Jump to content

MIDIO128 V3


TK.
 Share

Recommended Posts

  • Replies 87
  • Created
  • Last Reply

Top Posters In This Topic

Pete,

Currently,  the LPC is USB powered from the Artisan sound engine computer which is running jorgan.  The SE computer boots up in linux ubuntu which takes a 10-15 seconds after powwer is applied,  After Ubuntu is loaded jorgan auto boots up in to the disposition.   since the Soundengine, 12vdc supply,, 5vdc supply for the core 8 all come on at the same time as the computer, I would have to say that the 12vdc is applied long before the shift register chain is innitialized.  this is definitely something that is new to me.  since I have a core 8 with DOUTS driving SAMS also  i suspect the same problem exists. 

Question;  So what happens if the 12vdc comes on before the DOUT shift register is initialized. I suspect that i could switch the 12vdc and 5vdc power separately, or time delay with a relay which would be preferred, on switch turns everything on!

 

The LPC has only 6 DOUTs connected, which can be in excess of one ampere.    The 5 conductor ribbon diasy chain from the LPC on down through the MIO pcbs and DOUT cards is in place.  I understand about the 18awg ground from each card, and i assume you mean from the ground plane of the PCB to the common ground point, leaving the ribbon cable interconnect in place?

With that done, is it necessary to put another ground on the LPC?

 

To be sure that i have this right, i will post a block diagram of how it exists now, i how i understand your suggestions.

 

Question:  There are 2 dout4x pcbs connected to the LPC as described above, but only two ULN2803 are installed on the last DOUT4x(last 2 sockets are empty), however there are 4, 74HC595 installed.  i don't need the last two ULNs.  is not having pull down/up resistors on the last two 595's a problem?

Thanks for the help,

johnc

Link to comment
Share on other sites

Question; So what happens if the 12vdc comes on before the DOUT shift register is initialized.

One indication is the SAM's might change state when power comes on. With the 74HC595's reset line tied high, there is no guarantee what the outputs will look like. On my LPC17 I have LED's on the Dout board and see the lights come on for about 1 second if I pull the USB cable then plug it back in after waiting about 5 seconds. If the LPC hasn't been plugged in for some time, I see the LED's are generally off, but not every time.

 

With that done, is it necessary to put another ground on the LPC?

This is where I said BUT in my previous message. I would add this extra wire to the LPC17 and if it works, leave it in. I do understand there is potential for ground loops with this due to the ground in the ribbon cable going to the Dout chain. At least this will keep any noise introduced by connecting the PC ground to your LPC17 from going thru the LPC17 to the Dout chain. With it grounded the noise from the PC will be directed to your common ground (star) connection. In a perfect world I would say leave the whole ground signals floating with reference to each other, but with a PC on one side of the LPC17 and a high current 12v power supply on the other, it ain’t going to be perfect no mater what you do.

 

As far as the resistors on the output of the 595's, just don't connect anything to them. These are outputs and not sensitive to being unloaded.

 

PS-I think it would have been better to have this thread placed in a separate topic rather than the generic Midio128-V3 area.

 

Pete

Edited by kpete
Link to comment
Share on other sites

Pete,

Thanks for the explanations.

Finding a place on the LPC to connect a ground pad is difficult - any suggestions?  Bare in mind that all the pcb's are  firmly attached to a back plane.

An underlying problem that may be connected is that when the  J1/j2 jumper is plugged in to the last DOUT4x,  multiple SAMS begin turning on and off randomly.

Johnc

Link to comment
Share on other sites

Most of the front and back of the LPC17 PCB is ground which should be a good ground almost anywhere. But I would put it on the minus side of big capacitor C8. With it laying down, the extra ground wire can be soldered to the cap wire going to the PCB.

 

As far as random Dout's turning on and off, this is why you need all these extra ground wires going to all of the Dout boards. They must not be daisy chained like they are normally hooked up. Something else that worries me is that you have a DIO-Matrix board between the LPC17 and your Dout chain. It so happens that just before reading the Din chain, the shift registers are loaded with the RC strobe signal. This same signal also strobes the Dout shift registers into the output holding registers. So when the outputs are changed to the SAM's, the input shift registers are being loaded at the same time. Could there be noise getting into the Din chain of the matrix board?

