latigid on Posted June 29, 2013 Report Share Posted June 29, 2013 If you can't beat 'em, join 'em. Following a discovery that the I am working on a revised design based on SN7407 level shifters. (If you want a board from the the buffers work as expected, please PM me for details.) Simple enough circuit but I have several questions. First off: any problem with cascading the buffers this way? The idea is to have a switchable output between +5 (also serving as Vdd for the ICs) and an adjustable voltage which is derived from a low-dropout regulator (say 12 V or higher). It is more simple to route when the output of the first buffer connects to the input of the second, but would it be better to tie the inputs of both buffers together instead? What should the handling of MIDI IN 3 and MIDI OUT 3 be? Presumably MIDI IN 3 should go through an optocoupler with a 3.3 V pull-up on the Core side: http://ucapps.de/mbhp/mbhp_core_lpc17_midi3_midi4_extension.pdf Hence, no level shifting is necessary I think. But for MIDI OUT 3? From the Quad-IIC schematic it is not clear whether the signal needs to be level-shifted. Will a 3.3 V output be compatible with the 8-bit core of the BLM? In this case, MIDI OUT 3 is level shifted: http://ucapps.de/mbhp/mbhp_core_lpc17_output_buffers.pdf For these two, it isn't: http://ucapps.de/mbhp/mbhp_core_stm32_midi3_extension.pdf http://ucapps.de/midibox_blm/blm_connector_mbseq.pdf There is a 3.3 V pull-up for the LPC core: http://ucapps.de/mbhp/mbhp_core_lpc17_midi3_midi4_extension.pdf Even though I will not use the LPC core, I plan to make it compatible with both for the greatest flexibility. Some clarification would be welcome. Cheers, Quote Link to comment Share on other sites More sharing options...
latigid on Posted June 30, 2013 Author Report Share Posted June 30, 2013 I think that this can be more efficient with only two 7407 chips and a switchable pull-up voltage instead. Working on this now. Quote Link to comment Share on other sites More sharing options...
latigid on Posted July 11, 2013 Author Report Share Posted July 11, 2013 Okay, I think I'm done now. I think that it looks okay but I'm not 100% sure on the regulator circuit. If somebody could check that would be great. Also: no answer as of yet on the MIDI IN/OUT 3 situation. In any case, I have brought out the 3.3 V and 5.0 V connections. Quote Link to comment Share on other sites More sharing options...
latigid on Posted July 12, 2013 Author Report Share Posted July 12, 2013 I can squeeze in one more connector on the top left side for people who will send the signals from J19 to an external AOUT module. Should the three data lines be buffered also? This should be doable with an SMT chip. Quote Link to comment Share on other sites More sharing options...
latigid on Posted July 15, 2013 Author Report Share Posted July 15, 2013 (edited) This revision has a connector for J19, for ocassions when the AOUT module is outside the case. The three data lines are buffered by a CD4050, unfortunately it's SMD as I couldn't find a part that would fit easily. One of the additional buffers is available, there wasn't room to route up the other two so they are tied to ground. The connectors share the same pins as J4 from the AOUT. EDIT: Pic Edited July 15, 2013 by latigid on Quote Link to comment Share on other sites More sharing options...
TK. Posted July 23, 2013 Report Share Posted July 23, 2013 Hi Andy, STM32 is a bit different compared to LPC17, therefore probably the confusion. In distance to LPC17, the STM32F1xx family supports open drain mode for UART outputs, which is the preferred option for MIDI (because a MIDI line is a current loop). Accordingly, no buffer should be used, instead just create exactly this circuit: http://ucapps.de/mbhp/mbhp_core_stm32_midi3_extension.pdf Btw.: you will also notice, that MIDI OUT3 is on J5B.A6 for STM32, while it's on J5B.A7 for LPC17. For LPC17 actually no buffer is required as well, but certain (vintage) MIDI equipment doesn't work properly at 3.3V voltage level (although MIDI data transmission shouldn't rely on the level, only on the current draw...). So, for the case that you want to support LPC17, and if a free buffer is available, then connect the input to J5B.A7, the output to the "data pin" of the MIDI OUT socket via 220 Ohm, and the "supply pin" of the MIDI OUT socket via 220 Ohm to 5V. Best Regards, Thorsten. Quote Link to comment Share on other sites More sharing options...
