Jump to content

MIDI interference


sideshowmark

Recommended Posts

Although the problem was originally with an LX desk :blink: - I've put this in sound as it's to do with my audio control PC and MIDI...

 

We have been experiencing LX desk crashes – which always happened at the same point in the show – after two months of going round in circles with lots of desks being swapped around I traced it back to the MIDI – My air con causes interference with the MIDI – PC USB interface which causes it to send out junk This junk causes the LX desk MIDI unit to get confused and crash the next time a coherent command is sent.

 

It’s not just the air con though – you can flick switches all over the place and it makes up its own commands.

 

I’ve run the whole system from a UPS battery and it still happens so It’s not coming up the mains – the only thing I can think of is that the PC or USB cable/interface are picking this stuff up from thin air as MIDI itself is very resilient to interference.

 

I’m using a standard PC tower with an M-Audio MIDISPORT 2X2 USB MIDI interface.

 

I hear nothing unusual over audio

 

I'm going to try different ways of generating the MIDI signal - like using a laptop instead - or a PCI MIDI card - or a different USB interface (someone suggested a MOTU unit) to try and eliminate this.

 

I have a print screen of the nonsense MIDI commands which I won't post here as I'll probably get told off. PM me if you think taking a look will help resolve this.

 

Has anyone ever come across this before? The manufacturers of all the gear are all claiming ignorance of this 'phenomenon' but I'm sure I can't be alone out here.

Link to comment
Share on other sites

I’ve run the whole system from a UPS battery and it still happens so It’s not coming up the mains...

not true - it'll depend on your UPS and how well filtered it is. Is it an on-line or off-line UPS? Also, the earth will come all the way from the input mains to the output.

 

– the only thing I can think of is that the PC or USB cable/interface are picking this stuff up from thin air as MIDI itself is very resilient to interference.

possibly although any interference on the USB signal will not generate valid USB messages

 

The manufacturers of all the gear are all claiming ignorance of this 'phenomenon'...

Of course they do.

 

Inside the MIDISPORT is a small 8-bit microprocessor, it's possible that noise is causing it to crash and send out rubbish.

 

My money is on an earthing issue.

Link to comment
Share on other sites

I'd like to see the printscreen of the nonsense MIDI commands, and how exactly did you capture the MIDI data? Nonesense can mean many things. Is it valid nonesense, or invalid nonsense.

 

The differential on this problem is where the duff MIDI is being generated. I'd guess that the cure would be to put an ordinary MIDI card in the tower computer, on the guess its somethying to do with the USB bus or the USB interface, but its just a guess. If the bad MIDI is being generated in the computer for whatever reason, then it is likely that it will pop out irrespective of which interface you use.

 

I note Bruce mentiones earthing; your MIDI cables etc are all the proper jobbies, yaas? MIDI is designed to isolate earths, to prevent exactlky this sort of problem.

 

On the wider issue, just what lighting desk is it that goes so confused by invalid MIDI that it crashes? That is appalling behaviour.

Link to comment
Share on other sites

I note Bruce mentiones earthing; your MIDI cables etc are all the proper jobbies, yaas?

 

Nope - that was Brian. :blink:

 

I come from the west of Scotland, where "proper jobbies" has a completely different meaning...

Link to comment
Share on other sites

both PC and Lx desk were via the UPS. MIDI Sport is USB powered. It's just a standard PC UPS but I tested it with the UPS completely unplugged from the wall - so the earth was obviously also detatched. the only way I can see it could physically some in would be along DMX to the LX desk or through the VDU's which were still plugged in to mains. so airwaves interference looks the most likely candidate.

 

Midi stream was captured by running the THRU from the LX desk back into the MIDISPORT and monitoring the MIDI input montor included in the show control software whilst flicking switches - I also ran it without flicking swiches and nothing happened.

 

Desk is an ETC Express 24/48 96 channel - problem occurs on my desk and the replacement ETC sent when crashes first apeared. My 01v96 V2 which is also plugged into it all seems unaffected.

