Hydra v3.2 developments

With the latest work on OpenEVSE II, I've made some exciting (well, to me anyway) developments on the Hydra.

Hydra v3.2 unifies the PCB design - the "splitter" variant PCB was taken from the v3.2 EVSE variant design, and the splitter variant changes back ported from v2.2.

Additionally, there's a new "mezzanine" board with all of the AC components. And there is also a new relay test system (also courtesy of OpenEVSE) that will report an error on the UI if the power does not turn on and off as expected with the respective relay.

The mezzanine board has opto-isolated contactor control triac circuits. There's simply no point to supporting L1 (120 volts) with the Hydra, so contactors with 208/240 volt coils can be used. DigiKey sells a variant of the same relays of the previous reference design that have 208/240 volt coils,

The AC board connects to either variant with an 8 pin FFC. The AC board has jacks for AC input, contactor coil outputs and two AC test inputs for the load side of each contactor.

The logic boards have an FTDI port, ISP, a 3 pin connector for the UI switch and ammeter and J1772 connectors for each car. For the splitter variant, the per-car connector has pilot and proximity lines and there is an inlet pilot/proximity connector and a power switch connector. For the EVSE variant, the per-car connector has only pilot and there is a 3 pin GFI terminal and a CR1220 battery holder for the RTC chip.

So let's take a look at the result:

That's the AC board. It's the same for both variants. The actual board is the exact same size as the logic board and has standoff holes in the same locations. The idea is that you mount the logic/display board on inch-and-a-half standoffs above this one.

Each logic/display board is in 3 sheets. Let's look at each variant side by side, one sheet at a time:

Each has identical +5 and -12 volt DC-DC converter modules based on the MC34063 controller. Thanks for that go to Dave Jones over at EEVBlog. It was his video on the subject that introduced me to this chip. Other than a couple other designs that have had requirements beyond this chip's abilities, this chip has become my new one way to change DC voltages.

For the EVSE variant, the GFI circuit is from OpenEVSE (who in turn got it from a CR Magnetics application note), but with a couple of minor changes. OpenEVSE added a cap to the first feedback resistor, but theirs is 0.1 µF, which results in a roll-off for the low-pass filter that's too low. I've changed it to 0.01 µF. I've also reduced the voltage divider on the detection side, changing the 20k resistor to a 16k. This increases the GFI sensitivity.

For the splitter variant, there's a proximity transition detector that looks for someone pushing the proximity button on the inlet. If your J1772 inlet lacks the 2.7k resistor to ground, then you should close the INLET_PROX jumper, which will supply a replacement. The inlet pilot is wired much like the spec indicates - a 2.7k resistor to ground (after the required diode) permanently anchors the inlet in state B. A 1.3k resistor can be added in two ways. First, the "switch" connector can be used directly to connect it to ground. Alternatively, if either relay output is turned on, a transistor will ground it. The concept is that if you connect the AC power connector to 120 volt power supplied externally, then you can switch your EVSE on or off on demand. As before, if you supply AC power separately, then you should leave the switch connector open. If you simply connect the AC power input to the line side of one of the contactors (to take power from the inlet), then you should jumper the switch connector with a wire.

The two car signaling sheets are almost the same. The only difference is the addition of the outlet proximity signal. The idea is that when the inlet proximity button is pressed, that same action can be repeated on each car to insure that a rapid removal of charge current happens to prevent any arcing.

The only difference here is the addition of the RTC clock chip, battery and crystal for the EVSE variant, and the three digital pins that that change from GFI and GFI test to inlet pilot, proximity and outlet proximity. In each case, there's a two-way jumper for selecting the polarity of the common pin for the LCD backlight, and a jumper that allows DTR transitions on the FTDI port to reset the controller (which would allow a traditional arduino bootloader to upload sketches over serial). New to v3.2 is a 4 pin header to bring the unused analog input pins (ADC6 and 7) along with +5 and ground out for potential enhancements in the future.

International orders

A couple of folks from outside the United States have wanted to buy stuff from the store. Unfortunately, Square doesn't have any plumbing for that. But I'm happy to entertain international orders manually with PayPal. Here are the details:

I generally ship stuff with USPS flat-rate boxes, and internationally those cost $25 instead of $6. I will fill out an honest customs declaration form, so if that results in import duties for you, then that's something you'll need to deal with.

Of special note for those in the EU: My products are NOT RoHS compliant. I use tin-lead solder paste and have no plans to switch to RoHS solder any time soon. My understanding is that that means my products may not be imported into Europe. I'm not an expert on RoHS regulations, so whether my understanding is correct or not, I have no idea.

