SEQ V2->V3 Migration inc older V2 problems in post too.

[dcer10]

Hi everyone,

After nearly a year ive finally finished putting my midibox sequencer into its box and closed the lid {perhaps pre maturely!}. Everything is working great except for a few small bugs being:

1} When I connect up a bankstick board the seq wont turn on, or wont register the memory, this is the case even when using a single chip soldered directly to J4 with some wires in an IC holder.

2} Some of the encoders are doing unexpected things ie encoder 10 will sometimes generate a value itself, and when I turn it will sometimes change values of other steps, or can sometimes only generate a - signal instead of - or + ie it only goes down not up in note number. Im getting a similar thing with encoder 14 which  seems to generate the highest note number {G8 I think} without touching it randomly and also only generates a - signal not a +.

3} All of the encoders {all bought from smash TV, the ones without the button} are a bit unstable in that it usually will not like small movements and often flickers between the previous value and the current value instead of the next value. This makes exact notes difficult to get, and also makes getting to some menu settings hard.

Heres some photos too of the sequencer, I might put up a bunch in a seperate post as I have taken lots during the construction. Some of those may help people but these are just to give an idea of how its turning out.

If anyone can help with some suggestions Id appreciate it.

Thanks in advance,

John

[stryd_one]

1) Fried?

2) Fried?

3) Right encoder type/speed settings for these encoders?

[TK.]

Really great frontpanel work! :)

Is this ready for the MIDIbox of the Week section? :)

To 3): you have to select the encoder type "MIOS_ENC_MODE_DETENTED2" in setup_mbseq_v2.asm

In MBSEQ V3 I will make this as the default encoder type, since this is the right one for Voti's and SmashTV's encoders

To 2): it makes sense to debug the DINX4 modules, I wrote down some informations about this here: http://www.midibox.org/dokuwiki/doku.php?id=din_module

Best Regards, Thorsten.

[MTE]

WOW great Work !  :o

Regards

MTE

[Artesia]

thats frickin' sweet :)

[stryd_one]

Could you tell us about how the frontpanel was done? It sure is pretty :)

[dcer10]

Hi everyone,

Thanks for all your positive comments. I cant take credit for the front panel as it is TK's design, I have just tweaked it a bit. The main differences between this design and the one you download from the midibox sequencer page for manufacture with front panel designer are:

1} I requested a green anodised metal from them

2} I added my artist name Ivy Mike etc to it for a personal touch {would have liked to have the midibox logo etched in too, but was too hard to get a front panel designer object of the logo, can someone make it for download??}

3} I added the LTC module and mounted the in/out lights for it on the front {the out light is not working, the LED is ok, the wires are ok, the LTC has been fine in other projects and shows in light but not out}

4} I used expensive ultra bright LED's which I bought from aztronics here in Adelaide and chose some complimentary colours for the different modes

5} I pumped up the contrast on the LCD's using the trim pots on the core

6} I chose some nice trimpot covers

Apart from those minor changes the panel is the same as TK's. I will post a seperate topic with the making of this sequencer with detailed photos of every step of the constucution esp the panel so people can see the inside and copy it if they like. I had to try to figure out what to do with the panel as there were no clear photos showing the panel setup, and the panel itself did not fit my rack case, or any of the pots I ordered so maybe this could save people some time on theirs??

In regard to the technical bugs...

stryd_one, when you say fried, do you mean for 1} the core or memory board? Ive had this same problem with each core I have built, and have only fixed it once by replacing the whole core and memory board and memory chips all at once. Is there a program suitable for testing memory chips one by one as I have now got around 30 of them, some wont work im sure, but i would like to be able use the working ones. Kind of hard at the moment without a working memory module in the seq, and my sid is sealed up tight and too hard to open. Also for 2, when you say fried, do you mean the din board or the chips? I will try to go thru TK's suggestion re the DIN board. I hope its something simple as its too much work to rewire any of those boards!!

I will continue to investigate the faults and let you all know whats going on in hope of getting this box finished :}

On a very good note {excuse the pun!} I had a great jam session with the sequencer last nite, and it exceeded my expectations, even with the current bugs and no memory!!! All hail MIOS!

