The T.I. bq78pl116 and bq78pl116EVM

[auraslip]

The T.I. bq78pl116 is a newish battery management IC from texas instruments.It has up to 16 channels and is user programmable via usb for compatibility with any lithium chemistry up 72v. It has user friendly software for configuring your pack, and has an output for a small led/lcd display. It's very powerful with features like monitoring internal resistance growth, cycle fade, and temperature of the cells.

bq78pl116 tech sheet.
bq78pl116 manual.

Of course, this is only an IC, so you'd have to build up a board. BUT, and this is exciting, they sell demo boards that are completely built up and functional for $99.

It is anticipated that users of this EVM may fall into two general categories:

• Those who wish to use the cell simulator (resistors and power supply) to explore the basic functionality
of the bq78PL116 measurement and control system. This can be referred to as the Quick-Start
approach.

• Those who wish to use a battery pack and treat the EVM as a possible interim design solution until a
custom PCB can be fabricated. This can be referred to as the Full-Featured approach.

And they mean it; any one can buy the EVM boards from their website or mouser. They also have an active support forum on their website, and any one is free to join. I like this sort of approach they are taking. I think I'll order one of their boards. I e-mailed about availability, and they should be stocked in a few weeks but you might be able to get them by contacting T.I. sales direct.

bq78pl116EVM user manual

 

[amberwolf]

There is a thread somewhere here on ES about that chipset, or another one in the same family. Might even be two threads; it or they were about building a BMS from it, but I don't recall the outcome.

 

[BatteryMooch]

That other thread is about the bq78PL114, the pin-for-pin predecessor of the bq78PL116. I've been using the '114 in a BMS and it's an amazing chip.

I've been playing with the bq78PL116 for a few weeks now and it's a nice step up from the '114. The bugs are pretty well gone from the firmware and you don't need to update the firmware to use the latest features (like you need to do with the '114).

I'm told that the evaluation boards and the chips should be available in a few weeks. A PCB isn't too bad to lay out but you will need to be careful with the balancing circuitry since it's essentially each channel is a small switcher (active charge-transfer balancing). The tough part is getting the hang of the over 150 settings in the firmware. The Technical Reference manual and the bqWizard software (run it in demo mode to check out the settings) will give you TONs of info on how to use the chip.

Be aware that if you're purchasing the evaluation module (EVM) for the '116 that you'll also need to buy the GPIO-to-USB adapter in order to communicate with the EVM. The adapter costs $49 IIRC.

 

[auraslip]

Very cool camlight. I bet these new BMS are great for discharging battery packs

Do you build battery packs with the ti boards, or do you just use them to build dischargers?

They specifically say these boards are for ebikes, I can't wait to see how people here can put these to use.

 

[BatteryMooch]

I designed the bq78pl114 BMS for a pack assembler who then connects a pack and sells the smart-battery to his clients. Other versions of the BMS are in the pipeline, from 3A to 600A current ratings. All I have to do it change the FETs and FET drivers to handle just about any voltage and current level. Oh, and the firmware settings too, of course.

The levels of protection you can set in the firmware are extraordinary. And they've never once failed. The short-circuit protection is my favorite because it's so much fun to demonstrate. Take a big pack (with 4AWG leads) and touch the leads together without a BMS....BIG sparks. Add the BMS and all you get is a <1mS baby spark (less than 60A peak). Looks like a tiny static spark.

 

[BatteryMooch]

The TI eStore shows that the bq78PL116EVM is in stock!

 

[auraslip]

Cam -

do you think this would be a good bms for the average endless sphere guy? I ask because @ $100 it's cheaper than the goodrum/fetcher and it's pre-built.

Throw in another $50 for the programming adapter and you got a bms that can handle any chemistry, and is user tunable for your preference of LVC/HVC. A lot of people spend a lot of money and time to take care of their batteries (especially lipo!), and this board seems like it could take a lot of the grunt work out of lipo maintenance as well as get rid of the need for balancing chargers/custom lvc.

I think I'd gladly pay $150-$300 if I could use it with lipo like I use a signal lab BMS with a ping!

 

[grindz145]

Awesome, hopefully they are not as twitchy as the 114 Eval boards were...

