Weber & Coulomb's LyteFyba BMS

How do you store and manage your electricity?
User avatar
weber
Site Admin
Posts: 2252
Joined: Fri, 23 Jan 2009, 17:27
Real Name: Dave Keenan
Location: Brisbane
Contact:

Weber & Coulomb's LyteFyba BMS

Post by weber » Sun, 08 Feb 2015, 18:30

1. The heat generated by the CMU during bypass is insignificant. 0.8 amps at 3.5 volts is 2.8 watts. A typical 180 Ah cell with 1 milliohm internal resistance will generate 2.9 watts internally when supplying only 54 amps (0.3 C) -- quite a low rate of discharge. But the CMU is on top of the cell, not inside it, so the vast majority of its heat is convected upward, away from the cell.

2. Unless you fuse every one of those fine wires as it leaves a cell terminal, with a fuse capable of breaking the prospective fault current at the full battery voltage, you are, in my opinion, in far worse violation of AS3000 (Section 2.5 Protection against overcurrent) than you might be by including a PCB in a bolted connection using a belleville washer, particularly in a vehicle with a 720 V battery composed of 218 cells in series. Apart from the high current cables which are appropriately fused, and capable of withstanding a 1 second short circuit, nothing but optic fibre goes in or out of our battery boxes.
One of the fathers of MeXy the electric MX-5, along with Coulomb and Newton (Jeff Owen).

User avatar
jateureka
Noobie
Posts: 14
Joined: Thu, 19 Mar 2015, 12:04
Real Name: Jenny Theiss
Location: Brisbane

Weber & Coulomb's LyteFyba BMS

Post by jateureka » Tue, 24 Mar 2015, 20:06

Are these compatible for use with the Nissan Leaf battery modules also?
E-Bikes: FreeGo; eLation V2; Bafang 8FUN; MAC; Ezee
E-Motorcycles: ZEV 7100 LR; Zero S 2011
E-Mower: Ego push mower & line trimmer
E-Power: 3.3kW solar panels and Xantrex inveter

User avatar
weber
Site Admin
Posts: 2252
Joined: Fri, 23 Jan 2009, 17:27
Real Name: Dave Keenan
Location: Brisbane
Contact:

Weber & Coulomb's LyteFyba BMS

Post by weber » Tue, 24 Mar 2015, 20:46

Funny you should ask. We've just looked into that. It turns out that the Leaf module terminals are too close together to bolt our CMUs at both ends onto a single Leaf module. But if you are paralleling at least 3 of the modules then the CMUs fit on a "knight's move" diagonal (one across and two down) and bolt straight on.

Of course you need two CMUs per parallel group of leaf modules, since a leaf module has two series cells.

The 3.7 V as compared to 3.2 V for LiFePO4 is just a change of some constants at the start of the software. If you want to be able to measure voltages beyond the existing 4.095 volts, then it's a resistor change.

While I'm here, I should mention that we are working on the next version which may eliminate all 16 bypass resistors and replace them with a single 1.4 metre long PCB track that zig-zags through the area previously used as heatsink copper for the bypass resistors. It would be 0.2 mm (8 thou) wide and 35 um (1 oz/ft^2) thick.

We also plan to use thinner, higher-glass-temperature PCB material to minimise plastic creep under compression by terminal bolts. Not that we've found any evidence of it occurring yet, after more than a year in the MX-5.
Last edited by weber on Tue, 24 Mar 2015, 09:47, edited 1 time in total.
One of the fathers of MeXy the electric MX-5, along with Coulomb and Newton (Jeff Owen).

User avatar
Richo
Senior Member
Posts: 3043
Joined: Mon, 16 Jun 2008, 00:19
Real Name: Richard
Location: Perth, WA

Weber & Coulomb's LyteFyba BMS

Post by Richo » Tue, 24 Mar 2015, 20:53

tonyw wrote: I just have one niggle - I seem to recall that in Oz, at least, we are bound by AS3000 for all wiring.

Nah that doesn't seem right.
NCOP14 wrote: Before starting construction of an electric vehicle, some knowledge can be gained by reading Australian/New Zealand Standard AS/NZS 3000:2000
Seems you can get "some knowledge" from AS3000 but doesn't imply that these ARE the codes and standards for an ev conversion.

I suspect an "engineer" could over rule this in the design anyway.

