valerun
Well-known member
our attempts to source CHADEMO plug from UCHEN were not successful. yet, anyway.
Mini-QC Rapid-Charger (RC) Project for LEAF QC Port
- Thread starter garygid
- Start date
Help Support My Nissan Leaf Forum:
garygid
Well-known member
OpenMiniQC and Collaboration:
Open source of part of a design, like the software and not the hardware details,
after the design is finished is one form of "openness".
However, there are other forms, as most of you know.
Being open during the design process in order to achieve a better end result is
a different type of "open". But, sadly, the concept of sharing seems to be
almost un-American, at least in some business circles. Too often, greed triumphs.
Sometimes ideas are just a pen on paper version, which one might scan and email.
Yes, it is possible that it is still all in one's head, but that makes it very difficult
for a meaningful collaboration with others. When one wants help from other
experts, or more experienced people, it is often better to make it easier for them.
If one wants constructive suggestions on a book in progress, it is
often beneficial to offer the whole text, rather than just the chapter titles.
I prefer to work toward good end results, that others can use easily, and safely.
I have posted a lot of info and hints for this project, and several have jumped in.
I also understand that many have a real job, responsibilities, and even a real life.
If you have a power supply that you feel meets the criteria for doing QC on a LEAF,
please feel invited to send me some logs to review.
The voltage up and down ramp test would be the first test.
If you don't, and need help finding or doing the next step, you can post
and explain where you are, and most likely you will get help here. Several
have ordered, or even received the Arduino Due and a touch Display,
and I have sent them my in-progress sketches and libraries.
If you wish to chat with me, you may try after 9 AM please, since my wife
might be sleeping, at my home in Laguna Hills, in Southern California.
PM your number and a time to call, if you prefer that I call you.
I wish to be constructive, and helpful, not argue about why you are doing
whatever you are doing. I might make suggestions, but you decide to follow
them or not. I will ask questions, but you decide to answer or not.
Sometimes I think that I know the answer, but sometimes I think that
just knowing the question can be beneficial.
There are a lot of things that I do not know, since most things were
discovered after I finished schooling. However, sharing, kindness,
and Please and Thank You... I learned early, even if they are not
considered to be important by many these days. Yes, I am an old guy.
Thanks everybody, sincerely Gary
Open source of part of a design, like the software and not the hardware details,
after the design is finished is one form of "openness".
However, there are other forms, as most of you know.
Being open during the design process in order to achieve a better end result is
a different type of "open". But, sadly, the concept of sharing seems to be
almost un-American, at least in some business circles. Too often, greed triumphs.
Sometimes ideas are just a pen on paper version, which one might scan and email.
Yes, it is possible that it is still all in one's head, but that makes it very difficult
for a meaningful collaboration with others. When one wants help from other
experts, or more experienced people, it is often better to make it easier for them.
If one wants constructive suggestions on a book in progress, it is
often beneficial to offer the whole text, rather than just the chapter titles.
I prefer to work toward good end results, that others can use easily, and safely.
I have posted a lot of info and hints for this project, and several have jumped in.
I also understand that many have a real job, responsibilities, and even a real life.
If you have a power supply that you feel meets the criteria for doing QC on a LEAF,
please feel invited to send me some logs to review.
The voltage up and down ramp test would be the first test.
If you don't, and need help finding or doing the next step, you can post
and explain where you are, and most likely you will get help here. Several
have ordered, or even received the Arduino Due and a touch Display,
and I have sent them my in-progress sketches and libraries.
If you wish to chat with me, you may try after 9 AM please, since my wife
might be sleeping, at my home in Laguna Hills, in Southern California.
PM your number and a time to call, if you prefer that I call you.
I wish to be constructive, and helpful, not argue about why you are doing
whatever you are doing. I might make suggestions, but you decide to follow
them or not. I will ask questions, but you decide to answer or not.
Sometimes I think that I know the answer, but sometimes I think that
just knowing the question can be beneficial.
There are a lot of things that I do not know, since most things were
discovered after I finished schooling. However, sharing, kindness,
and Please and Thank You... I learned early, even if they are not
considered to be important by many these days. Yes, I am an old guy.
Thanks everybody, sincerely Gary
valerun
Well-known member
major update on the isolation stage.
With some improved cooling of the transformer from the last pictures posted, we had our first continuous run at 12kW charging one of our converted BMWs (340V CV point, LiFePo4 pack)!
Some stats (steady state):
* Switching at 14kHz
* Efficiency: 94%
* Ambient: 22C
* IGBT: <40C
* Secondary rectifier diodes: <40C (note how tiny their common heatsink is! those SiC diodes are amazing!)
