Headway Pack Builders thread

I got some more time to work on the pack this weekend. I also had some time to go pickup some parts I had set aside at a local scrap yard. I got really lucky and found some 175 AMP Anderson power pole connectors from some scrapped battery backup systems. They were significantly cheaper used than they would have been if purchased new. Now I've got some safely insulated disconnects for my battery pack to use!

View attachment 5

The AMP connecters were rated for 175A and they were expected to carry both supply and return current on different pins. I'm going to wire them so that both pins are in parallel so the "rated" current should be closer to 350A instead of 175. In the pictures below you'll see how I've get them connected to the pack (at least how they are connected for now).

We left off at having the pack assembled but no cross bars, molex connectors or bus bars attached. Those were my first goals for the weekend. I first assembled the cross-bars. Here's a shot of the cross bars attached to the non-bus-bar side of the pack. My cross bars are cut from copper pipe, hammered flat, machined clean, drilled on a CNC drill press and then sanded to de-burr the edges. This is similar to how another thread on this forum suggests how to get cheap copper interconnects.

View attachment 4

Here's the side where the high-current bus bars will go.

View attachment endCrossBarsAttached-1.JPG

Once the cross bars were attached the next order of business was to connect the Molex connectors. I'm using 5-pin 0.062" molex connectors to connect the balancing taps from my BMS boards to the batteries. The "female" side of the molex connectors are on the battery side and the male side of the Molex connectors are on the BMS side. This way I can place extensions in the connectors if I want to (in case the BMS ends up not being co-located with the battery pack). The wire gauge is more than enough to handle the current of the shunt resistors in the BMS.

View attachment 2

From there I went in to attach the bus-bars on the most positive and most-negative ends of the pack. I'm using 0.75" x 0.25" bus bar with 1/4" holes drilled to access the mounts to the cells. The arrangement I have holding the bus bars on will have to change soon. The set-screw lengths are barely long enough to make it through the 1/4" bus bar to the nut. I'll have to replace these with 3/4" or 1" M6 bolts. I'll see if I can track down some brass or brass-plated bolts to make the connections even better.

View attachment busBarsAndCrossBarsConnected.JPG

Once the bus bars were connected it was time to play with connecting the AMP connectors. Here's the AMP connectors mounted on the bus bars.

View attachment busBarsAndAmpConnectorsAttached.JPG

That's it for now.. up next: Hooking up the G/F BMS and Charging! I've already hooked it up; don't have time to write about it now, but here's a link to a you tube video showing the setup.

[youtube]6tN9aSJwx0U[/youtube]

Cheers,
--Adam
 
This is shaping up to a very clean build!

Are you making a blog of your progress with the vehicle conversion? It is nice to look back on.

-JD
 
Otanet, I was making a blog of the conversion... it's just fallen behind. I loved the thread you did on the VW bus - it helped me out a ton and I think I should do the same so that others can use my experiences as a reference too. I started off writing about it on my own WordPress site at http://flashbone.serveblog.net/wordpress/. I've found that it is really hard to keep my own blog up to date. Plus, most of the reader's eyes and people who would be interested in it end up at either the diyelectriccar.com, evalbum.com, or endless-sphere.com forums. It seemed silly to try to keep my own blog updated with entries for those sites.

I guess right now my build documentation is spread across 4 different sites: ES, DiyEC, my blog and EV Album (http://evalbum.com/2619). I need to consolidate all the posts into a single build-tracking thread. That would probably be helpful for people doing lithium EVs and converting Cabiolets (although I'm far from the first Cabrio converter). I do have all of the photographs from my conversion logged at http://flashbone.serveblog.net/cabrioletPictures/.

I intend to take those photographs and make an end-to-end build thread somewhere that documents my progress. I'm not sure if endless-sphere or DiyElectricCar would be a better forum for that content. So far I've gotten more positive feedback from ES but ES is more bicycle centric where DiyEC is more car centric. I'm sure either site would be good to start a build thread - I just haven't decided which one wastes the fewest people's time! :) I might be the first Car-EV that's using the G/F BMS, but I could be wrong there. I've seen lots of e-bike and e-motorcycle conversions using it but not many e-car conversions mentioning it. Any other e-car G/F users out there?


