Tuesday, 28 October 2014

EIG Cell Testing continued

So seeing as the EV was out of action and my mechanic had pushed back the start of the RX8 build due to his business partner upping sticks and declaring he wanted to sell up and move on, I got busy with the testing of the new EIG cells, and this time I was going to make sure that there was no chance of failure.
So I bought a brand new 12V battery to compliment the better of my two existing 12v lead acid batteries and dumped the older/weaker one just in case this had anything to do with the failure (although Steve assures me that the Powerlab will compensate for any weakness in the Pb cells and simply fail the test rather than going bang so I still feel the last power lab must have been faulty) and I also replaced all the 12v battery terminal connections with brand new bright and shiny connectors again just in case a bad connection had caused any sort of spiking.
Voltage spikes are the most likely cause of electrolytic failure so I am taking no chances, also I reduced my test cell to a single 20Ah EIG cell rather than 2 in series even though the powerlab should be well able to handle 2S of these cells (we had already done the power calculations to check we were not pushing it beyond the 1344W limit and again if we did try to it should just fail the test not explode!) I was not willing to take any chances this time and am going to stick to single cell tests despite the huge amount of time this will eat.
So to get the ball rolling proper I configured my BC168 balance charger to charge at 6S (8A) so I could top up 6 EIG’s at once and once I had the first six charged the ball really started to roll, it turns out that a single cell test takes around 28 minutes @ 40A to discharge and it takes the BC168 around 3 hours to charge 6 cells so all in all a pretty good balance I can charge 6 whilst I am discharging 6 (albeit one at a time) and then rotate.
So pushing hard to keep this turnaround going I managed to get 30 cells done in 1 weekend that is 1/10th of the entire pack so assuming I only get weekends to do it (I will be doing some in evenings) it would take me 10 weekends to complete the job.

Brush/Bearing wear on the Kostov K9!

Ok so here comes another tale of woe due to my own inexperience of EV related matters, I was driving to work the other day with my daughter to drop her at school, on the short divergence to do the school drop I have to pass a lane with a large regular flood puddle and being an EV driver I always take it slow through this puddle, it is not enough to swamp the motor thankfully but if I hit it with any speed it would give it a damn good wash, Still on this particular day as soon as I had passed the puddle a new and very disconcerting noise appeared and sounded distinctly like brush failure!
So I limped it home and took the fat fryer to work, Guessing that there was more than a clean going to be required here I dropped it down to my mechanic and waited for him to have the time to investigate, a few days later I had the bad news “new brushes required” and the following day a call to say that one of the bearings was also failing (the one outside nearest the wheel).
After thinking over the issue for a day or two and during a chat with a friend the penny dropped, RPM Sensor! I Still have no RPM sensor “WHAAAT” I hear all you EV experts shout, Yes I am a fool and have been driving the little Honda Beat EV for over a year with no RPM sensor and no sensor means no RPM limiting, So now the penny has dropped it seems almost obvious that if you don’t limit the revs somehow then at some point you will be over-speeding the motor, And on a similar note I also have no motor temperature monitoring either.
So I guess the key words here are MONITORING and LIMITING!
So a new set of brushes and bearings are on order and my EV sits looking sorry for itself waiting for the parts to arrive from Bulgaria!

Testing the EIG cells

I began testing the EIG cells by balancing them to 3.8v and then balance charging them up to the 4.15v max Voltage (slightly lower than the normal 4.2v of standard LIPO if there is such a thing as standard LIPO!) shown on the DataSheet.