* Transformer (outer windings): <60C
* Input caps (elcaps): ~30C
* Blocking cap (large film): ~40C
* Output inductor: <30C
No signs of core saturation. Which proves our hypothesis that previous slight signs of core saturation were caused by the core heatup given inadequate cooling. This time we had a large 40W AC blowing directly into the transformer and then into the heatsinks. As you can see, helped quite a bit. After shutting down, temp of transformer surface never exceeded 75C, which probably means that the core did not go above 80-90C.
What does this all mean? It means that our core selection and sizing was not that bad after all. Running a few more strands of somewhat finer wire for the primary and secondary should get us to higher power (as I mentioned before, I think we can get to 20kW with this setup largely unchanged now that we get the transformer run cooler).
This also means that we now have an alternative to $750/kW BRUSA chargers.
This will be released as open source for non-commercial use - same as our non-isolated charger products. For us, open source means that you don't have to beg for source code, PCB files, etc and get stonewalled in the process. Look at our public non-isolated charger page to get an idea: http://emotorwerks.com/VMcharger_V12P/" onclick="window.open(this.href);return false;.
As with our other kits, we will be pricing this around the total cost you would have to pay to procure all the components (including shipping etc). We think it's a pretty good deal. For reference, our 12kW PFC charger kits are ~$1,300 (non-isolated). This is pretty close to the total parts cost + shipping you'd have to pay to 20+ different suppliers ordering in single volume (and some suppliers won't even talk to you for that quantity). That's ~$110/kW. The isolation stage kit, based on our current BOM, would run at less than that. So the complete 12kW isolated charger kit would likely be around $2,500, which is ~$200/kW. Maybe less...
Of course, once we get this stage up to 20kW and connect it to our 25kW controlled PFC stage, economics become even better.
I am pretty excited about this. Are you? ;-)
Gary - we are ready to run tests of your code on this hardware if you like. Let me know.
Thanks,
Valery
With some improved cooling of the transformer from the last pictures posted, we had our first continuous run at 12kW charging one of our converted BMWs (340V CV point, LiFePo4 pack)!
Some stats (steady state):
* Switching at 14kHz
* Efficiency: 94%
* Ambient: 22C
* IGBT: <40C
* Secondary rectifier diodes: <40C (note how tiny their common heatsink is! those SiC diodes are amazing!)
* Transformer (outer windings): <60C
* Input caps (elcaps): ~30C
* Blocking cap (large film): ~40C
* Output inductor: <30C
No signs of core saturation. Which proves our hypothesis that previous slight signs of core saturation were caused by the core heatup given inadequate cooling. This time we had a large 40W AC blowing directly into the transformer and then into the heatsinks. As you can see, helped quite a bit. After shutting down, temp of transformer surface never exceeded 75C, which probably means that the core did not go above 80-90C.
What does this all mean? It means that our core selection and sizing was not that bad after all. Running a few more strands of somewhat finer wire for the primary and secondary should get us to higher power (as I mentioned before, I think we can get to 20kW with this setup largely unchanged now that we get the transformer run cooler).
This also means that we now have an alternative to $750/kW BRUSA chargers.
This will be released as open source for non-commercial use - same as our non-isolated charger products. For us, open source means that you don't have to beg for source code, PCB files, etc and get stonewalled in the process. Look at our public non-isolated charger page to get an idea: http://emotorwerks.com/VMcharger_V12P/" onclick="window.open(this.href);return false;.
As with our other kits, we will be pricing this around the total cost you would have to pay to procure all the components (including shipping etc). We think it's a pretty good deal. For reference, our 12kW PFC charger kits are ~$1,300 (non-isolated). This is pretty close to the total parts cost + shipping you'd have to pay to 20+ different suppliers ordering in single volume (and some suppliers won't even talk to you for that quantity). That's ~$110/kW. The isolation stage kit, based on our current BOM, would run at less than that. So the complete 12kW isolated charger kit would likely be around $2,500, which is ~$200/kW. Maybe less...
Of course, once we get this stage up to 20kW and connect it to our 25kW controlled PFC stage, economics become even better.
I am pretty excited about this. Are you? ;-)
Gary - we are ready to run tests of your code on this hardware if you like. Let me know.
Thanks,
Valery
Attachments
garygid
Well-known member
Sounds like great progress.
Thanks for the update, very encouraging.
To be able to control your new power supply, I need to
understand a bit more about what one needs to control.
I realize that these are all obvious to you, but I do not want
to make incorrect assumptions.
So, this 94% includes a PFC front end booster, deriving DC from 240v AC,
from a 50 amp breaker, and the new isolated regulator stage?
I do not really understand what combination of "stages" you
are testing. Does it fit in the box in the 3rd picture above?
Are there two stages, similar to your non-isolated kit, just with
the isolated regulator replacing the non-isolated regulator?
Is there a final output diode to prevent back-current into
the power supply?
