Jump to content

Craig Shelley

Member
  • Posts

    60
  • Joined

  • Last visited

Everything posted by Craig Shelley

  1. Hi, I'm not sure I'd buy into the greasing of contacts on sealed flat-terminal batteries. As far as I see it, any contamination between the contact points is unwanted. On conical posts seizing can be a problem, and where there are corrosive vapours a protective barrier to the exposed metal parts is needed. As far as I'm aware, LiFePO4 shouldn't be producing vapour, and in dry conditions flat surfaces shouldn't be seizing or corroding together. I might be wrong about this, as far as i'm aware, the best dry electrical contact is achieved by maximising the metal to metal contact area. This can be achieved by polishing the two surfaces so that they are as flat as possible. By polishing I just mean working up to 800 or 1200 grit wet and dry paper with a solid block behind it to ensure flatness is obtained. As soon as the finish looks good enough i.e.major imperfections flattened out, wipe it clean to ensure no grit is left behind, and immediately bolt the two surfaces together. The exception is that if the contacts are plated with nickel and are in brand new condition, sanding can only make the surface finish worse. Nickel naturally plates to a very smooth mirror finish. The lugs on our battery pack are tin plated, which is significantly less durable. As a result, they have needed a bit of work to flatten down the contact area. On the copper terminals it's easier to see where the physical contact between the surfaces has been poor; sometimes evident by the formation of oxide in localised patches. Also where the copper and aluminium contacts, which were previously clamped together, are separated it's sometimes possible to observe an amalgam layer on one of the surfaces - I forget which one. I can fully appreciate why grease is necessary in a variety of other situations e.g. on connectors where the contacts slide together, or any connection point that is regularly connected and disconnected and/or exposed to environmental contamination/corrosion. -- Craig
  2. Some clean data at last. The same data is plotted on both graphs. Charging was from solar, discharge was from various fixed loads. On the fourth day, charging was stopped due to upper voltage limit being met.
  3. Just to add a bit of confusion to the conversation, the design of the main rectifier varies somewhat too. The traditional design for a 3 phase rectifier uses 6 diodes. However, the waveform shape at the input to the rectifier doesn't look anything like the sine waves we see in the textbooks. As a result the neutral point at the centre of the Y doesn't stay nicely centered at a mid point voltage. Instead it bounces around all over the place, and at times exceeds the supply rail voltages. So the designers capitalised on this by adding a 4th pair of diodes connected to the neutral. This small tweak can apparently recover up to an additional 10% more power. This scheme sometimes referred to as "3rd harmonic" rectification, presumably because the neutral bounces around at 3 times the fundamental frequency. Interestingly, the last time i replaced the diode pack on our alternator, was because this 4th pair of diodes had decided to phase change to the 4th state of matter. We only happened to notice the blackened remains of this diode pair when we had the alternator stripped down for an entirely unrelated reason. It had probably been like that for years. -- Craig
  4. I have a vague recollection that back in the days when we used to use a handheld dmm to take cell measurements. The spreadsheet we recorded the results on would sometimes indicate an anomalous reading based on the sums of the cell voltages. I think we concluded that probing the nut gave unreliable results. It's been a while since we've had to take manual measurements and my memory is a bit hazy. Probing the terminal post propper started to cause damage to the insulating terminal covers. Then we decided enough was enough and automated the process. I suspect there might be some form of thermal junction or galvanic effect between dissimilar metals at the interface of the nut and terminal threads that causes voltage measurement error on the DMM. I did read somewhere (unable to find the link) that the nuts might have thread locking compound, but the fact that you've been able to rotate it suggests otherwise. I believe the nuts are made of zinc plated steel, and the terminal post is aluminium or copper.
  5. Interesting teardown photos and description: https://liionbms.com/php/prismatic_cells.php Youtube video CALB Autopsy: Youtube video of CALB Resuscitation!
