Down to my last used Jonny. Only the Anglia 105E left.

[Rust Collector]

30 minutes ago, jonny69 said:

Honda UK gave me a phone number but I was put off by the language barrier. I tried to get in touch by other means but the Honda Japan site is in Japanese so I sort of gave up there.

I'm still friends with a chap I used to buy cars from in Japan. I could ask him if he'd mind getting in touch on your/our behalf?

[jonny69]

4 hours ago, Rust Collector said:

@jonny69 I'd be really interested to know more about testing the sticks though as my pack needs rebuilding at some point!

Ok, cool, you're familiar with the basics then. None of this is difficult, it's just very time-consuming. You need to determine a couple of bits of information for each stick in the pack: the capacity of the stick and the voltage drop when you put a big load on it.

First make sure the cells are all active. So if it's a pack which hasn't been used for a while, charge it up and discharge it a couple of times to re-activate and balance the cells. You'll need some big resistors and a digital voltmeter for the next bit.

To measure the capacity, take a fully charged stick out the pack and put something like a 8 Ohm resistor across it so it's discharging at around 1A. Measure the voltage at the beginning and at 30 minute intervals until it has fully discharged. You can do some V=IR calculations to work out the rough capacity. A new stick should be 6.5AH.

Repeat that on all the sticks, number them as you go and make a list with the capacities. You'll have a list of high to low capacities. Then put them all back in the pack and charge them up again using the grid charger.

To measure the voltage drop under load, which is probably the more important of the measurements, take a fully charged stick and put a 0.5 Ohm load across it so it's discharging at around 15A. This is a lot of power to dissipate, so use 4 cheap 50W eBay resistors mounted on something that can dissipate the heat. Ideally you actually want to pull about 30A out of it but this is a hell of a lot of current and power to deal with. You don't need to fully discharge the stick like this but you want to note the voltage drop when you connect the load. Again, repeat this with all the sticks and make a list. This tells you how strong the stick is (how it's effectively performing when the car asks for assist) and if there are any dead cells in the stick. Any stick with a dead cell will be obvious because it will drop a lot more volts than the other sticks - you should discard those ones.

Repeat all that on a second pack and you can basically make up a set of 20 of the strongest, best-matched sticks.

2 hours ago, Rust Collector said:

I'm still friends with a chap I used to buy cars from in Japan. I could ask him if he'd mind getting in touch on your/our behalf?

Could give it a try! I'm guessing he'll basically need my VIN and Honda Japan will be able to access the information if they have it?

[jonny69]

Another Civic battery has entered the building. Let the stick matching commence.

So the rough plan of action here is to get an idea of the capacity of the sticks and to see what they do under load. I had a root through my resistors in the loft and I’ve got these, which will do me a selection of loads from almost nothing up to a couple of amps:

They’ll be fine for voltage readings and light discharge. Then to pull some serious current out of them to highlight any bad cells, I’ve bought five 1 Ohm 100W resistors. The stick voltage is around 8V when charged, so each of these will pull 8A. That’s about 64W going through the resistor. With all of them in parallel, I’ll be able to pull 40A out of a stick for a couple of seconds and basically see what the stick does under heavy load. That’s a lot of current, so I’ll be switching it with a spare solenoid I found in my shed stash:

Handy. I’ve also got some heavy gauge wire and M8 ring terminals that can handle that sort of current. I should be able to sort all the sticks and pick the best ones and make a good pack.

I’m stressing a little less about this now because I’ve sorted out a tiny replacement. That means I don’t have to worry too much about timescales or insurance, I’ll just sort that out ad-hoc for test drives, MOT etc.

[Asimo]

Watching with interest; I had planned to do just the same with my Insight if needed but I stayed lucky.

[Tickman]

On 8/30/2023 at 10:13 AM, jonny69 said:

Another Civic battery has entered the building. Let the stick matching commence.

So the rough plan of action here is to get an idea of the capacity of the sticks and to see what they do under load. 

If you wouldn't mind doing some extra pictures that would be ace as almost anything on Insightcenteral is ages old and pictures are no longer visible.

[Tickman]

16 hours ago, Asimo said:

Watching with interest; I had planned to do just the same with my Insight if needed but I stayed lucky.

From this I'm assuming that the battery pack is original. 

It is a bit unbalanced at present so I'm guessing some of the cells are not doing great.

[Asimo]

19 minutes ago, Tickman said:

From this I'm assuming that the battery pack is original.

I believe that it was original when I had it. 23 year old batteries! Amazing they work at all really.

[jonny69]

They must have been absolutely amazing batteries. If you think that up until a couple of months ago mine was still happy delivering 100A at full assist until just one or two of them in the pack couldn’t do it. A hundred amps; that is in excess of MIG welding currents out of an 18 year old battery!

[captain_70s]

On 8/30/2023 at 10:13 AM, jonny69 said:

. That’s a lot of current, so I’ll be switching it with a spare solenoid I found in my shed stash:

Looks an awful lot like a 1970s Triumph starter solenoid...

