[dtbaker]

...been thinking a lot about what I WISH was available as far as a charger goes, and not being an EE, I want to just put the functional requirements out, and ask if anyone can put it together, or maybe already has?

first some assumptions/goals:
- as a mechanical DIY guy, I am not interested in huge amounts of flexibility, data recording, etc. I want to 'set it, and forget it'.
- primary goal is a safe, reliable on-board charger designed for LIFePO4 prismatic cells (TS, CALB, and the like).
- huge amount of flexibility is not required because once installed, I hope it lasts at least 10 years along with the batteries.
- minimum cost while meeting functional requirements is primary, but also important to minimize size, weight, and be sealed against dust, water durable with regard to cycle life, vibration, temperature, etc.

major functional specs:

- plug into 110v, or 220v, and self limit to no more than 15 amps so as not to blow the typical house breaker. I think the major slice of the DYI market is fine with a 6 to 10 hour charge time, and not about to install a dedicated high-amp line to the garage.

- have a built in interlock terminals that would pass thru the usual 13v key-on juice to the main contactor solenoid when not charging, and open the circuit when plugged in to prevent key-on closing the typical circuit. Eliminating the need for one external component.

- have multiple (perhaps 50) terminals to accept 24ga wire monitoring individual cells. one wire per cell, plus one. 50 would allow enough for typical packs up to 48 cells.

(this is the tricky part)
- have internal brains to scan thru all cell monitor voltages, and switch from CA mode to CV mode at a cell level as soon as any cell hits target voltage. When in CV cell mode, output amps would be ramped down as needed to hold cell(s) at or below target voltage. No real attempt made to shunt or balance, just monitor and ramp the charge down per the highest cell.

possible? is this kind of 'scanning' the problem, or is that not really very expensive to build?

The least expensive internal design would probably be to not even attempt to ramp down, but use a fixed target finish of something like 3.70v and just stop the charge as soon as the first cell hits the limit. One extra that would be nice if it could be done without too much cost would be user-settable finish voltage... perhaps pre-set to 3.70, but able to move up or down a little in .05v increments. Extras would be some external display of which cell hit the limit first with something as simple as an external LED. Slightly more complex might be internal settings to show 'yellow' when cell is within .05v of target, and then 'green' when at target. It would be a quick visual of balance to see if the first high cell is more than .05 higher than MANY cells, then it may need to be balanced down a little.

....so my question is whether there is already a simple foolproof relatively inexpensive charger like this I don't know about, or if ya'll think something like this would even appeal to the typical DYI as a minimum charger/monitor system at moderate cost.

I know I for one don't feel I need or want a super adjustable super smart charger with completely settable voltage rnges and data bus, nor do I want to spend $1000-$2000 for a BMS. What I really want is a simple charger guarenteed not to overcharge any cell with simple single-strand monitoring loom terminal block built in.

 

[Tahoe Tim]

+1

a true "consumer friendly" charger

 

[DIYguy]

No one wants to jump in ...yet... I think it's a great idea Dan. For those who don't want to run with a BMS... ya, others may throw a few stones. .. but there are definitely some advantages of what you suggest. It can definitely be built. I haven't seen anything like it. It's kind of like a "top-only" monitoring or CMS. (charge monitoring system)
I would suggest though...that perhaps the best way is to build a pony-box or separate box... and just send output to charger to switch modes. (like some do now with contactors). This way, you could use it with any charger.

Of course the charger features you speak of would have to be there...but I think these exist now...

Just my $.02

Sorry Tim...lol, u pulled the trigger b4 me...lol

 

[tomofreno]

so my question is whether there is already a simple foolproof relatively inexpensive charger like this I don't know about

Yes, A minibms (around $500 for your 38 cells) and a Zivan or Chinoz charger. And you don't need 50 connections. Just one wire between cell level boards and to the main board, and connection of the main board to 12VDC and to an AC relay to shut off power to the charger at an HVC event. Schematics are on Dimitri's website. Downside is you get no choice on the cell limit voltage - which would be nice I agree. Seems to me what you are describing is essentially a charger with built in bms.