Back to the build - I spent the afternoon at the Houston Dyanmos MLS soccer playoff game. Good fun, but it prevented me from doing much further work on the pack. I have learned a few more things, however. In light of the BMS shenanigans in this thread earlier, this might be more appropriate for the "Even Newer..." BMS board thread, but it's getting posted here first. If people complain, I'll move it - but it's definitely part of my HW pack build! If anyone else considers the Zivan charger this info might be useful. I'll start with a little background...

I'm using a Goodrum/Fechter 24 channel BMS board, version 2.2, outfitted with the higher current shunt resistors (500 ma, I believe?). I am using the board with a Zivan NG-3 charger re-programmed by Zivan USA to support a 24S LiFePO4 pack. The charger is set to do a 17 amp CC/CV up to 86.3 and then to trickle charge up to 91.2 volts. This should be just enough to handle latching voltage of 24*3.7 = ~88.8 volts of the G/F BMS board. The 17 amps charge rate should be about 1/3 C for my 50 AHr pack - which is a reasonably conservative charging rate.

The G/F BMS board starts interrupting the current to the pack through its FET when the first set of cells reaches its ~3.65 volt high-voltage-cutoff. When this happens the LED starts to flicker for that "channel" and the FET interrupts current from the charger. When the Zivan senses this current interrupt it trips into some high spike alarm mode and, flashes red on its indicator and plays a minor-tone alarm on its buzzers. I found that there needs to be a buffer between the Zivan and the G/F BMS board. I placed a 250V, 3300mF capacitor (another scrap-yard special) in between the Zivan and the G/F BMS and this seems to have solved the problem. With the capacitor the BMS choking charging current no longer trips the Zivan's alarm.

Once the Zivan moves past the 86.3 volt threshold it starts "trickle charging" the pack. This can be heard, audibly, because the Zivan slows or sometimes outright turns off it's cooling fans. At this points you can see the main LED on the G/F BMS board flicker from red to green while the channel LEDs start to turn orange. The pulse frequency coming out of the Zivan is initially slow. You can see the frequency in the LED... I am sure it is measured in 10s of hertz.

The Zivan has a 30-minute or so time out after switching to trickle mode. After that it flags the pack as full, stops the charge, and kicks off it's green "done" LED. When this happens the G/F BMS still has orange channel LEDs and a RED main LED. I haven't had the G/F BMS do it's latching thing yet - where all 24 channel LEDs turn on and the main LED goes green. This is presumably because the "weaker" or "more empty" channels haven't caught up yet.

At this point I'm on my third trickle-charge cycle trying to get the weaker channels caught up. There are 6 more channels that still have yet turn on their shunt signals to the G/F BMS - I'll post an update if/when I can finally get the weaker channels caught up and observe the "latch off" that the G/F BMS is supposed to have.


Cheers,
--Adam
 
I guess that's a good point! I'll have to replace the bolts on the bus bars anyway. I'll move the connectors when I have more space on the bolts to hold them to the bus bars. Good call on the current distribution.

Also you don't need the paralleling strips under the bus bars.
Click to expand...

Another good point. Since they're already in place I'm going to leave them. I don't want to have to snip them off right now. On the next pack I'll be smart enough to simply patch the voltage sensor connector over to the bus bar.

Cheers,
--Adam
 
I would definitely add another couple of planes of orange cell holders. As it is NVH, (noise,harmonic,vibration) in a vehicle will bite you in the ass.

I know PITA. :?

Also consider a logitudinal plate or rails stiff enough to prevent any sagging/flexing of the battery pack logitudinally, though a battery box with a stiff bottom will work also.

http://wikileaks.org/leak/toyota-prius-a123-car-fire-investigation-report-2008.pdfis well worth the read on page 22-23 regarding the need to properly support cells, and prevent bad /loose connections at all cost.

Safety is a biggy in big packs. :wink:
 
Brief update.

Once all the cell sets were balanced the G/F BMS correctly latches, stops the charging and turns its main LED green. It's pretty neat to see all 24 orange LEDs start to go solid, then the main board lock in a Green LED, and then the orange LEDs burn off back down to dark. Very cool setup.