User avatar
weber
Site Admin
Posts: 2252
Joined: Fri, 23 Jan 2009, 17:27
Real Name: Dave Keenan
Location: Brisbane
Contact:

Weber & Coulomb's LyteFyba BMS

Post by weber » Sun, 07 Jun 2015, 04:56

Coulomb and I have both been busy with other things lately, but Coulomb came over today and, among other things, we fitted some of our LyteFyba CMUs to a Lishen brand battery module belonging to a client. You'd think they were made for them.

Image

This is a 48 V (16S) 80 Ah (4P x 20 Ah) LiFePO4 module. The links are spot-welded to the terminals and rated for 100 A max continuous (25 A per 20 Ah cell).
Last edited by weber on Sun, 28 Jun 2015, 08:46, edited 1 time in total.
One of the fathers of MeXy the electric MX-5, along with Coulomb and Newton (Jeff Owen).

User avatar
jateureka
Noobie
Posts: 14
Joined: Thu, 19 Mar 2015, 12:04
Real Name: Jenny Theiss
Location: Brisbane

Weber & Coulomb's LyteFyba BMS

Post by jateureka » Sun, 07 Jun 2015, 14:35

looking good Image
E-Bikes: FreeGo; eLation V2; Bafang 8FUN; MAC; Ezee
E-Motorcycles: ZEV 7100 LR; Zero S 2011
E-Mower: Ego push mower & line trimmer
E-Power: 3.3kW solar panels and Xantrex inveter

User avatar
offgridQLD
Senior Member
Posts: 1783
Joined: Tue, 23 Jul 2013, 16:05
Real Name: Kurt
Location: Brisbane

Weber & Coulomb's LyteFyba BMS

Post by offgridQLD » Tue, 09 Jun 2015, 16:57

That's a nice solid looking cradle - cell enclosure. Sure looks like a perfect fit. Image

Kurt

User avatar
weber
Site Admin
Posts: 2252
Joined: Fri, 23 Jan 2009, 17:27
Real Name: Dave Keenan
Location: Brisbane
Contact:

Weber & Coulomb's LyteFyba BMS

Post by weber » Mon, 29 Jun 2015, 21:15

After our experience using the open source LyteFyba* BMS in the Black Monolith solar power system, Coulomb and I have been working on expanding the IMU (current and insulation monitoring unit) to become a full BMU (battery management (master) unit). * Note minor change of spelling, from Lite to Lyte.

We're using an MSP430G2553 microcontroller which has more resources, and giving it isolated control of up to four contactors, and isolated RS232 comms with both a charger (or inverter/charger) and a system control unit (e.g. a Raspberry Pi or Beagle-Bone Black for internet access and data logging). Optic fibre comms is optional.

We'd love your feedback on our low cost isolated RS232 design, shown in the first image below. There will be two of these on the BMU. The LM334 is a constant current device to allow it to work with input voltages from +-3 V to +- 25 V. That's a male DE9 connector on the left and the signals on the right are 3.3 V logic signals to and from the microcontroller.

Image Image

Edit:

The second image shows the isolated-current-loop comms with the CMUs, with optional optic fibre. This differs slightly from the implementation on the IMU and CMUs because the microcontroller on the BMU has a 3.3 V supply instead of 2.5 V, and there is also a 5 V supply. However I'm a little concerned that maybe I should not be pulling the fibre LED up to the 5 V supply when it is being pulled down by the 3.3 V micro. Its datasheet is here:
http://i-fiberoptics.com/pdf/if-e96edatasheet.pdf

Unfortunately the datasheet has no IV curve and so can't tell us whether the fibre LED is guaranteed to have less than 2 mA with 1.7 V (=5-3.3) across it.
Last edited by weber on Mon, 29 Jun 2015, 18:44, edited 1 time in total.
One of the fathers of MeXy the electric MX-5, along with Coulomb and Newton (Jeff Owen).

User avatar
weber
Site Admin
Posts: 2252
Joined: Fri, 23 Jan 2009, 17:27
Real Name: Dave Keenan
Location: Brisbane
Contact:

Weber & Coulomb's LyteFyba BMS

Post by weber » Mon, 29 Jun 2015, 22:16

Here's the isolated contactor-coil drive section for your questions and suggestions. The signals on the left are 3.3 V microcontroller logic. The terminals on the right will be plug-in screw terminals with 3.5 mm spacing.

