Converting the stock EVSE to an OpenEVSE unit
- Thread starter GlennD
- Start date
Help Support My Nissan Leaf Forum:
chris1howell
Well-known member
GlennD said:
Hi Glenn. I am trying to figure out the datasheet http://www.fairchildsemi.com/ds/MI/MID400.pdf to find the optimum value. I see Figure 1 which is what I believe you are referring to. I am confused by the units of Kv. If Kv was mislabeled and should be kOhm the chart seems to conflict with the formula and data.
All of the datasheets and the application notes http://www.fairchildsemi.com/an/AN/AN-3007.pdf use 22k for 120v.
In the Transfer Characteristics sections it lists the Min input voltage of 90v with a 22k resistor. The formula matches.
Rin = Vin - Vf / Iin
The data sheet lists Vf as 1.5v and Iin at 4.0ma minimum
so for the 90v minimum listed in the datasheet...
Rin = 90v - 1.5v / 4
Rin = 22125
22k is the current value for OpenEVSE PLUS
For our international friends
Rin = 200v -1.5v
200v = 49625
47k is the current value for International OpenEVSE PLUS
Figure 4 also shows a lower Iin based on capacitance from aux to ground, it’s not currently connected but it looks like the trip point can be from 2.4 to as low as .8. If we try 2.5…
then 90v - 1.5v / 2.5
Rin 35400
I think the minimum current may have a lot of margin, it would be nice to know what the minimum current for our application is and find the highest value maintaining enough margin for reliable operation and lowest leakage current. Also, I wish it listed Max Iin so we knew if the same value could be used up to 230v as well...
I considered actually measuring the pick up and drop out voltages. I have a variac and could do it. The only reason I did not was the change worked fine.
To test it properly I would need to separately power the OpenEvse Plus with 120V. This seemed like to much work for me.
The resistor values assume a full 16 MA load, the Atmel chip is a micro power device. I do not know if this affects the device. It would on a normal optical isolator.
My EVSE is still open on my bench since I siliconed in the LED and it is drying. If a get a whole lot of ambition I will test it.
To test it properly I would need to separately power the OpenEvse Plus with 120V. This seemed like to much work for me.
The resistor values assume a full 16 MA load, the Atmel chip is a micro power device. I do not know if this affects the device. It would on a normal optical isolator.
My EVSE is still open on my bench since I siliconed in the LED and it is drying. If a get a whole lot of ambition I will test it.
Chris, I will test it tonight. The graph shows 22K at 50v and your calcs at 90V. I was just lazy and let good enough rule.
Since the chips are referenced to ground they will never see 240V and I have no way of testing beyond a variac's max voltage which as I recall is in the 130V range. A variac is something I use once in a blue moon.
Since the chips are referenced to ground they will never see 240V and I have no way of testing beyond a variac's max voltage which as I recall is in the 130V range. A variac is something I use once in a blue moon.
First off I replaced the resistors with 4 22K resistors since a single one is what Chris uses. With two in series I could easily short out 1 letting me test stock and GFCI friendly.
Both the 22K and the doubled 44K picked up at about 45V. In my tests both were identical. It could be that there is a difference under full fanout. For our use a 47K (5% value ) would work fine.
Assuming enough current is available, the voltage is the determining factor. I plan to leave in my test resistors instead of the 39K ones I chose from the chart.
Thanks for pushing me Chris! The resistors I used were left overs from the job I did for you. I wanted to compare the stock setup.
Both the 22K and the doubled 44K picked up at about 45V. In my tests both were identical. It could be that there is a difference under full fanout. For our use a 47K (5% value ) would work fine.
Assuming enough current is available, the voltage is the determining factor. I plan to leave in my test resistors instead of the 39K ones I chose from the chart.
Thanks for pushing me Chris! The resistors I used were left overs from the job I did for you. I wanted to compare the stock setup.
Ingineer
Well-known member
This is not true. Some commercial 240V outlets are on delta power systems, and thus, one leg is often the "Stinger" leg, which is usually around ~208V over ground.GlennD said:
Also, if it's ever used on power systems outside the US, it's common to have 240V over "earth".
-Phil
I will revise my statement to say that for my use they will never see 240V.
At work I had it connected to a delta 240 source plus ground. I had to move it to the 208V wye circuit to get 120V to power a TED since the city wanted to monitor my use. It is also connected to ground.
