Teardown of a HomePlug Adapter


#61

Brilliant, thanks Stefan.

I'll be giving this a go over the next day or so, hopefully will have some positive results to feedback with.

I'm taking a different direction with my own design of ROV, plus using a Raspberry Pi and an Arduino. I'll be supplying some power via the tether (two pair cable), so hopefully should get things communicating!

Jon


#62

Yes, Walt, tether lenght and transmission rate are surely most important. For some reason, the TP-Link power line utiltity that comes with the adapters doesn't recognize the modified adapters, so I couldn't examine the transmission rate using this software. However, I'm able to transmit 1080p at 30fps from the ROV with no problem. The Windows performance monitor tells me that this is at a transmission rate of some 50 MBits/s. My tether still is the orange one that came with the kit (50m), but I would really be interested in the maximum tether length the TP-Link adapters support.


#63

Jon, good luck for this! Let me know how you are doing with your modification and if you can get it working.

-Stefan


#64

Hi All, this is awesome work going on here!! We learned a fair bit about the Homeplug AV adapters as they were in the Intellon INT6300/INT6400 (more power hungry!) days in the SCINI ROV. In 2009, we performed a dive at 301m and still maintained successful video transmission (http://scini2009.mlml.calstate.edu/2009/12/southernmost-deepest-rov-dive-ever/).

I tested a 400m tether in the lab with the same equipment on each end but was unsuccessful in establishing a link. Of course, SCINI sent AC down the tether and required noise filters to achieve 300m.

A powerline adapter engineer that I spoke with in 2007 stated that all Homeplug devices transmit at -50dBm/Hz. I assume this would remain the same for IEEE P1901 devices since that is for FCC regulations. He also confirmed that he believed Devolo's claim on their website that they could transmit up to 750m over coaxial (non-powered) cable, but that they had not tested at this distance. Good prospects for sending high throughput very deep!

Cheers all,

David


#65

Hi All,

I've been doing a bunch of testing with EOP over a dry pair and I was able to get a connection over 1000 feet of tether.

Also, the one kind of 200 Mbs adapters I have worked better than any 500 Mbs adapters I had.

Hopefully I'll have time in the next few days to do a real write up about it.

-Jake


#66

Jake,

That's great to hear and I can't wait to see more details on your testing! Was the 200Mbps adapter you used that had such good performance the same one we've been talking about (the MHP-EA200 or the Tenda P200)?

Eric


#67


Hi,

i received a set of adapter today and check it. Only thing i hat to do is to remove the capacitor you marked with x an connect the tether to the original power connection.

WORKS


#68

Hi,

Great work on the teardown Walt!

I just got the Tenda HomePlug adapter myself, and it seems to be an upgraded version of the adapter (V2.0). The board now has 6 and 4 pins connecting the two board together.

So my question is, how should I connect the tether and the power to the board?

Any help with this will be greatly appreciated.

Cheers,

Ola





#69

Ola,

This is an extremely interesting development! I was unaware that there is now a v.2 of the board.

I'm actually a bit worried, because we just committed to a very large order of custom OpenROV Controller Boards that are specifically built to fit the v1 header style, and I'm not sure how easy it will be to adapt the new layout.

If you still have it, would you mind taking a photo of the box this came in? Also, I'd be very interested to hear when and where you got the board from.

As for what you've got, I imagine the board will work just fine (and may even work faster or more reliably) but we'll have to try to get one our selfs to do some testing for you.

Thank you very much for making this post- we probably wouldn't have known about this for quite a while (perhaps until it was too late and we had already made purchase commitments) without this project being open source with a bunch of eyes out there.

Thank you!

Eric


#70

Okay- I just got off the phone with Tenda and learned a bit about he changes between v1 and v2. It looks like the reason for the change has to do with power management when the board is idle. This will have no effect in performance for our application, but the differences in physical layout will likely mean we'll have to redesign the layout of our Controller Board for the next rev:( According to the guy I talked to, the change between versions happened in July.

I'm getting a few samples of the new version sent to us at OpenROV, so once we get them, we should be able to document how to hook it up properly.

Stay posted!

E


#71

I am planning to make this mod this weekend, using Tenda P500 pairs.