Then using my newly purchased Powerlab 8 I did a discharge and graph test at 20A and everything looked great a nice looking graph formed over the course of around an hour at which point the Laptops power saving mode kicked in and put the PC to sleep ruining the test!
Bugger! So I began again and at this point things started to go wrong, firstly I could not repeat the test without the Powerlab stating that I had disconnected the battery and then quitting, after much dicking around suspecting bad voltage sensing leads and trying other batteries (the original Turnigy’s I had laying around) I eventually discovered that the hibernation mode kicking in whist the previous test was running had corrupted the AppData in the profile on the laptop and this was preventing me from running another test! (Talk about tenuous links) thankfully I am an IT engineer otherwise I might have gone mad with that one.
So finally I had gotten back to where I was and thought I would attempt a higher power discharge using 2 Lead Acid batteries in series providing a 40A current dumping capability (BAD IDEA)
Now according to Steve my local expert (and the instruction manual) the Powerlab 8 is perfectly able to do a 40A discharge and Steve tells me he has done many hundreds of 40A tests using his own Powerlab 8,And although he did warn me that seeing as I was intending to test the whole 20+KW pack using this unit and that they are not really designed for such extensive repetitive use I would need to watch the connections carefully for any signs of localised heating around the connections, especially the banana plugs.
Fine I thought as I began my very first 40A discharge and graph, first of all everything seemed fine and the 40A discharge graph began to appear as expected on the laptop screen, then within a second or two of starting the test a bad burning smell appeared and I frantically searched the connections for the tell-tale signs of heating but in less than a second there was a loud hissing and steam began gushing out of the rear fan of the Powerlab, power was disconnected less than 1 second later, I did not wait for the bang!
On closer examination I could see the blown electrolytic capacitor from the rear of the Powerlab 8 its tin can blown up like a balloon!
So bugger again another failed bit of kit, I know I am pushing the boundaries here but a 40A discharge is what this Powerlab device is rated at and so should have easily managed this test without failing, after speaking with Steve he feels that the unit must have been faulty and advised me to send it back for a complete replacement!
I know from my electronics experience if an electronic product is going to fail it will fail in the first 3 months of regular use and in my line of work (IT support) you see it time and time again around 3 in a hundred new computers will have a component fail in the first 3 months, some are critical and kill the PC most of the time though they go unnoticed until the faulty component is next used it might be a CD drive or the hard disk, audio or video but guaranteed you will find fault if you dig deep enough!
So today much to my annoyance I have packed up the Revolectrix Powerlab 8 a second time (They sent me a Powerlab 6 by mistake the first time I ordered! Still Anthea from Revolectrix UK called me to apologise personally and had the correct one out within a day or two so I can’t complain about the service it was second to none, Thanks Anthea)
So I wait for my replacement Powerlab and I am now fearful of doing 40A discharges despite Steve’s reassurance I don’t fancy blowing up another Powerlab for any reason even if they are supposed to do a 40A discharge.
I may not have any choice however so I may just get Steve in to double check that I am not doing anything daft before I attempt another 40A test, The one and only 20A test of the EIG cells I did manage showed extreme promise however as it took a whole hour to drop down to 3.8V from 4.15v and the low voltage discharge cut off for these cells is 3V (although I may set my cut off a little higher than this 3.3v for example) so it should take around 2 hours to do a 20A discharge.
Hence the need for the 40A discharge as I have 350 of these cells to test and at 2 hours for 2 cells (I am testing these in a 2S1P configuration) we are talking about 175 hours of test time never mind the battery swapping and connecting etc.

Monday, 8 September 2014

Brighton Mini Maker Fair (BMMF) 2014

Phew that was one hell of a busy day on Saturday, I spent the whole day from around 9:00AM till 6:00PM jabbering like a crazy man about the wonders of Electric Vehicles had a lot of interest and some serious contacts to so well worth the effort, I also showed the 3D printers again this year but was unable to run them as we were outside and the printers don’t like draughts.

We were so busy I did not even go inside the fair once and below is the only photo we managed to take.

It was absolutely heaving inside, thankfully we were outside and enjoying the good weather at least, also someone mentioned that the Brighton Marina Drive speed trials were on the same day! Whose idea was it to do these events on the same day, Bummer I would love to have done the trials in the EV, maybe next year.

Cycle Analyst finally fitted!

So seeing as the BMMF is drawing in very close now my maker genes are kicking into overdrive and last night I finally stole the time from my family to fit the Cycle Analyst into the “ElectroBeat”
I had previously prepared for the fitting by threading three wires through the bulkhead behind the seats and routing them into the central trunk where the brake cable electric windows & gear stick are located, Once I had it all apart and had a look where I left off it was a fairly simple matter of connecting one of the three wires to the –ve side of the shunt (already fitted), The –ve side of the shunt is connected to the Traction Packs –ve terminal the +ve side of the shunt is connected to the –ve that goes to the Soliton Jr controller, The Shunt goes inline between the controllers –ve terminal and the batterys –ve terminal, The only other connection (the third one) goes to the +ve side of the Traction Pack, I connected it to the controller side of the HV contactor so that it is switched off when the ignition is turned off.

A Shunt is just a very accurate high current resistor that allows for exact measurement of current flow through it by the Cycle Analyst or another current (Amp) counter.