I'm willing to ship to any country that qualifies for the USPS flat rate $25 price. So far as I know, that's pretty much any country that isn't on the "naughty list," like Cuba, Iran or North Korea.

Next steps for Toast-R-Reflow

Toast-R-Reflow was a success, but I think I could take it to the next level. I think there's a market for a hobbyist tabletop reflow oven that would cost $200 or less, be able to handle up to a square foot of PCB, and be able to successfully keep up with a RoHS paste profile. 

The trouble is that the idea isn't quite to a KickStarter project level. There are two important parts of the project that I don't know how to scope: custom heating elements and custom sheet metal enclosures. 

Everything else is doable. But without those two, there's no product. 

Who does one even talk to about those sorts of things?

Raspberry Pi power supply

I've been playing with an RPI. So far there's one thing in particular that irks me: its power arrangement. 1A+ @ 5V over µUSB? Really?

No me gusta.

So I've got an idea to design a "cape" (or whatever you call RPI add-ons) to sit on the expansion interface. The board will be as small as possible, but will have a 2.1mm barrel jack and take in 6-12 VDC and supply up to 2A @ 5V regulated by a LM3485 buck converter. Mischief managed.

The board will have a "long tail" DIP jack to mount on the bottom which will allow you to stack other expansion boards on top.

OpenEVSE II - high powered designs

A look ahead past the OpenEVSE II prototype.

I got a hint from someone on the OpenEVSE group to consider looking into FPC/FFC instead of ribbon cables. I like what I found, so the 0.2 version at this point is being designed to use FPC as the mezzanine cable.

The only other change will be that the relay pins will be directly connected to the controller (along with the AC Test pins), and for the HV+Relay board, the switching transistors will move down to the HV board. This facilitates the design for...

The relays on the relay board are only rated for 30A. If you want to make a beefier EVSE than that, in the past you've had to do a custom design with pilot relays and contactors. I designed a contactor adapter triac board to try and ease that a little bit, but for OpenEVSE II I've designed to offer two options for the HV board. The HV+Relay board will work at both L1 and L2 (since the relay coils are DC), and be useful for designs up to 30A. For higher powered designs, there will be a contactor driver HV board instead.

The contactor version of the HV board will have the same FPC/FFC mezzanine connector, but on the HV end, there will be 3 2 pin screw terminals: AC Line, AC Test and Contactor. The first two are connected to the line and load sides of the contactor, and the third is connected to the contactor coil.

Most contactors have .25" QD terminals, plus for the line and load points, large box-screw terminals for the high current path. You would assemble a high-powered EVSE by running 22 gauge wire with QD terminals from the contactor to those points on the HV board. The rest is the same as the relay version: You run a bonding wire from one of the grounded chassis mount screws over to the ground bus, you run the FPC/FFC mezzanine cable to the logic board, connect the CTs and the J1772 pilot line up to the logic board, then wire the line and load side of the contactor, passing the J1772 hot lines both through the GFI CT and one (only) through the ammeter CT.

The board is much smaller than the relay board - it's just under 2 1/2 x 3 inches. It should make it much easier to put higher powered EVSEs in smaller chassis than was previously possible. In fact, I wouldn't be at all surprised if a 50A rated unit couldn't be constructed in Chris' chassis!

Hydra v3.1 a success

The Hydra v3.1 boards came back from OSHPark, and they're a success. Since they had both the new DC-DC converter circuitry for the +5 and -12 supply, and the new 4 transistor pilot generator circuitry, it was quite a risk, since both are fairly critical subsystems and both were completely untested.

The power supplies work perfectly. The pilot generator works, but there is one little caveat - If you attempt to program the controller while 12 volt power is present, the TTL pilot output line will float, which will wind up turning both transistors on, shorting +12 to -12. The workaround is to not program the controller with power applied, but the fix is to add a pull-up resistor to insure that the floating bases aren't allowed. That will be the design going forward for both OpenEVSE II and the Hydra.

Hydra v3.2's design is well underway. It's going to be a two-board system, like OpenEVSE II. It's going to have an inter-board FFC cable. The mezzanine board has all of the AC systems - the 12 volt power supply, 2 contractor triac circuits, and two MID400 AC testers. The top board is the logic/display board that it more or less always has been. Going forward, I'm dropping support for 12 volt relay coils. DigiKey has 208/240 volt coil 30A contactors with the same form factor. It's not going to support L1 (120v) charging anymore, but nobody in their right mind would build a Hydra for L1 anyway.

Meanwhile, the OpenEVSE II display/logic board prototypes have been ordered. The prototype HV boards should arrive this week.