 

I would also try to put a 10uf (or larger) across the 5v of the problem Dout chain and see if it improves things.

 

One thing that I would check is that pin 10 of all the uln2803's are connected to your +12 power supply. I am sure you did this since you are not blowing up any of the 2803 chips. Just so you know, there are current spikes of 2 amps on pin 10 of the uln2803 when 4 of the SAM magnets on the same chip are turned off at the same time.

 

Pete

Link to comment
Share on other sites

Pete,

You are just full of good news!  but tell me more, Please

 

 

you said:

"Something else that worries me is that you have a DIO-Matrix board between the LPC17 and your Dout chain. It so happens that just before reading the Din chain, the shift registers are loaded with the RC strobe signal. This same signal also strobes the Dout shift registers into the output holding registers. So when the outputs are changed to the SAM's, the input shift registers are being loaded at the same time. Could there be noise getting into the Din chain of the matrix board?"

Ok -  So what is the solutiion?   The mio pcb uses Din and Dout, and the SAMS Dout boards start with D5, D1 thru D4 being used by the mio. Would reversing the do order do any good.  Can I connect the DOUTs for the SAMS to a different connector on the LPC?

 

Wow, 2 amperes on pin 10 of the 2803.  Didn't know that either, nor have i ever read any warning.  i suspect  the  connections from the pin 10 back to the +12v should be #18awg also.

 

Would have helped to know all thie good stuff before screwing all the pcbs to the backplane.

 

Thanks for the input, and please pass anything else along.

johnc

Link to comment
Share on other sites

Pete, TK,

Please feel free to move this thread to Design concepts or someother topic.

I will install the revised grounding including the LPC and the core 8,  test it then report back.

if you have more thoughts in the meantime, please post them.  i am in the late stages of construction on the console, and i hate to remove working systems, with the allways preasent possibility of inserting new issues.

In general though,  IMHO, I have implemented the DIN, DOUT. and matrix IO  cards within the posted guidelines.  How should I have done it to avoid this problem, including the grounding issue?

Thanks,

johnc

Link to comment
Share on other sites

Hi johnc,

 

As far as the pin 10 current, I just wanted to make yourself aware of this. Its not continuous, just a spike due to the SAM coils being turned off. Using the ribbon cable with #28 wire doesn't seem large enough but because of the short duration, it should be OK. Problem is that the currents can couple into other wires that parallel the wires going to the SAM's.

 

As far as the MIO-Matrix issue, I don't know if switching them around would help. And do we know that we have a problem here. It could be an issue like bundling the SAM switches and magnet wires together in one bundle. I don't know what software is driving the SAM chain but I would try to disable anything that reads the SAM switches and outputs Midi messages. This way you can use the J-organ panels on the PC to send Midi messages to your SAM processor to physically change the SAM state. With the SAM switches disabled, there wouldn't be an issue of the SAM processor echoing back unrelated switch states that were triggered by any noise. I suspect this is your DIO-Matrix but don't know for sure.

 

Pete

Link to comment
Share on other sites

Pete,

As for the pin10 spike, it could be a problem as  i assumed wrongly that there was minimum current flow, and the connectiion from the pin to 12vdc is no more then 26-28ga.

 

As for the SAMS,  i am using the sAMS extension in jorgan which utilizes the reed switch on the SAM in the time delay process where by the pulse to the magnet lasts for only 500 ms (adjustable).  I have not noticed any problem therein.

Srr the attached block diagram

Organ block Diagram - Ground.bmp

Link to comment
Share on other sites

Hi John,

You have gone thru a lot of work getting to where you are now. The block diagram speaks for itself. What you might do is totally isolate the SAM contacts from going to J-organ/Artisan sound engine by pulling the Midi1 cable from the LPC17 board. Then start punching stops on the J-organ panel and see if all of the stops work with your setup. Even try pushing pistons on the screen to see if the proper stops get set. What I want to eliminate is issues with the input switches sending spurious or wrong values.