Any comment? Will it cause a problem if the Tenda 3.3V board is not isolating supply? or the fiber modem's switching P.S. is not isolating supply?

Although my purpose is slightly different from ROV robots, but also using PLC board. Please forgive me for posting semi-out-of-topic here.

I have a few issues uncertain:

  1. When the circuit power up, the initial charging current of 2 remote switching P.S. will cause a surge. How to reduce this surge to prevent the thin twist pair overheat?
  2. Is Greinacher circuit better for voltage doubling? (48Vac to 96VDC r.m.s.)
  3. What is the optimal value for the 2 capacitors at the voltage doubling circuit?
  4. Do I need to add any Band-pass-filter anywhere?
  5. I assume that, if the fiber modem DC12v p.s. accepts 100Vac will also accept 100Vdc, as most switching P.S. have full bridge rectifier at the very first stage. Am I correct?


#72

I got a pair of Tenda P500 at around US$33.

http://item.taobao.com/item.htm?id=35856736774


#73

oh no, after teardown, P500 which pretend to be 500Mbps phy speed.

using 200Mbps phy speed BCM60321 chipset !!!!!!

Although to got 91.25M MAC speed, near the 100M theoretic maximal,

but they seem to be cheating me.

Here is another link for Tenda P200, US$25.50

http://item.taobao.com/item.htm?spm=a230r.1.14.11.xaABR2&id=22480448311


#74

Done!

I decided to use Ring type transformer at both side. CNY45 each.

http://item.taobao.com/item.htm?id=14186729512

I choose 28V-0V-28V secondary coil, 56Vac just below the 57V limit of Power over Ethernet standard, in order to deliver maximum watt over the tether.

  • REMEMBER: 56V can cause heart fibrillation and death, treat as MAIN.

Alternative: Isolated AC56V to DC5V converter.

http://item.taobao.com/item.htm?id=9391873906


Packaged in a TV booster box Cost CNY36.

http://item.taobao.com/item.htm?id=36211204723


90Mbps download / 82Mbps upload.

1ms latency only if connected to fiber modem direct.


#75

I must say that using PowerLAN/HomePlug adapters for combining data and power on the same single pair cable is a stroke of absolute genius - I seriously jumped out of my chair and threw my arms in the air shouting "OH YES BABY!" when I first read about it. Powering the ROV reliably and for long duration dives was the big missing piece in the puzzle for me and this solution fits so perfectly it's like it was designed for it - and not some hack by (very clever) DIYers :) Excellent work chaps!

Only, there is one thing that bothers me slightly... I was hoping to use a capacitor bank to buffer the line in order to handle the starting currents of my thrusters while keeping cable area (and hence tether weight) to an absolute minimum - but sticking capacitors on the line will kill the data signal dead :( Anyone have any suggestions how to work around this? I plan to send 48V down (which by the time it reaches the ROV will have fallen significantly) and feed a DC/DC converter to deliver a stabilised 12V output. Any thoughts?

Edit: What happens to a DC/DC converter when the input voltage drops below the rated minimum? Does it stop working alltogether, or continue to provide a (lower) voltage? Since the converter is isolated, maybe I could buffer the output instead? Would that perhaps even let me use a converter with a slightly lower power rating? I really can't see myself running all the thrusters, lights and thingymajigs on my ROV all at the same time and at full blast - but I don't want it to die a death if I ever did...


#76

Hi Lomax

Using a capacitor will not cause an issue for Data loss if you are planing on placing these in the ROV after the MediaLink has stripped off the data. My question would be why use a Cap with very limited storage when you could use a Lipo battery which is a very big capacitor anyway unless you are trying to reduce the weight in which case I would try using a small S3 Lipo and topping it off with voltage through the tether.

Regards

Dave


#77

Hi Dave,

Many thanks for your reply! I thought that placing a (large) capacitor in parallel with the load would suppress AC on the line? I'll take your word for it though as my electronics skills are more than a little rusty :S

I did originally design for an on-board LiPo battery pack, but after some discussion with a guy who works with battery management systems I got the impression that balancing the two power sources would be complicated and expensive (apparently it's not enough with a LiPo charger?).

