E1.20 RDM (Remote Device Management) Protocol Forums

E1.20 RDM (Remote Device Management) Protocol Forums (http://www.rdmprotocol.org/forums/index.php)
-   RDM General Implementation Discussion (http://www.rdmprotocol.org/forums/forumdisplay.php?f=4)
-   -   Sub-Device Personalities (http://www.rdmprotocol.org/forums/showthread.php?t=1124)

owaits January 4th, 2012 10:23 AM

Sub-Device Personalities
 
So I have found a post which answers my first question, can sub-devices have different personalities to which the answer is yes. My second questions is how should you report the personality index and count.

My interpretation is that each sub device is completely separate e.g.

Each sub device only supports a single personality mode but has a different set of channels.

Sub-Device:1 Personality:1 PersonalityCount:1 Footprint: 6
Sub-Device:2 Personality:1 PersonalityCount:1 Footprint: 30
Sub-Device:3 Personality:1 PersonalityCount:1 Footprint: 30

Is this correct?

ericthegeek January 4th, 2012 12:27 PM

The standard leaves a some room for interpretation as to how different one sub-device can be from another. Sub-devices aren't completely separate, but they are mostly separate.

All that's specified is that sub-devices must all report the same list of PIDS in SUPPORTED_PARAMETERS, the same sensor count in DEVICE_INFO, and the same SENSOR_DEFINITION for each sensor.

If you're building a responder, your safest bet it to keep your personalities identical between all sub devices IFF you can. But if the device's nature doesn't allow it, then yes, each sub-device can have a completely different set of personalities.

hamish January 4th, 2012 01:21 PM

I'm not sure if there is a right or wrong answer to your question.
To maintain a pseudo symmetry between sub devices, the implementation may look more like this:

Sub-Device:1 Personality:1 PersonalityCount:2 Footprint: 6
Sub-Device:2 Personality:2 PersonalityCount:2 Footprint: 30
Sub-Device:3 Personality:2 PersonalityCount:2 Footprint: 30

When personalities that are not supported for a particular sub device are elected, the responder may then NACK NR_DATA_OUT_OF_RANGE

I expect this approach would give some controllers a wee bit of a headache, but may be less so than having asymmetric sub devices.

sblair January 4th, 2012 01:21 PM

I agree with Eric's assesment. The intention of sub-devices is for an array of devices of "like" things. This doesn't mean they all have to be identical. i.e. it is okay to have a Non-Dim module in a dimmer rack.

I'd say the general expectation is that the Personality list for the sub-devices is all the same. Each sub-device can have a different Personality, but the reported list of sub-device options is generally expected to be the same across all the sub-devices. This is not a hard requirement, but rather what I think is a general expectation many controllers may have.

In a large system, no one wants to querty n number of personality descriptions across 500 sub-devices.

It is perfectally reasonable and generally expected that the options for the Root device would be different than the Sub-Devices though.

owaits January 4th, 2012 01:39 PM

The actual sub devices I am using are as follows:

Master (3 Channels)
Audio Layer (6 Channels)
Layer 1 (30 Channels)
Layer 2 (30 Channels)
Layer 3 (30 Channels)
Layer 4 (30 Channels)

It would seem odd for the audio layer to report that it has a personality for the video layer and then later reject it.

My concern is that different people might interpret this different ways and so some devices might assume each sub-device has the same personalities and others might not.

It would seem the safest approach is for each sub device to report all the personalities of each sub-device but then may encounter issues when it tries to change the personality to one that is not supported.

At the moment I am using ArtNet, are there any free pieces of test software that I could use to test this other than DMX Workshop? I think the best approach would be to see what works with most other implementations. (Regrettably I can not afford the ticket to the states for a plug-fest at the moment :-()

nomis52 January 4th, 2012 02:26 PM

Quote:

Originally Posted by owaits (Post 2288)
At the moment I am using ArtNet, are there any free pieces of test software that I could use to test this other than DMX Workshop?

OLA supports RDM over ArtNet. I haven't added sub devices to the web UI but now that I have a device to test with I can bump that up the list.

Simon

owaits January 6th, 2012 02:29 AM

I have a working beta now so if you wanted some software to test with I can give you a copy. If you contact me separately then I will give you details on how to get the beta (that goes for anyone else wanting test software).

I hope it does not prove a problem but I have gone with each sub device acting as a separate device (with seperate personality sets) as in my first post. After reading through the standard a number of times this seemed the most logical route. This means that it wont work with DMX Workshop but I think it will be the best long term approach. If it proves an issue I may need to change it in the future.

ericthegeek January 6th, 2012 09:41 AM

If you're going to go this route, it might be a good idea to have the the master device act as a managed proxy, and represent the Audio and Video Layers act as proxied devices.

owaits January 6th, 2012 10:01 AM

I am not sure I understand why this would be a better approach. It seems to me that standard sub-devices are perfect for a media server application with multiple layers. Does not going down the managed proxy route over complicate things?

I thought that managed proxies where designed for when a device acts as a gateway between two networks and it wants to report distinct devices on that network. Would it not be confusing that my devices are actually part of the same device.

ericthegeek January 6th, 2012 03:17 PM

I'm sorry, I may have misunderstood you. When you said "I have gone with each sub device acting as a separate device", I thought you meant that each layer was being assigned a different UID, and therefor acting as if it were a totally separate RDM responder (not using sub-devices). Compared to this, using the proxy model is a better solution.

But, if you're using sub-devices, then I agree that sub-devices are the best method to use.

prwatE120 March 2nd, 2012 04:04 AM

It looks from this discussion that I should add a discussion session on the use and care of sub-devices to the forthcoming Developers conference in April !

Peter

owaits April 25th, 2012 09:08 AM

I will be bringing my implementation of the sub-devices along to the conference so will lookforward to chatting more then.


All times are GMT -6. The time now is 07:12 AM.

Powered by vBulletin® Version 3.8.7
Copyright ©2000 - 2024, vBulletin Solutions, Inc.