Yeah, I did a crap job of labeling those things in the BOM. They just inherited their part names - but aren’t really tied to what you need to do. That is up to you. Just look at the number of pins, rows and gender.
I made the initial test TXo with standard headers all around. For the production run, I switched the board interconnects to precision headers (including the teensy one). This slimmed the module. You can also just solder the boards together and skip the female headers, but this makes makes it hard to troubleshoot.
If you need more help from here, I can go through and improve the list in the BOM. Might take me a couple days to get to it…
I’d love an improved BOM too if at all possible. I’ve a few builds under my belt now and feel that I could make a good job of building a couple. I’m still too green to understand the headers required but I’m sure I’ll find a way.
hi @bpcmusic – i’m also DIYing a TXi and TXo. I’m thinking of just using front panel express. Just curious if you happened to have a FPD file for the front panels from prototyping stages etc? (currently attempting to convert ai/pdf -> hpgl(plt) -> FPE software = PITA )
No worries if you don’t, just thought i’d ask!
I haven’t used FPD in several years. I’ve not figured out a good workflow to deal with hpgl files. I thought that the new FPE software was better at this stuff.
On another note, I pushed an update for the TXi and TXo BOM files that adds more clarification about the different types of headers and some links to other parts (including the LEDs which are important to match the specs if you want proper brightness and power consumption with the resistors in the circuit).
I didn’t link to specific pre-cut headers at Mouser as I find that just getting some cheap full-row headers and cutting them to suit is far less expensive. I also suck at finding them using Mouser’s search. Like, I just get totally lost in a sea of weird crap.
I’ve always been using Front Panel Designer with OSX, never had an issue…
The only issue I have is with my design abilities!
While I’m thinking about it, there is one thing I have to do after the export from Inkscape (or, to put it differently, before importing the HPGL files to FPD) : add the “.plt” extension to the HPGL files (surprisingly, Inkscape doesn’t do it).
Thanks again for reporting this and sending the video. It was super-helpful in me isolating the issue. I also used my ER-301 and O’Tool to monitor. That was a helpful tip (re: ER-301) as well.
I was able to replicate for envelopes where the attack phase was less than 1 second. I was able to trace it to a stupid counting bug. Added a “+1” and all was good with the world. Sorry. Sigh.
ANNOUNCING NEW BETA TELEX FIRMWARE
Refactoring of codebase to support some Future Fun Firmware1. For those hacking around, the development switch is now in the “defines.h” file.
Repaired the “Envelopes Sometimes Dip Below Zero Due to a Stupid Counting Thing” error from above. Now, envelopes always end at their origin (and hit their peak) regardless of the attack and decay times.
CV LEDs now always represent an envelope’s peak brightness; this used to sometimes not get bright enough for fast envelopes - esp with log curves.
Wavetableapalooza: the TXo running with the Teensy 3.2 now has 45 different waveforms. You can access them with the TO.OSC.WAVE parameter, which now increments waveforms every 100 steps. For example: 0 = SINE, 100 = TRIANGLE, 200 = SAW, 300 = SQUARE, …, 4500 = RANDOM/NOISE. Waveforms are taken from the AdventureKid free waveforms pack. I arbitrarily selected a handful for now; this may change in the future.
1 The “Future Fun Firmware” mentioned above is the support of the Teensy 3.6 on the TXo’s hardware. I’ll be posting a much longer bit about this in the future. Don’t go and grab a 3.6 just yet; be patient until I’ve got all of the details worked out. It is a more complex than just slapping the new processor on.
I am, however, very close and very confident that the upgrade will work. For example, here is a video of the TXo running with the Teensy 3.6. It is scanning through the wavetables that are supported in 3.6 mode - which in this demo is 327.
BTW: for the 3.6, sampling rate is now set to 24k with interpolation active all of the time (interpolation turns off if you start doing any complex interactions on the TXo running a Teensy 3.2; its sample rate is set at 15.625k).
A little update with minor improvements for both modules.
Additional smoothing for potentiometers vs. CV input jacks. The result is that the pots will be smoother and settle at a position. The CV values will continue to be super-responsive and may “twitter” a bit between values.
Enhanced the Envelope Generator to handle changes in peak voltage and offset during an active envelope.
A lot of stuff to parse below. Hope this is the best way to share with everyone. To those that hate long posts, sorry.
I’ve been banking up a whole bunch of stuff over the last several months. I’m going to officially dump it on you all at once.
What is Happening with the Expanders?
I’m currently working on a third run of the TXo output expander. I will also be making a small number of TXi input expanders. All of the parts for this run have arrived and I’m about to start the production process.
The units will be made available in the store when all are built, tested, and ready for delivery. That means that we are still a ways away from these being made available. If you are interested and want to be notified when they are ready to go, please click on the “Express Interest” button up at my store.
Additionally, I’ve added product variants for the expanders so that those who are DIY-inclined can order panel and PCB sets. Select the “Panel and PCB Set” from the type dropdown to register your interest. These should be packaged and ready to go in a few weeks.
What is a TELEXo+??
