A Power Distribution and Switching Board for OpenROV


#1

So in a couple of previous forum posts, I've hinted that at OpenROV HQ we were playing around with a circuit board design to clean up the wiring in the ROV and provide a power switch. We just went to fab with a new version of this board that is (hopefully) suitable for wide distribution, so it's time to spill the details!

Things all started a couple of months ago here, where Brian, Eric, and Thomas were looking at ways of measuring the current draw of the entire ROV (as opposed to just the Cape), as well as cleaning up some of the wiring around the ESCs. Thomas went ahead with that design, and at the same time Eric and I started discussing ways to expand upon the concept. We decided fit the board into the 3"x3" bay beside the ESCs, which would allow us to integrate the Molex connector for the vehicle wiring harness, and add some other functions as well- the first priority being a power switch for the vehicle.

I laid out a first prototype of the board, which came back from the fab a couple of months ago. Here's a picture of it in an E-chassis at OpenROV HQ:


The initial design had a bunch of different options on it for power switching. After playing with the prototypes for a while, the design we settled on was to turn the vehicle on by placing a small DC voltage on the tether. By choosing this, I was able to simplify the board and add some additional features.

So I've attached a schematic of the revised board, as well as a layout. As soon as I figure out how, I'll get the Gerbers and the Eagle CAD files posted up on GitHub.

The board has the following functions:

1.) It cleans up the wiring on the back of the E-chassis.

2.) It provides a power switch for the ROV. The principal switch method will be to place ~5V DC on the tether to turn the vehicle on. A set of solder pads (J4) is provided if the user wants to use a bidirectional reed switch (like here), or some other type of switch.

3.) An Allegro ACS711 current sensor is used to measure the power draw of the ROV.

4.) The board has an 8-channel 12-bit Analog-to-Digital converter on it. One channel is used to measure the battery current and one to measure the battery voltage, leaving six spare channels for the user.

5.) The board extends the I2C bus from the cape, sending it out the 18-pin Molex connector to the outside world, as well as providing solder pads on the board itself. The A/D converter runs on the I2C bus as well.

6.) The board uses an ISO1541 chip to isolate the I2C bus from the cape. Power for the sensors comes from a local voltage regulator as opposed to from the cape. So if you do something to muck up your sensors, you don't fry your Cape at the same time.

7.) Some circuitry is provided to allow us to start playing with power-over-tether. It's pretty experimental right now, but you've got to start somewhere.

In about two weeks we'll have a couple of assembled boards to start playing with. Take a look at this, and let us know if you have any ideas to improve on it, or would like to get your hands on one. If enough people think this is a good idea, maybe I can convince David to put the bare boards in the OpenROV store.

-Walt

1382-PowerController2.1FinalSchematic.pdf (32.6 KB) 1383-PowerController2.1FinalLayout.pdf (29.3 KB)

#2

Great work Walt, I would love to get one to play with.

I was going to use a Solidstate relay to power up the ROV over the tether in much that same way but with all the added features, I like your solution much better. how long will Q1 stay on if there is a momentary power glich? I am using a slip ring on the tether and if there is any noise on the power over the tether it would be nice if Q1 would stay on for 1 sec or so without power from the tether. I was thinking of a capacitor on the base of Q1. that would charge up and hold it on for a short time. Actually a 2nd. look at the schematic it looks like pulling the base low turns Q1 on so that would not work in that location. do you have any thoughts on this.

Regards

Dave


#3

Hi David:

Good idea to put some filtering in the power switch feature- I hadn't thought about the use of a tether slip ring, and I don't believe we've been seeing any dropouts on the prototypes, so I didn't add anything.

As you first guessed, you can just put a capacitor between the gate of Q1 and ground, or it would be easier to put it across pins 4 and 5 of the optocoupler, or just use the J4 solder pads. The capacitor would get discharged on power-up of the ROV, then when you remove power from the tether, the capacitor would charge through R2 until Q1 turns off. The turn-off threshold of Q1 varies a bit, but as a wag say it takes 2 RC time-constants to turn the thing off. A 4.7 uF tantalum capacitor would give you just shy of 200 msec of holdover. I would think that would be great plenty.

-W


#4

Thanks Walt, if this becomes an issue that would be a very easy fix and 200msec,I think would be lots of filter time on that. Plus maybe add a resistor in series with the cap so as not to damage the OC1


#5

Hi David:

If you're concerned about the surge current when the capacitor is discharging, then find a convenient spot to hook it to the gate of Q1- then it discharges through R3 when the optocoupler turns on. Unfortunately there's not a convenient set of pads to do this.

