Jump to content
Strawberry Orange Banana Lime Leaf Slate Sky Blueberry Grape Watermelon Chocolate Marble
Strawberry Orange Banana Lime Leaf Slate Sky Blueberry Grape Watermelon Chocolate Marble


  • Content Count

  • Joined

  • Last visited

  • Days Won


MoominPapa last won the day on January 23 2017

MoominPapa had the most liked content!

Community Reputation

405 Excellent


About MoominPapa

  • Birthday 12/11/1964

Profile Information

  • Gender
  • Location

Previous Fields

  • Occupation
  • Boat Name

Contact Methods

  • Website URL

Recent Profile Visitors

13150 profile views
  1. MoominPapa

    Braunston tunnel horse path.

    Change of plan: will be walking back to Braunston. I cycled the new, red, folding bike from Braunston back to Dunchurch Pools today and met the CRT hedge-bashing team, who were covering the towpath with thorns. Predictably, I got Dunchurch with two flat tyres. Grrr. MP.
  2. MoominPapa

    Upware Marina

    Have you fallen out with Bill Fen? I'd have thought that would be ideal. MP.
  3. Looks like the creative ambiguity at Castle Mill Stream has come to an end, and the boats mooring there are going to have problems. From our agent in the Oxford area comes this photo. MP.
  4. MoominPapa

    Another Lithium battery thread

    My estimate of imbalance was as based more on time. I looked in the logs at when the high cell started the fast "hockey stick" voltage climb at the end of charging, relative to the others. For the first couple of charges, this only gave a lower bound, as the other cells didn't start to climb at all. Once I had it sufficiently in balance that I could see the other cells starting to climb before the charge ended, I could look at the time difference between when the first high one started to go up, and when the others did, Multiply that by the charge current, and that gives you the number of Ah to remove from the high cell. My guesstimate came out at around 20, but that's pretty rough. As I was iterating and testing, that's not a problem. I have the formerly high cell the same as another two now, I think that the the fourth is still down a bit, so could probably do with taking a bit more off the top three to balance them to that. As I don't intend to visit the highly-discharged region, where the single low cell will hit the low voltage first, I'm not too bothered about that. The worry about connecting in parallel was precisely the jumper wire thing. For 3P4s, the cells are blocked together with three in one direction and three in the other, so you can't just link them all with straps, and anyway I didn't have any extra straps. I have a load of battery jumper cables from the old installation, but the idea of trying to bolt those on without losing control of any free ends was a bit scary. Your ohms law looks right, but note that a 300m-ohm 30W resistor is quite a beast, and may not be easy to acquire. Just connecting in parallel either takes a very long time, or doesn't work at all, depending on who you believe. You need to charge them in parallel, and for that you need a voltage adjustable, stabiliised, current limited power supply. I did buy such a thing for this project, and in hindsight I should have used it before I installed the cells. MP
  5. MoominPapa

    Braunston tunnel horse path.

    Quick question. Is the path over t he top of Braunston tunnel easily traversable by a bike? MP.
  6. MoominPapa

    Another Lithium battery thread

    An update. The first time I charged the battery, it was obvious that is was well out of balance. The supplier kind-of implied that it had been balanced, and the cell voltages were all close, so I believed him. Resting cell voltages don't tell you anything much. What happened was that the first time I fully charged the battery, one cell (the one with most charge) hit full first, climbed the "hockey stick" portion of the charge-voltage curve and tripped cell over-voltage well before the whole battery charge-termination conditions were reached. This is a very good illustration of why cell-level voltage monitoring is necessary. I considered breaking down the installation, connecting all 12 cells in parallel and doing a balancing charge, but that seemed like a whole lot of work, and lots of risky wiring with possibility of shorts. Instead, I took advantage of the fact that the system makes comprehensive logs, so it was easy to see which cells were out a balance and by how much. I added a BMS function to allow me to tell it to remove x Ah from cell y, using the built-in balancing system. This is quite slow: it can only balance by about 3Ah per day, so it's taken some time. I adopted the approach of charging the battery once per day. Noting what cells started to rise first as SOC approached 100%, and telling the BMS to balance those whilst we ran the boat from the bank for the rest of the day. Next morning I charged again and repeated the process. After two iterations, the charge started ending with the correct charge termination conditions (voltage > 13,9, current < C/20) rather than cell over voltage, but it's taken almost a week to get the balance right. Fortunately, this morning it was correct, as we've left out winter mooring today, so no longer have the mains power that I was using to do this routine. The software continues to need tweaking, Almost every case where action is taken when a voltage or current crosses a threshold has needed to be re-coded to check that the threshold is crossed for a minimum time, typically a few seconds to a minute. Otherwise noise can cause the odd rogue reading to trip a state-change long before it should happen. I also noted today that once the charge is terminated, when the engine is stopped and restarted, charge starts again and another few Ah gets put in before the charge termination conditions are reached again. I've coded something which waits until the battery has been discharged at least 10Ah from full before allowing recharge, to avoid this. We're now using the battery for real, so we'll see how it goes. First indications are good: The charge efficiency is indistinguishable from 100% given the accuracy of my current measurement and integration. It's extraordinary to watch the charge current pegged to the maximum the alternator can provide until the SOC is beyond 99%. Main lesson is that, if using the bare cells, you really do need per-cell voltage monitoring, and most likely also at least an alarm for cell overvoltage or undervoltage, Lack of balance can easily lead to one cell getting overcharged even if the whole battery is still well inside the 14v cut-off. MP.
  7. MoominPapa