The default use of the four contactors is intended to be:
PREcharge
BATtery
CharGe sources
DIScharge or DIScretionary loads

Image
One of the fathers of MeXy the electric MX-5, along with Coulomb and Newton (Jeff Owen).

User avatar
weber
Site Admin
Posts: 2252
Joined: Fri, 23 Jan 2009, 17:27
Real Name: Dave Keenan
Location: Brisbane
Contact:

Weber & Coulomb's LyteFyba BMS

Post by weber » Tue, 30 Jun 2015, 00:41

Here's the microcontroller, showing the proposed pin assignments. The four on the upper right port, whose first word ends in "V" for voltage, are all analog inputs. Pins marked Rly+ and Rly- are outputs that control the reed relays for the insulation (touch) test, whose results are measured by "TouchV".

Right-click and choose "View image" to see any of these schematics at higher resolution.

Image

The rest of the schematic is much the same as the IMU it is based on, which can be seen here:
viewtopic.php?title=weber-and-coulombs- ... 473#p48473

The differences are that the positive touch-test input (now called "Touch") is separated from the HazV+ input (now called "Batt"), and there is a second voltage-sensing input like the HazV+ input (called "Array"). And either of these voltage-sense inputs can be configured to read either positive or negative voltages.
Last edited by weber on Mon, 29 Jun 2015, 14:43, edited 1 time in total.
One of the fathers of MeXy the electric MX-5, along with Coulomb and Newton (Jeff Owen).

User avatar
weber
Site Admin
Posts: 2252
Joined: Fri, 23 Jan 2009, 17:27
Real Name: Dave Keenan
Location: Brisbane
Contact:

Weber & Coulomb's LyteFyba BMS

Post by weber » Tue, 30 Jun 2015, 02:23

Here's the analog section, as described in the previous post. That completes the schematics for this draft version of the BMU. As always, your comments are requested.

Image

Please also see the schematic for the isolated-current-loop comms with the CMUs, that I've added as an edit to the first BMU schematic post -- the one that shows the RS232 comms, here:
viewtopic.php?title=weber-coulombs-lyte ... 302#p57714
One of the fathers of MeXy the electric MX-5, along with Coulomb and Newton (Jeff Owen).

neilg
Groupie
Posts: 35
Joined: Sat, 07 Apr 2012, 06:24
Real Name: Neil Gibbs
Location: Perth

Weber & Coulomb's LyteFyba BMS

Post by neilg » Tue, 30 Jun 2015, 05:58

With respect to the low cost RS232 - can the Lm334 in series with the Opto LED take -25 volts? I believe most LEDs have a reverse voltage of about 5v. (haven't checked the specs of the EL3H7 or reverse charastics of the LM334).
With the relay drivers you could just use a Clare CPC1002 SSR - 2mA LED, 60v 700mA and eliminate the resistors and FETs.
Neilg

User avatar
weber
Site Admin
Posts: 2252
Joined: Fri, 23 Jan 2009, 17:27
Real Name: Dave Keenan
Location: Brisbane
Contact:

Weber & Coulomb's LyteFyba BMS

Post by weber » Tue, 30 Jun 2015, 06:43

Thanks neilg. The LM334 has Abs Max reverse voltage 20 V (with tens of microamps). The optocoupler LED has Abs Max reverse voltage 6 V. The LM334 has Abs Max forward voltage 30 V, power 200 mW.

But perhaps we should lower the spec to +-15 V which is the most any transmitter should produce anyway. Being isolated, we don't have to cope with ground offsets that can lead to higher voltages at receivers.

Do you think we'd still need a reverse diode across the LED?

Dunno why we didn't think of SSRs. However that particular one won't do the job, as we want to be able to drive EV200 contactors. They have a 3.8 A inrush for 130 ms before dropping back to 130 mA holding. Or GX14 contactors which are very similar (3.9 A for 75 ms then 230 mA). But I will look for suitable SSRs. Thanks.
One of the fathers of MeXy the electric MX-5, along with Coulomb and Newton (Jeff Owen).

neilg
Groupie
Posts: 35
Joined: Sat, 07 Apr 2012, 06:24
Real Name: Neil Gibbs
Location: Perth

Weber & Coulomb's LyteFyba BMS

Post by neilg » Wed, 01 Jul 2015, 05:44

The reverse current of the LED maybe similar to the leakage of the LM334 and keep the LED to <6v but I would err on the side of safety and put a reverse protection diode in. You wouldn't want it failing!

