10kW / 60A DIY charger open source design - Page 2 - DIY Electric Car Forums
Go Back  

DIY Electric Car Forums > EV Conversions and Builds > Batteries and Charging

Register Blogs FAQ Members List Social Groups Calendar Search Today's Posts Mark Forums Read

Reply
 
Thread Tools Display Modes
  #11  
Old 06-08-2011, 09:16 AM
dtbaker's Avatar
dtbaker dtbaker is offline
Senior Member
 
Join Date: Jan 2008
Location: santa fe, nm
Posts: 3,999
dtbaker is on a distinguished road
Default Re: 10kW / 60A DIY charger open source design

Quote:
Originally Posted by ElectriCar View Post
...I have mine behind the seat of my S10 which barely clears my Zivan NG5.
wow, that must heat the cabin up nicely after a charge! good in the winter.
__________________
Dan
http://www.envirokarma.org/ev - '97 Suzuki Swift
http://www.envirokarma.org/ev2_mx5e - '94 Mazda MX5 Miata
Reply With Quote
Sponsored Links
Advertisement
 
  #12  
Old 06-08-2011, 10:46 AM
ElectriCar's Avatar
ElectriCar ElectriCar is offline
Administrator
 
Join Date: Jun 2008
Location: Reidville, SC
Posts: 1,849
ElectriCar is on a distinguished road
Default Re: 10kW / 60A DIY charger open source design

dtb, I keep the windows open in the summer but yes in the winter it helps.
Reply With Quote
  #13  
Old 06-08-2011, 02:30 PM
valerun valerun is offline
Senior Member
 
Join Date: Nov 2010
Posts: 1,422
Blog Entries: 1
valerun is on a distinguished road
Default Re: 10kW / 60A DIY charger open source design

Quote:
Originally Posted by ElectriCar View Post
I have a question about the precision of the voltage at the shutdown point. Ie if it's set for 170V, how much can it deviate from 170V? I may be a candidate for the 10Kw in a kit form so I can get more use of my truck.

Another question, what are the dimensions of the unit as I have mine behind the seat of my S10 which barely clears my Zivan NG5.
Hi ElectriCar,

the voltage resolution is ~0.5-1V on the full pack voltage (10-bit ADC measuring the output of a hall sensor with ~5mV/V sensitivity to pack voltage). for a 170V pack, this will mean ~0.01-0.02V per cell. I normally set up the chargers for 3.5V CV cutoff for a few reasons:
1. Enables reliable end of charge detection for a bottom-balanced pack without driving any cell above 3.6V (for a pack with capacity matched to ~2% - which has been the case in all my dealings with CALB batteries).
2. Leaves 1-2% of capacity untapped at the top - which is good for battery longevity
3. Allows for any voltage sensor deviations due to limited resolution, unusual temperature profile, etc.

Dimensions are ~10x15x6". You will also need some space on the sides of the unit for connections and cooling.

Valery.
Reply With Quote
  #14  
Old 06-18-2011, 01:09 AM
valerun valerun is offline
Senior Member
 
Join Date: Nov 2010
Posts: 1,422
Blog Entries: 1
valerun is on a distinguished road
Default Re: 10kW / 60A DIY charger open source design

testing different PFC circuits now. Boost topology seems to be working but two stages do hit the efficiency...

so what do you guys think of going single-stage discontinuous conduction mode buck-boost? From what I read up, if it is kept in DCM, that topology naturally has high Power Factor.

A couple of questions I am struggling with a bit though (perhaps some lurking power management gurus can comment ;-):

1. Is what I read about naturally high PF correct?

2. How feasible is to let a microcontroller manage switching frequency based on output current (to keep things in DCM)? Max update rate is going to be ~1kHz so frequency will have to be calculated (as opposed to controlled based on real-time current sensing).

3. Since there is no bulk cap pre-switch, there will be 120Hz ripple on output. Probably uncorrectable by my previous approach of jacking up duty cycle and letting current limiter shave off the peaks. Not sure how big of a deal that 120Hz ripple, though.

Any ideas?
Valery.
Reply With Quote
  #15  
Old 06-18-2011, 02:44 AM
steven4601 steven4601 is offline
Senior Member
 
Join Date: Nov 2010
Location: the Netherlands, Overijssel
Posts: 506
steven4601 is an unknown quantity at this point
Send a message via MSN to steven4601
Default Re: 10kW / 60A DIY charger open source design

1. Power Factor depends strongly on the power. Discontinious COnduction (Critical conduction) mode PFC's often are suited for lower powers. (<1kw)

Reason: large flux swings require ferrite. When using ferrite to its max (full flux swing & high frequencies) it starts to dissipate substantially more than the copper windings.