OpenEVSE / Hydra IEC 62196 type 2 ("Mennekes") adapter

I took some time out at lunch to sketch out how to best adapt OpenEVSE and Hydra relay outputs for Europeean IEC 92196 Type 2 connectors.

These differ from J1772 Type 1 connectors in four ways, two of which I'll call out for this discussion:

1. They have a locking actuator that the EVSE can use to prevent removal of the plug while charging.
2. They can take up to three phase (plus neutral and ground) power, and so typically require a contactor.

OpenEVSE and the Hydra today only have 12 volt DC relay control outputs. So how do you bridge that gap?

This circuit has an input jack and two output jacks. The input jack comes from the Hydra and/or OpenEVSE board and carries 12 volts and a relay output (assumed to be an open collector). The two outputs are an AC contactor switching output and the 3 pin lock actuator output.

The lock actuator is rather simple - there are two lines that are simply fed into a 12v motor. You apply a 12 volt, 300 ms pulse to lock it, and a reverse polarity pulse to unlock. The third line is a switch that is closed when the lock is engaged (the other side of the switch is shared with one of the motor lines).

The circuit above uses an ATTiny85. The controller watches the relay line for changes. On an off->on transition, it will engage the lock, verify that the lock switch is closed, and then turn on the power. On an on->off transition, the contactor is turned off, then there is a brief delay, and the actuator is unlocked.

The circuit above is a 4 transistor "H bridge" for reversing a motor. The bases on each side are pulled up to 12 volts and an open collector driven from the TTL output of the controller is used. This insures that the bases of the H bridge never see other than +12 volts or 0 volts. Intermediate states could potentially turn on both transistors, creating a dead short, which would not be good.

The controller simply raises one of the lines to lock, and raises the other to unlock. To sense the lock state, the controller sets both outputs low and then reads the sense line. That line is configured in the controller as INPUT_PULLUP, so when the switch is closed, the input will read LOW. A zener clamp is used to prevent the full 12 volts from being seen at the controller when the motor is active.

The contactor controller is the same circuit as the contactor adapter I sell in the store today.

The remaining significant difference with the Mennekes connector is the handling of the PP signal. This line should have a resistor to pull it up to +5 and then be fed into an unused analog input on the ATMega. The Hydra has two left - ADC6 and 7. OpenEVSE is not quite so constrained - ADC2 can be used. The trouble is that you'd have to patch the OpenEVSE board both to get to the unused ADC input as well as pull out the 12 volt power supply (the relay output jack has +12 on it).

Chris is working on a new OpenEVSE board specifically for the European market with all of these items addressed, but if anyone wishes to retrofit their old OpenEVSE, that's what you'd need to do.

OpenEVSE II layout

I've got a layout set up for OpenEVSE II. The goal is to make something that will be compatible with the OpenEVSE Chassis. And I think I've come up with just that.

The logic/display board is almost a backpack (that is, 80x36 mm intended to mount behind 2x16 character LCD module), but  it will cantilever off the two "short" ends by .6 inches. It will still fit that way onto the mounting pins for the i2c display module originally intended to mount on the lid.

You'll have to drill 4 new holes in the plastic interior mounting panel to accept the HV/relay board, but that should not pose any particular difficulty. In fact, you'll be able to use the mounting hardware that comes with the enclosure that's intended to mount the OpenEVSE board.

Like Chris' DIY 30A board, there are two PCB mounted relays on the board, so the overall amount of wiring is reduced. You just wire the HV path through the relays. The AC Power and AC Test wiring is taken care of by the board.

OpenEVSE II

I sat down today and sketched out a new design for OpenEVSE. I call it OpenEVSE II. It's completely compatible with the existing OpenEVSE firmware. It is designed as two boards - the first is an HV + Relay board that has the AC power supply, two 30A relays and the AC test functionality. You connect the AC power and J1772 cables to QD terminals in the tops of the relays. The board itself has a 4x2 DIP header that connects via a cable to the controller / display board, which is a semi-backpack board (it's slightly larger than the 16x2 LCD that's it's sandwiched to).

Here are the schematics:

The design isn't yet proven, so don't be in a hurry to go build one just yet... But I hope to have it ready shortly.

Toast-R-Reflow and PowerSwitchTail

I just discovered the PowerSwitchTail on AdaFruit. For what it's worth, it appears to be an almost "consumer friendly" solution to the problem of controlling reflow ovens or hot plates with a micro controller.

If you wanted to build a reflow oven without actually taking the oven apart, you could connect its power cord through one of these, and then connect one of the two Toast-R-Reflow outputs to the control input, turn the oven's own controls on as high as they will go, and go.