 

[Franky.EV]

rwaudio, A guy in this forum who is converting a porsche, build a charger whith lot of DC/DC, charger design

 

[rwaudio]

...been thinking a lot about what I WISH was available as far as a charger goes, and not being an EE, I want to just put the functional requirements out, and ask if anyone can put it together, or maybe already has?

first some assumptions/goals:
- as a mechanical DIY guy, I am not interested in huge amounts of flexibility, data recording, etc. I want to 'set it, and forget it'.
- primary goal is a safe, reliable on-board charger designed for LIFePO4 prismatic cells (TS, CALB, and the like).
- huge amount of flexibility is not required because once installed, I hope it lasts at least 10 years along with the batteries.
- minimum cost while meeting functional requirements is primary, but also important to minimize size, weight, and be sealed against dust, water durable with regard to cycle life, vibration, temperature, etc.

major functional specs:

- plug into 110v, or 220v, and self limit to no more than 15 amps so as not to blow the typical house breaker. I think the major slice of the DYI market is fine with a 6 to 10 hour charge time, and not about to install a dedicated high-amp line to the garage.

- have a built in interlock terminals that would pass thru the usual 13v key-on juice to the main contactor solenoid when not charging, and open the circuit when plugged in to prevent key-on closing the typical circuit. Eliminating the need for one external component.

- have multiple (perhaps 50) terminals to accept 24ga wire monitoring individual cells. one wire per cell, plus one. 50 would allow enough for typical packs up to 48 cells.

(this is the tricky part)
- have internal brains to scan thru all cell monitor voltages, and switch from CA mode to CV mode at a cell level as soon as any cell hits target voltage. When in CV cell mode, output amps would be ramped down as needed to hold cell(s) at or below target voltage. No real attempt made to shunt or balance, just monitor and ramp the charge down per the highest cell.

possible? is this kind of 'scanning' the problem, or is that not really very expensive to build?

The least expensive internal design would probably be to not even attempt to ramp down, but use a fixed target finish of something like 3.70v and just stop the charge as soon as the first cell hits the limit. One extra that would be nice if it could be done without too much cost would be user-settable finish voltage... perhaps pre-set to 3.70, but able to move up or down a little in .05v increments. Extras would be some external display of which cell hit the limit first with something as simple as an external LED. Slightly more complex might be internal settings to show 'yellow' when cell is within .05v of target, and then 'green' when at target. It would be a quick visual of balance to see if the first high cell is more than .05 higher than MANY cells, then it may need to be balanced down a little.

....so my question is whether there is already a simple foolproof relatively inexpensive charger like this I don't know about, or if ya'll think something like this would even appeal to the typical DYI as a minimum charger/monitor system at moderate cost.

I know I for one don't feel I need or want a super adjustable super smart charger with completely settable voltage rnges and data bus, nor do I want to spend $1000-$2000 for a BMS. What I really want is a simple charger guarenteed not to overcharge any cell with simple single-strand monitoring loom terminal block built in.

My goals are the same as yours but my implimentation is a bit different, but I did cover the monitoring, indivdual current ramp, adjustable finish voltage etc. The only thing left to prove is reliability. If it works as well as it should I will make it available to others.
http://electricporsche.rwaudio.com/2010/11/electric-porsche-charger-design/
There is also the capability of SOC, high cell, low cell, avg cell, total pack voltage, current etc.

 

[dtbaker]

Yes, A minibms (around $500 for your 38 cells) and a Zivan or Chinoz charger. And you don't need 50 connections. Just one wire between cell level boards and to the main board, and connection of the main board to 12VDC and to an AC relay to shut off power to the charger at an HVC event. Schematics are on Dimitri's website. Downside is you get no choice on the cell limit voltage - which would be nice I agree. Seems to me what you are describing is essentially a charger with built in bms.

exactly... either a new simple charger ground up w/ integrated HVC cutoff, or a nice clean add-on black box with a simplistic cell limit stuff all inside that would be installed basically as an interlock or pass-thru on the AC power to the charger and kill it when the first cell hit limit. A settable cell limit would be nice, but not required if it makes it much more expensive... just some reasonable setpoint like 3.70 would be fine. Visual indication of which cell was first, and which ones were close would also be nice, but not required.

