Sign in to follow this  
Followers 0
freddy

MBSEQv4 unable to drive TR-9090 clone?

11 posts in this topic

Folks,

while having fun with my shiny MBSEQv4 I'm able to drive MB6582 with no issues on both OUT1 and OUT2 but my TR-9090 clone doesn't seem to receive any messages. It seems to be MBSEQ issue because (I think) I've ruled everything else out:

- used different MIDI cables of course,

- using OUT1 or OUT2 makes no difference,

- connecting MIDI Monitor (in MIOS Studio) to MBSEQ output shows the notes and proper MIDI channel,

- connecting TR-9090 directly to MIDI Interface (not via MBSEQ) works as a charm,

- connecting MIDI Monitor (in MIOS Studio) to TR-9090's MIDI Thru doesn't show any messages flowing through TR-9090,

- connecting TR-9090 to the same output as MB6582 (and changing MIDI channel to 10) makes no difference,

- "amplifying" MIDI signal via an external MIDI interface (routing OUT to IN) does work as a charm again.

Could this be possible, in theory at least? Did anyone experience similar issue (maybe with an other equipment) as well? I'd suspect some part of my setup being faulty because I found other references to 808/909(0) in the SEQ forum but no clue what could it be...

Thanks,

freddy

Share this post


Link to post
Share on other sites

I have had this on DIY MIDI circuits, what opto is the TR-9090 using?

At a guess the opto is not as sensitive as others.

Assuming it is the same issue, three choices:

1) Replace the opto circuit with a 6N139 circuit. Will require a bit of work to get it right.

2) Connect the Out via a MIDI Thru/Merge unit to as you say, "amplify" the signal.

3) Change the resistor in the TR-9090, I wouldn't but its up to you.

Best choice is number 2, but these are just thoughts and I may be writing rubbish!

S

Share this post


Link to post
Share on other sites

It seems to be MBSEQ issue because (I think) I've ruled everything else out:

- connecting MIDI Monitor (in MIOS Studio) to MBSEQ output shows the notes and proper MIDI channel,

...

- "amplifying" MIDI signal via an external MIDI interface (routing OUT to IN) does work as a charm again.

You realize that these two statements are contradicting each other, right?

The 9090 is the problem here, follow the above advice and check the optocoupler!

Good luck!

Share this post


Link to post
Share on other sites

You realize that these two statements are contradicting each other, right?

Not really (unless I'm entirely missing the point in here), of course this was done as two separate steps:

1) watching OUT of MBSEQ showing the notes being played, to confirm the MBSEQ itself works as expected (no doubt, of course, but still strange),

2) connecting MBSEQ OUT to TR-9090 IN via an external interface in 'through' mode showing the TR-9090 input works after the signal is "amplified".

The 9090 is the problem here, follow the above advice and check the optocoupler!

Optocoupler used in TR-9090 is 6N136. That must be it, with non-amplified signal (straight from MBSEQ) I see no signs of life on optocoupler's pin 6 (optocoupler transistor's collector) - steady H there, with amplified one I can see H dropping to L. I have no logic analyzer to confirm the signal but since it works I guess it's just correct.

I'll consider either optocoupler and optocoupler circuit replacement, thanks a lot for advice(s)!

Share this post


Link to post
Share on other sites

Not really (unless I'm entirely missing the point in here), of course this was done as two separate steps:

1) watching OUT of MBSEQ showing the notes being played, to confirm the MBSEQ itself works as expected (no doubt, of course, but still strange),

2) connecting MBSEQ OUT to TR-9090 IN via an external interface in 'through' mode showing the TR-9090 input works after the signal is "amplified".

Optocoupler used in TR-9090 is 6N136. That must be it, with non-amplified signal (straight from MBSEQ) I see no signs of life on optocoupler's pin 6 (optocoupler transistor's collector) - steady H there, with amplified one I can see H dropping to L. I have no logic analyzer to confirm the signal but since it works I guess it's just correct.

I'll consider either optocoupler and optocoupler circuit replacement, thanks a lot for advice(s)!

