus shipping.

However, as of this writing, I find no salvage battery assemblies on eBay Motors. Who is scooping them up?

 

LOL! Guess we'll have to 'dig up the landfills with all those old Prius batteries.'

Bob Wilson

 

One interesting technical question is how do we refurbish potentially good NiMH modules that simply self-discharged too low?

This isn't the type of experiment one wants to use their only traction battery. I have six modules to use for testing:

	Column 1	Column 2
0	SN	V start
1	21ZBBB01447D	0.216
2	21ZBBB01448D	0.094
3	03XBH904012C	0.012
4	03XBH904013C	0.010
5	03XBH904020C	0.012
6	03XBH904027C	0.010

Unused for ~2 yrs., from bad pack

I'll put them on my MRC 989 and we'll see what a straight charge accomplishes. Protocol:

• two, single modules - do individuals just to see how well it works
• four modules in parallel - to shorten refurbishment charge time, assuming the single modules work

Suggestions?

In the meanwhile, you might pull the traction battery out and put it on the bench. Trying to work on the battery in the car is a pain. BTW, it is a heavy unit. I've moved one by myself but I also have a hand truck. This is something you might consider asking a friend to help.

FYI, be sure and have 4-5 cups or containers to hold the fasteners. There will be a bunch.

Bob Wilson

 

Ahh, thinking about getting a minimum bulk charge to get the car started and letting the ICE finish it off? Good idea.

Let's get the basics out of the way:

• 1.2 V - single, NiMH cell
• 6 cells, 7.2 V - a module
• 2 modules, 14.4 V - nominal charge

Looking at:

14.4 - 12.5 = 1.9 V > 1.2 V. a potentially dead cell
​

The problem is if we charge a module with a dead cell, the other five cells will 'charge.' Then during the discharge, the dead cell will see a reverse charging current. This is reported in the literature as 'death to the cell.'

I don't have a good answer. The differential voltage between the module pairs is not supposed to exceed 0.3 V ... they really need to be balanced. Nothing jumps out in the manual.

Bob Wilson

 

Who is "us"? Would that be you?

I'm sure there are some folks on this forum who are quite good at it. But with all due respect, you tried to help diagnose this vehicle twice, and came up empty both times. Indeed, most of your guesses were well off the mark, and leaning towards components which had no fault. To be frank, I didn't think it was that hard to diagnose: inverter to HV ECU circuit, bad scan data, bad battery, done.

There's no reason why you should feel bad for being unable to diagnose this particular vehicle. None whatsoever. People are here to learn. But I find it less than sincere to patronizingly offer diagnostic "help" to someone who was able to do so what you could not do, and do so with little fuss.

Your words, not mine.

 

The first module required an initial charge before it could autocycle. Initial results suggest the latent capacity is ~1% of the expected capacity. The charge voltage, 9.1-9.4 V is way higher than expected. The module is showing pressure stress.

I'll let it sit for a bit and look at the other module.

Bob Wilson

 

Interesting, 21ZBBB01447D completed the first 8 cycles and I found:

• 30 mAhr - charge cycles, ending at ~9.4 V.
• 17-19 mAhr - discharge cycles, ending at ~6.4 V.
• GOOD NEWS! No evidence of a shorted cell! - this means we should be able recover the original 6.5 Ahr capacity!!!

So I reprogrammed the MRC 989 for:

• 20 mAhr - charge cycles, still reaching to ~9.4 V.

It is still running but at least I'm not running excess charge over what the module handles. I don't like the extra voltage but we'll let 'er run tonight and check again in the morning.

What I'll do in the morning is take that module out and add water. I'll leave the fill hole open and see how much capacity we can recover. Then I'll use a Dremil tool to mechanically weld the plugs.

MORNING NOTE: The module is still under pressure from the generated gas. I'll have to let it sit longer to allow the gas to recombine. The Charge/Discharge ratios were: 20/15, 20/15, 20/14, and 20/14. AM voltage, 2.20 V. Knowing the initial recovery values, I can program the MRC 989 to avoid excess gas generation.
​

One of the things I like about the MRC 989 is the ability to run 'autocycles' and later read out the last eight results:

1. discharge to a finite voltage - programmable discharge low voltage compared to the expected SOC level.
2. charge to a finite aHr level - the voltage can range up but the aHr limit means we don't run 'open loop.' Best of all, if the dV threshold shows up, 5 mv. for NiMH chemistry, it stops.
3. temperature probe - ensures the battery doesn't go through thermal runaway.