The hope I would have is that if you have a reliable cutoff based on the cell limit, the charger might be a lot simpler, really with no settings for different pack voltage or curves... assuming you are ok with a simple CA charge up to some limit, and off.

an extra $500 for a mini-BMS add-on is getting approachable, but still too expensive for my taste. My question is if it were simpler and more hard-coded, could it be sold for significantly less?


I suggested a set of 50 terminal blocks just to cover the likely config of up to one wire per 48 cells, plus one.

 

[Tesseract]

an extra $500 for a mini-BMS add-on is getting approachable, but still too expensive for my taste. My question is if it were simpler and more hard-coded, could it be sold for significantly less?

It is highly unlikely the miniBMS could be made any less expensive. Unless it were mass-produced with all SMT components, no "flying" ring terminal and/or "some assembly required" by the end user.

Your idea is to basically put a BMS inside the charger which, of course, makes the charger more expensive. How much more expensive? Oh, about the same as having a separate BMS module on each cell.

So, same difference.

Integrating the BMS and the charger makes sense - to me, anyway - if it is for single-cells or, at most, 4 cells in series.

Making a series pack charger "set and forget" is certainly doable. Making it so it draws a maximum of 15A from 120V and, say, 30A from 240V is also doable. Making it so it doesn't change setpoint when going from 120 to 240V is also doable (as well as not blowing up if accidentally disconnected from the battery pack... )

So, many of your goals/desires make perfect sense to me, just not integrating the BMS and charger in one box (as long as there is an input that shuts off the charger if a cell hits HVC).

 

[todayican]

Hmmm "Chargiton 1"?

 

[Elithion]

.. what I WISH ... charger... monitoring individual cells.

Chargers with built in Li-Ion BMS to protect from cell over-voltage do exist. The BMS runs only when the charger's BMS runs.

To protect against cell under-voltage you need a second BMS, one that runs when discharging. So, instead, you might as well have just a single BMS to protects cells from both under-voltage and over-voltage, and that runs both when charging and discharging.

Davide

 

[rwaudio]

Hmmm "Chargiton 1"?

Lol I was thinking the same thing. But lets start off with the Chargiton 0.01, no way to power a 300kw charger yet

 

[LiFe]

I agree with Tesseract -you'll just have leave out the CV/CC control based on cell level monitoring.


IMHO what a conversion builder wants is a comprehensive package that includes amp hour counting, cell level monitoring, easily configurable, easily installed, and can be integrated into an existing vehicle instrument panel (SOC gauge, temp, etc), along with an integrated 120v/240v charger that can accept most any source and not pop a breaker. And exhibits some tolerance to installer mistakes. Example; an optical coupled BMS that configures itself as a daisy chain repeater, easily identifying the problem link, besides complete rejection of EMF.

Bonus points awarded for a standardized I/O package to communicate with motor controllers. Such as current limiting based on SOC/pack temp, charge interlocking, etc.

What you see in the market now are engineering projects being passed off as engineered products.

They're produced by an individual, with mediocre hardware design discipline, or a bit twiddler who integrates what they believes are important features. Sometimes the features have nothing to do with market reality, and distract from a solid and validated system design.

Then when you include less-than-adequate social skills it often results in a somewhat alienated customer base.

I've watched these characters throw darts for awhile now.


Ready to belly up there Mr. Tesseract?

 

[dtbaker]

(Just don't call it a "BMS", because that term seems to have negative connotations in some circles.)

Its not the term 'BMS' that bothers me, its the cost....

I pretty think that the ignoring undervoltage issue is plain driver error. I think any active shut-down of the system versus a simple idiot light has pros and cons with unexpected limp mode, etc, and I would rather rely on SOME awareness of the SOC with a decent SOC meter that counts ah for the typical Li setup.

I am thinking a standalone box that just watches for HV on any input leads you plug into it, and is maybe a pass-thru for the AC power with a relay to the charger might be the simplest and cheapest. Many of the cheaper chargers don't have any way to shut them down externally, so perhaps just a relay on the AC input would 'fit' the most chargers and be an easy aftermarket.

How about this? Given these very modest design goals, what do you think?