 

[BatteryMooch]

Awesome, hopefully they are not as twitchy as the 114 Eval boards were...

What kind of problems were you having?
While they were definitely not laid out for anything remotely resembling EMI immunity, I never had any problems with them.
But I never brought one out on an e-bike though.

 

[BatteryMooch]

Cam -

do you think this would be a good bms for the average endless sphere guy? I ask because @ $100 it's cheaper than the goodrum/fetcher and it's pre-built.

Throw in another $50 for the programming adapter and you got a bms that can handle any chemistry, and is user tunable for your preference of LVC/HVC. A lot of people spend a lot of money and time to take care of their batteries (especially lipo!), and this board seems like it could take a lot of the grunt work out of lipo maintenance as well as get rid of the need for balancing chargers/custom lvc.

I think I'd gladly pay $150-$300 if I could use it with lipo like I use a signal lab BMS with a ping!

The EVM board only handles 10A continuous so you'll have to work with that limitation.
And getting all those firmware parameters set can be a daunting challenge. It certainly can be done though, especially with help of the ES forum folks. It just might take a few trips back to the computer to tweak things.

You can download the bqWizard software and run it in demo mode to take a look at the available settings (Start > Run... > bqwizard /demo). Also, I highly recommend downloading the bq78PL116 Technical Reference Manual and looking at the available settings.

[Edit] Regarding the 10A limit for the EVM....
You could mod the unit to enable/disable your own set of FETs to raise the current rating.

 

[auraslip]

[Edit] Regarding the 10A limit for the EVM....
You could mod the unit to enable/disable your own set of FETs to raise the current rati

Exactly what I had in mind Use the board to switch some high powered relays

 

[grindz145]

I had an earlier version, but I consistently blew them up. It was in the transistors in the charge pump circuit. Not sure what the deal was. TI eventually admitted to having widespead problems with them.

 

[BatteryMooch]

I had an earlier version, but I consistently blew them up. It was in the transistors in the charge pump circuit. Not sure what the deal was. TI eventually admitted to having widespead problems with them.

Wow, that would have driven me crazy!
Sorry you had to deal with that...glad I didn't.

[Edit] Aha...I just remembered the EVM suddenly wasn't available when I wanted to order a couple. They said there was a new board rev being worked on and it took a few weeks. I guess that's when they fixed that problem.

 

[charithjperera]

I designed the bq78pl114 BMS for a pack assembler who then connects a pack and sells the smart-battery to his clients. Other versions of the BMS are in the pipeline, from 3A to 600A current ratings. All I have to do it change the FETs and FET drivers to handle just about any voltage and current level. Oh, and the firmware settings too, of course.

The levels of protection you can set in the firmware are extraordinary. And they've never once failed. The short-circuit protection is my favorite because it's so much fun to demonstrate. Take a big pack (with 4AWG leads) and touch the leads together without a BMS....BIG sparks. Add the BMS and all you get is a <1mS baby spark (less than 60A peak). Looks like a tiny static spark.

Hi John, what was the voltage and amp hours of this pack you shorted and what sort of FET's did you use for protection?
I was wondering if MOSFET based protection would work on a 120V, 40Ah pack for a dead short?

Thanks!

 

[BatteryMooch]

Hi John, what was the voltage and amp hours of this pack you shorted and what sort of FET's did you use for protection?
I was wondering if MOSFET based protection would work on a 120V, 40Ah pack for a dead short?

Thanks!

I used a 4S2P M1 A123 pack and five paralleled tiny TO-262-cased discharge FETs (same five FETs for the charge FETs). The protection, in my circuit, shut off the FETs in a bit over 300usec and limited the current to 57A max. Fewer FETs, or ones with smaller gate charge, would allow the protection to shut off the current even faster.

The pack is easily capable of three hundred amps when shorted so as long as the FETs work for my continuous and surge current requirements, they'll work for any short circuit. The current level during a short circuit is actually a LOT lower than the 150A-200A surges the pack sees during normal use.