  6. Yes, it doesn't seem logical, unless the terminal post is fractured within the nut. Or your DMM is acting up. Contaminated probe tips? I've read that the nuts can sometimes have thread locking compound applied, and therefore shouldn't be relied upon for an electrical connection. E.g when probing with a dmm. It would be interesting to see a cut away diagram of the inside of one of these cells. It might shed some light on the mystery. Thanks for posting... We'll bear it in mind in case anything similar happens to ours. Regards, Craig
  7. That's really strange. I always assumed the hex part was just a nut to stabilise the terminal against the case of the cell. So you saw a voltage drop between the two points arrowed below?
  8. I just noticed that you have some additional lugs for your instrumentation on the 'dodgy' terminal. That's something we considered doing, but didn't like the idea of because it jacks the bolt out of the battery terminal, and causes the mechanical pressure of the bolt to be applied to a small area on the strap. In the end we ended up going for crocodile clips for the instrument wires, which certainly aren't great.
  9. Get polishing! Hey, we've had a similar issue. It was especially visible on the temperature readout. From memory, the CALB manual does say that the terminals need to be polished each time the pack is reassembled. I think our issue started when we striped the pack down to relocate it. The terminals looked good enough so we just put it back together. The best method we found in the end was fine wet and dry paper backed with a block of wood to ensure the polished surface is flat. The copper at the centre of the terminals becomes a mirror finish. Same goes for the straps.
  10. Hi, The link below is to a ZIP containing the HTML document and Javascript files. I was a little reluctant to share this via google drive as I cannot guarantee how long the link will remain live. Also I'm not too sure about the licencing terms. I'm aware that the Leaflet library is distributed under a BSD permissive licence. Since this off-line page loads its data from the feeds published by IWA, if ever those feeds change it will stop working. https://drive.google.com/file/d/1KaBVmV8ua3Le-KHoSwlUvdTBAtwso7hQ This is exactly the same map as is published on IWA, but with some personalisation of the style settings, different line colours etc. in map-v2.js I've only ever tested this on a PC using Firefox. Just make sure the files "IWA Map.html" "map-vendor.js" "map-v2.js" and "leaflet-2.css" are in the same directory on your computer, and open "IWA Map.html" in your web browser. I suppose theoretically, I could embed the CSS, and Javascript within the HTML document to possibly make it easier to use. It might work on a smartphone - I've not tested that though. Regards, Craig
  11. This map is great. I've had it on my wall for several years now. We also have one on the boat, but had to cut it down to a more practical size. It's possible with a bit of snipping to get it down to A2, whilst still having the enlargement sections not obscuring anything. One thing that makes this map particularly useful is the inclusion of major roads. Also, it is possible with a bit of script modification to use the very nice map that the IWA publish via their website as a full-screen webpage with scroll-wheel zooming. Unfortunately, it is not possible to attach the .html .css and .js files here due to file-type restrictions. I've attached some screenshots, which ironically are about 10x bigger than the html/javascript. Edit: I forgot to mention that it also supports hovering over the waterways to show their name. The map display is just using a Javascript library called Leaflet.
  12. Had another letter through the door today, stating that the policy expires on 4th of December 2021. Starting to get a bit confusing now.