 

[DavidDymaxion]

My goals are the same as yours but my implimentation is a bit different, but I did cover the monitoring, indivdual current ramp, adjustable finish voltage etc. The only thing left to prove is reliability. If it works as well as it should I will make it available to others.
http://electricporsche.rwaudio.com/2010/11/electric-porsche-charger-design/
There is also the capability of SOC, high cell, low cell, avg cell, total pack voltage, current etc.

Great stuff, thanks for posting that on a web page.

Lol I was thinking the same thing. But lets start off with the Chargiton 0.01, no way to power a 300kw charger yet

Except to tow a car with regen behind a truck or use a dump pack. That's still a lofty goal!

 

[tomofreno]

IMHO what a conversion builder wants is a comprehensive package that includes amp hour counting...

You left out one important point...they want all that for less than $1000.00

 

[dtbaker]

You left out one important point...they want all that for less than $1000.00

yes, like $500 would be good. Just as a market reference point and considering the cost of controllers, motors, and batteries..... $500 feels about right for a charger that would truly do the job in a worry-free way.

Just the basics of charging, and protecting against overcharge of any single cell would be great for the DIY person using large format cells, or at the string level for the smaller headway-ish cells.

Instrumentation, over-discharge warnings, ah in/out can all be handled by meters and displays.

 

[Tesseract]

What you see in the market now are engineering projects being passed off as engineered products.

That's a somewhat incendiary way of putting it, but I tend to agree with your characterization. There are two reasons for this state of affairs, however, which I feel it is only fair to point out: 1) very few people in the entire world convert their vehicles to electric every year so there isn't much volume for EV products; 2) most of those people are terminal cheapskates that balk at even paying even $10 per cell for a BMS.

Point 1 prevents there ever being the economies of scale to satisfy Point 2, hence why most (all?) BMS offerings are either science projects or priced like they were for the military.

Then when you include less-than-adequate social skills it often results in a somewhat alienated customer base.

I probably fit this description. That's just typical of engineers, I've noticed, but I, at least, don't feel particularly apologetic about it. Which is also typical of engineers...

Ready to belly up there Mr. Tesseract?

Yeah, a smart charger is in the pipeline but we aren't real interested in making a BMS, mainly because there is so little profit to offset so much liability, but also because I truly feel that Jack Rickard's active campaigning against using a BMS is depressing any further innovation here. I know I certainly don't feel like butting heads with him over this again.

...$500 feels about right for a charger that would truly do the job in a worry-free way...

Maybe if the annual sales volume were in the tens of thousands, which is definitely not the case now - something along the lines of 100's per year is more realistic. There simply aren't enough economies of scale to push pricing down from "boutique" to "commodity", and you are definitely thinking "commodity".

Even $1000 might be too low unless the power rating is very modest.

Also keep in mind that anything that connects to the AC mains is supposed to be UL and FCC/CE listed. Testing for compliance with these safety and EMC regulations can cost more than developing the product in the first place.

 

[dtbaker]

Also keep in mind that anything that connects to the AC mains is supposed to be UL and FCC/CE listed. Testing for compliance with these safety and EMC regulations can cost more than developing the product in the first place.

if we scale the project to a black box that has an AC plug to the wall, and an AC plug thru to the charger... and the relay inside that is connected to the AC is UL listed.... does the black box have to be UL as a whole new thing?

All the little stuff monitoring voltage, and deciding to cut the power is not 'really' connected to the AC, so I am wondering if it really needs UL?

I guess what I am looking for is the cheapest possible safety net so that if any one cell suddenly becomes unbalanced and goes ballistic, the charger shuts down. Worst case failure for the charge monitor is that you are no worse off than with just the charger... with no chance of the monitor affecting batteries other than turning off the charger.

 

[tomofreno]

yes, like $500 would be good.

Dan, look at the price of the cheapest Charger you can get from China which has none of the features you are asking for, then estimate the cost of those added components.

That's just typical of engineers, I've noticed, but I, at least, don't feel particularly apologetic about it. Which is also typical of engineers...

Speak for yourself Jeff.

 

[todayican]

Engineer = Someone who dreams up ways of doing things, then does them.
Ungineer = Someone who gets in the way of an Engineer with speculation and detractory