Syncing OXI ONE to Midronome

[Jandt]

I'm having trouble syncing my OXI ONE sequencer, firmware 3.9.3 (slave) to Midronome, firmware v1.1 (master) via MIDI (DIN) and using a direct connection from Midronome port 1 to the OXI and its TRS MIDI IN (using the correct adapter). OXI configured to sync to MIDI, happy to answer other potential questions about specific setings on the OXI. On the Midronome, it doesn't matter how it's configured (currently pla = BOT). Observed behavior: OXI shows BPM roughly at around one sixth of what Midronome is set to (say, OXI hovers between 20 and 25bpm for Midronome set to 120), with OXI's tempo pretty instable and the OXI even falling back to internal clock from time to time.

I suspect an OXI firmware issue, and will also ask on OXI's discord and/or forum. Just thought I might ask here as well, is there any other OXI + Midronome user? Workaround likely syncing via analog connection, which might even be a bit more precise. Will test that next.

Edit: syncing via analog connection works very well, with OXI very quickly adapting to Midronome's tempo (not quite, but almost instantly). So that's an option. Still puzzled by OXI refusing Midronome's MIDI clock.

[Jandt]

Crazy: just tried Midronome's MIDI OUT 2, and OXI syncs instantly. Something's not right with MIDI OUT 1. Hardware fault?

[Simon]

Hi Jandt,

Could you try that output 1 on other MIDI gear?

Cheers
Simon

[Jandt]

Hi Simon,

thanks for getting back so quickly, much appreciated. Will try that next, a fun little practical problem with this is the fact that the hardware synths I had in mind (DSI Mopho, Audiothingies Micromonsta2) don't seem to show the tempo they're synced to on their display, so it's going to be a bit more work. It's late, I should go to bed now but will report back as soon as I know more.

[Jandt]

OK, this is getting really crazy now. Simplest idea to check with another device was to connect Midronome to MIDI IN of my Motu M4 and see what Ableton Live 11 tells me about tempo when setting it to EXT mode (not planning to slave Ableton that way in practice, just meant to check the two Midronome outputs). With Midronome OUT 2 and OUT 1 same result: very good sync after a few seconds! Using the exact same MIDI cable (to rule out cable problems) I connected the OXI again. OUT 2 perfectly fine, OUT 1 not (with Midronome set to 120bpm, OXI showed between 27 and 34 bpm for OUT 1). Back to the Motu to see what MIDI Monitor shows me, MIDI events looked identical (clock = F8 every 0.021 seconds, if I'm not mistaken). No visible difference between OUT 1 and 2 anyway.

Totally lost here, and lacking the means for a deeper analysis (like an oscilloscope). Is it maybe some electrical difference between OUT 1 and 2, enough to confuse OXI, but not enough to throw off the Motu??

OK, I can use OUT 2 or analog for syncing OXI, so no real showstopper, but I'd love to know what's going on here.

[Simon]

Thanks Jandt for investigating

Very strange indeed!
There are zero difference between MIDI out 1 and 2 in terms of electrical circuit, it's the exact same circuit. I cannot be noise either since MIDI-DIN connections are fully isolated (it basically transmits info using light - so no directly electrical connection betweent the two devices)

I guess you cannot do that but the next steps would be to look at the Midronome's MIDI outputs with an oscilloscope or a multimeter? Do you happen to have one of those?

Let me loop in Manu from OXI here who maybe has an idea - he has an OXI and a Midronome so at least he can try to reproduce it

Simon

[Jandt]

Thanks for getting back. Very weird.

I still may have an old oscilloscope in the basement somewhere, collecting dust - if so, the question will be if I'm competent enough to use it properly. I'm not really a hardware guy. Also, it's analogue, so even in the best of all cases I will only be able to take "screenshots" using my cellphone

Will let you know... - Great to hear you can talk to Manuel, and that he owns a Midronome, wasn't aware of that. He knows how to use an oscilloscope

[Jandt]

Ha! Nerd alert... when looking for the oscilloscope, I found a rather old logic analyzer in a drawer. A miracle happened, I managed to create a capture of OUT 1 and 2 with it. Kudos to Saleae the company for a) still existing and b) supporting a 12y+ old device with their latest software. Fantastic.

Now, I hope the screenshot gives an idea and is sort of readable. I also hope I didn't do the capture all wrong. If so: please guide me. If not: it really looks like the two outputs are behaving differently, only OUT 2 seems to be emitting 0xF8 (see on the right). Even I as a non-hardware person can see that the signal's not the same for the two outputs, whatever the exact problem may be. Midronome was set to 120bpm.

Saleae's capture file available on request, I couldn't upload it here. It should be readable with their free "Pulse 2" software. DM me if necessary.