Even though this was a fairly simple fit it still took me 2 ½ hours to fit what with the fiddling with crimps (making sure the crimps were good) carefully routing the cables protecting the 300v +ve signal that is routed to the dashboard (an obvious safety concern) and redoing the shunt connections when I realised I had routed them the wrong way out of the heat shrink sleeve that I slid back to expose the shunt (I have not shrunk this as yet as I was awaiting the fitting of this Cycle Analyst thereby completing the shunt connections) also whilst I was redoing the shunt connections the right way I broke the screw that connects the signal wire to one side of the shunt and I note that they seem to be deliberately placed as close to the resistive junction as possible, Seeing as I broke the screw off in the hole that is placed close to the junction I had no choice but to put that sides signal wire onto the 50A battery connection, I may have to redo this again and remove the shunt for drilling out so I can fit the signal wire where it is supposed to be mounted closest to the resistive junction.

Still once it was all back together and safely zip tied in place (yes the meter is held in with zip ties at the moment I am currently 3D printing a dashboard mount for it.)  And all the cables were safely zip tied to the (immovable parts of the) brake cables and electric window cables I put the covers back on just in time as darkness fell at around 8:30 (Winter is on its way!) fired up the ignition and was greeted with the lovely site of the Cycle Analyst V2.23 message (I had previously tried to fit the V3 Cycle Analyst but as it turned out the design had changed so much that they had connected the –ve Battery terminal to the Chassis! A big NO-NO in High Voltage EV’s but then this Cycle analyst is designed for e-bikes that don’t generally require 200+ volts!  

BMMF 2014 is nearly here!

I have been trying to concentrate on a second 3D Printer build so that I have something new to show at the BMMF but it looks like I will not be able to complete this before next Saturday now so I will probably just concentrate on the EV and just see how it pans out with the printer, reason being as it depends on whether I have to leave the car outside as if so I won’t be able to have a stall inside and 3D printers don’t like draughts so the printer may well have to be just static this year, Bummer ;-(
Still it should be fun and that is really what it is all about, also I had planned to concentrate on the car this year anyhow so I will probably just do that.

Planning the RX8 build

The first Monday of my hols I visited the Mechanics to find out when they could get the RX8 in for its strip down and motor fitting, it is booked for mid-September so I paid Steve for the Kostov 11 inch and Soliton 1 controller and advised Brent of the need to get Steve’s MX5 in for the motor & controller removal, Steve has assured me that he will cover any additional “removal” costs although I might well take some other parts from his MX5 EV for my RX8 EV.

So it’s chocks away for the start of the RX8 build and I am seriously looking forward to it, The battery is bought and paid for as is the motor & controller, I have been planning the battery compartment layout it will be laid out in rows of 5 (5S) and blocks of 5 (5P) I will arrange the 70S in 14 rows of 5 starting in one corner of the battery box and going to the other side up or down a row and back again, continuing on until I reach the other corner, The reason for this configuration is to simplify the buss bar manufacture as it will mean that all I will need to do to fabricate the buss bars is cut (or get cut) some 3 mm copper plate into squares and rectangles in order to join all the +ve’s and –ve’s correctly no special shaping will be required as I can modify the plastic cassettes by removing some small plastic tabs that currently separate the cells +ve’s and –ve’s bridge them with a plate that joins 5 +ve’s together in parallel same with the –ve’s and then make the bar wide enough to reach over to the next 5 +ve’s in that row, when I reach the end of a row of 5 I will make a buss bar that reached up or down to the next row and orientate that row in the other direction so I can continue the series arrangement of the 70S string.

I will knock up a diagram to explain this, I will also be modifying the cassettes in another way as Steve suggests that the latest advice from the manufacturers themselves is to hold the battery pouches in compression, not to actually compress them but to avoid any potential swelling from separating the layers in the cell pouches, Steve suggested putting balloons at each end to put them under a little pressure but I think I may just go for a tight fit in the boxes as in the Honda.

Steve has also suggested I should lose the aluminium separators on the cassettes and replace them with a sheet of silicone between every cell, I would be inclined to agree with this good advice as the cell pouches are made from Mylar which is itself a conductor and aluminium is also a conductor so the whole battery bank is only very slightly protected from shorting by the small air gaps between cell pouches, a sheet of silicone as a separator would provide both an insulator between cells and also a method of holding the whole cell block in compression.

The only issue I have wit this current layout plan is that the cassettes seem to support the pouches vertically and the blocks of cells are arranged with the tabs at the top so other than the supporting plastic cassette the cell pouches would be hanging from there tabs.

I was planning on cutting off the cassette frame completely just retaining the top piece that provides support for the screw terminals and allows the top tab parts to be bolted in blocks and separate the cells using silicone sheet, however this would appear to mean that the cells would literally be hanging by their tabs (not good) so the obvious answer would be to arrange the battery pack so that the pouches lay flat rather than tabs at the top, this will complicate the wiring arrangement but should not cause a major problem, I will try and get photos and drawings of the proposed arrangement uploaded soon.