All the best,

John

[dcer10]

Hi All,

For anyone interested I have made a new post showing many photos during the construction of the sequencer here:

http://www.midibox.org/forum/index.php?topic=7770.0

I still however have the problems with the sequencer in this thread, and I also am having trouble with my SID as mentioned here:

http://www.midibox.org/forum/index.php?topic=7311.0

And I would really appreciate all of your continued support to get these projects finished.

All the best,

John

[stryd_one]

stryd_one, when you say fried, do you mean for 1} the core or memory board?

Well it was just a quick guess that the memory might be no good... But given your ongoing problems, I'd say there's something bigger going on there.

for 2, when you say fried, do you mean the din board or the chips? I will try to go thru TK's suggestion re the DIN board. I hope its something simple as its too much work to rewire any of those boards!!

Definitely take TK's advice first :) Well actually I meant the encoders... If you constantly see problems with just the two encoders, I would try swapping them around to see what happens. You need to isolate the fault. If it was me, I'd do something like this:

Move encoder 10 to encoder 1 (or some other encoder that works OK) and vice versa.

Run and check.

If encoder 1 is playing up, it's the encoder itself.

If encoder 10 is playing up, it's not the encoder.

Now swap around the IC's in the DIN board.

Run and check

if encoder 10 is still playing up, then it's not IC because it's been moved, and we know it's not the encoder from before, so it must be wiring (or soldering or pcb fault or some other kind of circuitry issue).

If encoder 10 is OK now but another encoder is playing up, then it's the IC

That's a rough outline but you get the idea?

[dcer10]

Hi everyone,

TK, I have uploaded the midiIO128 application to the box. All of the encoders do show a value on the screen as expected. The encoders in question being number 10 and 14 show the following:

encoder 10 shows note B_6 with a value of zero, when turned in any direction shows 127 and quickly returns to zero

encoder 14 shows CC# 19 with a value of zero, when turned in any direction shows a almost random change between CC# 18 and CC#19 and  shows changing values of 0 or 127. It makes no difference if turned left or right, sometimes it shows 18, sometimes 19. The speed of the turning has a difference on which number will come up.

I removed encoder 14s connection and wired up a fresh encoder of the same kind and got the same results.

I also changed the encoder type to "MIOS_ENC_MODE_DETENTED2" for all of the encoders. This has changed the way the encoders work in relation to my original problem:

"2} Some of the encoders are doing unexpected things ie encoder 10 will sometimes generate a value itself, and when I turn it will sometimes change values of other steps, or can sometimes only generate a - signal instead of - or + ie it only goes down not up in note number. Im getting a similar thing with encoder 14 which  seems to generate the highest note number {G8 I think} without touching it randomly and also only generates a - signal not a +."

In that they seem not to generate the random values or change the surrounding encoders as much, encoder 10 seems to do note G8 from power up and not work after that, on some bootups it will have no value then not work. Also if encoder 10 goes to G8 on boot up then changing the note number on that step with the main datawheel is sometimes temporary and it will quickly jump back to G8 again, other times the new value will stay. Also encoder 10 can land up giving a value to encoder 14 when doing that, generally value A#2, other times moving other encoders may give a value to encoder 10.

Encoder 14 generally does work but seems to be in reverse, probably my bad on that one, will reverse it and also check out replacing encoder 10 to see whats happeining there. Strange tho as midiIO128 showed activity on encoder 10, but when you load the sequencer back up it doesnt do anything except its initial value going to G8 or nothing and then it does nothing.

The change to "MIOS_ENC_MODE_DETENTED2" makes all of the encoders a bit more stable but there is still a lack of real control as they sometimes do a negative value when turned slowly. Is this something that can be fixed? I have seen many other posts showing this but no real answer on how to get around it. I can live with it if it cant be changed but I may think about fitting + and - buttons onto a remote for finer control for some elements if there is no real solution.

stryd_one thanks for your comments also, I will be following your suggestions tonight and tomorrow nite and looking at pulling out the core to see whats going on with the memory problem. I have several other cores sitting about but no working pics for them {my last pic is cactus from my JDM experiments} and I have no "fresh" memory chips left to try, and my other cores have all suffered from the same exact problem so troubleshooting is going to be a nightmare.

I liked your suggestion on swapping the DIN boards chips etc and will let you know how I go with that too. Any other thoughts will be appreciated by you guys or anyone else :}

Thanks for the ongoing support! Any more would be appreciated very much!