As in the previous kit,...
1. There is no control to the PFC stage, correct?
2. The only control to the regulator (2nd) stage is the
duty cycle of the IGBT, right?
Are there any other controls?
We had to add a switched shunt to discharge the
output stage.
What values are you measuring?
Output voltage, current, some temperatures?
How often are you sampling, and are you logging them?
If not yet logging the data, then the first addition to your code
would be the logging, I would think, right?
Is this using a Due or AVR-CAN for the controller?
When idle, do you keep the output voltage at essentially zero?
Under no load, can you ramp the voltage up to around 450v,
hold briefly, and ramp back down to essentially zero again?
If so, good, let's log that function.
If not, then someone needs to work on that function,
which is required by the QC process.
We had to make substantial modifications to our kit
to get to where we could charge a LEAF.
Basically, we had to start over with the control software
and most of its connections to the outside world.
To experimentally adjust parameters and "command"
the power supply, we added a "serial" connection
to a PC so that we could ask the controller to
perform different tasks.
Until we were convinced that the power supply, as
"manually" controlled, could do what we thought
the QC car would request via the CAN, we did not
try to hook up to the expensive car.
Do you have a high power dummy load to
use in testing?
No, I do not yet know all the details, but I will be
learning as we try to program the Due, which
appears to be more difficult than we hoped.
Once we get a schematic and exact parts descriptions,
we hope to be able to help you better. Or, you can add
the control functions and we can go step by step.
In any case, monitoring and logging come first.
I suggest using CAN-Do, but others used their own
tools, but converted their data to CAN-Do format
so that we could collaborate.
See the Arduino Due thread for some of the
problems I am stumbling across with the Due.
---------
We are trying to add a power-off relay, for main power
On/Off, and for an emergency off. Normally the 12v power
would remain on when the power supply is plugged in.
How have you handled limiting the plug-in current inrush?
In our experience, your two thermisters are insufficient.
We are going to add another relay to short out an inrush resistor.
Any suggestions, please?
Thanks for the update, very encouraging.
To be able to control your new power supply, I need to
understand a bit more about what one needs to control.
I realize that these are all obvious to you, but I do not want
to make incorrect assumptions.
So, this 94% includes a PFC front end booster, deriving DC from 240v AC,
from a 50 amp breaker, and the new isolated regulator stage?
I do not really understand what combination of "stages" you
are testing. Does it fit in the box in the 3rd picture above?
Are there two stages, similar to your non-isolated kit, just with
the isolated regulator replacing the non-isolated regulator?
Is there a final output diode to prevent back-current into
the power supply?
As in the previous kit,...
1. There is no control to the PFC stage, correct?
2. The only control to the regulator (2nd) stage is the
duty cycle of the IGBT, right?
Are there any other controls?
We had to add a switched shunt to discharge the
output stage.
What values are you measuring?
Output voltage, current, some temperatures?
How often are you sampling, and are you logging them?
If not yet logging the data, then the first addition to your code
would be the logging, I would think, right?
Is this using a Due or AVR-CAN for the controller?
When idle, do you keep the output voltage at essentially zero?
Under no load, can you ramp the voltage up to around 450v,
hold briefly, and ramp back down to essentially zero again?
If so, good, let's log that function.
If not, then someone needs to work on that function,
which is required by the QC process.
We had to make substantial modifications to our kit
to get to where we could charge a LEAF.
Basically, we had to start over with the control software
and most of its connections to the outside world.
To experimentally adjust parameters and "command"
the power supply, we added a "serial" connection
to a PC so that we could ask the controller to
perform different tasks.
Until we were convinced that the power supply, as
"manually" controlled, could do what we thought
the QC car would request via the CAN, we did not
try to hook up to the expensive car.
Do you have a high power dummy load to
use in testing?
No, I do not yet know all the details, but I will be
learning as we try to program the Due, which
appears to be more difficult than we hoped.
Once we get a schematic and exact parts descriptions,
we hope to be able to help you better. Or, you can add
the control functions and we can go step by step.
In any case, monitoring and logging come first.
I suggest using CAN-Do, but others used their own
tools, but converted their data to CAN-Do format
so that we could collaborate.
See the Arduino Due thread for some of the
problems I am stumbling across with the Due.
---------
We are trying to add a power-off relay, for main power
On/Off, and for an emergency off. Normally the 12v power
would remain on when the power supply is plugged in.
How have you handled limiting the plug-in current inrush?
In our experience, your two thermisters are insufficient.
We are going to add another relay to short out an inrush resistor.