 

This is a new system in the theatre but I know that people running samplers in the same room from the same MIDI interface have also noticed some (admittedly VERY intermittent - I think it happened twice in three months) odd behaviour that we couldn't replicate which may have been the same thing.

 

I'll try and add the MIDI capture picture to the Photo section of my profile - failing that I'll email it to you. the first line (yellow) was the software generated 'GO' for the LX desk. everything after that is all interference. - it is odd as it is appearing in MIDI format - as in 'Pitch bend channel 16' but as this signal goes though the Lx desk maybe that is just the LX desk's MIDI module trying to make sense of what it is receiving - I'm no MIDI expert I'm afraid.

 

All off the shelf MIDI cables - no technician soldering or bodging to be seen!

Link to comment
Share on other sites

And here is the MIDI analyser screenshot

 

http://www.davidbuckley.name/pix/marks_midi_problem_sshot.jpg

 

Doesnt look at all like real MIDI, it looks like electrical interference, as almost always the bits in the byte are ones. When the top bit is one then its a start of a MIDI message, and most messages are one byters. Its junk.

 

What is interesting is that the reception UARTS have presumably seen valid start bit. Which I'd have thought was odd as it doesnt look like enough zeros in the rest of the scrambled data (status column) to make sense.

 

Next question is if you plug the PC MIDI out stright into the MIDI in (thus cutting the ETCD out of the loop) does it still happen?

 

My guess would be the culprit is the USB/MIDI box, and that you should try a PCI MIDI card in the tower.

 

And throw out the ETC. If a bit of dodgy MIDI causes it to lose its cool, what else is wrong wioth its firmware....

 

(Or maybe all lighting desks are this bad...?)

Link to comment
Share on other sites

The LX desk shouldn't be adding anything. A MIDI thru connection is just an electrical copy of what is on the MIDI in. So like David I'd remove the desk and just feed the output of the MIDIsport back to it's input and run the same tests.

 

I'm wondering where your setup is earthed. IIRC the Express runs with an external 'laptop' style PSU? Does the earth follow through that? When running on just the UPS your setup would be earthed via the DMX run or maybe the VDU.

Link to comment
Share on other sites

A MIDI thru connection is just an electrical copy of what is on the MIDI in

Not always. Originally, it was exactly the case, with the classic In - Thru - Out 3 socket layout. The was the cost revolution of using just two sockets, so we had In - Thru/Out on 2 sockets, and thats when Thru stopped being an electrical copy but a software copy. Get a few software thru in a line and the delays start to build up. Common mistake when first plugging a rackful of sound modules together...

Link to comment
Share on other sites

I know that the ETC desk is used at a few theme parks with multiple desks all running through MIDI with no issues whatsoever - The interesting thing I saw was the SYSEX: F0 line - F0 signifies the start of a MIDI Show Control String does it not? if there is no F7 end command would the desk treat the whole lot as one uber long cammand with no end point? I've used identical desks in other venues for years trouble-free

 

I would love to know if other desks are immune to this as it is a new desk to me and I still have the option of sending it back and choosing another one. 'twas ex rental so finding a similar desk for a similar price will be difficult though. I'd rather just eliminate the interference and be done with it!

 

I've just tried running the MIDI straight back into itself and there is no junk - also when I run it into my sound desk and back out via the thru port I also get no junk - I only get junk along the MIDI when it's run through the LX desk - but unplug the MIDI and messing with the air con has no effect on theLX desk.... Wierd.

 

All I can think of is that messing with the air con affects how the lx desk receives its MIDI - as it still crashes when I unplug the thru output from the lx desk - or that the way the LX desk is wired/programmed converts normally ignored interference into confused MIDI signals.

 

The plot thickens...

 

 

The LX desk shouldn't be adding anything. A MIDI thru connection is just an electrical copy of what is on the MIDI in. So like David I'd remove the desk and just feed the output of the MIDIsport back to it's input and run the same tests.