latigid on Posted September 8, 2014 Author Report Share Posted September 8, 2014 (edited) Okay, it works! Sorta... So this is my one year old level shifter/DB-25 breakout project. It works! +/- 12 V is coming in off the Eurorack bussboard and powers both the open collector 7407 buffers and the AOUT_NG mounted directly above. This simplifies the SEQ power supply a great deal. +5 V is taken from J5C at the moment, but there's also a 2-pin connector for this. 8 channels of CV connect directly to the DB-25 port; I used a 16-way ribbon, but a samich connection is also possible as the spacing lines up perfectly. The CV and gates are wired to two rows of 3.5 mm sockets. You can adjust the gate voltage near to the +12 V rail (10.3 V seems to be the maximum you can get out of an LM317) and the triggers fire as expected. Although I just soldered in links in my case, you can switch between +5 and +V gates with jumper pins. I didn't test MIDI OUT 3, DIN sync or the LPC connector as of yet, but they are wired up according to schematics. There's room for a buffer (SMT) if somebody wants to try it with AOUT J19 data. TK recommends differential line drivers however. Silly me, I neglected to put the correct spacing for the J5 connectors :rolleyes: Fortunately I left 100 mil spacing, so it works fine with one 34-way IDC from a STM32 F1 Core comprising J5A, B and C. It might also work with one 10-way cable underneath the board? I have nine spare boards, so please PM if you would like one or more. They will cost 6 EUR each, plus standard letter postage of about 1 EUR for Europe and 3 EUR to the US. Other postage options can be discussed of course. Edited September 14, 2014 by latigid on Quote Link to comment Share on other sites More sharing options...
latigid on Posted September 13, 2014 Author Report Share Posted September 13, 2014 (edited) Parts list: Resistors: 11 220R 19 10k I used metal film 1% but the tolerance is not critical. 1 5k single turn trimmer (horizontal adjust) Caps: 1 1uF tantalum (C2) *** please observe the correct polarity! *** 3 100n monolithic (C1, 3 and 4, non polarised although C1 is marked as so) Semis: 1 LM317 variable voltage regulator (TO-92 package) 2 SN7407 open collector hex non-inverting buffer 2 14 pin DIP sockets if you wish Connectors: 100 mil SIL headers: 5 2-pin 11 3-pin 8 2-way jumpers, or replace with wire links if you don't want to select the gate voltage 100 mil DIL headers: 1 16-pin 3 10-pin As noted, the clearance is too tight for three 10-pin DIL sockets. Sorry about that! I used a 34-pin with two gaps of two pins. 1 DB-25 connector I used this one from RS. It seems as there are two standards: 0.318" and 0.590". You want the 0.318". EDIT: I just noticed that the + sign for C2 went wandering in the board picture above! Please observe proper polarity, which is correctly marked on the PCBs Note: +/-12 or 15 V can be used. Edited September 14, 2014 by latigid on Quote Link to comment Share on other sites More sharing options...
ilmenator Posted September 13, 2014 Report Share Posted September 13, 2014 It might also work with one 10-way cable underneath the board? That should reverse the pinout, so you need to be extra careful to get it right (unless you do the same on the target board as well). Quote Link to comment Share on other sites More sharing options...
latigid on Posted September 14, 2014 Author Report Share Posted September 14, 2014 That should reverse the pinout, so you need to be extra careful to get it right (unless you do the same on the target board as well). Absolutely! Quote Link to comment Share on other sites More sharing options...
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.