    EE 30Gb Data sim - £15/month Rolling contract.

    You could always go back to Lyn's WiFi MP.
  8. MoominPapa

    EE 30Gb Data sim - £15/month Rolling contract.

    Yes, one SIM. Getting this took some time and haggling with the EE call centre. Stiff drink beforehand recommended! The tariff is called "Mobile broadband" or some such. MP.
  9. MoominPapa

    EE 30Gb Data sim - £15/month Rolling contract.

    Don't pay that much. We're in the second year of an EE 24 month contract, 60Gb per month, and paying almost exactly half what you are. MP.
  10. MoominPapa

    Another Lithium battery thread

    Under bed operations complete, and battery covers all in place. The wires are the USB cable to the Arduino, and a remote reset switch, as it's necessary to be able to reset the processor to upload new code. Second photo is after the usual junk has been returned to the under-bed space Third photo is control panel. The new BMS interface is the dial on the right. MP.
  11. MoominPapa

    Another Lithium battery thread

    If the circuit Tom linked to is beyond you, then my design is likely to be also. It does the same sort of thing, but uses CMOS logic ICs rather than discrete transistors. MP.
  12. MoominPapa

    Canal lockage feels last summer’s heat

    Last time we were in Chester I walked the Dee branch. The lock onto the river was shut, but the lock from the basin down to the first pound was open (there are rather nice moorings there) and the the next lock down also looked fine. Is the Graving lock the top of the flight, next to the drydock? MP.
  13. MoominPapa

    Another Lithium battery thread

    I'll try and put something down this evening. On question 3), what are the outputs you're trying to combine. Active high or active low voltages, volt-free contacts, contact closures to ground?. Also what sort of signal does the BEP switch want? MP.
  14. MoominPapa

    Another Lithium battery thread

    I have compression plates fitted. The arrangement I have is a set of six cells mechanically clamped together. The six cells are two sets of three in parallel, with the two sets in series to make a 6v "block". One of these blocks is roughly equivalent in size and weight to a 12v 110Ah lead-acid battery - about as much as you'd want to be lifting and moving in one go. I have two of the 6v blocks next to each other in the battery box, connected in series to form the final 12v battery. The compression arrangement, which came as part of the deal on the cells, looks like factory parts. It consists of aluminium plates across the ends of the cells, joined by substantial straps along the long sides of the cell stack.
  15. MoominPapa

    Another Lithium battery thread

    Update on this. The precharge function is not necessary, my first design was right in that respect. Even bridging the contactor with a wire doesn't generate serious sparks, so there's no much energy moving when the lithiums come on line. I don't have the test equipment to diagnose this absolutely, but my guess is that the change in voltage to the Arduino supply is not a problem, by dV/dt is. The supply is picked of right at the contactor, so the impedance to the batterries is very low. I'm using a little switch mode 12v to 5v PSU for the Arduino. The specs say it doesn't need any filtering on the input, but the very rapid change in voltage looks to be triggering the protection circuits. I put an RC filter on the PSU input, consisting of a 7 ohm resistor, and 1000uF and 0.1uF capacitors to ground. The Ardunio only uses about 60mA so I can easily afford the voltage drop over the resistor. This has fixed the processor resetting when going online problem completely. Phew. Onwards! MP.

Important Information

We have placed cookies on your device to help make this website better. You can adjust your cookie settings, otherwise we'll assume you're okay to continue.