There's nothing that's preventing MOSFETs from protecting a 120V, 40Ah pack during a short circuit. Or even a much higher voltage, larger capacity pack. I don't know how fast the current would rise in your circuit but, worst case, the FETs would turn off in a millisecond (assuming lots of big FETs) so you just have to make sure the FETs can each handle their share of the short circuit current for one mSec. As I mentioned earlier, the short circuit current for my pack easily exceeded 300A (probably a LOT more for the first few hundred mSec) but the protection limited the current to 57A. Probably due to the protection turning on in 100uSec and starting to limit the current. It then took another 200uSec to totally stop the current flow.

 

[neptronix]

I think I'd gladly pay $150-$300 if I could use it with lipo like I use a signal lab BMS with a ping!

Oh man, i think you're kinda nuts.

I think once you go over $100 for the BMS, the value it provides starts to erode rapidly.
I mean, here we are spending 1/2-1/3rd the money of the lipo packs themselves for protection.

And in reality you cannot protect from a lot of bad things that lipo does.
The big thing that worries me is the balance leads on these packs. Sometimes they are not soldered on correctly or have cold solder joints. A few of my packs exhibit this behavior. This would reak havoc on a BMS.

Nonetheless, good on you for trying to find a good solution. I think the root cause of many lipo problems, is the qualities batteries themselves. I think BMS malfunctions can be way more dangerous on a lipo pack ( say the BMS drains a few cells below LVC, which seems to be common ) because the end user expects everything to be peachy keen when they install the BMS.

Again, good luck.

 

[auraslip]

I didn't realize quality was such a problem. That certainly makes lipo much LESS attractive.

In any case the fact that you could use it with other chemistrys still makes it a valuable tool. One day you could be using it on your daily commuter with headways. On the weekend you could be using it with your lipo pack.

As far as the price: $150 is about what most people pay for there first lipo charging/lvc setup. And that only does 8s at a time @ like 5 amps. This could replace all that and take care of a whole 16s pack. That's the biggest attraction. Getting the low price and high power of LIPO in a worry free, "peachy keen" package. Some of us charge at school, work, or like last night, at some random house party.

But as you said, if lipo has some quality issues, than maybe it wouldn't be such a good idea. Even if I had to monitor the pack with some cell logs - I think it would be worth it. Just for the ability to fast charge and balance the whole pack ANYWHERE by just plugging in and flipping a switch. And also not having to rig a ghetto LVC by disassembling cell-logs.

In any case, LIPO is cheap compared to even headways: 2.5 wh/$ vs 1.5 wh/$ respectively. And then the c-rates...

I think you get understand already lol

Eventually I'll be building a multi-kw/h pack, and it just makes sense to use lipo. But maybe the QC will make it better to use headways or a123....

 

[neptronix]

See my charging simplification solution above.. .. 20 amp hours is good for 40 miles on my rig, maybe 30 miles with some major, major hills. You could do the same.

This fits in my center triangle, and i don't see myself going further than 20 miles round trip anywhere.
On a long road with few stops you get better miles per amp hour too.. my testing was in city miles mostly.

8s is not the highest charger you can get.. i have a 10s charger and it balances my whole 10s pack, and i only plug in the power main power cord 90% of the time and don't even need to balance. I paid $121 for my charger and $30 for my power supply so 10s is possible on a $150 budget.

10S on the MAC motor does up to 29mph, and 10-15mph up really steep hills, so 10s ain't a bad voltage..
You could even do 14S lipo and charge with a hyperion for balances, or modify a meanwell to do the opportunity.

You really don't need a BMS for HVC charging because of this.

Discharge down to 3.5v per cell to be safe and you won't need a BMS for LVC either.

I see how a BMS was important for so long. 10s isn't even 36v nominal on lifepo4 and 14s chargers just appeared a few months ago. There was really no good way to balance charge even a 36v pack until now.

 

[neptronix]

Oh and sorry to derail your thread. Good luck with finding a good BMS.

 

[auraslip]

mmmmm....... nice battery porn

What is the charge process like? Switch all the plugs around, power up charger, select settings, and walk away?

I'd like my $1000 battery pack to be drunk proof - or girlfriend proof for that matter

 

[neptronix]

I have serial and parallel harnesses that those will be plugged into.

So, octopus of wiring, with one plug to charge the entire thing.
When i want to balance charge, i plug the two 5s leads into my iCharger as well as the power plug.
Balance leads would be paralleled, thus balancing the voltages of all the cells in parallel..