Still a hardware problem with my particular Midronome?

[Simon]

Hi Jandt

Thanks for checking that, and well done with setting up the signal analyzer!
Can you send me the Saleae file by email? To contact@midronome.com.
How did you connect the outputs? I forgot that MIDI is a current-based (not voltage based) signal so it's a bit tricky to connect to an oscilloscope or digital analyzer. Ideally you would need at least a 470ohm or so resistor between pins 4 and 5 of the connector, and then measuring the voltage at pin 5.

But yes the fact you get two different results seems strange, it does not necessarily mean there is something wrong with your device.

Could you try the clock with other MIDI devices?

[Jandt]

Hi Simon,

email is on the way. To me it really looks like OUT 1 emits 0x00 / 0xFF (alternating), and OUT 2 the expected 0xF8. - I'll see if I can find some other gear that can be sync'ed and shows me the bpm. It's slowly starting to bother me, I should spend my time making music, not playing with Logic Analyzers. Oh well.

[Jandt]

I can offer two more inconclusive results. Connected an Elektron Digitone to OUT 1, synced fine and also reacted to Play as expected.

Connected it to a Meeblip Cubit Go MIDI interface on my main workstation using OUT 1 - Ableton could not be synced. OUT 2 however worked fine.

I have some more gear, but these are I think all devices that don't show the bpm, so are a bit more tricky to test.

Summary so far:
- Oxi One syncs fine with OUT 2, but very unreliably with funny OUT 1 clock
- Motu M4 MIDI IN accepts funny OUT 1 clock and Ableton could sync
- Meeblip Cubit Go MIDI IN does not accept funny OUT 1 clock, Ableton cannot sync (in fact doesn't even see the clock signal at all), OUT 2 fine
- amateur capture using a logic analyzer shows different output, suggests OUT 1 emits alternating 0x00 / 0xFF

We're reaching a point where I may have to ask for how you handle warranty cases, or even your return policy. I still really hope it's something simple I could potentially fix myself or is even a software thing. Meanwhile, I might try another one or two synths or FX devices.

[Simon]

Thanks Jan, I received the files.

Yes the device is under guarantee of course, I'll repair or switch your device if there is something actually wrong with it.
It could potentially be a hardware design issue, but then someone else would have found out about it? I know a few people use the OXI. Let's wait and see what Manu says - he has not replied yet.

And let me know if you manage to test it on other synths. The Cubit Go has a green LED lighting up when receiving MIDI - is it lighting up when connecting to the OUT 1?
Also - could you maybe use a MIDI monitor software for example MIDI View or MIDI Monitor (mac only) to see what you receive on MIDI Out1?

Thanks.

I'll also make some measurements next week to see if the hardware design has any flaws.

Simon

EDIT: I may have an explanation for your measurements results - warning, this is technical
Basically, there is a slight difference between the two outputs, MIDI Out 1 uses an external transistor, while MIDI Out 2 uses the MCU's transistor. That means MIDI Out 2 has a protection diode which pulls it to 3.3V, while Midi Out 1 does not. So basically, when Midi Out 1 wants to write "1", your measurement could be anything (why is it regular though? could be many reasons). When Midi Out 2 wants to write "1", that little diode is enough to pull it high and you actually read "1".
If you have not understood a word of that tech stuff, in short it means the measurement for Midi Out 1 is invalid (so inconclusive), which also makes sense since your MOTU does receive the 0xF8 messages.
If you feel like doing more measurements I can send you an email with how to do it properly

[Jandt]

Thanks for trying to help. Maybe my measurement was crap.

Last news on this for today, it's late already:

- Meeblip blinking when connected to OUT 2, not blinking when connected to OUT 1 (blinking speed apparently unrelated to clock tempo)
- another device that does not like OUT 1's funny clock: 1010music Nanobox Lemondrop

The latter doesn't seem to show a bpm figure, but there's that little pie chart-like icon top right. Turns with last known tempo no matter what I do with Midronome when connected to OUT 1, while the tempo gets visibly adjusted when connected to OUT 2 and I change tempo on Midronome. It's pretty clear the Lemondrop doesn't like Midronome's OUT 1 either. For completeness - its manual states:

If the lemondrop receives a clock signal over MIDI or the Clock In, it will follow that signal. The nanobox will follow the first clock available from the following list:
1. Clock In
2. MIDI Clock
3. Internal BPM from the Preset Config screen.

So there's nothing that can be configured.

[Simon]

Hi

I investigated, made some tests, and I can confirm there is nothing wrong with the design.
Manu has also got back to me and tested his Midronome with his OXI and he says everything works fine on both MIDI ports.