 

And I notice there is a cable on the LPC Midi2 out connector. What is this for? Maybe just disconnect it since all I want to check is the output functions from J-organ to the SAM's. Don't want to cloud any issues with inputs going into J-organ at this time.

 

Something else that you should also be aware of is that there are 2 different types of SAM's. +common and – common types. The coils between the 2 types are wired in reverse as far as polarity. You have the SAM wired in a “+common†mode. If you use the wrong SAM type for your system, the SAM's will not snap from 1 state to the other efficiently. They may work but not with as much power due to the magnet orientations. Another issue is that the SAM's are more prone to having cross-talk issues between a neighboring SAM. This means that when you active the coil on SAM-1 that the switch contact on SAM-2 might open or close for just an instant and give bad signals to the input board.

 

Pete

Link to comment
Share on other sites

pete,

As of today, all SAMS work with the jOrgan either by picking with the mouse on the desktop, or by using a piston, with exception of the last addition to the DOUT train that is connected to the LPC. specifically it drives all but 3 of the SAMS on the pedal.  This particular DOUT4X is the one where i am only useing 16 douts, Two uln2803s are not installed.    Without the last DOUT4x plugged in to the 5/c LPC chain,  everything works.  When I plug in the last DOUT4X, initially the SAMS tied to it are not responding to the screen picks or pistons.  now the strange part,  if I put my finger in close proximity to the back of the first DOUT stage,74hc595 and Uln2803, the sams connected begin to randomly jump back and forth from on to off.  Removing the finger, it stops.

So I:

1.  Removed the 5/c jumper from the upstream DOUT4x and verified continuity through the jumber - no problem.

2.  I inspected the solder connections on the back side, in particular the trace connecting to the middle SO pin starting on J1 through the first DOUT then the second, third and forth.

3.  I replaced the 595 and the uln chips.

Thie above did nothing.

 

The other midi out routes the note on /note off Stop messages from jorgan to the Soundengine, as well as key press info from the keyboards. Actually, it is not installed, although shown on the diagram. I haven't got it yhe SE producing sound  yet, only  running  jOrgan.

The router on the LPC is configured to route only select channels to the SE.

 

I think that everything is fine, with exception of the last DOUT4X.  If I could figure out what is wrong with it, all would be well.

I appreciate being able to bounce my problems off of you!

I do have a lot of time and effort in this project and thias little glitch is holding up the show!

 

PS: The last DOUT4X has a dedicated lead from the ground plane back to the common -12vdc  ground point.

johnc

 

So as it stands, with the last DOUT4x on the LPC chain disconnected,  everything works.

I am puzzled by the proximity issue  ( not touching any pins)?

Any comments.

 

As for the SAMS. they are Syndyne with positive coils, and one reed switch.  They worked fine on initial bench checkout and appear to work quite satisfactorly.  They were in service previously driving pipe magnets on my now defunct 5 rank pipe organ.  The way the SAMS extension works in jOrgan is that  upon receipt of a note on message, it energizes the on coil then intercepts the reed switch closure (positive mech. feedback) which means that the armature moved to the on position and is magnetically (permanent mag on the SAM)  held, once the reed switch closes, jorgan terminates the pulse to the on coil. There is also a 500ms time out which terminates the pulse in case the armature sticks, etc.  on receipt of a off message, it performs the same way bu works with the off magnet.  Complicated, but it works.

Edited by John_W._Couvillon
Link to comment
Share on other sites

if I put my finger in close proximity to the back of the first DOUT stage,74hc595 and Uln2803, the sams connected begin to randomly jump back and forth from on to off. Removing the finger, it stops.

Boy this sure sounds like you have a floating input pin somewhere.

 

Pete

Link to comment
Share on other sites

Pete,

The ic's involved with the proximity issue are on the DOUT4X that is down stream from the two matrix boards.  you say floating input pins, but the only input pins feeding into the LPC are in the 4 matricies, (three keyboards and  SAMS reed contacts) inputs. could it be an open circuit in one of the Matricies.  I haven't noticed any notes not playing, or pistons not working.