Then I realised that I could pull the current I need over a thinner cable than I at first thought - even on a 100m tether I can manage all I need on a 2x2.5mm2 cable with a bit of margin. Sure, it will be heavy, but not terribly so - and I have come up with an ingenious plan for making it neutrally buoyant :) So I revised my plan (not for the first time - and not for the last time either!) and got rid of on-board power altogether. I now have a design which (with a 10-15% margin) can pull all its power from top-side via a DC/DC converter, which although expensive ($80) ends up being cheaper than a thinner cable, smaller converter + on-board power. This also greatly simplifies my design and gets rid of a lot of bulky (and heavy) parts - some of which I would have had to cram into the precious space in the pressurized part of the hull.

This is very much a learning process for me though and I'm always open to advice!

Btw - not all caps are that limited in storage :)

Edit: This is the DC/DC converter I've been looking at. My pressurised hull will be made of die-cast aluminium (I hope), so correctly mounted I should have no trouble dissipating the heat through the baseplate.


#78

I have not worked with these PowerLAN adapters but the usual way to handle this is to put a choke coil between the point where the data is extracted and the capacitor bank.

Just as a reference for power management, the small commercial ROVs I am used to working with get 150VDC through 18AWG tether conductors. The 150V is at the ROV and it is remote sense regulated from the topside power supply. The topside supply can provide up to 300V, so up to half the voltage can be lost in the tether. At these power levels the tether must be unspooled or it may melt, or sometimes if space is tight we put a small water sprinkled on the spool to cool it.


#79

It will be very interesting to the group to see how you make out with this as there has been much discussion on this topic. At this point, all the work has been going into getting the control board and software design to a stable level where other interesting developments for this project can then start to take place. Looks like you have a good choice on DC\DC converter. I assume that you would just use batteries in series on the top side or did you have something else in mind like a dc\dc converter from 12v up to what you need. I went looking for such a device a few months ago but never came up with anything. Please keep us posted on your developments. I had a peak at your die-cast aluminum boxes, keep in mine that square boxes do not take kindly to hi pressure forces exerted on them from the water, so depending on how deep you plan to go, you may want to do some testing on it before before putting in any electronics. good luck with your project.


#80

Thanks Douglas, tapping the data off with a coil was the sort of thing I was imagining, great answer! But I'd need to know the "frequency" of the data signal to calculate the RLC network that results - it will basically look like an oscillator?

150V would be great but I'm just too safety conscious :) 48V is the highest I'd feel safe using at sea (salt water is both conductive and corrosive!) and it can easily be produced from 4x12V batteries. I will not use remote sensing, but allow the voltage to drop. The DC/DC converter I'm looking at keeps going down to 36V and I'm dimensioning to have that arriving at the ROV end even at full draw, which I've estimated to be 8A with a healthy margin (20%) - the capacitors are there just to provide that little extra, and because we can :)

And thanks David, I'll be documenting my design and sharing the details - once I have an actual design that is! Currently working on the thrusters (which also differ from the norm!) and the power supply side, including the tether (for which I plan to use speaker cable for it's thin insulation, and a strength member of some sort). Only when I have a grip on these two will I progress to hull design. Power and thrust dictates everything else! Needless to say, I am not going for the OpenROV hull, though I'm hoping to make use of a lot of your software and some of your hardware, like the IMU for example. So the control side will be the same as the OpenROV, but pretty much everything else will be different :)

Edit: Re. the hull - yes, a tubular, or even spherical, hull would of course withstand much greater pressure. The idea to use a die-cast aluminium box (or boxes even, if my connector design proves viable) springs from several thoughts: 1) a lot of people are successfully using OtterBox enclosures, even at 30m depth (my target) 2) aluminium solves the heat dissipation issues, and thereby hopefully condensation problems as well 3) it would be much easier to work with, both in terms of mounting stuff inside and in making pressure tight connections to the outside 4) it's cheap! The current plan is to get one of the Gainta enclosures (they look quite sturdy!) and beef up the seal with an additional flat rubber gasket, make up two dummy connectors with a loop between them, head out to sea and lower it to progressive depths (10-50m) to check if it remains dry inside. This won't happen until sometime this spring though!