Neilg

PlanB
Senior Member
Posts: 465
Joined: Sat, 16 Jan 2010, 15:24
Real Name: Kris McLean
Location: Freemans reach
Contact:

Weber & Coulomb's LyteFyba BMS

Post by PlanB » Mon, 06 Jul 2015, 21:48

Have you ever thought about wireless Dave? A small RF or bluetooth module on board?

User avatar
weber
Site Admin
Posts: 2252
Joined: Fri, 23 Jan 2009, 17:27
Real Name: Dave Keenan
Location: Brisbane
Contact:

Weber & Coulomb's LyteFyba BMS

Post by weber » Mon, 06 Jul 2015, 23:43

Thanks neilg, I have added reverse protection diodes to the RX opto LEDs as you suggested. I also looked into suitable SSRs. It turns out that to get the grunt we need, it's about half the cost to go with separate optocouplers and MOSFETs rather than use the all-in-one SSRs.

Hi Kris, we will be revisiting the option of wireless communication between CMUs, but this would use infrared in a line-of-sight daisy chain. And because the master unit would not normally have line-of-sight, we would still use optic fibre to connect to the CMUs at the two ends of the chain, and between rows or shelves of cells. RF options just look too difficult to implement, too much PCB area, too expensive and too insecure at the moment.
Last edited by weber on Thu, 09 Jul 2015, 06:44, edited 1 time in total.
One of the fathers of MeXy the electric MX-5, along with Coulomb and Newton (Jeff Owen).

User avatar
weber
Site Admin
Posts: 2252
Joined: Fri, 23 Jan 2009, 17:27
Real Name: Dave Keenan
Location: Brisbane
Contact:

Weber & Coulomb's LyteFyba BMS

Post by weber » Sat, 25 Jul 2015, 03:54

It's been quite a slog finishing the PCB for the new LyteFybe multi-purpose BMU (Bms Master Unit or Battery Management Unit). The most enjoyable times were "pair programming" with Coulomb when he came over on a Saturday or Sunday. Coulomb also did lots of work on it remotely. I have finally sent it off to OSH Park to get 3 made.

I already realised a problem with it. No date or version on it. Ah well. It's only a prototype. There's bound to be something important that needs fixing.

The case it is designed to fit is a "black monolith", 30 x 80 x 140 mm. As you can see, it is fairly bristling with connectors, although, as with the CMUs, the fibre optic connectors on all three serial ports are optional. There are two male DE9 connectors, but most of the connectors are plug-in screw terminals.

Image

Now to determine precise component values and order the parts.

Edit: To see more detail, right click and choose "View Image".
Last edited by weber on Fri, 24 Jul 2015, 17:57, edited 1 time in total.
One of the fathers of MeXy the electric MX-5, along with Coulomb and Newton (Jeff Owen).

User avatar
weber
Site Admin
Posts: 2252
Joined: Fri, 23 Jan 2009, 17:27
Real Name: Dave Keenan
Location: Brisbane
Contact:

Weber & Coulomb's LyteFyba BMS

Post by weber » Sat, 25 Jul 2015, 04:12

Here's the schematic, in four parts. To see more detail, right click each one and choose "View Image".

Image
Image
Image
Image
Last edited by weber on Fri, 24 Jul 2015, 19:17, edited 1 time in total.
One of the fathers of MeXy the electric MX-5, along with Coulomb and Newton (Jeff Owen).

User avatar
coulomb
Site Admin
Posts: 3036
Joined: Thu, 22 Jan 2009, 20:32
Real Name: Mike Van Emmerik
Location: Brisbane
Contact:

Weber & Coulomb's LyteFyba BMS

Post by coulomb » Sat, 25 Jul 2015, 16:05

All you BMS builders that are keen to make one of these [ edit: before it's been tested Image ], be aware that SourceForge's SVN service has been down for around two weeks, so we can't check in the latest design files. I've been getting the latest files from Weber over email. The ETA for the restoration process is today (but US time, so later today), and they say it may take a few days to validate the data.
Last edited by coulomb on Sat, 25 Jul 2015, 07:04, edited 1 time in total.
Learning how to patch and repair PIP-4048 inverter-chargers and Elcon chargers.