Would a lose connection or bad solder joint on an  input pin on the 595 shift register cause what we are seeing?

i'll touch my soldering iron to all the input pins on the 595's tomorrow.

All the other inputs of which there are almost 64 (pedal keyboard and pistons)are on the DIN train that is connected to the core 8.

Thanks,

Johnc

Link to comment
Share on other sites

you say floating input pins, but the only input pins feeding into the LPC are in the 4 matricies . . .

When I said floating I mean it sounds like there is a floating pin somewhere on the Dout board or the chain.  Maybe its a curled up pin or a socket pin thats not soldered.  I noticed on the matrix J1 & J2 board connectors the RC signal is not passed thru on all of the pins.  The pin 9's are missing on one side but this is on the Din side of the chain of the matrix board so it shouldn't be a problem.  It uses the signal from the Dout RC signal for this.

 

I haven't noticed any notes not playing, or pistons not working.  This is the first time that you ever talked about the keyboards working properly.  Thats great!     I was worried that these might be affected by the firing of the SAM magnets.

Link to comment
Share on other sites

Pete,

you are correct on the J1/j2 issue of not continuing some of the  DIN wiring  past j2.  Since i did not intend any DINS beyond that point, and since all bu the SI and SO pins on j2 are duplicated for Dout and DIN I used only a 5 pin connector.

Also involved in this project are quite a few old DIN4X and DOUT4X boards which  have been around since i built my first Virtual instrument  with revision 0 of SMASHTV pcbs.  The DOUT4X I am having the problem with is a modified DOUT which pre-dates SMASH's revision of the board to accept the ULN2803, just guessing i would say 2003 or so.  So this board has seen a better day.  This morning I decided to give up on it and order a new Dout4X and start fresh,  I have already spent many multiples of $15.00 in time.  In the mean time I'll work on the soundengine part and get some sound going.

Thanks for the input, comments, etc.  its nice to have someone to discuss this stuff with, especially the hardware side.

johnc

 

p.s.  I touch soldered all of the 595 ics socket pins but nothing changed, except now touching anyone of the 4 makes the SAMS on the  DOUT pcb flip up and down.  I agree, somehow the clock pulses are not getting through. to the PCB.  With a new DOUT4X that works, either the problem will go away, or it will move upstream.  I do know that the DOUT4X that connnects to the last matrix bd works, so that narrows the search range.

Link to comment
Share on other sites

  • 2 weeks later...

Pete, TK,

Looking closely at wiring,  I found that the ground in the 5/c ribbon cable to the last DOUT4X pcb had separated so the PCB had no grond connection back to the LPC.  it did have the new ground connection from the ground plane on the DOUT pcb to the common system ground point, but not to the -dc from the USB power source.  That probably caused damage to the ics on the DOUT.

Reading through the forum posts I noticed I was behind on updates, and  felt it a good time to update all software, boot loader, midio128ver.3 and mios studio.

With so many things not working correctly, I unplugged everything and started over introducing items.

First is the LPC with the 2 KB matrix pcbs, no other DOUTS or DINS.

First problem came with mios studio.  Avoiding the MSD access to the SD card, i tried the file browser which lists the default.mio, and the Update button which is highlighted.  As i understand the revised mios studio, the .mio file should be displayed so it can be edited.  Picking the file name brings up text at the bottom of the window  that indicates the % of the file loaded,  The progression of the load stops with an error message of an error in the file.  This is the same .mio file I have been using which is puzzling.

in any event it must be  loaded, because depressing keys on the KBs produces an input on the mios studio monitor window.

If using the MSD on the control surface is not recommended anymore, how else can I load the mio file, except with the file browser?

Also,  What should i look for in the .mio file that would cause an error?

The first test of the KB matrices is not a complete success. Inputs on the monitor from key presses show the correct note on and note off messages, but there are additional messages for notes using channels not on the matrix configs on the default.mio file.  This was happening before i did the updates.

Strangely,  of the 3 kbs which are on channels 1,2,3  The offending messages are on channel 4, and the same 5 or so notes appear on 2 of the three kb inputs.