This ability to do a module-by-module analysis takes time BUT it ensures the final parts are good with know specifications. A bulk charger won't do a module-by-module charge and analysis.

Derk,

Are you under a time constraint?

The reason I ask is in engineering we usually have to pick two from the following three options: Good, Fast, Cheap. My experience has been that to refurbish an individual module takes about 2-3 days. For 38 modules, we're looking at just under 120 days, worst case. If you need to get the car rolling again faster, per-module refurbishment may not be feasible.

However, I plan to test gang restore in this series. Initially four modules in parallel. This has the potential of reducing the restore time to closer to 30 days. But again, it depends upon your service restore time.

Bob Wilson

 

21ZBBB01448D log

9:00 AM - Starting with a flat charge, limited to 8 V, 0.5A. What I'm seeing is the current into the NiMH is a gentle oscillation:

7.85 - 8.20 V, -0.01 - 0.07 A. - the scant literature mentions using pulses to activate NiMH cells. This unintended oscillation somewhat resembles a series of pulses. One nice thing, if I stop the charge, the module continues to hold 8.00 V.
​

11:00 AM - I found that voltage limiting the module charger to 8 V., a short rest of 30-60 seconds, and another cycle, seems to be bringing the module capacity back up. But I have to do this manually.

What I'm thinking of doing is taking our MRC 989 and adding some computer interfaces so I can control it from a laptop. This means I can automate the recovery cycling and come up with a much safer approach to module refurbishment.

Bob Wilson

 

I'm starting to see a way that might work. If you can rig up a cyclical charge system:

• X seconds ON - target for 8 V per module
• Y seconds OFF - to let modules rest

Early testing (see my latest update) with fully discharged modules suggest this may be the key to getting the traction battery reconditioned and safely charged.

If you pull the traction battery, please post a table with:

• module serial number
• initial rest voltage

A "table" is created by:

• start with a table tag - the word "table" with "[" and "]"
• column labels with "|" separators
• next rows have data field separated by "|"
• last row has the negated table tag - "/table" and the tag characters

If my testing with previously discharged modules works, this may turn out to be a lot easier than expected.

I don't know, yet. The real problem is pressure inside the modules. I need to find a pressure sensor that won't be "eaten" by the KOH and a load cell. Then I can measure how these modules deal with elevated internal pressure and temperature effects.

Bob Wilson

 

Upper and lower limits help. The general rule is any module running 1.2 V low relative to the rest is a candidate for a bad cell. This means charging will appear to be brief and discharging will happen even faster:


If you want to do it manually, you could charge up a weak module recording the time to reach 7.9-8.0 V. Then put a resistive load, say a tail light, on the module and plot the discharge voltage curve. It it looks like the above graph, it probably is a shorted cell. Note, I recorded this data by hand.

The literature reports that cell shorting normally happens with local heating that melts the plastic separator. It can not be repaired.

Two years ago, I shopped for a smart, RC battery charger and got the MRC 989. However, there are others including one that interfaces with a PC to plot the charge and discharge curves. I paid about $130 for mine and I remember the quality ones running $100 and above. I would loan you mine but it will be busy for a couple more weeks.

New to me. I'll check the manual tonight.

Bob Wilson

 

Hi Dirk,

Are you referring to this?

And this?

Voltage sensing wires that plug into the battery ECU?

 

I have ten Gen 1 batteries that I've been keeping charged somewhat with a regular bench supply. I apply about 8.2-8.5VDC to the cells for about 1 min. X5times /day. This seems to keep the cell voltages anywhere from 6.3 to 7.3 V. When first applying current I see a 1A + spike and then it comes down to around 100 to 300ma within 1 minute. I plan on automating this procedure with 555 timers and series FET's (ea. cell) very soon. This way I will be able to adjust charge time and rest time much more precisely. I also have a couple completely dead cells that read 1v -3v. I think initially pulse charging these at 1Khz /8.5V limited to 1A may revive them as well. We'll soon see.
BTW Dirk, concerning "Autoenginuity came out with an update today that resolved data accuracy and some bugs." This new version requires that you use the "Proline" connector as well, not the older one. You failed to mention what hardware cable you're using.. T

 

I've started designing my 'penultimate' but it won't be ready anytime soon.

Excellent! you're 6/38 through the whole pack. Keep on Keeping On!

At this rate, you'll be ready to do a whole pack test in about 5-6 days. Cool!

The MRC 989 has a charge/discharge cycle built-in. Unfortunately, it is limited to a maximum of 9 cycles and my penultimate will be a little more flexible. <grins>

You're at the bleeding edge!

Bob Wilson