Telex: Teletype expanders

[jnoble]

try this on the output with the envelope. [… awesomeness redacted …]

Wow, I haven’t read the docs in a while, and I never even saw this! Thank you!!

What I miss most from my three now-departed Maths is the continuously variable expo-linear-log response: there is a ton of room for shaping tone and volume over time in there. I’m not asking you to do it, but something like TO.ENV.SHAPE x y that had a range of, say, -100 (full expo) to zero (linear) to 100 (full log) and respected the current TO.CV value would be the ultimate.

But I’m so stoked about having TO.CV.LOG pointed out to me that I don’t care much about pie in the sky stuff like that.

[jnoble]

But I’m so stoked about having TO.CV.LOG pointed out to me that I don’t care much about pie in the sky stuff like that.

Just a quick followup about this: it seems that TO.CV.OFF is applied before the log conversion. Is this the intended behavior?

[bpcmusic]

Just a quick followup about this: it seems that TO.CV.OFF is applied before the log conversion. Is this the intended behavior?

It clearly is how I added log to the system. The behavior … more unintended.

Really good find it a tricky, performance critical part of the code. Out of the country at the moment - but I’ll give it a peek when I get back to see if I can make the behavior more intentional.

Thx!!!

TO.ENV.SHAPE x y

Maybe? Cycles are almost exhausted on the TXo - not sure if I can fit this one in. But, when it is on the table, I’ll take a look!!

[a773]

It sounds really sick. Regardless of whether I built it or not, I’d be happy to take a look at it. PM if you wanna set that up.

Thanks, that’s a mighty fine offer, on which side of the pond are you located?

Anyways here’s an (very long) update

It dawned on me the that the two new txos are in fact txo+. I reflashed firmware in all units, no change. I cleared out a section of my case so there were absolutely no cables to touch the backside, things started working. I started switching the parts of the two modules (to see if it was a faulty teensy) and finally things started behaving as expected, provided I connected i2c to individual pins on teletype and not daisy-chain as recommended. Everything back in the case with cord round the modules. After happily watching my test scene run perfectly for about 10-15 mins, the whole case powers off, which to me is the worst thing that can happen pretty much. Although there should be more than enough power, I simply removed the two txo+ from the case, powered back up and everything was fine again.

At this point I have no more time to test, debug. The only reason I can think of as causing the power out is the case drawing too much power, but I should be well below the capabilities of my PSU (have written to the manufacturer for his opinion). My main conclusion was that the electrical connection between the two main boards are really, really flimsy and unreliable, maybe too unreliable for me.

Disillusion mode on: My hope for a i2c super brain (teletype, 4x txo, 1x txi, er-301), seems to have been to high. Unstable expanders, unstable i2c, I get the feeling (and comments from people high in the food chain) that the system is not really designed to handle that I was hoping to throw at it.

I have a gig to prepare, so I have to come up with something. Although maybe not out of the most ideal of reasons, something new came about. I started triggering teletype from 8 channels from my newly build swt16+ sequencer. This was a lot of fun, actually much more fun than the rigid triggering-scripts-from-M I did before. Not practical for all types of music, but I’m going for a more imprecise musical expression for this one, so it seems a fine.

I have to reconsider my next steps

1. Should I give the 2x txo+ a final chance?
2. Was I using the system (all things i2c) in a way they was never designed to work?
3. Is there in fact another (more stable + more fun + more open ended) way that either completely or partially could/should replace my flawed dream (sequencing the whole case from teletype M script), namely what I played with yesterday: triggering teletype scripts from swt16+?

If anyone is wondering how I was (ab)using the teletype before yesterday, and would like to weigh in on whether it was sane, many of my scenes including links to youtube video of me performing them is on github: https://github.com/attejensen/a773_teletype

[bpcmusic]

Disillusion mode on: My hope for a i2c super brain (teletype, 4x txo, 1x txi, er-301), seems to have been to high. Unstable expanders, unstable i2c, I get the feeling (and comments from people high in the food chain) that the system is not really designed to handle that I was hoping to throw at it.

I hope that I didn’t contribute to your impression. That system should be unbelievably capable and very stable as long as you have a stable PSU and a well-configured case.

My main conclusion was that the electrical connection between the two main boards are really, really flimsy and unreliable, maybe too unreliable for me.

Not sure the state of your second hand Telex. If I built them originally, I endeavored to ensure that the boards were fairly snug before shipping. That said, with time and abuse, it is possible for them to loosen. (The broken legs on one of your units suggests some abuse - perhaps during shipping.)

Zip ties, string, or dental floss will help make them more secure and road-worthy. Soldering the boards together (as @kbit demonstrated above) will guarantee it.

Was I using the system (all things i2c) in a way they was never designed to work?

Doesn’t seem like it to me. As long as you don’t poll inputs at too high of a rate (faster than what the Teletype and i2c bus can handle), you can do a lot.

how I was (ab)using the teletype before yesterday

What do you think was abusive about the way you were using it?