Troubleshooting work already done:  inspected the KB pcbs with a magnifying glass and touched up with a soldering iron any suspicious joint;  Replaced both ICs on both KB pcbs; checked continuity through the 5/c ribbon cable.  no change in results.

Several of the note numbers show up on the same row on the matrix input which would indicate to me that a key contact is stuck closed, so mynext task is to inspect all of the key contacts.

Any comments about any of the above are welcome.

I tried attaching the default .mio file as well as screen dumps of the monitor window showing the KB input stuff, but files are too big.  may i email them to you?

Johnc

Link to comment
Share on other sites

Which error message is print by the Filebrowser exactly?

And do you see an error message in the MIOS Terminal as well? Which one exactly?

 

.MIO file: please pack it into a .zip file, and then attach it to this thread.

 

Best Regards, Thorsten.

Link to comment
Share on other sites

Of course, you can use the remaining matrix pins for other purposes, but they can only send Note Events for obvious reasons.

 

.MIO file: please try with MIDIO128 V1.015 and the "single_usb" option enabled (has to be enabled from Bootloader Update App V1.012)

 

Changelog:

 

MIDIO128 V3.015
~~~~~~~~~~~~~~~

   o support for the "single_usb" option, which can be enabled with the
     bootloader update application V1.012 ("set usb_single 1")

     Use this option if you are working with a Windows PC which can't handle
     the 4 USB ports correctly (e.g. stucking MIDI, Filebrowser operations are failing, etc.)

 

 

Best Regards, Thorsten.

Link to comment
Share on other sites

TK,

Thanks for the specifics.

i am current with the uploader, and reloaded the latest ver of midio128.

Now when mios atudio boots up, it shows only one usb,  usb1.  before it showed usb 1 - 4.

I tried the file browser, and it performs as before.

The default.mio appears,  i select it with the mouse and nothing happens.  The loading messages appears at the bottom indicating %loaded which stops at about 50%, then the other message indicating an error appears.

I also pressed one key on the three keyboards and one SAM, and the off channel message still appears.  Should I start checking hardware? 

Thanks,

johnc

Link to comment
Share on other sites

It sounds more like a Windows (or computer) issue, for which I don't have diagnosis tools. And since I'm not a windows programmer, I also wouldn't be able to provide an alternative solution (such as a dedicated USB driver).

 

There are certain things you could try out:

- try to download the file multiple times, does it always stuck at 50%?

- install the GM5 USB MIDI driver, does it help?

- maybe a different SD Card?

- try it on a different computer - any luck?

 

If this doesn't help, then the only way how you can get reliable SD Card access is via an external SD Card reader.

Such readers typically cost ca. $5..$10, so it would be an inexpensive workaround.

 

Best Regards, Thorsten.

Link to comment
Share on other sites

TK,

i tried the file browser several times and it went up to 76%.

Fortunately the MSD still works, and it accesses the SD card with out a problem.

I am back to seeing 4 USB, at the top of the midimon, input and output.  What do i reload to get back to just USB 1?

Again when i enter help in the terminal window, i don't see the command to set USB to 1.

Johnc

Link to comment
Share on other sites

It's normal that the application still shows 4 USB ports (single_usb only affects the USB descriptors)

And it's normal that the single_usb command isn't accessible from the MIDIO128 application, it's only available in the bootloader update package.

However, I'm sure that it's still enabled.

 

Missing tests:

 

- install the GM5 USB MIDI driver, does it help?

- try it on a different computer - any luck?

 

Best Regards, Thorsten.

Link to comment
Share on other sites

TK,

Loaded the GM driver and now the midi in and midi out windows on mios studiio show GM1 thru 4.  The file browser still doesn't work.

This is an xp computer and i don't have another to use.  

In my application which is jOrgan,  i need midi 1 (in/out) and 2 (in/out),  I am able to set the router in the terminal with the set command.  Is midi 2 active, or does it have to be set..

Would still like to have mios studio running.

Johnc

Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

Loading...
 Share


×
×
  • Create New...