Any suggestions, please?
valerun
Well-known member
Thanks Gary.garygid said:
garygid said:
not quite. 94% is the isolation stage only. Complete system at this point would be:
1. 12kW PFC stage. uncontrolled, fixed output voltage
2. 12kW Isolation stage. uncontrolled, fixed *theoretical* transfer ration (1:1.2). under 12kw load, drops to ~1:1.06 due to various complex factors (losses being a minor part of that). These factors actually reduce fundamental 120Hz current ripple quote a bit which is a nice side effect.
3. 12kW Output buck stage. controlled in the same way as the one you have today
the 12kW test I reported on had these stages in a different order (1-3-2) which meant that we didn't have direct voltage and current measurement at the output. but it worked as a supervised test. This system definitely does not fit into one 10x10x8" box. If we do a good job packaging it up, it will fit into 2 of those.
For 'production' design, we would use the following system:
1. 20-25kW PFC stage. controlled with control loop bandwidth of ~20Hz. Which means that you should be able to ramp from 0A to full output and back in ~0.1-0.2 seconds.
2. 20kW Isolation stage. Initially uncontrolled, later with very narrow control between 1.1 and 1.2 transfer ratio - which should be sufficient for full-range current control on a 'stiff' pack like any production car's lithium packs. Current & voltage measurement would be connected to the output of the isolation stage.
This system will fit into the same space as the 12kW system above
garygid said:
not yet but is trivial to add. I actually think that the diode dissipation is too much waste - my ideal solution would be a series resistor with a relay shunt. We have a separate 2-channel precharge control board built up for this. It also has a dual-channel voltage measurement built-in. both channels are isolated. One will be used for output, one - for input. This should take care of the need for AC precharge, as well.
garygid said:
yes, Robert mentioned - that's why we have those shunt relays on the newer control boards now. See V13 control-driver board file at http://emotorwerks.com/VMcharger_V12P/00%20-%20Schematics%20and%20PCB%20layouts/" onclick="window.open(this.href);return false;
garygid said:
measuring all that, not logging yet. Once we fire up the Due based control board, we will have storage to log this into (SD). What is the sample rate needed - both 'ideal' (but reasonably justified with data) and minimally viable?
garygid said:
yes we can do that using the relays on the control board
garygid said:
understandably so. As I mentioned in one of my emails, we now have serial control built into the charger now. I need to know the timing SLA I need to meet (again, both 'ideal' and minimally viable)
garygid said:
of course. can't really build much without it.
garygid said:
can you point me to Due thread with detailed discussion on this?
garygid said:
getting there. As I mentioned, I'd like us to standardize on the same open source design that can be reused as a universal control board for various things. I am working on getting you the schematics as quickly as I can given other time constraints. The current board schematics (for non-Due board) is at http://emotorwerks.com/VMcharger_V12P/00%20-%20Schematics%20and%20PCB%20layouts/EMW_SmartCharge-12000_schematics-V12/" onclick="window.open(this.href);return false;. Driver board has voltage sensing circuitry you were asking about before.
garygid said:
you nailed it - this is the solution we are going to use for the higher-power units (~10R NTC thermistor shunted with a relay). 10R thermistor is sufficient to run the charger at idle (~50W max across all fan loads etc). We will likely use the same 2-pole relay for input and output and close it ONLY when both input and output are confirmed precharged via monitoring voltage rise. We will use an 80A relay we are using in the JuiceBox EVSE. Should be good up to 20kW.
So the next steps are:
1. I will post Due board schematics
2. Please send / link to Due discussion
3. Please send / post / link to timing specs for voltage / current slopes in response to all critical commands (apologies if this is posted somewhere already - please point)
Thanks,
Valery
garygid
Well-known member
My Names for the Jolomo pins (not official):
(looking at the open end of the plug)
(my guesses/interpretations, assuming that I got it correct, YMMV)
Use at your own risk.
GP = Ground, Protective
QR = QC Ready, QC applies 12v
EC = Enable Charger (permission to charge), car pulls to ground.
PP = Power, DC Plus, isolated
PM = Power, DC Minus, isolated
GD = Ground, Disable car motion, Plug Proximity
CH = CAN-High
CL = CAN-Low
BR = Enable Battery Relay closure, to begin charging, Ground from QC
This is patterned after the info in these documents:
The interface: http://chademo.com/05_interface.html" onclick="window.open(this.href);return false;
The protocol: http://chademo.com/05_protocol.html" onclick="window.open(this.href);return false;
Apparently this is the sequence of opperation (guesses) :
GP and GD are grounded when the plug is inserted.
CH and CL are biased but not yet being used.
PP and PM are unpowered, but will carry the charging voltage & current.
That leaves just 3 control lines (again my assumptions):
QR: the QC applies 12v if/when it is ready/available to do charging.
... (The car responds by starting CAN communication, sending car requirements.)
... (QC responds with charger capabilities.)
EC: the car pulls this line to ground to enable charging prestart testing.
... After successful pre-charging tests...
