Multiple signals on single Serial TX line

th 115kBaud

barrier

ected to the signal "TX"

short time. On the secondary side the HF signal is caught by the advance of a ripple counter (4017). The clock rate is significantly higher than the T X rate and higher than the reset signal, so when the short time slice with many pulses occurs it will set O3 output high for long enough to charge the "OtherSig" capacitor.

simpler idea? (I have looked at a microcontroller to decode the signal, bu t that takes to much current)

ones I have sketched up quickly has too many parts.

That was my first intention, but it takes a new microcontroller and the cod e for the serial protocol has already been written and this would not be we lcome by the sw guys

I could use a small microcontroller to act as the glitch detection, but tha t would add another microcontroller to be programmed which adds production costs

Cheers

Klaus

On a sunny day (Tue, 15 Jan 2013 13:45:42 -0800 (PST)) it happened Klaus Kragelund wrote in :

across the barrier

If that solution is more expensive / more complicated than an extra optocoupler, why do it?

across the barrier

optocoupler, why do it?

Because the optocoupler is a type with extra creepage, and it takes up 50mm2 of realestate. A couple of tiny logic ICs and some 0402s can be done in 10-20mm2

Cheers

Klaus

Check out this type, comes in dual:

--
Regards, Joerg 

http://www.analogconsultants.com/

115kBaud

barrier

to the signal "TX"

time. On the secondary side the HF signal is caught by the advance of a ripple counter (4017). The clock rate is significantly higher than the TX rate and higher than the reset signal, so when the short time slice with many pulses occurs it will set O3 output high for long enough to charge the "OtherSig" capacitor.

simpler idea? (I have looked at a microcontroller to decode the signal, but that takes to much current)

I have sketched up quickly has too many parts.

for the serial protocol has already been written and this would not be welcome by the sw guys

would add another microcontroller to be programmed which adds production costs

This should be transparent to any existing protocol, except in that it would add a small delay. I'm adding a pair of uC's across the barrier. All they're doing is converting and passing on the existing serial comms at a higher speed so that the extra digital line can be added.

Unless I've misunderstood the problem, which is quite likely, this solution is easy. I would say 'trivial' but every time I use that word, extra work seems to appear from nowhere.

AIUI now we have...

115k tx--->barrier--->115k rx

But we could have...

115k tx--->Syd's uC1--->barrier--->Syd's uC2-->115k rx

So uC1 receives the existing 115k serial data and sends it at a higher speed across the barrier. UC2 receives this higher speed serial signal and sends it on at 115k.

uC1 also has a digital input for the extra signal. uC2 has a corresponding digital output.

Say, for example, the new barrier rate is 500kbaud, then we could send alternate bytes across the barrier using a 9-bit protocol. With bit 9 set these could be re-transmitted transparently by uC2 at 115k, with bit

9 clear, these could be your digital data.

Cheers

--
Syd

So do the ADI iCouplers. Quads, even. RS-232 and RS-488, too. ;-)

The ADUM2201 is considered, has lower current than the Si8421

/Klaus

115kBaud

the signal "TX"

time. On the secondary side the HF signal is caught by the advance of a ripple counter (4017). The clock rate is significantly higher than the TX rate and higher than the reset signal, so when the short time slice with many pulses occurs it will set O3 output high for long enough to charge the "OtherSig" capacitor.

idea? (I have looked at a microcontroller to decode the signal, but that takes to much current)

have sketched up quickly has too many parts.

Klaus,

Are the two data streams as simplistic as shown on your drawing? Or can they assume other patterns? ...Jim Thompson

--
| James E.Thompson, CTO                            |    mens     | 
| Analog Innovations, Inc.                         |     et      | 
| Analog/Mixed-Signal ASIC's and Discrete Systems  |    manus    | 
| Phoenix, Arizona  85048    Skype: Contacts Only  |             | 
| Voice:(480)460-2350  Fax: Available upon request |  Brass Rat  | 
| E-mail Icon at http://www.analog-innovations.com |    1962     | 
              
I love to cook with wine.     Sometimes I even put it in the food.

You made the same mistake as those building the first trans-Atlantic telegraph cable. They tried to increase the signaling rate by increasing the voltage levels, finally destroying the cable.

The RS-232 model is a high voltage, high impedance voltage source driving a capacitive load. Ultimately, this will kill the performance quite rapidly.

Low voltage signaling, such as RS-422/485 use much lower signaling voltages and hence much less current is needed to charge and discharge the line capacitance. The low +/-200 mV receiver threshold would allow much higher line speeds even when looking at the situation as a simple voltage source and capacitive load case.

Of course, running 115k2 at 1 km relies on the transmission line approach.

If you do a frequency scan for some residential telephone pairs, there are going to be some deep notches at certain frequencies due to 1/4 wave stubs in the wiring system. The ADSL system will ignore these subcarriers as well as those subcarriers suffering from heavy external interference, such as local medium wave radio broadcasts. Thus the real ADSL throughput is less than advertized.

My suggestion to my customers is to use RS-232 only within the same equipment rack, partly due to the speed issues, but mainly due to ground potential issues.