Sollution: go to Continious conduction mode (ferrites still can be used, but also Iron powder) for high power (1kw>). CCM has a lower flux swing in the storage inductor, but your IGBT or mosfet will be 'hit' harder because it will be hard-switching. Hard-switching is not really an issue , just expect a few watt's extra there.


2. why use a micro when an 80 cent PFC chip can perform equally & proven technology. Missing feedback detection, easy to replicate, easier to get certification. A CCM converter chip will fall back to DCM automaggically when the output current is too low to sustain CCM.

3. No bulk-capacitor for your buck-converter? It will need some sort of bulk capacity. You else might get beat-frequencies causing regulation / loop instabilities.


CCM should be possible to do 97..98% without going overboard. Just use a fine stranded Litze wire for your storage inductor and keep the beginning and end away from eachother inside the bobbin. Using low drain-source capacitance capacitor will improve efficiency & SiC diodes.
99% requires tweaking and is likely only possible at reduced power levels.
Reply With Quote
  #16  
Old 06-19-2011, 12:47 AM
valerun valerun is offline
Senior Member
 
Join Date: Nov 2010
Posts: 1,422
Blog Entries: 1
valerun is on a distinguished road
Default Re: 10kW / 60A DIY charger open source design

Thanks Steven for you help! Really appreciate it!

Quote:
Originally Posted by steven4601 View Post
2. why use a micro when an 80 cent PFC chip can perform equally & proven technology. Missing feedback detection, easy to replicate, easier to get certification. A CCM converter chip will fall back to DCM automaggically when the output current is too low to sustain CCM.
mostly because one can't get access to these chip's inner control circuits. In order to make these work in a charger, I need to be able to command the chip to maintain certain constant current or voltage. With the chips I studied, it's far from trivial. I have bought some PWM chips with exposed error amplifier inputs but hasn't tried it yet.

Quote:
Originally Posted by steven4601 View Post
3. No bulk-capacitor for your buck-converter? It will need some sort of bulk capacity. You else might get beat-frequencies causing regulation / loop instabilities.
there are - both on input and output. I was referring to the buck-boost topology with high PF which would not have a bulk cap on input. And therefore would have large 120Hz ripple on output.

Quote:
Originally Posted by steven4601 View Post
CCM should be possible to do 97..98% without going overboard. Just use a fine stranded Litze wire for your storage inductor and keep the beginning and end away from eachother inside the bobbin. Using low drain-source capacitance capacitor will improve efficiency & SiC diodes.
99% requires tweaking and is likely only possible at reduced power levels.
you're probably right - maybe I just shouldn't stress out too much about these 2-5% losses and just go with a straightforward boost PFC followed by micro-controlled charger PWM converter. Have designed a PCB today based on IR1153 PFC chip (datasheet attached). Picked that for its low switching frequency as am planning to use a 200A IGBT module as a switch.

SiC diodes are hard to come by at this current level I think, no?

Thanks again!
Valery.
Attached Files
File Type: pdf ir1153.pdf (380.9 KB, 136 views)
Reply With Quote
  #17  
Old 06-19-2011, 01:17 AM
steven4601 steven4601 is offline
Senior Member
 
Join Date: Nov 2010
Location: the Netherlands, Overijssel
Posts: 506
steven4601 is an unknown quantity at this point
Send a message via MSN to steven4601
Default Re: 10kW / 60A DIY charger open source design

Quote:
Originally Posted by valerun
mostly because one can't get access to these chip's inner control circuits. In order to make these work in a charger, I need to be able to command the chip to maintain certain constant current or voltage. With the chips I studied, it's far from trivial. I have bought some PWM chips with exposed error amplifier inputs but hasn't tried it yet.
I think you can design the PFC stand-alone. No control lines what so ever. Also you do not need to controll the current or voltage for the PFC once you have chosen the resistor values. eg, 400Vdc out, ~ 25A limit


Quote:
Originally Posted by valerun
you're probably right - maybe I just shouldn't stress out too much about these 2-5% losses and just go with a straightforward boost PFC followed by micro-controlled charger PWM converter. Have designed a PCB today based on IR1153 PFC chip (datasheet attached). Picked that for its low switching frequency as am planning to use a 200A IGBT module as a switch.
200A IGBT for PFC ? Which brand and model are you planning on using? Those brick type IGBT's may not be as efficient for 20Khz than you like. Paralleling 3 or 4 smaller Mosfets/IGBT's is no shame either. ST & IRF has some nice 40+ Ampere@ 100C-j IGBT's under 5 euro that can do 40kHz with reasonable losses.