BR, the QC connects this line to ground to participate in closing the Battery Relay.
... The car closes the Battery Relay to "expose" the car's battery pack HV leads.
... (The car requests current, in amps.)
... (The QC "follows" the requests, and reports its status.)
If there are errors here, please let me know.
Further discussions are in the Arduino Due software thread:
http://www.mynissanleaf.com/viewtopic.php?f=44&t=14172" onclick="window.open(this.href);return false;
Cheers, Gary
(looking at the open end of the plug)
Code:
GP
x QR
x EC x
PP x x PM
x GD x
CL CH
BRUse at your own risk.
GP = Ground, Protective
QR = QC Ready, QC applies 12v
EC = Enable Charger (permission to charge), car pulls to ground.
PP = Power, DC Plus, isolated
PM = Power, DC Minus, isolated
GD = Ground, Disable car motion, Plug Proximity
CH = CAN-High
CL = CAN-Low
BR = Enable Battery Relay closure, to begin charging, Ground from QC
This is patterned after the info in these documents:
The interface: http://chademo.com/05_interface.html" onclick="window.open(this.href);return false;
The protocol: http://chademo.com/05_protocol.html" onclick="window.open(this.href);return false;
Apparently this is the sequence of opperation (guesses) :
GP and GD are grounded when the plug is inserted.
CH and CL are biased but not yet being used.
PP and PM are unpowered, but will carry the charging voltage & current.
That leaves just 3 control lines (again my assumptions):
QR: the QC applies 12v if/when it is ready/available to do charging.
... (The car responds by starting CAN communication, sending car requirements.)
... (QC responds with charger capabilities.)
EC: the car pulls this line to ground to enable charging prestart testing.
... After successful pre-charging tests...
BR, the QC connects this line to ground to participate in closing the Battery Relay.
... The car closes the Battery Relay to "expose" the car's battery pack HV leads.
... (The car requests current, in amps.)
... (The QC "follows" the requests, and reports its status.)
If there are errors here, please let me know.
Further discussions are in the Arduino Due software thread:
http://www.mynissanleaf.com/viewtopic.php?f=44&t=14172" onclick="window.open(this.href);return false;
Cheers, Gary
garygid
Well-known member
Serial can mean many things.
Do you mean a USB connection that "creates" a virtual Comm
Port in the PC?
Or, an RS232 serial port that is basically obsolete?
What speed, format, and handshake protocol?
What data flows each way?
Commands, and responses?
Without that information it is impossible to evaluate.
Thanks.
What is "timing SLA", I am not recognizing the SLA term?
Cheers, Gary
garygid
Well-known member
Mini-QC Interface circuitry:
(a first pass at circuit suggestions, not official)
Handshake Out:
QR = When the mini-QC is Ready (to handle a charging session),
the mini-QC applies 12v power to car's QR pin, and it appears
that this 12v power will later be used to actually energize a relay
coil, but the current and impedance requirements are unknown.
Probably, inductive surge protection should be included.
A 3.3v operated 12v optical isolated switch might be good.
Perhaps this 12v supply should be separated from the supply
used for the mini-QC control, drivers, the Due, and the Display.
For experimenting with possibly non-isolated power, we used
12v from an inexpensive isolated international-voltage AC adapter.
BR = Enable Battery Relay closure, to signal begin charging.
Pulled to Ground by the mini-QC. When grounded, this appears
to act as a status flag for the car (normally pulled up to 12v from QR),
as well as being the ground end of the relay coil power circuit,
with the car having a series switch to actually activate the relay.
Another opto-isolated "switch" would probably be good.
Probably, inductive surge protection should be included.
GD = Ground, a sense input to the car, apparently used to
Disable car motion when plugged in. Not switched, but
always grounded, pulling a resistor on the car's sensor
input. Also, possibly used by the car to sense "Plugged In".
Note: There are two different grounds involved, the Chassis ground
from the car, and the Protective ground from the AC Power source.
Normally, for safe operation, these two must be connected,
and the HV power source must be isolated. The car seems
to expect that the HV is close to centered around Chassis Ground.
However, for experimenting very carefully with non-isolated
supplies, we designed our mini-QC control (uP, CAN, and Display)
to be connected to the car's Chassis ground, and completely
opto-isolated from the mini-QC guts. In any case, this is probably
good design practice, and optically isolated sensors and "switches"
are common, and relatively inexpensive.
Handshake Input Sensor:
EC = Enable Charger (permission to charge), car pulls to ground,
expecting it to operate an opto-isolated input that is pulled up
(by an unknown value resistor in the mini-QC) to 12v.
CAN Communication Lines:
CH = CAN-High
CL = CAN-Low
The car end of these lines seems to be terminated, and the
mini-QC end must, it seems to us, provide line bias in addition
to being terminated. Thus, two extra appropriately 5k resistors
could be used to both provide the bias and also provide some
additional noise immunity.