 

I'm wondering where your setup is earthed. IIRC the Express runs with an external 'laptop' style PSU? Does the earth follow through that? When running on just the UPS your setup would be earthed via the DMX run or maybe the VDU.

 

 

The express does have an external laptop style PSU which has an uninterupted earth on it - have run it with DMX unplugged and then with VDU unplugged and still had issues - I could try running VDU's et al on UPS' as well but I don't have that many to hand.

 

I'm reluctant to do too much troubleshooting while I have this show on - we've abandoned the MIDI for this one but I would like to get it sorted for the next show - firing the LX desk was part of the reason I went for PC show control.

Link to comment
Share on other sites

I'm wondering where your setup is earthed. IIRC the Express runs with an external 'laptop' style PSU? Does the earth follow through that? When running on just the UPS your setup would be earthed via the DMX run or maybe the VDU.
Hmm, I asked a question in relation to earthing the PSU and PAT not that long ago. The screen of the DIN plug has a resistance to earth, but I'm not near it at the moment to check.
Link to comment
Share on other sites

I've just tried running the MIDI straight back into itself and there is no junk - also when I run it into my sound desk and back out via the thru port I also get no junk - I only get junk along the MIDI when it's run through the LX desk

Looks like Brian may have placed the smart money on this one...

 

The MIDI standard specifies exactly who supplies power and ground and who opto-isolates to keep earth problems out of the system. I think your system somehow violates the rules.

 

With a mutlimeter, check that your MIDI cables do not have shield to sheild connectivity.

 

Also with said meter, check that all pins of the Express's MIDI in socket are isolated from Express chassis.

Link to comment
Share on other sites

OK, Deep breath...

 

The shield contacts on the desk sockets are connected to each other and earth

 

The pins on the MIDI sockets in the desk appear isolated from each other and the shield around the socket and also the desk chassis...

 

The pins on my cables are isolated both from each other and the shield

 

but...

 

When I plug a MIDI cable into the MIDI socket of the LX desk - I get a current from PIN 3 to sheild and desk chassis etc.

 

also a similar thing happens from the sheild to pins 4,2,5 in the out/thru sockets - but again only when a cable is plugged in!

 

I have access to 2 of these desks and both give identical results

 

What does it all MEEEAAAAAAANNN?

 

OK so unless I'm mistaken none of the pins should ever be connecting straight to earth - so LX desk MIDI wiring is my problem no? excuse my ignorance in this matter. :blink:

Link to comment
Share on other sites

OK, Deep breath...

 

The shield contacts on the desk sockets are connected to each other and earth

The pins on the MIDI sockets in the desk appear isolated from each other and the shield around the socket and also the desk chassis...

The pins on my cables are isolated both from each other and the shield

 

but...

 

When I plug a MIDI cable into the MIDI socket of the LX desk - I get a current from PIN 3 to shield and desk chassis etc.

also a similar thing happens from the shield to pins 4,2,5 in the out/thru sockets - but again only when a cable is plugged in!

 

Ahh, understanding (maybe) dawns!

 

What you have is a 5 pin din lead, NOT a midi lead!

 

A correctly made midi lead connects pins 2,4,5 ONLY on the cable and the receiver should leave pin 2 floating. In fact a correctly implemented midi receiver should be totally floating wrt the local ground (it is really just a fastish opto isolator in series with the current loop).

 

Making leads with the metal screen on the din plug connected to pin 2 is another common error.

 

Try a lead that connects as follows:

 

sender receiver

2 -> screen -> open

4 -> wire -> 4

5 -> wire -> 5

 

That should work no matter how much of a misunderstanding about pin 2 applies at the receiver.

 

HTH.

 

Regards, Dan.

Link to comment
Share on other sites

Archived

This topic is now archived and is closed to further replies.

×
×
  • Create New...

Important Information

We have placed cookies on your device to help make this website better. You can adjust your cookie settings, otherwise we'll assume you're okay to continue.