[jonny69]

Right then, got pretty much everything I need. Essentially what I was planning to do was just take a voltage reading, manually trigger the solenoid and take another voltage reading after 5-6 seconds. This would be perfectly adequate, but I decided to complicate things a little by adding some Arduino control. I can get the Arduino to automatically trigger the solenoid and read/log the voltage for 5-6 seconds, then disconnect it again. So below I’ve got:

Spare bit of aluminium extrusion construction kit stuff, two Insight sticks, bag of high current M8 ring terminals, Arduino UNO, 4x relay board for the Arduino, bag of 100W 1R resistors, and a 1970s Triumph starter solenoid. Well spotted @captain_70s

Quick mock-up with some bolts and washers and T-slot nuts for the extrusion stuff and it’s going to look something like this:

The Arduino will be controlling the board with 4x relays. Will get one of those relays to switch on the starter solenoid from a 12V supply. It pulls about 3A at 12V DC so it’s probably going to need a flywheel diode across the solenoid coil to protect the relay. Relay is rated to 10A at 30V DC but my experience is switching DC tends to make those sorts of relays stick. I found some 1N5401 diodes in my stash at work which are pretty big. They ought to handle anything the coil spits back out. 

So then the Arduino reads the voltage. It only reads up to 5V and the stick voltage is 8-9V depending on state of charge, so I’ll put it on a 10k potential divider with a 50/50 split. My pal also suggested I put a 5V Zener diode across the Arduino to protect it from getting fried when the solenoid disconnects the batteries. I had another root in my stash and found some 1N5338B Zeners which are 5.1V. Voltage reading range will be up to around 4.2V so they’re ideal. 

I am of course wondering if I’m over-complicating things. It’s not too late to back out of this and just do it manually…

[jonny69]

Lunchtime shens. I soldered up the resistors and connected the solenoid up to my 12V supply so I could trigger it manually and have a look at what the battery sticks do. Started with an 8 Ohm resistor because not much can go wrong there and then added in the 1 Ohm ones one by one as I got braver. So with all five 1 Ohm resistors in parallel, that's 0.2 Ohms, which is about 40A at 8V and 35A at 7V depending on the state of charge. Just for reference, the car commands about 10A for cranking and 35-40A at around 50% hybrid assist when driving.

This stick was sitting around 7.7V resting voltage, and pulled down to 7V with the full load attached:

Emergency snips on hand in case anything bad happens. Snip through that thick red wire on the left.

The voltage actually settled after a few seconds so I don’t think I’ll need to bother with the Arduino. I think I'm just interested to see what it drops to under heavy load, to see if there are any weak cells and I'm pretty sure this will tell me what I need to know. It’ll be pretty obvious which sticks have a weak or bad cell in them because the voltage will drop further.

I actually brought two sticks in to play with and, based on the above, it looks like I might have one good one and one bad one on the bench. MUCH EXCITE. 'Bad' stick was 7.3V resting and 5.2V under load - a much bigger drop. However, I do need to repeat this after a full charge because these Insight sticks haven’t been charged since July and are probably a bit self-discharged. Then I can rattle through these of a lunchtime. Takes very little time per stick, but there are 60 of them to do!

[Asimo]

As you've got the 'scope meter, why not set the 'scope up single shot with timebase say = 30 seconds.

Serial number on the stick, run the test, photograph the plot with it's serialised stick. and repeat.  

Unmuddleable* record of each stick's 30 second discharge characteristic!

[jonny69]

1 hour ago, Asimo said:

As you've got the 'scope meter, why not set the 'scope up single shot with timebase say = 30 seconds.

Serial number on the stick, run the test, photograph the plot with it's serialised stick. and repeat.  

Unmuddleable* record of each stick's 30 second discharge characteristic!

That’s not a bad shout. I’m not an expert on that scope but I’m pretty sure it’ll do what you’ve described and hold the trace on the screen. 

[Tickman]

This is all just words but very interesting words.

[jonny69]

I had a look with the scope attached but it didn't actually tell me anything extra that was useful. Good stick followed by 'bad' stick:

I guess what it has told me is that the voltage drop when the load is connected is almost instantaneous, as it happens in about 5ms. I was expecting it to be slower than that. So having cleared that up, I tested all the sticks using the basic voltmeter method, making a note of the resting voltage and the voltage after 5 seconds with the load connected. That resulted in 17 sticks which were almost identical and 3 outliers which dropped more voltage. So what I did on the outliers was got a leather punch and made some small holes in the heatshrink so I could measure the voltage across the individual cells while they had a load attached. I did the same on a good stick so I had something to compare it to. It basically showed that the bad sticks had a single weak cell in them that was dropping a lot more voltage under load:

Good one, voltages written above the holes, all about the same:

Bad one, with one weak cell showing about 0.3V instead of 1.3V:

The other two bad sticks looked exactly the same. So that leaves me with what look like 17 good sticks and 3 bad ones:

What I really need to do now, is get the other pack out the car and do the same on those sticks, and get the third pack charged and cycled a few times so I can do the same on that one. In the meantime, I've also rigged up a capacity checker with a cheap Chinese eBay current and voltmeter. It's got a shunt so it can measure current and will log battery capacity. I've got it wired up on a good stick with an 8 Ohm resistor doing the discharging (about 1A when the stick is charged) to have a look at what it does:

[jonny69]

Tidied up the rig a bit, checked it didn't explode or overheat and left it discharging overnight:

That suggests a remaining capacity of 6.4AH. Original sticks were supposedly 6.5AH so that looks pretty promising. I've now got one of the bad sticks discharging to see what it does. The starting voltage was about 0.1V down on the known bad cell. After about 1.4AH discharged from it, the bad cell was empty and showing a negative voltage while the others were still holding 1.26V. I'm not purporting to be an expert on this, but it’s looking like the ultimate bench test for these sticks is to charge them fully, leave them for at least a month to weed out the high self-discharge cells, then test them.

[Noel Tidybeard]

i supose the next step would be to see if you can re-make a stick with all good  cells

[Dave_Q]

1 hour ago, jonny69 said:

I'm not purporting to be an expert on this

I think whether you like it or not, it looks like you are

[RayMK]

1 minute ago, Dave_Q said:

I think whether you like it or not, it looks like you are

Agreed. Shockingly obvious.

 

[jonny69]

You guys are funny 😛

28 minutes ago, Noel Tidybeard said:

i supose the next step would be to see if you can re-make a stick with all good  cells

It's possible in theory. The cells are welded together but I understand they break apart with a bit of force. Then it would be a matter of doing the Youtube spot welding trick with nails and a car battery to stick them back together. I think I've got enough sticks to be able to avoid doing that though!

[Dave_Q]

You can also get little spot welders off of your preferred Chinese tat site, people use them for building electro-push iron batteries.

[jonny69]

Still working on this. I got the 3rd pack apart and tested on Friday. More of a mixed bag in that pack but plenty of strong-looking sticks and I’m more confident I can make a good pack now with what I’ve got. 

I’m planning to make the pack up today, so overnight I’ve re-charged one of the sticks I’m planning to use that I’d fully discharged. I borrowed the bench supply from work, set the current control to 0.5A and put it on a mechanical timer overnight (in the socket):

Bit hokey but nothing exploded overnight. That also means nobody got hurt and nobody died, so I should be good to go. 

[jonny69]

Comfort break update. It’s taking a spectacularly long time to do this today because I’m having to do this on the kitchen table and I’m getting a lot of help from little interested hands. Screws and tools keep moving and interested little faces keep getting in the way. It’s frustrating but at the same time I secretly want them to be involved so I can’t get too angry at them. 

So, the pack is now assembled and ready to put in the car. I chose to swap the best six sticks from the second HCH pack with the weakest six sticks from the original Insight pack. That seemed to be where the crossover point was in terms of stick performance. Then I put them all in order of voltage drop under load and paired them up one from each end until I got to the middle. The logic here is the car monitors them in pairs, so I want the pair voltages to match as closely as I can get them. Then it was just a matter of double check I got the stick pairs in the right places in the pack based on the connections on the board in the back of the pack which links the sticks together. 

[jonny69]

It’s in. Top-up charge overnight and test drive tomorrow…

[jonny69]

Calling that fixed. Battery is working the best I think I’ve ever seen it. No issues on a 10 mile test drive with a few restarts thrown in the mix. Motorway fine and it’s healthily dipping in and out of the battery on acceleration and deceleration. Got a clear stretch on a local bypass where I could nail it for a good 15 seconds before I had to smash the brakes and it was happy putting out the full 100A of assist for all of that. It’s also using full regen on hard braking which it won’t always do if the battery is out of whack. I’m happy/relieved/drained

Now need to clean it up because it’s absolutely caked in shite from the last 3-odd months of neglect. 

[jonny69]

Need to cheeky bump this as it’s technically for sale now the battery is fixed. Looking for £xxxx. Got like 4 days left on the insurance so if you want it and want to go on a razz and you’re somewhere near Heathrow give me a shout asap otherwise I’ll have to sling it on SORN and test drives will have to be restricted to the local roads where it’s out of view of anpr and the filth 👍

[jonny69]

No for sale thread because I’m hurriedly posting this while hiding in the shitter because there’s pandemonium downstairs, kids screaming etc. This is what my life has turned into. Buy my car please. 

[jonny69]

One last trip to the tip to make some space in the house. Bye bye knackered microwave, dead hifi, old hoover, two Civic hybrid battery cases and some old clothes. Hello rusty old French bike. Er, What? That wasn’t supposed to happen  

[Kringle]

I find it to be an abject failure of my cockney DNA if I return from the tip empty handed.

[jonny69]

8 hours ago, Kringle said:

I find it to be an abject failure of my cockney DNA if I return from the tip empty handed.

They usually won’t let you take anything. I always ask but I was pretty surprised when the bloke said he’d look the other way