Internal Sensors:
PP = Voltage, variable DC, usually plus, isolated
PM = Voltage, variable DC, usually Minus, isolated
DC Current
AC Voltage
AC Current
Temperatures of coils, and heat sink, possibly others.
Notes:
This is not a complete list, but intended to become complete.
(a first pass at circuit suggestions, not official)
Handshake Out:
QR = When the mini-QC is Ready (to handle a charging session),
the mini-QC applies 12v power to car's QR pin, and it appears
that this 12v power will later be used to actually energize a relay
coil, but the current and impedance requirements are unknown.
Probably, inductive surge protection should be included.
A 3.3v operated 12v optical isolated switch might be good.
Perhaps this 12v supply should be separated from the supply
used for the mini-QC control, drivers, the Due, and the Display.
For experimenting with possibly non-isolated power, we used
12v from an inexpensive isolated international-voltage AC adapter.
BR = Enable Battery Relay closure, to signal begin charging.
Pulled to Ground by the mini-QC. When grounded, this appears
to act as a status flag for the car (normally pulled up to 12v from QR),
as well as being the ground end of the relay coil power circuit,
with the car having a series switch to actually activate the relay.
Another opto-isolated "switch" would probably be good.
Probably, inductive surge protection should be included.
GD = Ground, a sense input to the car, apparently used to
Disable car motion when plugged in. Not switched, but
always grounded, pulling a resistor on the car's sensor
input. Also, possibly used by the car to sense "Plugged In".
Note: There are two different grounds involved, the Chassis ground
from the car, and the Protective ground from the AC Power source.
Normally, for safe operation, these two must be connected,
and the HV power source must be isolated. The car seems
to expect that the HV is close to centered around Chassis Ground.
However, for experimenting very carefully with non-isolated
supplies, we designed our mini-QC control (uP, CAN, and Display)
to be connected to the car's Chassis ground, and completely
opto-isolated from the mini-QC guts. In any case, this is probably
good design practice, and optically isolated sensors and "switches"
are common, and relatively inexpensive.
Handshake Input Sensor:
EC = Enable Charger (permission to charge), car pulls to ground,
expecting it to operate an opto-isolated input that is pulled up
(by an unknown value resistor in the mini-QC) to 12v.
CAN Communication Lines:
CH = CAN-High
CL = CAN-Low
The car end of these lines seems to be terminated, and the
mini-QC end must, it seems to us, provide line bias in addition
to being terminated. Thus, two extra appropriately 5k resistors
could be used to both provide the bias and also provide some
additional noise immunity.
Internal Sensors:
PP = Voltage, variable DC, usually plus, isolated
PM = Voltage, variable DC, usually Minus, isolated
DC Current
AC Voltage
AC Current
Temperatures of coils, and heat sink, possibly others.
Notes:
This is not a complete list, but intended to become complete.
jclemens
Well-known member
I wrote an update on the brass pins we discussed here.
It is on my blog along with exactly 1 metric butt-load of photos.
Blog post here: http://erroneus.myevblog.com/2013/10/19/9mm-pins-for-jewelry-or-perhaps-rapid-charging-a-car/
Here is one
My time is up for today, gotta run!
It is on my blog along with exactly 1 metric butt-load of photos.
Blog post here: http://erroneus.myevblog.com/2013/10/19/9mm-pins-for-jewelry-or-perhaps-rapid-charging-a-car/
Here is one
My time is up for today, gotta run!
garygid
Well-known member
Last picture looks like pins for a QC-MF-LA, right?
Nice!
Nice!
valerun
Well-known member
garygid said:
TTL UART. Can be connected to PC through FTDI cable. Or direct to another TTL UART device.
Description of the [minimal] command set is at http://emotorwerks.com/VMcharger_V12P/EMW_SmartCharge-12000_Quick-Start-Guide_V12.pdf" onclick="window.open(this.href);return false; (towards the end of this 4-page doc). The command set can be changed / augmented, of course.
SLA is 'service level agreement', sorry. Basically what do I have to guarantee on the hardware side as to response rates to your commands. E.g, less than 100ms between 'M,000,450,E' command and actually putting 450VDC on the output, or less than X00ms before current falls between 90-110% range of requested, etc.
garygid
Well-known member
We do not know what the QC response-time specs are.
Through experiments, and logging the data, we have found
things that seem to work fairly well, by controlling the
PWM to the output stage, and some that do not work.
In our next experiments, we hope to achieve better control
by updating the PWM value more often, and by having finer
control over the actual PWM value.
We hope to work toward a 2nd stage that is both isolation
and regulation, rather than have three sets of rectifiers
and transformer/inductors.