  13. Hi, Sorry to reignite an old thread, but it looks like my dad has ended up being caught out by this trap. The letter (attached) arrived today 04/02/2021, saying that his car insurance had been cancelled yesterday 03/02/2021. With the poor weather recently, I though perhaps the letter had been delayed in the post? ... but the envelope was franked on 01/02/2021. Some weeks ago, my dad had to phone his car insurers for a completely unrelated reason but ended up mentioning that he was living on his narrowboat. He's been staying on the boat permanently recently because his girlfriend is working on the covid wards and he narrowly escaped getting infected at home during the first wave. He does have a permanent address, where he often does keep the car because the car becomes an inconvenience to him while he's cruising. During the lockdown it's been a bit different; boat movement has virtually ceased and with not wanting to use the trains and buses as much, he's preferred to keep the car close by. I guess he was telling the truth for the situation at this point in time. To be fair, he was told over the phone that his policy would be cancelled, and confirmation would be in writing. After a follow-up call to clarify his situation, they said they would forward his clarification to the underwriters. I'm guessing this letter states the underwriters final decision? As you can probably imagine this has now left him in a bit of a predicament. Any recommendations on a new insurer as we're guessing it's not going to be easy to find one now that he has had a policy cancelled? Best Regards, Craig
  14. After digging back through the photos.. (memory not as good as it used to be) I can vaguely remember cutting through the transparent surface film with a scalpel. I can also remember it being possible to thread the wire in from the ends alongside the copper foil. I think the copper foil is only adhered to one side of the transparent film. I remember cutting a hole in the film part way along, to allow the wire to be soldered, either to the foil or back onto itself. You can see on the photo that I did a staggered cut of the sheet so that the connection point wasn't located directly beneath the batteries. The batteries sit directly on top of the heating film, with insulation beneath that. The sides of the batteries are also insulated, as are the lids of the enclosures. Until recently we didn't have much in the form of battery temperature monitoring, we just used to use the read-out on the solar charge controller, which had a single sensor on a long lead. In my previous post, I described the fact that we had wired all 3 heaters in series to limit the maximum power and make the system "inherently safe" I recently determined that this winter the series heater configuration would not be sufficient to heat the batteries in a sensible time period, and heating the batteries over a long period of time wasn't very efficient due thermal loss from our enclosures. We now have the heaters wired in a parallel configuration, with fuses rated at 7.5A on each of +ve and -ve as before. Each heater measures as approx 4.9 Ohms. Powering from 13V, with approx 0.4 Ohms of wiring resistance, we are able to dump about 22W of heating power into each battery. Approx 6.4A total. The attached graph shows the the temperature rise. The heaters were run for 2 hours, then turned off for 1 hour before charging the batteries using the alternator. The latter stage of charging was done at a lower engine RPM and the lower charging current this caused is evident in the graph. The graph shows the self-heating of the batteries during charging at a similar rate to what the heaters achieve. The thermal resistance per enclosure (leakage) was determined to be about 1.16 °C per W, and the heat capacity per enclosure was determined to be about 35.5kJ per °C. From these values it is possible to predict how much heating time is needed to raise the temperature by a given amount, and to determine the maximum achievable temperature etc... Depending on the design of your enclosures, wiring the heaters in parallel like this might not be inherently safe i.e. it might be possible to heat the batteries beyond their specification. I'd recommend installing an additional protection system such as thermal fuses, clockwork timer unit etc.. With our system, it is theoretically possible to heat the batteries to about 19 °C above the engine room temperature. We'll certainly be removing the fuses once the winter is behind us. -- Craig
  15. I've tried to implement this, but due to various effects our long-term Ah counting has turned out to be not good enough. I'm hoping to make improvements over the next few weeks, but there's no guarantee that will overcome all of the long-term drift error which is inherent with this method. Charging to 100% periodically to reset the meter is one obvious solution, but you might not want to do that as often as is needed. The problem we found when using estimated SoC to stop the charging was that the numbers ended up looking too trustworthy and we lost our mental reference of SoC. We only realised that the SoC gauge had drifted out when the vacuum cleaner tripped the low voltage protection switch. This is despite the fact that we still had a trust-worthy volt meter. For some reason it was more satisfying to trust the Ah based SoC gauge. We then decided to charge to a known point just to see how far down they really were. The engine ended up running for over 6 hours! It was reminiscent the feeling of "suddenly realising you're lost" after figuring out that you've been misreading a map... for most of the day. In the end we opted for a simple approach of charging until a predetermined voltage, and then resetting the counter. This method is not perfect, but it is easy to live with the "known imperfection". The result is a variation in cut-off point due to temperature, charge current, and other factors but at least this is a variation on a fixed datum as opposed to something which can gradually drift with time. I was hoping to implement a SoC percentage gauge based on the battery voltage, which could then be used to null out the long-term drift in the Ah measurement. It turns out that this is non-trivial since the various factors are involved, primarily the history of charge/discharge currents. There are countless papers on this topic in the engineering/science journals. It's easy to overlook the fact that we've got by using just a simple volt meter for the first couple of years. My father used to take a glance at the meter first thing in the morning before the solar started to kick in. Using this approach he somehow got an approximate feel for the SoC. Also without realising it he also used to wait for an appropriate moment in the fridge compressor cycle. I imagine this is possibly the simplest approach to implement in software as instead of having a complex model to compensate for the effect of current, it's easier just to eliminate the effect by taking the voltage measurement at a carefully chosen moment.