I’ve finished the R+D work on a supercharged version of the TXo that swaps out the Teensy 3.2 for a Teensy 3.6 and adds some interesting capability. The module stays the same (so you’ll need a TXo to start with) - we’re just upgrading the processor.
Improved Sampling Rate - increased to 25kHz from the TXo’s 15.625kHz
Improved Quality when Combining Features with Oscillation - the stock TXo disables interpolation if more than one advanced feature is used at a time for a CV output (envelopes, wave morphing, pitch slew, etc). The TXo+ always uses interpolation which increases the accuracy and lowers distortion.
More Morphing Wavetables - the TXo+ has 327 different waveforms on board (up from the TXo’s 45).
Future Advanced Features Only Supported by the TXo+
Now, you can’t just go and grab a Teensy 3.6 to make this happen. Due to some peculiar quirks of the 3.6, it needs a tiny daughterboard to be soldered underneath in order to ensure that the processor boots when connected to Eurorack power-supplies. It will also need to be outfitted with the same precision headers as your existing Teensy so that it mates properly with the module.
I’m going to make an upgrade available at my store that will include a Teensy 3.6 outfitted with the proper headers, the daughterboard, and the latest firmware at the time of shipping.
Price is still up in the air - I’ve put a ballpark on the site for now. Register your interest at the URL below by clicking the “EXPRESS INTEREST” button. I will most likely be making these available when the current run of TXo is complete.
What is this TXff thing you made??
With the enhanced sampling rate of the TXo+, I wanted to put a little module together that cleaned up some of the aliasing that @Leverkusenpointed out above. This is so that we can get pristine sine-waves out of the unit at some reasonably-high frequencies.
The TXff is a little 2HP module that has a a set of four, three-pole filters that do a post-DAC scrub on the output of the TXo+ (augmenting the unit’s internal 1-pole filter with a cutoff of around 12kHz). Jacks are oriented so that the four CV outputs of your TXo can be patched to four adjacent inputs; the outputs are at the top of the module.
Here is a comparison of the RAW signal out of a TXo+ (green) and the filtered output of that same signal processed by the TXff (red):
Now, I want to be pretty clear on this, the TXo/TXo+ will still have lots of aliasing if you use waveforms other than the sine wave. The filter helps with these, but much of the aliasing is happening in software (due to the sampling rate of the unit) and cannot be corrected after the output.
That means that this is really a luxury module for those that care about using the sine waves at audio rates in circumstances that would benefit from the cleaner signal. You most likely don’t need it.
I’m planning to do some band-limiting tests on the SAW, SQUARE, and TRIANGLE waveforms for the TXo+. If it works out - and I’m pretty confident that they will - then they should also be enhanced by the TXff. At that point, the TXo+ with the TXff will give you four pretty solid oscillators for the primitive waveforms. I just want to be clear that this is still in R&D.
If you are interested, go and let me know up in my store by clicking on the “Express Interest” button. These will be available after the completion of the TXo build - but I’ll be using the interest check here to decide on final pricing and order quantities.
The TXb is the third module I had planned when I set out to do these expanders a few years ago (you can see early shots of it way up at the top of this thread). In final form, it is a remix of the open-source monome “tt-busboard” in 2HP with a pair of stereo 3.5mm i2c jacks on the faceplate. These jacks make it possible to connect multiple TXb units together or to connect to the forthcoming 16n fader bank over i2c. It is also handy for folks who don’t have the depth in their case to support the existing busboard mounted to the back of their Teletype.
Here it is with a black prototype panel next to a Teletype (final panel will match the material and style of the TELEX):
Now, you will need to be mindful to have only a single busboard connected to power. You will also need to keep your cable runs as short as possible. Long stereo i2c cables will cause the bus to fail and your Teletype to lock up. Finally, you will only want to carefully make connections when things are powered off.
If you are interested, express interest up at the store. I’ll be building these folks after my TXo run is complete.
Also, I’ve had enough requests for the type of jumper cables that I shipped with the TELEX that I’ve made them available up at the store as well.
I’ve also put up a version of the i2c cables for the ER-301 and Matrixarchate that swap the order of the clock and signal pins:
What is the TXn?
I also did a 2HP remix of @billyhologram’s amazing 7 input panning mixer “nearness". I’m planning a run using the same satin aluminum, knurled nuts, and panel treatment as the TELEX. If you are interested, head up to the store to express it:
Find out more about nearness (and the awesome 3HP canonical version) here:
Here is a family picture of the prototypes; the final versions will have panels that match the TELEX:
If you made it this far in the post, I’m impressed. Sorry to dump it on you all at once.
If any of what I talked about above interests you, head on over to the bpcmusic store and click the “Express Interest” button to let me know. This will help in planning, setting of final pricing, and the ordering of parts. I plan to do that while I’m building the TXo so that I can move on to these units immediately afterwards.
I’ve registered interest on your site
Very cool! Lots of teensy stuff around but I’ve never built anything with a 3.6
I don’t even own a teletype (yet) but was looking at getting boards made just yesterday for these, to hack for use as stand alone i2c expanders for Ansible and JF. Teletype probably on the horizon though