On a sunny day (Wed, 16 Jan 2013 10:20:50 +0200) it happened snipped-for-privacy@downunder.com wrote in :

I did not make that mistake. I was called as consultant to help PLC guys that had a problem with communication in some factory. Took a day of travel, and 5 minutes with a scope to find the problem. Reported it, made a screenshot, showed and explained it to big boss, made some recommendations, and then they decided to replace it with glassfiber.

Interference is an other thing, especially in an industrial environment where large machines are switched on and of, and you get magnetic induction, ground currents too. AFAIK from that moment on they used glassfiber for all communications.

Sorry for my bad English, by "you" (plural) I was referring to the original situation.

When working with power plant people, they always demand fibers, even if 1 kV galvanic optoisolation would be quite sufficient.

115kBaud 115200 NRZ has a null in the spectrum at 115200Hz, so you can put yout sidechannel signal on that frequency.

if you dothe 115K2 BPS in software and add glitches into the stream when you need to generate the sidechannel signal.

eg if every bit is 16 "uart clocks" long during a 0 in the main signal send a 1 during the fourth clock and during a 1 send a 0 half a bit-time later - in the twelveth clock

that'll show up as a 115200Hz tone which is right in the first trough of the NRZ spectrum, you can detect it with a pulse detector, diode pump, or even a tuned circuit

Any decent UARt will ignore it, or you can filter it out using a low pass.

--
?? 100% natural 

--- news://freenews.netfront.net/ - complaints: news@netfront.net ---

h 115kBaud

arrier

cted to the signal "TX"

hort time. On the secondary side the HF signal is caught by the advance of a ripple counter (4017). The clock rate is significantly higher than the TX rate and higher than the reset signal, so when the short time slice with m any pulses occurs it will set O3 output high for long enough to charge the "OtherSig" capacitor.

simpler idea? (I have looked at a microcontroller to decode the signal, but that takes to much current)

ones I have sketched up quickly has too many parts.

The datastream is like the one from an UART TX channel. 11 bits, start bit, data and stop bits. Baudrate can be anything from 2400 to 115200.

The second signal that needs to be combined with the TX signal just needs a response time of max 1ms.

Cheers

Klaus

115kBaud

barrier

to the signal "TX"

time. On the secondary side the HF signal is caught by the advance of a ripple counter (4017). The clock rate is significantly higher than the TX rate and higher than the reset signal, so when the short time slice with many pulses occurs it will set O3 output high for long enough to charge the "OtherSig" capacitor.

simpler idea? (I have looked at a microcontroller to decode the signal, but that takes to much current)

I have sketched up quickly has too many parts.

data and stop bits. Baudrate can be anything from 2400 to 115200.

response time of max 1ms.

OK. ...Jim Thompson

--
| James E.Thompson, CTO                            |    mens     | 
| Analog Innovations, Inc.                         |     et      | 
| Analog/Mixed-Signal ASIC's and Discrete Systems  |    manus    | 
| Phoenix, Arizona  85048    Skype: Contacts Only  |             | 
| Voice:(480)460-2350  Fax: Available upon request |  Brass Rat  | 
| E-mail Icon at http://www.analog-innovations.com |    1962     | 
              
I love to cook with wine.     Sometimes I even put it in the food.

For the stuff I work with (power system protective relaying), copper based communications is only allowed inside a panel. If it leaves the panel, it gets converted to fibre.

115kBaud

barrier

Can you do a minor modulation of the width of the start or stop bits that is small enough that it doesn't affect the downstream UART. At 115 kbps there should be enough packets to send one bit every ms with a goodly pile of redundancy. Unfortunately, it might take more space than a second optocoupler.

115kBaud

barrier

...and money and power. I think these avenues have already been explored.

with 115kBaud

me barrier

a e

for

and

al,

the

eds

t

le

The alternative would be to try another direction removing the iCoupler and Opto and using a transformer instead (the Pulse PH9185 is approved for the correct isolation and has the same area as the iCoupler)

Signaling from primary to secondary would then be at two different frequenc ies with some filters to "decode" the signals and communication the other w ay could be a current modulation scheme or injection of a third frequency.

It does however quickly become complex, but the upside is that current cons umption may possibly be reduced.

Immunity to common mode signals is another matter, the ADUM2201m has immuni ty of 35kV/us. The PH9185 has 8pF capacitance from primary to secondary, so a dV/dt of this rate will inject more than 100mA into the rail of the prim ary side which could affect the communication related to how well the capac itance is coupled to the primary side and if it is symmetric coupled

Anyony been down this road, using a transformer for data transmission and o bservations on the effects of common mode transients?

Cheers

Klaus

er with 115kBaud

same barrier

g

or a

y

the

h for

r and

gnal,

t the

r

rt

2. >

needs

hat

s

pile

nd Opto and using a transformer instead (the Pulse PH9185 is approved for t he correct isolation and has the same area as the iCoupler)

ncies with some filters to "decode" the signals and communication the other way could be a current modulation scheme or injection of a third frequency .

nsumption may possibly be reduced.

nity of 35kV/us. The PH9185 has 8pF capacitance from primary to secondary, so a dV/dt of this rate will inject more than 100mA into the rail of the pr imary side which could affect the communication related to how well the cap acitance is coupled to the primary side and if it is symmetric coupled

observations on the effects of common mode transients?

Lower capacitance and space reduction can be athieved with a coreless PCB x former, although it needs to be subjected to a higher frequency to reduce m agnetizing losses.

Cheers

Klaus