  16. That is an impressively low idle power consumption. Is that also counting Ah in/out of the battery? I always wondered how the BMV achieved zero-drift. It's very difficult to achieve the required dynamic range i.e. being able to measure 300-400A at the top end whilst accurately measuring sub 10mA at the bottom end. The zero calibration can be thrown out by a few microvolts at the shunt resistor - easily created by a thermocouple junction given the different metals involved. This makes my RPI solution look rather laughable at 80mA!
  17. Attached is a "zoomed in" view of the charge profile from yesterday. I agree that the voltage we are tripping the alternator off at is a little low. It would be nice if the alternator controller could benefit from some of the monitoring improvements which are currently going in. I might add a comms link back to the Pi to enable a proper CV/CI charging profile. If it's useful, I can export the data from google sheets, or provide a link to the live sheet. From experience, increasing the alternator cut-off voltage further doesn't tend to lead to a great deal more charging time because the voltage at this point in the charge profile begins to ramp quite steeply. The data on the sheet and graphs is decimated down to a 10 minute sample interval, but is internally sampled at a much higher rate to enable accurate AHr counting. The cycling of the fridge can clearly be seen on the graph, and it is possible that the fridge compressor cutting off might have triggered the alternator to then trip off too. I did wonder if that might be the case. When I originally assembled the pack, I polished the mating faces of the terminals, perhaps too obsessively, using several grades of sand paper. The sand paper was wrapped around a wooden block to ensure flatness of the surfaces. They looked like mirrors when they were done! In late 2018 the pack was split up and installed into enclosures. I remember noticing that there was no longer a bright mirror finish to the copper, although we didn't have time to faff about re-polishing. They looked good enough. The temperature probes are tightly attached to the centre links of the packs, my rationale being that the copper terminal must surely extend deep inside the cell, and hence would give a closer indication of the internal temperature than sticking it onto the plastic casing. In the temperature graph, the battery heaters were running throughout the entire plot, although their effect is barely noticeable at that scale.
  18. Its nearly been a month since the new battery monitor was switched on. Still some work to do to move the current measurements to the new monitor so that the prototype system can be switched off. The graphs are relatively self explanatory. and are beginning to show the effect of temperature on performance. The near vertical steps on the blue Charge trace show when the alternator was used. This charges initially at about 100A, before dropping back to about 80A as it warms up. The charge current is engine RPM dependant, and quite possibly not consistent between charges. The more gradual charging is caused by solar, which can provide a maximum current of 30A but I doubt that has been realised in these graphs. Maximum value is probably around 20A, on a sunny day. The temperature graph shows a number of effects. The obvious one is the internal battery heating caused by charging, which appears as the steep temperature rises on the graph. As the engine room remains warm for several hours after running, the batteries continue to warm up after charging, albeit at a much slower rate. There are a couple of obvious instances on the temperature graph where the battery heaters were switched on. The first being at about midnight on 11th Nov, running till about 9am. The second being midnight of 29th Nov, still running now. A noteworthy point regarding the heaters is that my previous comment about their efficacy was wrong by quite a long way. I think i previously stated a temperature gain of about 20C above ambient. These graphs show it to be at best 10C. This is obviously just the equilibrium point where the enclosure is losing (to ambient) the same power as we are providing in heat. It should be possible, given a clean enough heating curve, to estimate the thermal resistance of the enclosure, and the heat capacity of the batteries. All rather academic, although it shows the insulation of the enclosures could be improved. The most interesting effect is how the charge trip-off point is effected by temperature, and these graphs are helping to unravel some of the mysteries. Firstly a quick note about the charging scheme; the alternator charges at full current until a voltage threshold is crossed, and then trips out. Hence constant current charging only. I hope to improve this now that we have a better monitoring