How about ramp up of the no-load voltage over about 1 second,
hold for about a second, and ramp back down over about
another second, or perhaps two.
The current ramp is more difficult, in our rather limited
experience, but the car request ramps at 2 amps per 0.1 second,
and getting too far behind the requests during the initial
current ramp up, usually to MaxAmps, causes the car to "barf".
So, a ramp up to 30 amps is requested over 1.5 seconds.
One needs to respond to a request for zero amps (a Stop
charging request) fairly quickly.
I believe that a diode on the output is "required", but we
have not done any experiments without a diode.
The car might not like current flowing out of the battery.
Cherrs, Gary
Through experiments, and logging the data, we have found
things that seem to work fairly well, by controlling the
PWM to the output stage, and some that do not work.
In our next experiments, we hope to achieve better control
by updating the PWM value more often, and by having finer
control over the actual PWM value.
We hope to work toward a 2nd stage that is both isolation
and regulation, rather than have three sets of rectifiers
and transformer/inductors.
How about ramp up of the no-load voltage over about 1 second,
hold for about a second, and ramp back down over about
another second, or perhaps two.
The current ramp is more difficult, in our rather limited
experience, but the car request ramps at 2 amps per 0.1 second,
and getting too far behind the requests during the initial
current ramp up, usually to MaxAmps, causes the car to "barf".
So, a ramp up to 30 amps is requested over 1.5 seconds.
One needs to respond to a request for zero amps (a Stop
charging request) fairly quickly.
I believe that a diode on the output is "required", but we
have not done any experiments without a diode.
The car might not like current flowing out of the battery.
Cherrs, Gary
garygid
Well-known member
Commands and status once a second are unlikely
to be sufficient for QC type control.
When you can show some graphs of response to commands,
we will know better what the power supply and regulator can
accomplish.
I will try to add a basic log-data pseudo command to my
Sketch so that logging initial tests with the Due's Native USB
port connected to a Windows PC running my CAN-Do program
should be fairly painless, along with a screen to monitor
the values on the display.
Values would include Amps-Out, Voltage-Out, PWM value,
Limiting-Temperature,...?
Other values desired?
Five of the 8 data bytes are used with the above, and
millisecond time stamps are with each message.
If necessary, a second pseudo message could be used.
Suggestions welcome.
to be sufficient for QC type control.
When you can show some graphs of response to commands,
we will know better what the power supply and regulator can
accomplish.
I will try to add a basic log-data pseudo command to my
Sketch so that logging initial tests with the Due's Native USB
port connected to a Windows PC running my CAN-Do program
should be fairly painless, along with a screen to monitor
the values on the display.
Values would include Amps-Out, Voltage-Out, PWM value,
Limiting-Temperature,...?
Other values desired?
Five of the 8 data bytes are used with the above, and
millisecond time stamps are with each message.
If necessary, a second pseudo message could be used.
Suggestions welcome.
valerun
Well-known member
garygid said:
[EDITED as I saw your prev msg after writing an initial response]
Thanks for timing suggestions. Will test.
Yes, I understand 1 second reporting / commands will not be sufficient. That's why I asked for requirements. Seems like 100ms is the minimally viable command response time, correct? We will shoot for that in the initial demo
valerun
Well-known member
Note that you don't necessarily need to make the calculation loop faster to achieve this. The way PWM ramp was implemented in our chargers was adequate for EV conversions where there are no timing requirements. Specifically (as you know), we just simply incremented or decremented PWM depending on the binary error signal (lower or higher than target). This is VERY crude way to control things but was good enough for the application.garygid said:
We would not use it for a time critical application we are discussing now. What we would use is the same method as used in our motor controllers. Specifically, full PID loop (with or without D as will be determined by the dynamic response we get). We are getting current ramp rates of up to 10,000A/s in the motor controllers using that method. All this is done on the same Arduino Pro Mini... To get that, we run a tuned PID loop at ~1kHz, triggering recalc of duty parameters on every 8th interrupt of the main PWM timer (phase correct so it interrupts in the middle of the pulse - the most correct way to measure current in a switching supply). ADC conversions are done in the other 7 slots in the 8-cycle period. ADC clock prescaler is reprogrammed for ~13us conversion time which leaves plenty of time between interrupts.
This approach is able to command much larger PWM swings in just one control cycle (1ms) - up to a full range from 0 to 100%. Stability of the loop is determined by the P, I, and D constants. It is also inherently able to 'dither' the duty to meet current target - PWM duty will quickly oscillate between the two adjacent values to create an effective higher PWM duty resiolution.