[scanner_darkly]

at some point i did some pretty heavy testing of teletype + 4TXos (plus iirc more on the i2c bus) with things running at super high metro rate for like 40+ hours with no issues. this is not a proof that such setups won’t have any issues but it does show that it is possible to have a stable setup with many devices on the i2c bus.

as mentioned in the other thread, there could be many variables contributing to the issue you’re seeing. have you tried only having teletype and TXos in the case with nothing else being powered? power shutting down in the whole case indicates a separate issue, not i2c related. it’s either it’s severely underpowered, or something is possibly shortcircuiting. the only way to rule this out is to put just the teletype and TXos in a case, make absolutely sure nothing can touch the back of the modules and leave it like that with no i2c running and see if power still shuts down. if that works, then try i2c. there is a small chance i2c commands just create a condition where modules draw more power, so if this happens try just leaving one teletype and one TXo and see if you still get the issue.

have you tried the test teletype builds i posted?

[a773]

Thanks for the reply @bpcmusic and @scanner_darkly you infuse some hope in me. Lots of good suggestions, will spend som time when I get back to the case going over them, but there’s a gig coming up, and I need to prepare that pretty soon.

In the missing triggers thread, I got the impression that my problems was simply due to overloading the i2c. And please understand it from my side of the table: everybody’s saying the problem is somewhere else, which I totally understand, but doesn’t get me closer to a solution.

No I didn’t have time to install the new tt firmware.

The psu is 5A for 6u 126 hp, i can ensure you the power draw is below that. Psu is trogotronic, I know it’s a brand a lot of people love to hate, but besides the power outage in question it’s been running steady here for years. They suggested that a power conditioner might clean up the power and improve stability on the txos.

Edit: I have some time Monday, will follow all the great suggestions here very systematic, and report back.

[foz]

I also have trogotronic power for about a year (18u*168hp). I had a problem a while back, funny things happened with some modules. After contacting them, I tested all the power brickes with a volt meter. One of them was putting out low voltage. They replaced it (and sent an extra spare) and all has been good. I know more advice but just to say, don’t assume your power is okay, test it to be sure!

[desolationjones]

[kbit]

Well that looks great. How many hp?

[desolationjones]

3u x 4hp 1u x 22hp (Intellijel)
I’m not particularly interested in organizing a group buy, but I will for certain share my Front Panel Designer file before I order! Looks like ~$40 for a single.
I also made this (OT, sorry)

[Nordseele]

@desolationjones I could use a 1U Disting Mk4 panel like this if you decide to share the files I never found the exact dimensions… (Made some mockups for txo last year Telex: Teletype expanders , never tested though)

[jnoble]

Another idea to toss in the stew:

TO.ENV.RET x y

Where ‘x’ is the CV output and ‘y’ is 0 for no retrigger and 1 for retrigger (current behavior is to retrigger from zero, and the default). The applications are probably obvious, but right now I’m working with a grid-based keyboard and would like a little smoother legato action.

[dianus]

I was looking at the Gerbers from the Git, and couldn’t get them to appear properly without adjusting some of the file extensions. I didn’t realize Telexi and Telexo are 4 layer boards. Can anyone confirm this is correct?

[bpcmusic]

Telexi and Telexo are 4 layer boards

Yep. It was a bit of a routing nightmare.

TO.ENV.RET x y

Sounds tasty!! Let me see what happens when I get firmware hacking.

[Olivier]

I’m getting some strange behavior with my TXi, and specifically with the returned range of the param knobs no matter what calibration steps I take (TI.PARAM.CALIB X Y, TI.PARAM.INIT X, TI.INIT 1 and TI.RESET 1).

When I call TI.PRM, the returned range seems extremely limited (between 6500 and 7500) no matter what the position of the knob is. The same thing happens if I map to a different range (for example, if I map the knob 1-20, the returned values are only between 8-10). Stranger still, if I leave the knob at a set position and call TI.PRM, the returned value fluctuates anywhere within the range.

Any obvious steps I may have missed?

[bpcmusic]

That is really weird. Please make sure your boards a properly seated. On first blush, it sounds like things might be disconnected causing the values to float.

LMK what you find. We’ll certainly find a way to resolve this.

For all, when calibrating on the TX modules, don’t forget to call the STORE operator when you have it set the way you like it. Without that, you will lose your calibration on power cycle.

[Olivier]

And…we’re back in business Indeed, when I pulled the module out, one of the boards had completely disconnected. Fastened everything with a bit of floss, and it all seems to be working great. Thanks!

[desolationjones]

Knob skirts were bigger than I estimated, so I moved the text. And I can confirm that the fit is good! Here’s 1U TXi, as promised. If you have multiple TXi, I suggest editing the panel text to reflect their assignments or making one monster panel for all of them!
TXi_1u.fpd (1.1 KB)

[thopa]

Hi!

So im about to order some boards from the pcb fab since they seem to be out of stock and i dont mind smd…i wanted to ask a few things:

-the txo is 2 boards (4 layers each) ? Since in the gerber folder there are two zip files.

I get confused when opening the brd file in eagle and seeing a third board (the one of the same size as the teensy) i know it says “ignore me”

-any one has a mouser cart for the expanders? (I know i can copy and paste) just wondering if someone had recently placed a mouser order and had a cart available.

Many thanks!.