E1.20 RDM (Remote Device Management) Protocol Forums  

Go Back   E1.20 RDM (Remote Device Management) Protocol Forums > RDM Developer Forums > RDM Physical Layer/Hardware Discussion

RDM Physical Layer/Hardware Discussion Discussion and questions relating to the physical or hardware layer of RDM.

Reply
 
Thread Tools Search this Thread Display Modes
Old August 4th, 2008   #1
berntd
Senior Member
 
Join Date: Jan 2008
Posts: 102
Default Line biasing Flaws - need solutions

Hello,

The line biasing method in the RDM spec is flawed in my opinion.
The reason is that the voltages on the DMX lines are allowed to rise way further than the power supply that biases the lines.

When something drives the line to a higher level, current starts flowing back onto the power supply of the master and depending on th capabilities of the driver, this may cause the supply to rise above its VCC level and damage can occur to the electronics on that line.

I am currently using a 3V3 low power system system and I just noticed that the data resoponses from a responder are quite capable of driving up the voltage on the Vcc towards 5V. this will result in damage.

In fact, data on the line (not normal operation) is even able to power the master through the biasing network, resulting in complete malfunction.

I am now scratching my head on how to solve this problem and still operate within the specification.

I am considering adding a diode into the biasing network. tjhis will upset the impedance quite badly when the diode become reverse biased.

Another option is to clamp the supply rail to 3.3V but there seem no Tranzorbs on the market for this voltage. Maybe a zener but that is not great either.

The diode may well be the only answer but I would love to hear other suggesstions.

Perhaps I am missing something?

Kind regards
Bernt
berntd is offline   Reply With Quote
Old August 5th, 2008   #2
ericthegeek
Task Group Member
 
Join Date: Aug 2008
Posts: 358
Default

The R1 bias resistor is recommended to be 562 ohm. If the 485 line is driven to 5v, you're looking at (5v-3.3v)/562ohm = 3mA. As for pulling your power supply out of spec, 3mA is no worse than a pullup resistor. It's hard to see this causing the problems you're describing unless you're building an extremely low power device or you have a very high ESR power supply.

Could you have some other path between your 5v and 3.3 volt supplies? Are you using a true RS485 driver? A 3.3v line driver with data lines driven to 5v is no different than a 5v line driver driven to 12v (at max 485 common mode voltage)

Are you using ESD protection diodes on your 485 data lines? Are those installed backwards? That seems like the most likely suspect.

Last edited by ericthegeek; August 5th, 2008 at 12:40 AM.
ericthegeek is offline   Reply With Quote
Old August 5th, 2008   #3
berntd
Senior Member
 
Join Date: Jan 2008
Posts: 102
Default

Hi Eric and thanks for the reply.

To meet the spec, The pullup resistor is actually 390R for a 3V3 supply.
If the dataline is now pulled to +12V (example but perfectly legal) by the responder driver, we get (12-3V3 / 390 ) = 22mA into the power supply.

That is quite high in my opinion!

the power dissipation on the pullup resistor is then 0.194W !

Worse, if the power of the master happens to be off for whatever reason, it is quite possible to power up the master throughthis single pullup.

Yes, we are building lower power devices but not that low and it is still a problem.

ESD protection does not help here because that only kicks in at +12 and -7V, whch are the ~ allowable limits according to the spec.

Kind regards
Bernt
berntd is offline   Reply With Quote
Old August 5th, 2008   #4
ericthegeek
Task Group Member
 
Join Date: Aug 2008
Posts: 358
Default

In your initial post, you described the problem at 5v. Are you seeing the problem with with 5v line drivers, or only under common mode conditions?

One option might be to build a separate "termination power" supply on your design. This would prevent line conditions from coupling back into your logic supply.

Under high common mode situations, the bias and termination resistors will dissipate a non-trivial amount of power. Certainly enough to damage 1206 and smaller resistors.