I'll post a video later on this week if I have time.

Cheers,
--Adam
 
Since the pack is assembled and charges so nicely I figured I should show it actually driving something. Here's a video of the bench test of the pack.

http://www.youtube.com/watch?v=-TEckyUPwOw
[youtube]-TEckyUPwOw[/youtube]

I'll end up rebuilding the pack to address some of the suggestions added here. :p

Here's the outline of what's left to do when rebuilding the pack:

- Replace the steel/iron set M6 screws with Brass, if I can find the brass set screws.
- Replace the end-bolts and screws with Brass (1/2" for the cross bars and 1" for the bus bars).
- Add the plastic cell caps between layers 1-2 and 3-4 so every intersection of cells is supported by more than just the set screws.
- Remove the balance taps from under the bus bars.

I'll also be building a second one of these - to bring my car up to ~144V x 50 AHrs. I went ahead and ordered the mounting plates to finish out this pack and to build out the second pack. I'll need to get the set screws and interconnects for the cells. Oh, that and I need to save up another couple grand for the next 120 cells. Sigh. At least the car will be rolling soon!

Cheers,
--Adam
 
Looking at the pictures again, are you bringing all of the charging leads out the same side, or are you alternating them? If they are on the same side, the cell closest to the leads has the lowest resistance because current has very little "charging tap" to travel through, and the the cell furthest from the leads has the most resistance because current has to travel through most of both "charging taps". However, if you alternate the leads, then current travels through the same amount of 'charging tap' for every cell, so they all have roughly equivalent resistance.

Look at the picture of my sloppy pack - while I bring all charging leads off the left side, notice how some connect directly on the left side, but every alternating row other row has a long red wire connected to the right side. The leftmost cell has the shortest run over the top charging tap, but the longest run on the bottom charging tap. The rightmost cell has the longest run from the top charging tap, but the shortest run from the bottom charging tap. As a result current has to run through the same amount of material for each cell.

I maintain this alternation all the way through, including the Positve power takeoff and the Negative power takeoff, so no matter what path current travels through the pack, it should be roughly the same resistance as any other path. If I correctly understand Doc's cross-current concerns, this alternation should prevent them.

-JD

file.php
 
I do have all the wire leads coming off the same side. I am fairly certain this is okay for my pack because the leads in my case are balancing, sensing, and shunting leads and not charging leads. If I recall correctly in your setup otanet you are using individual DC/DC converters to drive charging current down each junction between cells. That's why your taps have individual fuses and why they are a relatively high gauge of wire. In this case, since high current is traversing the balancing bar between the cells I understand why interleaving them would make a bit of a difference.

In my setup things are different. The full charging voltage and current are dropped across the entire pack by the Zivan charger. So the only high-current connects come from the bus bars at the most positive and most negative ends of the pack. The balancing and sensing leads (or "taps" or "tabs" as I call them) do not carry full charging current (a max of 18 amps in my case). They only carry up to the 0.5 to 1 Amp max that my BMS can shunt during charging. The leads are otherwise used to (a) make a small capacity but relatively low resistance parallel bus bar between cells and (b) to probe the voltage of that parallel set of cells for the BMS. That's why my balancing wires are relatively small gauge (16 AWG) wire.

Since they carry so little current I'm not worried about the resistance as much. 1 foot of 16 AWG copper wire has a resistance of .00473 ohms. The wire length between cell 1, where the sensing/balancing wire attaches back to the BMS, and cell 5 that is on the other side of the set of copper tabs is about 8 inches max. That ends up being (0.666 * 0.00473) = 0.00315 ohms difference between the furthest cell from the BMS and the nearest cell to the BMS. 3 milli-ohms on a wire that will carry only 1 AMP. I'm not worried about it. :)

Cheers,
--Adam
 
Hi Folks,

I've uploaded a few videos to you tube of my car's first day on the road (test runs really). I don't want to junk up this thread with notes about my conversion, but here is a video that shows the pack driving the car on the road. 3 passengers, ~20 to 25 mph and hooked up with only two pairs of 8 awg wire to the controller! :shock: My 2/0 wire is still being shipped in and this was the best we could rig up to get a proper test done. The headway cells took the load with almost no heating up. We put 9.7 miles on the car today "breaking in" the pack and generally geeking out over the whole bucket of bolts actually working.