At home I have normal residential service and my EVSE is connected to both legs plus ground.
At work I had it connected to a delta 240 source plus ground. I had to move it to the 208V wye circuit to get 120V to power a TED since the city wanted to monitor my use. It is also connected to ground.
At home I have normal residential service and my EVSE is connected to both legs plus ground.
I was thinking about comments while I was walking my dogs.
You are right about your comments about 240VAC delta. In that case the MID400s are exposed to 240v on 2 phases. The 120V sized MOVs are also wrong in that specific case. A 240V rated MOV from line to line would be better.
I just ordered a bunch of 240V MOVS. When they come in I will use 3 in my EVSE. While the odds are that I will never plug into a delta circuit its better to be safe than sorry.
You are right about your comments about 240VAC delta. In that case the MID400s are exposed to 240v on 2 phases. The 120V sized MOVs are also wrong in that specific case. A 240V rated MOV from line to line would be better.
I just ordered a bunch of 240V MOVS. When they come in I will use 3 in my EVSE. While the odds are that I will never plug into a delta circuit its better to be safe than sorry.
I have taken Ingineer's observations to heart. I have control of my EVSE and I would not knowingly plug into a 240 VAC delta circuit but it could happen. I replaced the MOV's with ones rated for 275VAC RMS and added one across the lines. I also doubled the value of the sense resistors. This helps greatly to avoid false tripping of 120V GFCI's and corrects the value for 240V. In my tests with this value the MID400 trips at less than 50V.
henrysunset
Member
- Joined
- Mar 13, 2013
- Messages
- 13
How does this OpenEVSE modification compare to the more common modification Phil offers to convert the stock unit to 240v?
Specifically:
- What additional features will I get with the OpenEVSE (based on current firmware - I understand that firmware updates can result in new capabilities)
- How does this modification compare in terms of cost and complexity to implement?
- Anyone try to support the recent Wifi capabilities of the evolving EVSE firmware with this modification?
Specifically:
- What additional features will I get with the OpenEVSE (based on current firmware - I understand that firmware updates can result in new capabilities)
- How does this modification compare in terms of cost and complexity to implement?
- Anyone try to support the recent Wifi capabilities of the evolving EVSE firmware with this modification?
henrysunset said:
I would not consider this a real alternative to Phil's EVSEUpgrade. The complexity to implement vs Phil's service is off the charts.
Cost wise it's probably reasonably similar if you have to buy everything to complete this vs just sending it to Phil.
AFAIC the reasons one would want to perform this mod vs just building an OpenEVSE from scratch or getting an EVSE Upgrade is they want to prove it can be done, they like building things and they've already built enough standard OpenEVSE and wanted a new challenge, etc.
If you wanted to build an OpenEVSE instead of using the EVSEUpgrade service I think I would recommend building one from scratch and then selling the Panasonic EVSE. It'll probably end up costing you nothing after all is said and done.
Ingineer
Well-known member
Reliability is probably also something you're going to lose in a DIY effort. I bet that one good drop onto pavement will likely render any DIY conversion of this type damaged. This is of the reasons Panasonic goes to great lengths to use a special energy-absorbing potting compound.
Also, it would be almost impossible to do this conversion in a 2nd generation (2013) EVSE due to the lack of room and the fact that the PCB is permanently potted into the bottom housing.
-Phil
Also, it would be almost impossible to do this conversion in a 2nd generation (2013) EVSE due to the lack of room and the fact that the PCB is permanently potted into the bottom housing.
-Phil
TonyWilliams
Well-known member
Ingineer said:
Apparently, some of the Tesla portable EVSE (Model S "UMC") are burning up. I wonder if its a lack of energy absorption from being dropped?
The interchangable plug adapters are also melting their connection pins (they are tiny). I'm modifying a few with J1772 plugs for Rav4 drivers, but I'm bumping into a new problem. Tesla decided to use two conductors in parallel per power pin, which causes a dilemma as to how to terminate them.
I wanted to just use a 4 indent die to crimp on the pins, but now I'm thinking that the two conductors should be combined to a single 6 - 8 gauge conductor to be crimped on the J1772 power pin. Of course, I still don't know how I would want to combine those two 12-ish gauge conductors into one 6 - 8 gauge.
Ingineer
Well-known member
I just did one of these conversions for Waidy:
The cool thing is Tesla's inclusion of a handy 3.3v source which is normally used for the charge door release transmitter. This enabled easy addition of a bright LED for nighttime, something not even Tesla thought of!
