grey
1539
I was wondering about that too. From my quick look at the Teensy documentation, there is an internal pull-up mode, but the source code isn’t using it, and it defaults to inverting the high low read behavior. I guess it’s something I could experiment with, but yeah, the schematic doesn’t have any resistors in that path. That was also my suspicion, but I also assume it should work because there are already so many units working out in the wild using the standard source code? (For the record, I am using the Pusherman boards, but there is nothing to indicate anything is different than the current GitHub versions of the files.)
I would probably just hardcode the addresses and move on, not that that’s helpful for your troubleshooting 
1 Like
grey
1541
I’m underslept at the moment and hadn’t considered that! So a welcome suggestion! 
That said, I would still like to get it working hack free, but that does give me a work around. I’m hoping to do some crow dev work with it this weekend, and that might do the trick.
1 Like
I had a problem with my TXi that @GoatFarmer built for me. Annoyingly, I can’t find my notes about what the problem was exactly, but I’m pretty sure it’s the same as you found. We ended up fixing it in the firmware (again, I’m so sorry I can’t find my notes of what we changed!). Now it works perfectly. It really baffled us for a while since we assumed it must be a hardware problem and it was only by looking carefully at the code that we discovered that it shouldn’t really work as it stands. Maybe something has changed about how the floating voltages on the teensy work?
If I can find the relevant file (or maybe @GoatFarmer will remember), I’ll let you know.
1 Like
Quick q, now that I’ve finally got my hands on a 16n, I was wondering if anyone makes a TXb? Or do I need to go hunting for the BOM and panel myself?
TXb Github
There you go. I saved you 10 seconds effort
I think BPC music is the sole maker of them. There’s a TXb commerce thread you could ask in.
grey
1545
2 Likes
If you don’t mind a tiny bit of DIY, the MSW F8R expander can be modified to suit your needs. It just needs 3.3V to power its I2C repeater chip, so a single regulator is all you technically need to add.
1 Like
FWIW, i already have a powered i2c board, it’s just the ease of connection from front panel that I’m lacking.
Probably at RTFM, but just bought boards for DIY Teletype and TXo and TXi Expanders.
Will I need also to source an ii backpack/busboard to connect both expanders, or can they be daisy-chained from the single header on the Teletype?
grey
1549
You shouldn’t need a backpack for just those.
Be sure to read over all of the documentation: https://github.com/bpcmusic/telex/blob/master/extras/TELEX%20Documentation.pdf
Don’t overlook the extras folder:
Does anyone know if it’s possible to do jack detection on the TXi in’s? I’d like to be able to use dummy cables to do some configuration, but I don’t actually know how modules that do jack detection work. 
I think to do reliable connection-detection, you need to send a known signal to the normalisation pin of the socket, then monitor the input from the socket to the MCU and compare the two.
This obviously relies on the correct circuitry to work. If the normalisation pins of the socket aren’t connected to the processor, it’s not going to be possible.
Sorry this is a bit vague. I’m not an electronic engineer myself, but remember reading @pichenettes explanation of the principle on the Mutable Instruments forum a while ago.
The jacks have an extra lug that is in contact with the tip/signal lug when no cable is inserted. When a cable is inserted the shaft of the plug pushes these two lugs apart, breaking the connection.
The most common ways to use this are:
- for something simple like “this jack normals to +5V”, wire the detect lug to +5V (probably through a weak pull-up so that 5V isn’t immediately shorted to ground or whatever’s on the other end when a cable is being inserted). Inserting a patch cable breaks the connection and instead the signal is connected to the tip of the patch cable plug.
- for detecting if a cable is plugged in, do some kind of impedance measurement between these two pins. If there is close to zero impedance, there is no cable plugged in, if there is a very high impedance a cable is present.
More complex or specialized schemes are sometimes used depending on the application.
My read of the TXi schematic indicates that the detect pin isn’t connected anywhere, so jack detection is probably not possible.
Much more detailed discussion here.
2 Likes
grey
1552
Something like that was my suspicion. Drat.
Thanks for taking the time.
@a773
Not sure if you have solved your LED problem. I’ve had a few units that had this issue when I built them.
I was able to resolve it consistently by a) making a bit more space for the LED by filing out some additional room in the panel and b) ensuring that the LED is cut to the proper size (not being smashed up against the panel). Tolerances.
You can also check the headers that pass the LED voltage from the back board to the front board. If they are loose (and sometimes can be if you are using the machined/precision headers), you could have spotty connectivity.
Agree. It shouldn’t work as designed - very embarrassed by it, but I was overwhelmed with details when I was putting these together and not surprising that I got tripped up by literally the first thing I learned when starting to work with microcontrollers. Isn’t that always the case? 
So weird that it worked for the gazillions of units that I built. That certainly lulled me into a false sense of security.
Something must have changed in either the Teensy itself or differences in the manufacture of the Pusherman boards (though, I doubt the latter as I’ve built it successfully with three different board manufacturers). Could just be that the state of the voltage in your case causes issues? Maybe the phase of the moon? Who is to say with a floating pin…the chaos of the universe is at play here.
If it doesn’t work as designed with your components, you can always change the firmware detection value or simply hard-code the address. Apologies. If I ever make a 2.0 version of these things, this is top on my list of corrections!!
2 Likes
grey
1555
I wish that I didn’t have another issue to post, but I do!
Both of my units have stopped communicating during use a few times. It seems to take a while, but I’m playing with my system so tracking time is… difficult. This time I dragged a computer over, and verified that druid could no longer recieve a response from TXi. The crow script is still running happily, but it can no longer read for updates. I’m hesistant to leave the TXi’s in debug mode as I’m unsure of how the added overhead effect things, but I guess that will need to be my next stop when I try to get to the bottom of this.
If anyone else has had issues with their modules ceasing to communicate after using them for a bit, I’d appreciate any insight you may have observed.
Has anybody sequenced their TXo through Crow/Max? I’d like to try and get something like that rolling.
1 Like
TELEXo C102: says in the BOM it should be 1206/180pF/1%.
Is it vital the part is 1%? I can only source 5% versions locally.
I had TXi drop off the bus in a multi-leader setup with crow and TT. So I can echo the stability concerns but my setup was obviously out of spec.