The only reason it might not work is if the oven's own temperature controls do not allow the oven to heat as hot as the reflow controller desires. Toast-R-Reflow's SnPb profile goes up to 225°C, which is 437°F. That's close, but not quite 450°F, so if the oven's controls go up to 500, and they're even in the ballpark, it should be ok.

The one gotcha is that Toast-R-Reflow's controller comes with the LED series resistors on the controller, while PowerSwitchTail has them as well. If you want to use a Toast-R-Reflow controller with PowerSwitchTail, you need to replace the output series resistors with either zero ohm resistors (yes, those exist) or wire jumpers. As a special option (and at no extra cost), if you want a Toast-R-Reflow controller with no output series resistors, just mention it in your order and I will do that (if you do, DO NOT use the controller with unmodified Toast-R-reflow power boards - you will destroy the opto-isolator!).

Hydra II EVSE v3.1 schematic

Here are the latest schematics of the Hydra II EVSE variant.

In the upcoming 3.1 board, MC34063 DC-DC converters will replace the two integrated modules that were present in previous designs. Also, the "Sziklai pair" switching transistor architecture has replaced the previous Op Amp and switch chip designs for the pilot generator. Lastly, provisions are in place for a future relay test input connector. The purpose of that is to be able to detect the presence of AC power on the J1772 hot lines for each car. If power is not present or absent when expected, then that's a (potentially serious) error.

I have what I hope is a finalized board design, but wish to check the layout of some of the new parts against what I get from DigiKey. Once that's done, the boards will be ordered from OSHPark, and then two weeks later, the prototypes will be built.

Similar changes have been made in preparation for the v2.3 board as well (which is the original "splitter" variant of the Hydra). If the changes to v3.1 work, then the process will be repeated for v2.3 as well.

Hydra HV mezzanine board ideas

I've got an idea for a helpful item for building high-powered Hydras.

The bad news is that it's a little pricey.

Here's the concept:

This board combines the 12 VDC power supply, 2 contactor adapters and stuck/open relay detectors for two cars onto what I call a "mezzanine" board. The board is slightly larger than the Hydra logic board, but there will be holes strategically placed so that you can mount the Hydra logic board on 1.5" standoffs right above this board. One side of the board is all high-voltage stuff, the other side (segregated by the optoisolators and by the power supply module) is low voltage - the intent is to keep those two portions as isolated from each other as possible.

The board includes separate fuses for each contactor adapter circuit, plus a slow-blow fuse for the power supply.

The relay test circuitry is designed to allow the Hydra to detect and complain about relays that are stuck open or closed - that is, if the state of AC power to the car does not agree with the state of the relay output from the logic board, then that's a serious error that requires immediate attention. This version of the relay test isn't as comprehensive as OpenEVSE, but it can't be because the GFI coil for the Hydra is located prior to this circuit, so attempting to use ground as a return will trigger a ground fault. It also can't be as comprehensive as OpenEVSE because there's only one relay per car for the Hydra.

Unfortunately, just the board by itself from OSH park is $21 because it's so large - and that's my price. The power supply module is $20... This board will wind up costing almost as much as the Hydra logic board if I sell it in the store.

Still not sure what to do about that.

DC-DC conversion, redux

After posting about this on the EEVBlog forum, I was pointed to a particularly neat chip for DC-DC conversion... The MC34063. It can do boost, buck or inverting conversions, and has a very wide voltage range. The next versions of the Hydra are going to be made with two of these - one to generate 5V@500mA and the other to generate -12v@50mA. J1772 pilot generation isn't so fussy that it requires perfectly balanced supplies or anything like that, so I'll just use the incoming +12 rail for the "high side" of the pilot square wave and the converted -12 volt supply for the "low side." With the Sziklai pair architecture (4 transistors per car) and the DC-DC converters, the BOM price for a Hydra board has probably dropped by almost $10.

I'm waiting for some of the new parts to come in the mail to verify the board footprints before ordering the next board version.

LCDuino on eevblog mailbag

Dave from EEVBlog checked out LCDuino on today's mailbag video!

I loaded AdaFruit's "Hunt the wumpus" sketch just so it wouldn't be just completely inert when he plugged it in. I don't know how to play that version of the game either. :)

For those looking for more information about LCDuino, try this link. It just searches this blog for the LCDuino label.

LCDuino is a mature design at this point, so don't delay, buy one today!

Oh, and "Go Cats!" was the Geelong Cats. I guess he wasn't expecting an American footy fan.