For the charging system, we would be running PWM at ~20khz, so we would operate PID loop at 1/16th or 1/32th of that.
garygid said:
We'd like that, too. As I mentioned before, this is the architecture of the final target product. However, we might stop at 3 stage design. There are always trade-offs that might make 2-stage design less attractive. For example, with multi-stage design, one can optimize each stage to work at its highest potential efficiency. As a result, a 3-stage design may *easily* be more efficient than a 3-stage design. I know, sounds counter-intuitive, but it's true. To provide a familiar analogy, think about a series hybrid car that operates its gasoline engine at best efficiency RPM and allows the 'second stage' (battery, inverter & electric motor) to soak up load pulses. It is well known that this 2-stage system is much more efficient than a single-stage system with gas engine directly coupled to drivetrain via a transmission.
this is easy with better control per above (and is even possible with the current level of control)garygid said:
ok let's test it. max current that can flow back from battery is ~2mA into the bleed-off resistor across output caps. I doubt the car will care. Everything else is blocked by the PFC diode inside the unit.garygid said:
Thanks,
Valery
garygid
Well-known member
When the output caps are zero volts, and the battery relay
closes, putting 350 volts on the outout, without a diode
there would be a huge "reverse" current flow, right?
Or, am I missing something?
closes, putting 350 volts on the outout, without a diode
there would be a huge "reverse" current flow, right?
Or, am I missing something?
valerun
Well-known member
garygid said:
There will be pre-charge relay in the charger on both input and output. So a possible sequence might be:
1. At the request of car to ramp to 450V (battery not connected), output relay closes, charger ramps
2. At the request of car to ramp down to 0V, output relay opens, discharge / pulldown relay closes. The caps are removed from circuit so this can be very fast
3. Car closes its battery contactor
4. Charger closes the precharge relay, waits until voltage rise on caps slows down OR preset time passes, when satisfied, closes the main output relay
Would this work? The main question I think is what kind of delay is acceptable in #4 - will the car expect some current right after it closes its battery contactor?
Valery.
garygid
Well-known member
Yes, it Might work, but
Will it work on all QC cars? Unknown.
I do not know how closely the car watches the
voltage ramp up, and then down. Just disconnecting
might not be accepted by the car, but we have not
tested disconnecting, since the kit has (I believe)
no output (or input) relay, or provision to control
such relays. But, I might be mistaken.
The car does not expect current in until it requests Amps.
The request comes soon after the battery Contactor closes.
On the order of one second delay, is my somewhat vague
recollection, but that is mostly a guess.
Much of our work centers around handling the currents
during this critical several seconds of time, from battery
closure to stable delivery of current into the battery.
Will you measure the HV both inside your output
relay and outside the relay?
You already have two diodes in the 1st stage, probably two
in the isolation stage, and perhaps one in the output stage, right?
Will it work on all QC cars? Unknown.
I do not know how closely the car watches the
voltage ramp up, and then down. Just disconnecting
might not be accepted by the car, but we have not
tested disconnecting, since the kit has (I believe)
no output (or input) relay, or provision to control
such relays. But, I might be mistaken.
The car does not expect current in until it requests Amps.
The request comes soon after the battery Contactor closes.
On the order of one second delay, is my somewhat vague
recollection, but that is mostly a guess.
Much of our work centers around handling the currents
during this critical several seconds of time, from battery
closure to stable delivery of current into the battery.
Will you measure the HV both inside your output
relay and outside the relay?
You already have two diodes in the 1st stage, probably two
in the isolation stage, and perhaps one in the output stage, right?
JeremyW
Well-known member
I think you may be able to get around this somewhat by telling the car over the CAN bus your qc is more limited (max available current is less) for the initial ramp, then gradually increase the max available current over a period of say ten seconds so that the ramp is easier to follow.
valerun
Well-known member
garygid said:
Hi Gary -
We would actively pull down output voltage (not just disconnect). We just won't have to discharge the caps since they will be disconnected by the open output relay.
You are correct that the current kits do not have precharge relays. The plan is to have these on for our 20kW units as inrush resistors just don't work at those power levels.
We will initially measure HV only on the charger side of the relay. I don't believe there is a real need for both as we can measure the rate of caps' voltage rise to control precharge. But can be done, of course - let me know if you think there is a need.
On the diodes in stages -
1. There is a diode in the front PFC stage. It feeds the PFC output caps providing interim energy storage
2. There is a diode in the isolation stage's secondary side (there are actually 2 sets of 2 diodes in parallel but that's not material - the assembly can be thought of as a single diode for all practical purposes). This diode feeds output caps.
So there is no existing diode protecting output caps. Adding such a diode is of course possible but would result in additional cost and 100W extra dissipation from conduction loss (0.5% hit in efficiency). The thinking was that since we already have to use an input relay for input pre-charge, we can reuse it for output, as well (one pole for input, one pole for output). This would require some logic to combine the precharges but I think it could be done.
Valery
