Teardown of a HomePlug Adapter


#81

So how can we know which version of the adapter we buy. Is it possible to use a v2?


#82

Hi Joel:

We just finished up testing the prototype of a new revision of the OROV controller board, that will accept either version of HomePlug adapter. It will probably be available in the store about the beginning of the new year.

If you buy a OROV kit, it will come with an appropriate set of adapters.

If you buy a current 2.5 controller board from the store now, it will only accept the V1 adapters. If you can't find a source for the V1 adapters, I would wait until the new controller boards are in the store. If you already have a 2.5 controller board and can't find a V1 adapter, send a letter to David or Eric - they may be able to find one for you.

Some retailers still have the older V1 adapters- perhaps people can post here their experiences of buying the adapters on-line and figuring out which one they receive. For instance, NewEgg had Tenda Adapters on sale during the USA Thanksgiving holiday, and I bought three pairs. I should be receiving them soon, and will find out whether they are V1 or V2.

I think there are some small differences in the front of the adapter between V1 and V2. Maybe Eric can post some pictures here if he's got photos of both?

I hope this helps.

-Walt


#83

Hi Lomax and all:

As Douglas said above, you can hang pretty much anything you want off of the tether without affecting the Homeplug adapters, as long as you isolate it from the RF signal of the adapters through the use of a choke / ferrite / inductor or whatever. If you surf the web for data on HomePlug AV (that's the protocol we use), you can find it's frequency band.

With regard to what happens to a DC/DC converter when you drop below the minimum input voltage: you have to look on a case-by-case basis, but most will shut down. Look through the documentation for something called under-voltage lockout (UVLO). That's the thing that shuts the converter down.

I agree with David above that the best way to deal with tether power is to have it charge an on-board battery pack, which then handles the peak current loads of the motors. This does add some complexity, but gives the most flexible overall solution. Last spring I had a design for this sketched out for OpenROV, but it got put on the back burner when we did some tank testing of motor/prop combinations and realized how inefficient the props we were using at that time were. Once we switched to the Graupner props, the battery life went way up, and we decided that power-over tether was not as important as some other things that were brewing at the time, such as switching the low-level I/O architecture from an Arduino Uno to an Arduino Mega.

We'll probably revisit the power-over tether question some time within the next year, so we're always curious to see what various builders are doing along those lines, and what kind of results they get.

If you're going to experiment with power-over-tether, please pay particular attention to safety, especially if you're going to be running in salt water with divers around.

-Walt


#84

Hi Walt,

Nice to get a reply from the man himself! Not to keep flogging this horse until it dies, just wanted to say that your support for the "on-board batteries with surface power feed" concept makes me think that maybe my initial plan wasn't so bad after all. I allowed myself to be dissuaded from this design by a friend who allegedly designs this kind of system for a living - his response was the typical "Oooh, that's very complicated, gonna cost ya! But your idea is fundamentally wrong, you should do X instead!". I merely wanted to know what type of device one would use to balance a static "low" current supply (say 12V 10A) with a LiPo battery pack for when current draw exceeds 10A - ideally in such a way that when power draw is below 10A (which it will be 90% of the time), the remaining capacity goes towards (re)charging the battery pack.

Having thought about the problem for some time this seemed to me the ideal solution for powering an ROV, the benefits being too numerous to go into here, but thin tether, low ROV weight and indefinite dive time are the main ones. Maybe I should go back to that plan - I count 4 votes for this as a good solution (including my own), and only one against (my "friend").

Do you, or anyone else here, have any suggestions what type of circuit would be required for this task? Clearly a LiPo charger would be required, but what decides when to charge and when to deliver? The only thing I can think of is monitoring the current draw and having a relay that switches between the two modes - which sounds a little too primitive!


#85

Ok, we'll se.

Actually I bought all the electronics independently from different web sites last week, including the controller board. Probably I will receive the homeplug adapter next week, and then be able to figure out which ver it is.

http://www.ebay.co.uk/itm/121174046436?ssPageName=STRK:MEWNX:IT&_trksid=p3984.m1497.l2649 - This is the one I bought, don't know which ver.


#86

For tether efficiency the higher the DC voltage the better, up to whatever safely limit you feel comfortable with. For adult hobbyists 48V is probably a good limit. Maybe 24V for kids.

Choosing the data choke value is easy if you have reasonable guesses for tether impedance and data signal frequency. The choke should be higher in impedance than the tether at the minimum data frequency. For conservative guesses of 300 Ohms and 100kHz the minimum inductance is:

L = Z / (2 * Pi * f) = 300 / (6.28 * 100k) = 477 uH

For 4A you are looking at something like this:

http://www.digikey.com/product-detail/en/2300HT-471-H-RC/M8933-ND/775472


#87

Gosh, great stuff Douglas, you just saved me a couple of hours trying to figure this out by myself - many thanks! But what about the (small) capacitance of the cable? Can this be ignored?

48V has indeed been chosen precisely because it's the highest voltage I feel relatively comfortable to use when mixing things with salt water. It'll still give you a jolt though - and can be dangerous if the current passes through your body (arm-arm or arm-leg) - this can very well be lethal! Also, don't underestimate the fireworks that will ensue if you accidentally short out a bank of large 12V batteries :o Even 12V of sufficient amperage should be treated with respect, especially at sea. And if you're using lead-acid batteries there is the added risk of hydrogen venting, which can produce even bigger bangs if your batteries are in an enclosed space. You have been warned!