The problem is that the MBSEQv4 midi is working at 3.3v instead of 5v like MOST midi instruments require. You need to build a 5v level shifter. You can find schematic for this circuit on the MBSEQ hardware page it's called "LPC17 output buffers"

Edited by smokestacksproductions

Share this post


Link to post
Share on other sites

The problem is that the MBSEQv4 midi is working at 3.3v instead of 5v like MOST midi instruments require. You need to build a 5v level shifter. You can find schematic for this circuit on the MBSEQ hardware page it's called "LPC17 output buffers"

Very true, this was discussed quite recently ago. Thanks for the pointer, again!

Share this post


Link to post
Share on other sites

Not going to build one of these, but can you provide a link to the LPC17 output buffers schematic please?

Share this post


Link to post
Share on other sites

6n136 is kinda weak, you want something that has more gain. 6n137 with a 10k pull up works for me, the 6N139 is the standard in the MHBP cores. You could also try a different brand 6N136 (this helped me with my 3030s which the same issue, I had a fairchild part in there and replacing it with a avago fixes it)

The buffer circuit is to shift the level on the analog gates, it has nothing to do with the midi section unless you plan on adding midi ports 3 and 4

the 3.3v thing is not really the core problem, you would probably issues with other 5v devices as well that send out a bit of a rounded square that gets rounded off a bit more by the crappy opto to the point that the uC does recognize it as a valid msg

Edited by Altitude

Share this post


Link to post
Share on other sites

Not going to build one of these, but can you provide a link to the LPC17 output buffers schematic please?

As smokestacksproductions suggested, it's referred from MBSEQv4 User Manual, "Hardware Options" chapter, "CV Outputs" section: http://www.ucapps.de/mbhp/mbhp_core_lpc17_output_buffers.pdf

6n136 is kinda weak, you want something that has more gain. 6n137 with a 10k pull up works for me, the 6N139 is the standard in the MHBP cores. You could also try a different brand 6N136 (this helped me with my 3030s which the same issue, I had a fairchild part in there and replacing it with a avago fixes it)

Okay, definitely will look around, replacing the input section won't be too hard, TTL pin has a separate connection connectable with veroboard-based replaced MIDI input circuitry. I'll look for another 6N136s first to ease my life.

The buffer circuit is to shift the level on the analog gates, it has nothing to do with the midi section unless you plan on adding midi ports 3 and 4

the 3.3v thing is not really the core problem, you would probably issues with other 5v devices as well that send out a bit of a rounded square that gets rounded off a bit more by the crappy opto to the point that the uC does recognize it as a valid msg

Ah, okay, thanks for explanation!

Edited by freddy

Share this post


Link to post
Share on other sites

I've just come across this thread, and seen a few of the ideas.

First thought: MIDI is a current loop interface, specified at 5mA loop current, though often running a bit higher to give leeway for differing opto isolator input LEDs. The voltage is not really the issue. 

Changing the input of the 9090 is a bodge, it works fine with other sequencers - you will only get the same problem with other devices, especially those using older optos like the CNY series.

The simplest solution to me, would be to reduce the value of the output resistors in the sequencer. Since there will always be a 220R in the input circuit of the target MIDI device, dropping the resistors at the sequencer end to 100R or 68R would get closer to the spec, and the max, shorted output current from the sequencer , (in case of a bad lead) would still be a fairly safe 16 - 25 mA. (standard short circuit current on 5V would be a round 11ma).

Oh, and if anybody is thinking of termination impedance matching: try to remember this is low KHz speed, not RF, and the lower output impedance would actually help keep the pulses square....

(added 05/11/2017)

The latest MIDI spec, (CA-033) MIDI 1.0 Electrical Specification Update [2014], suggests 33 Ohms from pin 4 to the 3.3 V rail, and 10 Ohms from the device output to pin 5. This makes for a short circuit current of 76mA, which I consider too high for safety on a microprocessor output. I think it's good  practise to use an output buffer: if you are working live, a spare buffer chip can be kept in the toolkit. A spare micro module is not so easy.

Just a few thoughts anyway

Mike

Edited by TheAncientOne
Added note on updated MIDI spec
2 people like this

Share this post


Link to post
Share on other sites

Hi everyone ! I experienced the same problem, and The Ancient One's solution works perfectly for me. I changed the 220R resistors to 68R for R21 and R22. Now the 9090 detects the signal without any issue. Thank you a lot Michael !

Théo

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!


Register a new account

Sign in

Already have an account? Sign in here.


Sign In Now
Sign in to follow this  
Followers 0