solution. The first observation is that the internal cell resistance is temperature dependant (negative coefficient) - no surprise. Cell voltage probably is too, but that is less easy to correlate from the graphs. At low temperatures, the internal resistance is higher, hence during charging the trip-out voltage threshold is crossed earlier. However the batteries warm up as they are charged. So if charging from a lower state of charge, there is more time to warm up, hence a higher temperature is achieved, lower internal resistance, and a higher state of charge is achieved. One final effect is that when charging for long periods of time, the engine room temperature rises further, and causes the alternator to become less efficient. The output might drop to 75A. Due to the effect of internal resistance, this causes more charge to be absorbed before tripping out. Clearly several inter-related temperature effects which are in a bit of a positive feedback loop. I've tried on several occasions to "model" the internal resistance of the battery; probably a bit of an over-technical term. This is largely academic, but has some value. If the internal resistance can be estimated, it would be possible to compensate for the effect of load/charge current on the measured voltage. This would also lead to better state of charge estimation, especially when it is not possible to take a clean voltage measurement with minimal load. Now granted, on a boat it is not really that necessary to precisely know the state of charge. A Volt meter is a bare minimum, but perfectly sufficient means of estimating SOC; this is all we had for the first couple of years. On my "quest" to understand the behaviour deeper, I discovered fairly early on that the resistance is non-ohmic. As the current is increased, the resistance reduces. The graphs also show another well-known effect, that the voltage is dependant, not just on the instantaneous current, but on the current in a window several hours prior. I've demonstrated that this can easily be modelled us an exponential decay function, and it is relatively trivial using a "goal-seek" to estimate the time constant and magnitude.This modelling has always been thrown out by the fact that the voltage/current relationship is non-linear. As always, in the data there are too many uncontrolled variables. Some experimentation under more controlled conditions might be needed.
  19. LOL the back bedroom already looks like a time machine from the '80s. The main interface to this is currently via smartphone/tablet, it's still undecided if we'll ever install a fixed display. The representation is a little experimental to see what works best. My father prefers to interpret an analogue dial over reading digital display. He finds the existing 7seg readouts need a lot more mental effort to understand. As for the 12 rather than 3 (I'm assuming you mean 4). Some of the decision is a little historical in how the battery packs have been assembled and installed. We've ended up inheriting some of the layout limitations from the original lead acid batteries. With hindsight we would have been far better off to have had a fresh start, and possibly install them into a more convenient location, but at the time we didn't have the tools or knowledge of how to make up the cable assemblies. We were very fortunate that Jeremy Bloomfield had offered to assemble a cable set with Anderson connectors to enable us to retrofit it into the existing connection points. Also at the time, it was my father's first year onboard, and the boat hadn't exactly been problem free. I've deliberated over adding the remaining links to the battery pack as only have 4 cells to monitor is far simpler, but it would now be a bit messy as there would need to three cables interlinking between the enclosures. Then there'd be the question of what fuses to install and to devise a means of disconnection to enable enclosures to be removed. The other thing which crossed my mind was regarding the fault tolerance of the pack. In the unlikely scenario that one cell began to break down, and I must stress that I don't know what the common failure modes of these cells are so this is speculative, but if there were some form of "non-catastrophic" membrane failure of one of the cells, I theorised that there might be a better chance of not damaging the other cells if the inter-connections weren't fitted. i.e. no cells directly in parallel with each other. Also with the cells separated, it might give some insight into how each cell is performing individually, and how they are contributing to the overall pack. There's still a lot we don't know about how these batteries perform in this application, and I certainly don't mind sharing what we learn as the project evolves. So the reason is partly historical, partly theoretical and partly educational.