User avatar
weber
Site Admin
Posts: 2252
Joined: Fri, 23 Jan 2009, 17:27
Real Name: Dave Keenan
Location: Brisbane
Contact:

Weber & Coulomb's LyteFyba BMS

Post by weber » Sat, 01 Aug 2015, 07:31

SourceForge did eventually restore access to our repository. But like many others, we have decided to desert SourceForge, like rats from a sinking ship, and move our project to GitHub. So long SourceForge and thanks for all the free hosting.

Here's the new GitHub project page.
http://dkeenan7.github.io/LyteFyba/
One of the fathers of MeXy the electric MX-5, along with Coulomb and Newton (Jeff Owen).

User avatar
weber
Site Admin
Posts: 2252
Joined: Fri, 23 Jan 2009, 17:27
Real Name: Dave Keenan
Location: Brisbane
Contact:

Weber & Coulomb's LyteFyba BMS

Post by weber » Thu, 13 Aug 2015, 03:17

Coulomb and I mostly-assembled one of the new BMUs on Saturday. I finished it off today when the last of the parts arrived, made the cutouts in the case and screwed the board in. There are a few clearances I will be improving for next time, but it did all go together.

The next job is to modify the Test-and-Calibrate software to work with the new processor and new pinout. Then to modify the IMU software that was written for the Black Monolith solar power system.

Image
One of the fathers of MeXy the electric MX-5, along with Coulomb and Newton (Jeff Owen).

User avatar
offgridQLD
Senior Member
Posts: 1783
Joined: Tue, 23 Jul 2013, 16:05
Real Name: Kurt
Location: Brisbane

Weber & Coulomb's LyteFyba BMS

Post by offgridQLD » Thu, 13 Aug 2015, 15:42

Looking good,
             Any chance some time (perhaps when you have more working samples) Of showing off a demo setup perhaps at one of the AEVA meetings. It Would be great to see it hooked up to a small bank that you could demonstrate it live under real conditions with a load and charger. Under different conditions. Balancing, live individual cell voltage, Stress and any other features or data it can report that perhaps could be triggered to show how it works.

I'm happy to donate a small - portable 48v bank made from 40ah calb cells and a connected PIP4048 if it would help for a demonstration setup.   

Last edited by offgridQLD on Thu, 13 Aug 2015, 05:45, edited 1 time in total.

User avatar
weber
Site Admin
Posts: 2252
Joined: Fri, 23 Jan 2009, 17:27
Real Name: Dave Keenan
Location: Brisbane
Contact:

Weber & Coulomb's LyteFyba BMS

Post by weber » Thu, 13 Aug 2015, 16:16

That's a great idea Kurt. Thanks.
One of the fathers of MeXy the electric MX-5, along with Coulomb and Newton (Jeff Owen).

User avatar
weber
Site Admin
Posts: 2252
Joined: Fri, 23 Jan 2009, 17:27
Real Name: Dave Keenan
Location: Brisbane
Contact:

Weber & Coulomb's LyteFyba BMS

Post by weber » Thu, 13 Aug 2015, 22:31

I've just finished writing a three page protocol manual for the LyteFyba BMS, for those who might want to talk to it from another computer such as a Tritium EV Driver Controls unit, a Raspberry Pi or a Beagle Bone Black.

LyteFyba_BMS_protocol.pdf

All suggestions or questions gratefully received.
One of the fathers of MeXy the electric MX-5, along with Coulomb and Newton (Jeff Owen).

User avatar
weber
Site Admin
Posts: 2252
Joined: Fri, 23 Jan 2009, 17:27
Real Name: Dave Keenan
Location: Brisbane
Contact:

Weber & Coulomb's LyteFyba BMS

Post by weber » Sun, 18 Oct 2015, 02:08

While testing the new BMU prototype a few weeks ago, I realised the reason the D9 serial ports weren't working was that the pinouts for the D9 connectors had been mirrored on the PCB layout. The library footprint that said it was for a male, was really for a female. It strangely fixed itself (with an attendant tangle of tracks that we then had to untangle) when we updated to a new version of DesignSpark.

Of course I had to fix the prototype by cutting tracks and running jumpers. But there was one track I needed to cut, that joined two pins of the connector, that was inaccessible because it was under the connector. So rather than desolder the 9 pin connectors to get at it, which is a slow and boring job, I decided to try something different.

Last edited by weber on Sat, 17 Oct 2015, 16:25, edited 1 time in total.
One of the fathers of MeXy the electric MX-5, along with Coulomb and Newton (Jeff Owen).

Post Reply