[youtube]66x3KVzl-Ak[/youtube]

Cheers,
--Adam
 
It's alive! :D

Congrats Adam!

-JD
 
Thanks, Otanet! :D

I am just now realizing that I thought I was obsessed before. Now that the project rolls the obsession only gets worse. :p

Cheers,
--Adam
 
I am running into some issues with my Freshly built Headway pack.

My Setup:
72v 10 amp/hr using Headway cells and headway 72v BMS.

My Problem:
I had the bike up and running in the shop. Everything was working well/tons of power. I strapped everything together/ tied up loose ends. Now the bike won't run. My battery reads 76.4 volts form +- terminals and 73.3 volts through the BMS. I am also reading 73.3 volts when connecting the multi between the aluminum heatsink on BMS and battery +. Is this normal.

I have looked everything over with no obvious problems. This setup is going on a Giant DH Team with regen. Can't wait :twisted:

What would a custom Headway master reccomend?

Thanks

Mike
 
mikebikerad said:
I am running into some issues with my Freshly built Headway pack.

My Setup:
72v 10 amp/hr using Headway cells and headway 72v BMS.

Now the bike won't run. My battery reads 76.4 volts form +- terminals and 73.3 volts through the BMS. I am also reading 73.3 volts when connecting the multi between the aluminum heatsink on BMS and battery +. Is this normal.
Click to expand...

Same cells, same BMS. I've had two cases where I had confusing voltages. The first was a loose connection where some solder broke from flexing. This gave me a full charge reading on my charge points, but when i switched on the controller my cycle analyst was a little lower. Easy fix.

The second was a misunderstanding on my part. I was reading the voltages on the flat pads underneath the BMS (mine is mounted upside down). Then when I read the voltages from the left most pad to the right most, it was 3.2v lower. Eventually I figured out that the positive most point doesn't have a pad which was throwing my reading off by exactly one cells amount due to only reading the sum of 23 cells. Make sure you're not doing the same.

The last thing I would check on your bike is whether all your cells balance leads are reading the correct voltage. If one is not connected it'll trip the BMS.

I've had 6 pages of back and forth in my headway BMS thread. Eventually Victoria sent me a new BMS and since then everything is perfect. The 72v headway BMS can apparently do 100 amps continuous and 400 peak so you're never likely to hit that limit. It also balances my cells so well that my multimeter can't pick up a difference after a run.

Edit: Also I've found that once my BMS trips I need to disconnect it from the controller. It might need to drop down to very low amp draw to get it to reset. I have a circuit breaker as a switch so I just trip that for about 30 seconds. Oh and what happens when you charge the cells? 76.4v is quite low for these cells, give them a charge and see if that helps...
 
News Update:

One of my cells is dead (0 volts). I tested all cells before building pack and they were all good v's. I discharged the pack maybe 1-2% while testing everything so I am surprised by this. Is this obviously a bad cell, or could the BMS be the culprit?

I am happy that I found a problem but it is not an easy fix :x .

thanks, Mike
 
You sure it's the cell and not the balance lead that's damaged? Did you check it at the cell itself?

Hopefully you have some spares available, otherwise order one quickly.

I doubt the BMS would drain a cell that quickly. What voltage were they at when you put them in the pack?
 
had you ever previously charged that cell? on a single cell charger or in a pack? the zero volts was after charging it up? where was it located? #1 or the other end? i let one go to .80V and bill let one go to 1.35V but they all charged back up with a single cell charger. still holding that charge.
 
The cell is in the middle of the pack. I charged it in the pack but without the BMS. BMS won't allow charging because of bad cell. I don't have a single cell charger. Now the cell has dropped to o.1 volts after sitting all day. I think it is cooked
 
if you can find a scrap cell phone charger, they will put out enuff voltage, so you can at least put current into it before trashing it. or you can use one of the other cells, or several of the other cells to charge it by connecting them in parallel. the BMS would not allow this to get so low so something is leaking the current off or maybe it shorted on something for a while. you can measure the voltage on that cell while it is charging too.
 
You must log in or register to reply here.
Back
Top