Telex: Teletype expanders

[csboling]

@EqualTemperament

X will almost never have the correct value. It will be something between what it was and what it’s supposed to be.

I have not personally used a TXi but from the code it seems this is pretty much how it works. TXi polls its analog inputs and quantizes them on a timer once every 1000 us == 1 ms, and sends back the cached value when it receives an I2C message. This is presumably done to make responding to an I2C query very fast, whereas waiting for ADC conversions and quantizer calculations could hold up the I2C bus and stall Teletype if the new value was acquired on demand.

By contrast Teletype’s IN op does an ADC read immediately when the op is called. This still requires communicating with an AD7923 chip over SPI but this is much faster than I2C communication (limited by bus capacitance, i.e. wire length, the TXi end doing its thing, and in a multi-leader setup potentially having to wait for the bus to be free).

[laborcamp]

When working with my tiny er-301/teletype/ansible box, I realized that I was often wishing for a manual gate/trigger gesture with which I could trigger scripts in TT or affect things directly in er-301. So today I quickly rolled my own “expander”.

I think the whole thing cost me $10 or so. For gates and triggers I have one TT output send a steady 10V to the input on my 2hp thingie. This is hardwired to 4 momentary switches, and 4 outputs. So, pushing the button sends the voltage to respective outputs. Of course I could also send some other CV stream to that input, and then manually direct it to the outputs, but for now, I just wanted some way to manually fire scripts (without any interface attached to the box.)

Anyhow: I kind of like the fact that it’s just wires and switches. A sort of very specifically laid out, pre-wired, passive patchbay.

[woofy]

Hello I had a question about the build, does the stripe o the opamp 4171 indicate top or bottom of the component?

[markel_m]

The stripe is always the top, so pin 1 is to the left of the stripe.

[woofy]

Thank you very much!

[grey]

Finally got pusherman boards and soldered up all the SMT on a few units. Potentiometers seem to be in short supply, but finally got a few today to finish up my first test unit. Teensy (3.2) programs okay (0.18 and 0.21B firmwares), but when I go to read from it (TI.PARAM 1), TT (3.1.0) is freezing. Any suggestions on how to approach debugging this lil’ dude?

[thopa]

I think its plugged wrong.
Where you have the ic wire go the adress jumpers.
Then you plug the ic cable to the connectors on the top left.
Hope this helps, i can take a picture later.

[grey]

Yeah, definitely seem to have that wrong! But the board isn’t labeled clearly, so I’m not sure what should be right offhand. I seem to have tried all of the possiblities, each resulting in a freeze or 0 for all values as if disconnected. But a step in the right direction! Thank you.

---

Edit: After checking Scanner Darkly’s post here:

scanner darkly

Now I think I have it right?

Still no dice on TT (plugged in via backpack), just getting 0 for all param’s and inputs, which is the same as if it weren’t connected at all… so. Will have to sleep on it.

[kasselvania]

I had some trouble as well. It’s weird, but try using the pins just to the right of the ones you have plugged in to, but keeping your actual plug aligned the same way. It’s a little strange. I found the i2c plug ins of the TT expanders from pusherman to be odd.

[iamkabuki]

This is an issue in general with quantizing voltages, and it has a lot to do with the distance between the voltages. You might need longer delay values when covering wider intervals such as a 4 octave jump compared to a minor second. Looking forward to the next episode of Teletype Talk!

[thopa]

Its that an old tt?
You might need a txn or teletype backpack ?
Just wondering…i had no problems with most length wires unless tgeir orientation was wrong.
Hope you get it fixed!

[scanner_darkly]

make sure no jumpers are in the address pins (so that it uses the first address), and the highlighted area is where you connect i2c. the line at the bottom is the ground, make sure to connect ground to ground. shouldn’t really matter if you use left or right pins - this was for daisy chaining multiple devices.

if you’re just getting all zeros it might mean 2 things - either your i2c is still not properly connected (try a different cable, try a different i2c device if you have one with the same cable etc) or the connections on the txi itself are loose. i seem to recall @bpcmusic saying that when it’s too “perfect”, some connectors might float without good contact - try gently moving the boards against each other.

you don’t need a txn or a backpack with just tt and txi.