  20. We're starting to get some interesting data from the new monitoring system. There are a few interruptions on these graphs where we disconnected cables to tidy up the installation, paint the boxes etc... The temperature graph shows some of the batteries leading/lagging, i guess due to differences in the physical locations of the enclosures. Interestingly, the effect of internal heating due to charging can be seen quite clearly. Engine was run on the 4th and 6th Nov. My understanding is that the linear(ish) portion of the heating is caused by resistive/chemical heating of/within the battery (temperature probe is attached to the battery terminals). The gradual ramp-up thereafter, I believe, is cause by the higher ambient temperature in the engine compartment. The gradient change on the temperature graph when the charging was switched on/off on the 6th Nov is quite noticeable. It gets even more interesting to see what the individual cell voltages are doing. The graph "Cell Balance - Battery A" shows that during discharge, cells A3 and A4 have a slightly higher voltage than A1 and A2, but intriguingly the roles reverse when the batteries are being charged. I can only guess that this is because cell A1, and to a lesser extent A2, have a higher internal resistance than A3 and A4. Packs B and C exhibit similar variation but with different cells appearing to be more dominant. These traces agree with the spot measurements I have carried out manually over the past few years. Note: the voltage ripple is caused by the fridge compressor cutting in/out. There's still more work to do on the graphing/dashboard side of things and the current shunts still need to be wired up. I'm hoping to eventually have it controlling stuff rather than just gathering data. Over the next week or so the data might begin to look a little cleaner with fewer interruptions. I'm sure, as always, there'll be something new to learn.
  21. Ah yes, i forgot about those. So that's 8 chunky rectifier diodes, and 3 little ones for the 100A iskra.
  22. Sounds right to me. You should be getting more than 30A - not sure what your alternator rating is. As with most things in the automotive world, the specifications are usually overstated. We have a 100A iskra (though some sources rate it as 90A) It only gives this current when stone cold, soon dropping down to about 80A as it warms up. This current is RPM dependant, so it is possible to compensate for the drop-off, to some extent, by increasing the revs. Failed diodes can also cause a drop off in available current. As most alternators usually have at least 3 phases, with a rectifier consisting of at least 6 diodes, a loss of a single diode does not necessarily cause a complete failure. Photo shows iskra diode pack where a pair of diodes presumably went short circuit for a while, caused some melting of the plastic, before blowing open circuit. In this case the current reduced to around 60-80A, and was more RPM dependant. A curious configuration of diodes in this rectifier - it has 8
  23. We just turn them on before charging with the alternator. Even when below zero C, as a domestic supply they perform quite adequately whilst discharging. From memory the solar controller is configured not to charge below zero C. A few years back, the recommended charge cut off temperature was zero C but that has been revised upward to 5 C more recently. If I remember correctly the winter of 2017/2018 dropped into sub zero temperatures for several weeks. At that time we didn't have any heaters, insulated enclosures or even temperature sensors. I'm sure we would have charged with the alternator whilst below zero C at some point. By contrast, last winter there was barely a frost so the heaters didn't get much (if any) use.
  24. I always find the term ”loss of capacity at low temperature" a bit of a curious conundrum. If the capacity of a charged battery goes down, where does the energy go? Even if the voltage at the terminals is lower at low temperatures, the energy stored cannot simply disappear. The main thing I have observed at low temperatures is the apparent increase in internal resistance. I.e. pulling a heavy load causes the terminal voltage to drop more when the batteries are cold. Likewise, our charger which charges until a voltage threshold is crossed, trips out notably earlier when the batteries are cold. In an EV with long strings of cells I could imagine this internal resistance increase being problematic as the load would have to be stopped as soon as the weakest cell voltage reached its minimum allowed value. This would definitely cause an apparent loss of range. I theorise that if the batteries were to be warned up, the the lost EV range would be recovered, hence energy conserved. One other observation worthy of note is that the internal resistance resistance of the cells is highly dependent on the state of charge. The effect of temperature and SOC seem to add together to influence the overall internal resistance.
  25. We currently have no thermostat. I cut the heater film into 3 separate sections, one for each battery enclosure. The heaters were then wired together in series instead of the original parallel configuration. That reduced the overall power to a level sufficient to provide a gentle heat which, from memory, eventually reaches an equilibrium temperature approx 20 degrees above ambient. This enables the heaters to be left on without any means of control. Fuses provide basic protection for over current, and the switch is a guarded and illuminates to give a very clear indication that the heaters are powered. As the new monitoring system gives us better visibility of battery temperature, it is theoretically possible to automatically control the heaters. I doubt we'll ever get around to making that happen. There are 3 separate enclosures, each containing 4 cells in series. The -ve and +ve wires from each enclosure connects to the busbars.
×
×
  • Create New...

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.