So I believe your unit must have a faulty port - we can sort this out over email

Simon

[Jandt]

Thanks, much appreciated. Looks like I'll have to return it.

[Scott M2]

Hi Simon, I was verrry frustrated with trying to sync my OXY-ONE to MIDRONONE too. It syncs just fine to my ERM MIDICLOCK.

I just did a net search and found a forum entry on the OXI site and here with the same problem.
From this info - I tried using MIDI OUT 2 on MIDRONOME and it works fine.

From MIDI OUT 1 - I get a MIDI tempo on the OXI at about a quarter or show with a fluctuation tempo.
The Play button does not start the OXI - except after many tries - and then it does not stop the OXI.

But - like Jandt's box, MIDI OUT 2 works fine.

I usually use SEP under PLAY - but I've just tested the other 2 modes and MIDI OUT 1 still won't work with the OXI ONE.

The MIDRONOME was ordered in Dec 2023 and immediately updated to FW 2.

Peculiar!
Scott M2

[Simon]

Hi Scott

Very sorry to hear - we investigated this issue heavily with Jandt (who created this topic), I also spoke about it with Manu (who designed the OXI One).

The conclusion is the original OXIs have a slightly different MIDI input design which is very sensitive to tolerances - if you get a newer OXI model instead it will work with both Midronome's MIDI output. In the meantime simply use MIDI out 2

Why it works with MIDI out 2 and not 1 is because of a slight electrical design difference. You can see it this way: the MIDI out 2 is right on the limit of "not working", so it just works, while the MIDI out 1 which has a different electrical design is on the other side of the tolerance, so it does not work.

And before you get there, no there is nothing wrong with the Midronome's MIDI out 1, it's way within tolerances of the MIDI specs and it will work with any MIDI input which respects the MIDI specs.

Simon

[bechillnow]

I'm experiencing a different OXI One sync issue, but I need to try this Midi 2 advice as I do have an early model OXI.

My particular issue it that the OXI is getting clock via DIN directly from the Midronome, but then sending clock back out to connected gear at exactly 2x the clock speed. It seems like a midi-loop, but the only midi-in in use is the Midronome for clock. The only other loop cause that I can think of is that I have both devices powered by the same USB power distributor, which is not sold as USB hub, and just sends power out to 8x ports for my pile of USB gear.

I've spent a lot of time in the OXI adjusting different midi settings without much luck, so now I'm wondering if the usb power distributor could be redistributing the Midronome usb clock signal to the other ports despite it not being built for that and has no dedicated input port.

As a trouble-shooting effort, I want to disable the midi-out via USB on the Midronome, but I couldn't find how to in the manual. So I came here to ask if it's possible and how to do it.

[Simon]

As a trouble-shooting effort, I want to disable the midi-out via USB on the Midronome, but I couldn't find how to in the manual. So I came here to ask if it's possible and how to do it.

I am not sure what you mean, you want to disable the USB communication, or the clock on the DIN-MIDI ports? None of them are possible unfortunately.

My particular issue it that the OXI is getting clock via DIN directly from the Midronome, but then sending clock back out to connected gear at exactly 2x the clock speed.

I remember hearing this before, it sounds like a bug in the OXI, or maybe it's related to the issue with the OXI MIDI input getting wrong data (in this case twice the speed). What speed is the OXI running at, simple or double?

I would strongly recommend that you do not get any devices sending clock except the Midronome, its main purpose is the generate a stable and precise clock, if you send it to a device and ask that device to "recreate" a clock to send further you are not using the Midronome clock as a base anymore but that device's clock, which might have jitter, latency and other issues.

You can either use a MIDI merger to merge clock and other MIDI data, or you can send MIDI data to the Midronome on its USB interfaces then it will merge the data for you, and without affecting the clock.
When the physical add-on modules will come out there will be a MIDI input module, so you will be able to do the same but with a DIN-MIDI input (instead of USB).

I hope that makes sense and clarifies your situation a bit

Simon

[bechillnow]

After more troubleshooting to try and repo the bug reliably, it really does seem like the Oxi is the culprit.

When I hit play on the Midronome, everything in the system starts and works as it should. When I stop the Midronome clock, it stops everything also as expected. But if I then hit play on just the Oxi, that's when the 2x clock starts coming out, even though the Oxi shows the initial tempo that it received from the Midronome. It also switches clock mode to internal when the Midronome is not sending clock, so there might be something happening in that switch over.

I didn't realize that the clock was being recreated when it was passing through the Oxi. I have a lot going on in my setup and separating clock and sequencer data for each piece of kit seems tricky. I'll try your suggestion of connecting the Oxi to the Midronome via usb, and then passing the Midronome-out signal to my splitter, instead of the current Midronome -> Oxi -> splitter routing.