The line biasing network is pretty tightly constrained. It has to match the 120ohm impedance of the line, but still operate across the full common mode range. In an idea world, you wouldn't need it (485 drivers are supposed to maintain state when there's no input), but real-word tests have shown it's required.
ericthegeek is offline   Reply With Quote
Old August 5th, 2008   #5
Nigel Worsley
Junior Member
 
Join Date: Jun 2006
Location: London
Posts: 13
Default

A capacitor after the diode will sort out the impedance issue, it doesn't need to be particularly large as the impedance is only relevant for higher frequencies - 100nF would probably be enough. Add a zener across it to prevent it charging up too much and changing the bias voltage.

BTW, 3.3V tranzorbs ARE available, NXP, Vishay and ST all do them.

Last edited by Nigel Worsley; August 5th, 2008 at 07:17 AM.
Nigel Worsley is offline   Reply With Quote
Old August 5th, 2008   #6
berntd
Senior Member
 
Join Date: Jan 2008
Posts: 102
Default

Hi Eric,

I did say 5V originally and I have seen the problem with a 5V external driver driving into my 3V3 master when the master power is dropped suddely during communication. The 5V is then able to couple from the data line through the pullup into my power and the circuit does not die off properly and interesting things happen at that moment.

As for the common mode and 12V issues, we have to cater for all conditions possible and still remain reliable.
The unit I am designing only has a 3V3 power supply and I can't add anything more because of other constraints.

I don't think this pullup method is the best solution. It seems more like an afterthought.

Kind regards
bernt

Quote:
Originally Posted by ericthegeek
In your initial post, you described the problem at 5v. Are you seeing the problem with with 5v line drivers, or only under common mode conditions?

One option might be to build a separate "termination power" supply on your design. This would prevent line conditions from coupling back into your logic supply.

Under high common mode situations, the bias and termination resistors will dissipate a non-trivial amount of power. Certainly enough to damage 1206 and smaller resistors.

The line biasing network is pretty tightly constrained. It has to match the 120ohm impedance of the line, but still operate across the full common mode range. In an idea world, you wouldn't need it (485 drivers are supposed to maintain state when there's no input), but real-word tests have shown it's required.
berntd is offline   Reply With Quote
Old August 5th, 2008   #7
berntd
Senior Member
 
Join Date: Jan 2008
Posts: 102
Default

Quote:
Originally Posted by Nigel Worsley
A capacitor after the diode will sort out the impedance issue, it doesn't need to be particularly large as the impedance is only relevant for higher frequencies - 100nF would probably be enough. Add a zener across it to prevent it charging up too much and changing the bias voltage.

BTW, 3.3V tranzorbs ARE available, NXP, Vishay and ST all do them.
Hello
I did look for tranzorbs but their breakdown voltage was not good enough at over 4V for most of them. The absolute maximum in my system is 4.0V for the processor.
I will check again.

The diode is the correct way to go here in my opinion, if we can make it work.

So we have 3V3, diode(a-k), pullup, dataline+.
Did you mean the capacitor needs to go from the 3V3 to gnd on the diode a side? How will this work if the diode is reverse biased?

Or do you mean the capacitor should go across the diode, providing an ac path to 3V3?

Kind regards
Bernt

Last edited by berntd; August 5th, 2008 at 07:01 PM.
berntd is offline   Reply With Quote
Old April 9th, 2009   #8
dangeross
Junior Member
 
Join Date: Feb 2009
Posts: 13
Default

Another option might be to use a chip like the Intersil ISL3296E with series resistors to limit the drive current to provide the line bias. You could even only enable the bias while the Controller is expecting a response to save some power if you device is battery operated.
dangeross is offline   Reply With Quote
Reply

Bookmarks

Thread Tools Search this Thread
Search this Thread:

Advanced Search
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump

Similar Threads
Thread Thread Starter Forum Replies Last Post
3.3V Biasing William CANO RDM Physical Layer/Hardware Discussion 3 December 5th, 2006 08:17 AM
Line biasing Andy Macdonald RDM Physical Layer/Hardware Discussion 1 October 26th, 2006 12:02 PM
4.2.1.1 In-line device Port Turnaround during Discovery prwatE120 RDM Timing Discussion 1 August 7th, 2006 11:52 PM


All times are GMT -6. The time now is 05:27 PM.


Powered by vBulletin® Version 3.8.7
Copyright ©2000 - 2019, vBulletin Solutions, Inc.