[grey]

Something is up with that Teensy. I soldered up a second one, and it’s reporting values now. Thanks for the tips!

Question: My CV in’s never 0 out, is that expected? They jitter around 60-70.

---

edit: After connecting it to crow and messing about trying to get a reading, it’s failing to read with TT again. So back to diagnostics: Tried a few i2c cables. Tried reflowing the joints for the i2c pins and the teensy pins and it went back to freezing. Something is clearly making this set of boards inconsistent; will consider again tomorrow.

---

Tried reflowing all of the joints for the pins again on the bottom board. Still freezing.

Finished a second module. Continuity seems to be good on power and the i2c lines to the Teensy. I can flash the Teensy no problem on either unit. Accidentally inverted the header placements between units, so sadly I cannot mix and match front and back boards as I was hoping in order to try to aid in diagnosis. However, both Teensy’s behave the same way in both units, so I’m not sure why the new Teensy momentariy worked in the one unit.

I’ve tried 3 cables, with and without the backback on TT for kicks, with no effect on behavior, so it would at least appear to be the units themselves. Also tried two firmwares, just to be sure.

That’s about all the dianostics I can handle for the moments. Will have to try again later.

Maybe I need to run compile the firmware so I can debug what is happening Teensy-side?

[grey]

Okay. I did a thorough review of both of my units and hit anything that looked remotely dubious with the iron for a reflow. Connected them, only getting 0’s for return values on Teletype. So I setup the firmware tool chain, and when I debug, everything is reading correctly!

ConfigID: 111
skipping - eprom not initialized
0=0; 1=0; 2=0; 3=0; 4=0; 5=0; 6=0; 7=0; 
0=9630; 1=8702; 2=10530; 3=9908; 4=-3345; 5=-3341; 6=-3345; 7=-3337; 
0=9632; 1=8702; 2=10532; 3=9910; 4=-3345; 5=-3341; 6=-3345; 7=-3341; 
0=9632; 1=8702; 2=10532; 3=9912; 4=-3345; 5=-3345; 6=-3345; 7=-3337; 
... etc

Is that the correct ConfigID for the first device? I presume so. I’m at kind of a loss here. When I mess with the param knobs values change as expected. I tried to do the calibration biz from the the manual, seemingly to no effect - which makes sense if i2c isn’t working.

The obvious problem is that despite being in debug mode, i2c commands are not showing up on the TXi. Continuity has be verified many times on the i2c pins for both of my TXi’s. I hid the flood of value data to be sure, and nothing else is tracing. I’ve tried crow and TT with no seeming variance. I feel pretty confident I have my cable orientation correct. (And even when I don’t I’m not getting an error seemingly?)

I’ll keep thinking on this, but if anyone has any suggestions on how I might further debug i2c communication, I’d welcome a suggestion. Here’s another photo for good measure, connected to crow.

[Brennanib]

Any possibility you would be into making more of these? Really exactly what I need for quick manual gates and such

[kasselvania]

Did you get this fixed? I had an issue with one of my TX’s and it came down to faulty pins on the teensy. I would get power and such, but the i2c would always come back as 0

Here’s a pic of how mine get messed up;

Now, I’m not suggesting yours got as goofed as mine, but I do believe this could be LOOSELY possible for your setup…

[Random]

Possibly silly question…

I’ve been kind of ignoring the whole Telex situation for a while… I had my Teletype unmounted.

I have a few of the original ones. Are they still ‘valid?’ Do I need to flash update them?

They are TXo and TXi’s. The first edition ones that are somewhat loose (you can pull them apart because they are simply plugged together).

[kasselvania]

They all still have that structure. The standoffs can be a little loosey goosey. I wouldn’t worry too much about it. Honestly, it sounds really silly, but if you’re gonna be having the telex modules unmourned, maybe just throw a rubber band around the modules to keep them sturdy.

[Random]

Thanks!

I ended up with two of each, I should probably just get rid of two since tbh I’m probably never going to use them all with the way I patch.

[kasselvania]

Yeah, I’ve found the best way to use the TX’s is to actually prepare a bunch of utility tools, either as scenes to load or printed out as a companion packet, so I can just create all kinds of utilities on the fly. Having 2 TXo’s and TXi’s would be WAY beyond me.