#88

The capacitance of the cable is part of the cable impedance. A quick Google tells me that 100Base T cable has a 150 Ohm characteristic impedance. So if that is the wire you are using in your tether it has half the impedance of my 300 Ohm guess. That means you only need half the inductance or 239 uH. More inductance is better, maybe 2 or 3 times, but that makes for a bigger heavier and more expensive inductor.

I also work with model trains. They commonly use 18VAC on the rails and are generally considered safe for children. Without using any metal you are not likely to notice the voltage unless you lick the rails.


#89

Ah yes, of course, I got confused since the capacitance is not part of the formula you posted :) Thanks again, you have been a great help!

18V won't be noticeable unless you increase the conductivity of your skin by wetting it with something conductive - such as salt water. 48V is a different matter, you may notice a tingle even on dry skin, and if your hands are dripping with salt water you'll get a small shock. It's still perfectly safe of course, but you do notice it. The danger isn't the shock though, it is the current passing through your heart that will cause problems. This only happens if the current passes through your chest, for example if you're standing bare-foot in a puddle of salt water and grab a naked 48V conductor with wet hands. DO NOT EVER TRY THIS! 300mA of current passing through the heart is enough to cause fibrillation, and possibly death. Low frequency AC is more dangerous than DC in most cases because it overcomes the resistance of your skin more easily, and you will notice AC at lower voltages. At very high frequencies AC becomes pretty much harmless as your body acts as a capacitor, preventing sufficient current to pass through to cause fibrillation of the heart. All this is academic of course, I'm sure everyone here knows better than to expose themselves to undue risk... ;)

Edit: I had to look this up, but according to Wikipedia deaths have occurred from voltages as low as 32V. This is extremely unlikely of course, and must have been from an AC source.


#90

Of course the electricity doesn't have to hurt you directly. It can startle you enough that you fall off the boat and drown or get into the propeller...


#91

Ok. I got the package today, So now I'm owner of a v2 homeplug adapter (pried it open to see, [before that, I asked the shop if the knew which ver, which they did not do]).

Too, bad I already ordered the Controller board.

So now I will wait for openrov team and others to post a guide how to use v2.

Is it due to physical limitations you cannot use it with the controller board? Cus I am planning putting the electronics in my own rov (which means I could put the adapter outside the controller board with wires going to the right spots on the controller board, if this is possible.

I will stay tuned

Joel


#92

Hi Joel:

If you don't mind adding some wires, then yes, you can use the 2.5 controller board with a V2 adapter.

Electrically the V1 and V2 adapters are pretty much the same, but mechanically they are sort-of mirror images of one another. So the V2 adapter won't directly plug into the 2.5 board, but by adding some wires it will work just fine.

As to where to put the wires, you can either work that out on your own by looking at the adapter and the schematics for the 2.5 controller board, or you can just wait a couple of weeks. Once we start selling the next revision of the controller board, we will be releasing the schematics for it, and it will show how the V2 adapter is attached.

-Walt


#93

Ok, thanks.

Surely I will need some help, but I have not received all parts yet, so my questions can wait for a while.


#94

How about using something like this to regulate charge/draw from an onboard LiPo battery pack with surface power supply: http://www.linear.com/solutions/LTC4000 It's only $6 but needs a couple of high power MOSFETs to do the actual switching. And a DC/DC converter is also required of course. Still a surprisingly cheap and powerful solution. This circuit example looks very close to what we need: http://www.linear.com/solutions/1700


#95

For Joel and all those looking for Tenda V1 homeplug adapters:

I just received the homeplug adapters I bought from NewEgg over the US Thanksgiving holiday. They are V1 models. Here's a link.

I don't know if there are any issues shipping them to addresses outside the US.

-W


#96

Yes, the LTC4000 is a good starting point for an on-board power control system. The design I sketched out in the springtime for OpenROV used a LTC4000-1, since that chip has a crude maximum-power-point (MPP) tracker on it. You probably need such a feature if you're going to run a long tether and don't want to burn up all your power in the resistance of the tether.

-W


#97

Unfortunately; http://gyazo.com/92e5bf9fadf02145c9498dc6c9011d25

I
think the best way is to learn how to use the v2.


#98

Use a wide voltage converter.

Example: (Prices in taobao.com are quoted in Chinese yuan.)

http://item.taobao.com/item.htm?id=9481705780

Concerning the need to buffer the output, I encounter a problem of high current surge and blow the fuse when things are first power up.


#99

http://gyazo.com/6e12110630aae27b4987b1506026e1a3

So how is the work with the v.2 going?


#100

Hello.
I have just done the tear down and testing on the TP-Link TL-PA411 home plug adapters.
These are 500Mbps adapters. But interestingly, before I removed the capacitor to add my own low voltage power source I measured the volts it was running at and found out that these devices run at 12v, not 3.3.
So I gave them 12v and everything works fine.So far I have only tested with a 1m mains power cord between them (without power ofc) and the software reports 330Mbit connection between them although the rj-45 connections are only 100Mbit on these. Further testing with lull length tether to be done later.