I'm guessing that the optocoupler can't turn on fast enough or have a high enough CTR to create a damaging current spike from a 4.7 uF capacitor. But it would be an interesting test with an oscilloscope to see.

-Walt


#6

I would love to test one out when you get them.

John Stager


#7

Regarding the circuitry for power-over-tether, is that just a rectifier network and coupling capacitor for smoothing? Seems like a good idea if we wanted to send AC down the tether instead of DC, however, for AC (or even DC) to be feasible over such a large distance, we'll have to use higher voltages to reduce the effects of voltage drop, which means there might need to be a switchmode power supply in addition to that rectifying circuit. Since the external tether power would only really need to be there for 'topping up' the ROV's batteries, high current source wouldn't be entirely necessary so a small power supply could be used.

Since the batteries are wired in parallel on the outside of the ROV, it might be important (if we ever decide to charge them) to separate each pack and give each its own separate charging circuit (Li-po style) which might then utilize that single spare bus on the MOLEX connector and give us extra control over battery pack usage.

These are all just ideas, and I'm sure they've been discussed before, but I guess that's where I'm heading right now. Once I get further along on my own ROV, I'd love to give a prototype board a shot. Great work!


#8

So you bring up some interesting points.

The rectifier circuitry for power-over-tether was not for sending AC, but instead to allow for an arbitrary tether polarity- I didn't want to have the user worrying about whether they had the tether connected backwards or not. The final design was going to have a boost converter in the topside box, providing maybe 36 volts or so (depending upon what we felt was safe). At the ROV, under high power draw conditions you'd only see half this voltage, and you would then buck it down to the 12.6V required to charge the 3S lithium battery pack.

When I first sketched out this circuitry, I sized it for sending 40-50 watts down the tether, since it appeared at the time that we were going to need that much power to run the vehicle. Given the work that we've been doing on optimizing the props, it now seems like we won't need nearly that much power to keep the batteries topped off, and in fact the power-over-tether work has been put on the back burner for now, in order to work on some higher-priority stuff for the Rev 2.5 design. With the lithium batteries and new props, one should be able to get 2-4 hours of run time on a charge, depending upon how aggressively you're driving.

For the Rev 2.5 design, we're looking at ditching the Molex connectors, and switching to a simple DB-25 connector. This handles less current per conductor than the Molex, but with the new props this will be less of an issue. The highest current draw is in the battery connection, which we're going to mitigate by doing exactly what you mention above- bring each battery pack in separately through the hull penetration. This will allow us to do some things such as monitor the performance of each pack separately, and over the longer term, cycle each pack separately.

With 25 conductors available as opposed to 18, and using 2 of those extra ones for the second battery pack, we'll have 5 more wires available for payload use than the 2.4 revision.

This is all in the early planning stages right now. I'll probably start schematic capture for the 2.5 electronics in the next week or so. Our target is to release the 2.5 design in September. One of the things on our near-term to-do list is to create a good "roadmap" page, so that everyone can see what we're working on at OpenROV HQ over the coming months.

-Walt


#9

I would love to try one out on my rig. I just got it running and am having issues with voltage drops when operating at any speed setting other than “1”. It would be cool to be able to monitor my battery condition.

I potted in 4 extra wires for I2c expantion.


#10

Hi Walt,

Firstly, thank you very much for sharing!

As I am new to power MOSFETs, I am quite curious about the function of the Zener diode being placed between the source and gate of the Q1 (the p-channel power MOSFET, correct me if I am wrong).
Is it for discharging of the MOSFET or is it for some sort of protection?

Thank you very much
Best regards
Sun


#11

@sunjianxiao12.08:

The gate-source zener diode is to protect the gate of the MOSFET from overvoltages. ESCs can create voltage spikes on the power buss, and the zener diode ensures that these spikes don’t destroy the gate.

Please note that this power board design is for the very early (Revision 2.3 and 2.4) OpenROVs. All OpenROVs made since the fall of 2013 (Revision 2.5 and above) have used a controller board that has this power switch functionality built-in.

-W


#12

Dear Walt,

Thank you very much for your prompt reply.
Didn’t notice that 4 years have passed and OpenROV has been upgrading all the way!
Kudos!
Thanks for your explanation, would you mind me asking you further how will the spikes created by ESC being applied to the MOSFET (particularly to the Gate)?
I could not imagine such a scenario. Could you please illustrate further with the schematic?
1382-PowerController2.1FinalSchematic.pdf (32.6 KB)

Thank you very much
look forward to your reply
best,
Sun