The only reason i would have to disconnect anything is if i noticed that a cell group was getting low.
A celllog 8s on each 5s parallel connector would monitor the cell group voltage, and beep at me if anything goes too low.

Go search for the 'why i stopped worrying and learned to love the LIPO' thread

Also, the new 8AH lipos make this job easier. You can do a 24AH pack with 25% less wiring than i have. Or do a 16AH pack.

Even 8AH on it's own is good for about 10-15 miles depending on your route and power level.

Or even easier, just modify a brick charger for the job. My girlfriend's store bought eBike has these li-mn cells that go to 4.2v. If you got the 36v route, you could get one of those and use it for opportunity charging.

And a 48v brick charger would work for a 13S pack ( 48v ). A resistor could be used on one of those to charge to 1 less volt, to give you sort of a buffer against cell imbalance.

 

[BatteryMooch]

I didn't realize quality was such a problem. That certainly makes lipo much LESS attractive.

Whoa, there are only quality problems in low quality BMS'.

There's certainly nothing inherently of low quality or low robustness in the bq78PL116 chip. I have done hundreds and hundreds of cycles with these chips, with the EVM and my own PCBs. The only problem I ever had was early on in testing of my first board. I blew out a FET driver transistor when I was doing 150A pulse testing into a coil of wire. Adding the same spike protection I was using on the output of the BKMS to another location fixed the problem. That was my design's fault and I caught it in early testing.

Not once, EVER, has a single board failed to shut off when it should have for an under- or over-voltage, under- or over-temperature, over-current, or short-circuit condition. This includes high-amp discharges then simultaneous short-circuiting of the pack, charging 4S LiFe packs at 25V, discharging the pack over and over after it reached LVC (and recovered and turned the discharge FETs back on), etc.

If you use a quality BMS, properly designed for its intended use, you get exactly that...quality.
Use an inexpensive BMS, that cuts corners at every turn, and you get exactly that...a cheap BMS.

I would have absolutely no hesitation in using the bq78PL114 or bq78PL116 chip with LiPo's. In fact, my next BMS board for a client is for LiPo packs using the '116 and neither of us is worried.

 

[auraslip]

Very cool cam- I would love to see pictures of that set up, if your client doesn't mind of course


I think nep was referring to poor quality sense cables on the lipo packs themselves. Hard for a BMS to do it's job it it can't get a good reading on the cells.

 

[Alan B]

Even bad sense wires or connections should not deter a quality BMS. It should do the best it can. Stop charging, discharging and report the error.

One of the problems with many simple BMS designs is they have no safeties in their design. They work in such a way that a broken wire or bad connection gets ignored. Load or charging continues until disaster.

The BMS should pass the info to the user and shut things down.

Now the user may choose to override the BMS. That is something they might need to do.

I had a Honda 4 stroke outboard years ago that had an automatic oil shutdown. They had double the amount of oil required for engine lubrication, but if it got too low it would shut the engine off to protect it. Years later I got a newer Honda 4 stroke outboard, an update of the same model. The system had been changed. It still has double the oil needed, and there is still a low oil detector but now it sets off light and sound indicators. It is up to the captain to choose whether to continue operating the motor, possibly ruining it, or shut it down. One can imagine the wind and tide drawing the boat into the rocks and the captain's choices.

The user needs good information. That is the most important function of a BMS. Simple BMS's don't do that. They do the minimum to protect, and things like a broken wire fool them.

But are we willing to pay the price required for a proper BMS? What is that cost? My microprocessor based designs are looking like about fifty bucks for 6S. Most of the commercial ones I see are double that or more. Quite a bit of work to make them, calibrate and test them, and deal with customer support. Members of ES such as methods and camlight are making quality BMS's and not selling them on ES, probably due to the cost being too high for this market.

 

[grindz145]

For those that aren't aware, this is a development tool, not a commercial product. The purpose is for evaluating the TI product. For the cost, it's obvious that this is not something for the average ebike hobbiest. But it is possible that an astute ebike revolutionary might use a tool like this to develop something practical for the every-day ebiker... 8) Frankly for the price, you actually get quite a bit.