I opened one of the UMC's up, but it's a pretty destructive process due to the housing overmold and it being glued shut. If not for this "one way trip", I'd have simply replaced the thin (too thin?) Tesla cable.
Tesla is really pushing the limits! They use two 2.5mm wires for each side, (Equivalent to #13AWG) and inside the box they use the exact same relay ClipperCreek uses on the LCS25, a single 30A rated relay. (at 40A!) It's arguable that splitting the high-current between 2 smaller conductors is better for heat-dissipation, but then you have the pesky termination issues. Inside the UMC, Tesla welds the 2 conductors to a little square terminal, then screws this to a PCB terminal. So this is a good termination and it looked well-done by my assessment. Having 4 smaller wires instead of 2 larger ones definitely makes for a smaller, more flexible cable overall, so regardless of any potential current handling gain, it's good ergonomically.
I definitely don't like their interchangeable right-angle plug connector design. First off, if you are going to do right-angle, why not make it symmetrical? That way you could flip it 180 degrees if your outlet is installed upside-down. (Patent-Pending! =)
Their "dongle" design makes for compact and simple plug adapters, but it ends up being a really big blob, especially when used on smaller outlets (NEMA 5-15). I think I prefer having a short length of cable for each adapter, as this makes the final plug smaller, more compact, more flexible, and spreads out heat.
Ok, onto my conversion. Sorry, I didn't think to take pictures of the pin connection detail until it was already all assembled. Here's the best I can do:
Since the "no-name" Chinese handle pins are made for much larger wire, you cannot crimp it by itself and get a reliable 40A connection. I took a crimp ferrule from a yellow butt-splice connector, and crimped the 2 Tesla wires into that, which were very well crimped, then crimped and soldered that into the no-name pin. I believe this is as good as I can make it.
I would not trust solder alone, there must be a solid mechanical bond first at these current levels with the relatively tiny wires. If you are soldering anything, it absolutely must not be allowed to move, as you will have stress-concentrations where the solder wicking inside the wire strands ends.
Needless to say, I'm not doing another one of these. It's a lot of work to do properly, I'm not sure I like the "no-name" Chinese handle, and the UMC itself isn't the hardiest design to begin with, though it is sexy!
-Phil
The cool thing is Tesla's inclusion of a handy 3.3v source which is normally used for the charge door release transmitter. This enabled easy addition of a bright LED for nighttime, something not even Tesla thought of!
I opened one of the UMC's up, but it's a pretty destructive process due to the housing overmold and it being glued shut. If not for this "one way trip", I'd have simply replaced the thin (too thin?) Tesla cable.
Tesla is really pushing the limits! They use two 2.5mm wires for each side, (Equivalent to #13AWG) and inside the box they use the exact same relay ClipperCreek uses on the LCS25, a single 30A rated relay. (at 40A!) It's arguable that splitting the high-current between 2 smaller conductors is better for heat-dissipation, but then you have the pesky termination issues. Inside the UMC, Tesla welds the 2 conductors to a little square terminal, then screws this to a PCB terminal. So this is a good termination and it looked well-done by my assessment. Having 4 smaller wires instead of 2 larger ones definitely makes for a smaller, more flexible cable overall, so regardless of any potential current handling gain, it's good ergonomically.
I definitely don't like their interchangeable right-angle plug connector design. First off, if you are going to do right-angle, why not make it symmetrical? That way you could flip it 180 degrees if your outlet is installed upside-down. (Patent-Pending! =)
Their "dongle" design makes for compact and simple plug adapters, but it ends up being a really big blob, especially when used on smaller outlets (NEMA 5-15). I think I prefer having a short length of cable for each adapter, as this makes the final plug smaller, more compact, more flexible, and spreads out heat.
Ok, onto my conversion. Sorry, I didn't think to take pictures of the pin connection detail until it was already all assembled. Here's the best I can do:
Since the "no-name" Chinese handle pins are made for much larger wire, you cannot crimp it by itself and get a reliable 40A connection. I took a crimp ferrule from a yellow butt-splice connector, and crimped the 2 Tesla wires into that, which were very well crimped, then crimped and soldered that into the no-name pin. I believe this is as good as I can make it.
I would not trust solder alone, there must be a solid mechanical bond first at these current levels with the relatively tiny wires. If you are soldering anything, it absolutely must not be allowed to move, as you will have stress-concentrations where the solder wicking inside the wire strands ends.