Quote:
Originally Posted by valerun
SiC diodes are hard to come by at this current level I think, no?
Thanks again!
Valery.
@ 22kHz SiC is a waste of funds :-)
HOw are you going to filter the input? 22kHz requires some serious input (X2) capacitors.
Reply With Quote
  #18  
Old 06-19-2011, 12:10 PM
valerun valerun is offline
Senior Member
 
Join Date: Nov 2010
Posts: 1,422
Blog Entries: 1
valerun is on a distinguished road
Default Re: 10kW / 60A DIY charger open source design

Quote:
Originally Posted by steven4601 View Post
200A IGBT for PFC ? Which brand and model are you planning on using? Those brick type IGBT's may not be as efficient for 20Khz than you like. Paralleling 3 or 4 smaller Mosfets/IGBT's is no shame either. ST & IRF has some nice 40+ Ampere@ 100C-j IGBT's under 5 euro that can do 40kHz with reasonable losses.
I was hoping to use something like the attached. 600ns turn-on/off times (Rg=10Ohm), 600nQ switching charge. <300ns reverse recovery on the diode (will use one of the IGBTs in the module as freewheeling diode). Hence 22kHz - should limit switching time to ~2% of the cycle max.

Of course, 5 50A MOSFETs like the attached could do the trick, as well - theoretically with much smaller losses but to take advantage of that, I'd have to REALLY pay attention to parasitics due to higher dV/dt. So far, my success rate on that beyond 32kHz was a bit spotty ;-))

V
Attached Files
File Type: pdf 600V 200A module.pdf (159.1 KB, 124 views)
Reply With Quote
  #19  
Old 06-19-2011, 01:26 PM
steven4601 steven4601 is offline
Senior Member
 
Join Date: Nov 2010
Location: the Netherlands, Overijssel
Posts: 506
steven4601 is an unknown quantity at this point
Send a message via MSN to steven4601
Default Re: 10kW / 60A DIY charger open source design

Just doing rough guess is that you will have about 100A inductor current worst case.

From the graph switching-losses vs current (mid way between 25 and 125tj):
Switching losses then are
Eon = 2.5mJ * 22kHz = 55 Watt for turning on
Eoff = 6.0mJ * 22kHz = 132 watt for turning off.
that is ~180..200 watts switching losses.

conduction losses 2V * 100A watts * 0.7 = 140

Total igbt loss 320 watts. or seen from 10kW output, 3.2% loss.

Not bad, but i think you might find better switchers with little effort.
A similar powerloss result but for 17 USD worth of switchers can be realized with the : STGW30NC60VD

Either way its going to work.
Reply With Quote
  #20  
Old 06-20-2011, 03:01 AM
valerun valerun is offline
Senior Member
 
Join Date: Nov 2010
Posts: 1,422
Blog Entries: 1
valerun is on a distinguished road
Default Re: 10kW / 60A DIY charger open source design

Thanks Steven - will try some new things next weekend & will post.

Some good progress on the charger itself:
1. Added automatic calibration menus - on power-up, the charger goes through self-calibration of all the sensors - zero point and (for all-important battery voltage sensor) sensitivity calibration. I am soooo happy I don't have to hardcode these things now for every unit I build!
2. Added battery type / size menus - again, only once per power-up.
3. Added output power memory so that the next time you connect the mains, the charger will begin charging using your last saved settings (after 10 sec timeout)
4. Added intelligent fan control based on heatsink temperature. Now can leave charger plugged into 12V circuit all the time - total draw ~2W
5. Added variable switching frequency control. The idea is to minimize switching losses while keeping output ripple current below 0.4. ripple factor ~ T * D (1-D) / Iout. The higher the current and the closer duty cycle is to 0 or 100%, the lower the frequency I can use. Limited the range to 2-12kHz for now - can see the effect very clearly - pushing 10kW into the 280V pack keeps the heatsink at 35C (with fans on)!
6. Did the thermal checks of all components in different regimes. Made a couple of design tweaks to bring everything close together (e.g., my input caps were heating up to 85C at 8kW before, etc). The limiting factor now is actually ...wires connecting components ;-) I guess that's a good problem to have.
7. Redesigned power stage to use a single 200A 600V IGBT module for both a switch and a freewheeling diode. Saving space while getting rid of isolation headaches (WHO THE F*K decided that on all diodes the cathode has to be connected to the backplate???). Will try to push 80-100A output current with this design. The limiting factor will be the inductor I think... Will probably try to use 2 in parallel - at that current level, half the inductance should be sufficient.

V
Reply With Quote
Reply

Share or Bookmark this

Thread Tools
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump

 
Support DIY Electric Car
Sponsors

All times are GMT -6. The time now is 03:26 AM.


Powered by vBulletin® Version 3.8.7
Copyright ©2000 - 2014, vBulletin Solutions, Inc.

Ad Management by RedTyger