Data over battery line

Do a google search for 1-wire. Maxim and others make inexpensive developers kits and interface products.

On a sunny day (Tue, 07 Feb 2012 10:03:12 +0100) it happened Marco Trapanese wrote in :

At 100kHz you would nee an inductor to prevent the RF from shorting. Its resistance would be prohibitive at 30A. Maybe better use some of the shelf 430 MHz modules and go wireless? Or at least use a much higher frequency, 100 MHz or up. Sometimes an extra wire (copper) is the simplest solution, bet there is space where there is a 30 A conductor.

Il 07/02/2012 11:42, Jan Panteltje ha scritto:

I agree.

I know this, but the customer requires neither extra conductor nor wireless... So I'm looking for a solution, if any.

Marco

Of course the battery will act as a capacitor.... that?s what they effectively are.

The issue is not the battery as that can be disconnected, it is the devices being power by the battery. They must have enough capacitance to deal with the added noise on the line.

You just need a way to disconnect both ends(because of bi-dir) and put your comm stubs inbetween.

device----+-------wire----------+----battery

where the + are the comm stubs.

Essentially you disconnect the wire when transmitting a low which causes the voltage to drop.

Dealing with the switching, the high currents, etc really isn't worth it though. It doesn't cost much do add a 30AWG comm wire. It will be much cheaper, easier, and safer.

(The above scheme is basically a TDM but 1 signal is actually power)

You can get around some of the issues with switch by not completely disconnecting the wire(so your low is not 0V but maybe Vcc/2). Since you are not transmitting much data and you can do it at relatively high rates you do not need much capacitance added after the stubs. Just enough to keep the voltage from drooping too much during transmission.

Several ways come to mind .. the cheapest could be to use the 'implied' third wire and drive the two battery lines against it.

Low bit rate?! I could do that at a distance of 2m for $.20 !!! Already did similar for a full-duplex wireless headset for the cellphone industry. Energy so low the test labs couldn't 'see' anything going on.

Best solution will be determined by cost/volume requirements.

Straightforward, is to insert band impedance between the battery and line and between the line and charging unit.

Subtle is to take advantage of the impedance inside the battery, measured [I HATE garage designs, but low volume production they work!] and work against that.

What is budget for parts? I know it can be done with less than $50/ link, but you may be thinking more a long the line of $.50/link.

What ever you do keep in the frequency domain.

You mean an inductor in the line?

I was wondering if he could use a resonant circuit. Put a series LC on the end and pick-off at the mid-point. Then amplitude modulate the 'signal'.

George H.

Il 07/02/2012 11:31, snipped-for-privacy@gmail.com ha scritto:

I know 1-wire protocol. I'm concerned about the battery connected to the bus. Do you think it isn't affected?

Marco

The cable may be less than 2m but the line of sight distance should be more like 1m. Using some 555s, an opamp active bandpass at the carrier frequency (~100KHz band), and some ferrites, you can make an NFMI transceiver. It's done all the time for shortrange communication and low throughput telemetry systems. NFMI stands for Near Field Magnetic Induction, which just means magnetic field generation and pickup, usually using ferrite rods. These systems are relatively low power too, it doesn't take much to communicate over 1m.

It would seem to me a 2m piece of wire and one-wire comm would be far easier.

Fred - you ever actually design/work with this kind of stuff? Way back when, I was involved with human performance research. We has a couple of rigs from a company in Finland called Polar. Had a strap that went around the chest, just under the pectoral muscles, that picked up an EKG signal then transmitted it to either a "watch" or a receiver module. They don't make this product any more, and I'd like to duplicate it. Also looking at longer distance device, will probably use 433/900 MHz off-the-shelf modules for that.

On Feb 8, 1:32=A0pm, Fred Bloggs wrote: (snippity-doo-dah)

Bot a complete answer but you might start looking here

also look up Nordic 'Ant' protocol.

There is not much margin in it

for $20

--
We have failed to address the fundamental truth that endless growth is  
impossible in a finite world.

use a device similar to a clamp on current probe to inject THE 100 kHz signal onto the batter circuit and receiver it at the other end with similar current probe. THis way you don't care how low the batty Z is.

Mark

Not really. I made a circuit 35 years ago that sent a signal up

300A+ charging leads to the charger, was a high temperature interlock.

Not at all. Batteries have a complex impedance you can play against at fast rise times.

Maybe? Technique I used basically shorted the battery for a very short time, resulted in pins on the DC line that could be picked up at the charger, a couple metres away down the charging leads.

IN a dirty noise environment as the DC converter was mains commutated SCR controlled bridge. A huge inductor in series with the charging lead probably helped, but tests showed the battery was not a short circuit for high frequency.

I remember the recovered pulse had to be lengthened to trigger a

555's reset pin, it was less than 500ns or so, long time ago.

No space for a special connector assembly with an extra wire?

Grant.

On a sunny day (Thu, 09 Feb 2012 19:23:37 +1100) it happened Grant wrote in :

Seems like amateur crap to me. Shorting a battery.

That sounds really good on the surface, but you're building a transformer that will have a high turns ratio. Works fine for the data, until you hit the starter and zap all the electronics on the other side of the transformer...or turn on the turn signals or...or...

SNR is gonna be horrible, even if you clamp all the voltages so the chips don't explode.

Quoting from your original post Cost is the primary target end quote. Nuff said.

We don't know all the constraints, but my advice is to tell your customer to add the extra wire. It's gonna be easier==cheaper than all the stuff you're gonna have to add to make it work over the existing battery wire.

Using a wireless link to replace a fixed 2m wire on a motorcycle frame, is insane.

Contrary to popular opinion, it's the system engineer's job to tell the customer when he wants the wrong thing.

Company got a patent on the technique, so it must've been bad ;)

Grant.

On a sunny day (Fri, 10 Feb 2012 00:02:53 +1100) it happened Grant wrote in :

There have been patents issued on perpetual motion machines.

You would not want to 'short' the sort of batteries I have been around, size of a room, the switch and wiring would evaporate immediately.

1 uS would be far too long, such a swicth would actually never switch off again, but just become some melted goo. I am talking from experience here... Was lucky not to get burned by flying hot copper.

Il 09/02/2012 12:59, mike ha scritto:

I don't, either ;)

I see.

Agree.

Yep. I always tell the customers when they're wrong. Sometimes they don't care, though. And I don't accept the job.

Marco