Needless to say, I'm not doing another one of these. It's a lot of work to do properly, I'm not sure I like the "no-name" Chinese handle, and the UMC itself isn't the hardiest design to begin with, though it is sexy!
-Phil
Ingineer
Well-known member
For those curious, here's some shots of the UMC. Note the 30A relay that is run at 40A:
-Phil
-Phil
I think comments about the OpenEVSE being fragile are off the mark. Certainly the potted stock unit is more rugged but I am sure that if I dropped my unit while pulling it out of the case it would survive. Neither unit would survive being run over.
I built this unit for a number of reasons:
I built it as a proof of concept. I wanted to see if the EVSE Plus would allow a full function EVSE to fit in the stock Panasonic housing.
I wanted a 240VAC unit for a backup for my home and work chargers. I have used my L1 unit maybe 3 or 4 times in the early days when I suffered from range anxiety. I can still use the unit as a L1 EVSE. I can never see me using it in a RV park. I have not ever built an adapter.
I constructed as apposed to buying Ingineer's unit solely because there is no information available on what the conversion consists of. I have built OpenEVSE's from the eatly days to current. I have a lot of experence with them. I try to have service manuals or information for most of my equipment.
There is no doubt Ingineer's conversion is safe and reliable but it is not for ME.
I built this unit for a number of reasons:
I built it as a proof of concept. I wanted to see if the EVSE Plus would allow a full function EVSE to fit in the stock Panasonic housing.
I wanted a 240VAC unit for a backup for my home and work chargers. I have used my L1 unit maybe 3 or 4 times in the early days when I suffered from range anxiety. I can still use the unit as a L1 EVSE. I can never see me using it in a RV park. I have not ever built an adapter.
I constructed as apposed to buying Ingineer's unit solely because there is no information available on what the conversion consists of. I have built OpenEVSE's from the eatly days to current. I have a lot of experence with them. I try to have service manuals or information for most of my equipment.
There is no doubt Ingineer's conversion is safe and reliable but it is not for ME.
Ingineer
Well-known member
I can assure you that the Panasonic unit easily survives drive-over. I've got a picture of my LEAF somewhere parked on one.
Out of thousands of units, The only time we've seen physically broken ones is when crazy things happen such as a dog literally chewing through the cable! (There were also teeth marks on the unit, but it hardly did anything!)
They just simply do not fail. I've seen a number of other makes with all-manner of problems related to physical design.
No offense Glenn, but I see numerous potential failure modes of your conversion if subjected to high-G loads. For instance; If it lands the right (wrong) way, the fuses will pop right out of their holders, but that's an easy fix. If it drops upside down, there's a very good chance the MPD power module will simply "unsolder" itself, or if it's a side impact, it'll shear off it's legs. Good designs using large devices mounted only by solder on PCB's supply adjunct physical mounting, as reliability is poor when using solder as a sole mounting strategy.
I can't tell you how many times I've seen large electrolytic caps fall off boards that are subjected to rough (or automotive) environments. Take a look at the Nissan (Nichicon) on-board charger. All large components are mounted off-board or have some form of auxiliary mounting system.
-Phil
Out of thousands of units, The only time we've seen physically broken ones is when crazy things happen such as a dog literally chewing through the cable! (There were also teeth marks on the unit, but it hardly did anything!)
They just simply do not fail. I've seen a number of other makes with all-manner of problems related to physical design.
No offense Glenn, but I see numerous potential failure modes of your conversion if subjected to high-G loads. For instance; If it lands the right (wrong) way, the fuses will pop right out of their holders, but that's an easy fix. If it drops upside down, there's a very good chance the MPD power module will simply "unsolder" itself, or if it's a side impact, it'll shear off it's legs. Good designs using large devices mounted only by solder on PCB's supply adjunct physical mounting, as reliability is poor when using solder as a sole mounting strategy.
I can't tell you how many times I've seen large electrolytic caps fall off boards that are subjected to rough (or automotive) environments. Take a look at the Nissan (Nichicon) on-board charger. All large components are mounted off-board or have some form of auxiliary mounting system.
-Phil
Ingineer you bring up a good point about the power module. I was able to use some JB Weld to securely afix the power module to the PCB. As for the low mass fuses jumping out of the powerful clips of a 30A fuse holder, well that one I will believe when I see it.