All the best,

John

[dcer10]

A quick update, I tried swapping over the encoder on step 10, same results. I reversed the wires on encoder 14 and it seems better but still a bit odd in that it jumps more than one step at a time some times, which none of the others do, its still progress tho.

Step 10 is no longer functioning as is in respect to the encoder, so I swapped out the IC for that SR, it was very random to begin with, then nothing happened at all so I put the original IC back and still nothing. Starting to think theres something odd on that board. I also swapped out the IC on step 14 and had the same thing, random high vaules, I put the original IC back in and now nothing on that step either. Every other step except 10 and 14 works fine. If the board/IC was dead, wouldnt that effect the other buttons and encoders on that SR? What could possibly only stuff up one encoder on a SR and not the others? A short on the board? Something else?

Its odd as I seem to get different results each time I turn the machine on!

Also I have had no luck with the memory yet either so Im missing out on all that extra midibox goodness of patterns, songs and morphing! DOH!

[stryd_one]

I would definitely be looking at the soldering... Had a look at the pics in your other thread, and I was thinking that maybe one of those tracks on the veroboard might not be broken?

If the board/IC was dead, wouldnt that effect the other buttons and encoders on that SR?

That kinda depends on how it was faulty

What could possibly only stuff up one encoder on a SR and not the others? A short on the board? Something else?

All kinds of things. Possibly, Possibly. Not very helpful am I?

Its odd as I seem to get different results each time I turn the machine on!

Yeh it's a weird one.

it was very random to begin with, then nothing happened at all so I put the original IC back and still nothing. Starting to think theres something odd on that board. I also swapped out the IC on step 14 and had the same thing, random high vaules, I put the original IC back in and now nothing on that step either.

That behaviour, putting the same IC back in and now it doesn't work at all... That's interesting. I'm guessing dry solder on the IC socket, and changing them over busted it completely loose?

All of this points to wiring problems hey...

[dcer10]

Thanks for the reply. This isnt going to be fun, pulling that panel off, or the DIN boards is going to be time consuming.

Ive ordered replacement DIN boards, a replacement core, replacement memory, and will pull the front panel off an look at the vero board, which I did cut and test, but could be a problem, Ill check out the bottom of the DIN boards that are there for now. Im ordering all that stuff because it usually seems to be quicker and easier than some of the troubleshooting, which I will attempt anyway while I wait for the parts.

Ill let you know if I have any good news :}

Thanks again for the help so far.

John

[TK.]

Here the same question: did you use a DIN_R3 board from SmashTV? Some pins are not where expected, see also the errata http://www.avishowtech.com/mbhp/mbhp_dinR3.html

Best Regards, Thorsten.

[dcer10]

Hiya TK,

I know I mentioned this to you before but for anyone else reading so they know this isnt the problem I bought my DIN boards from mike.

Thanks :}

John

[dcer10]

A friend and I had a big debugging session on the sequencer last nite, and got nowhere! However there were maybe a few hints to the problems with the step 10 and 14 encoders.

When the IC on SR 8 is removed there was no more random values being sent to either {usually they get random values from power up which land up going to G8 note eventaully}. I swapped the IC a bunch of times and we checked out the bottom of the board but it seemed fine, but still no solution. Step 10 and 14 still didnt work and encoders around them do, sometimes when turning the encoders around them it will move 10 or 14s values too. We tried swapping all of the ICs from SR 8 onwards. Basically we landed up where we started!

I had the idea that maybe since all the other buttons and encoders work except these ones that if I added an extra SR to the config in mios tables then I could assign the encoders to it and mount a new chip. I edited and compiled the code, connected up the wires, uploaded the hex but im still getting the same problems. Im guessing that either theres a much bigger problem, or that my code didnt work.

Ive put the code online at:

http://www.dcerecords.dnsalias.com/ivy_mod.zip

and a PDF showing what I have changed at:

http://www.dcerecords.dnsalias.com/mbseq_din_v2_mod1.pdf

Can someone have a quick look over the code and pdf to verify I didnt mess it up?

Thanks again!!!

John

[stryd_one]

Sooo weird. Different board, different IC, different wiring... The only common points left there are the core and the encoders themselves. Did you use different encoders too?

[dcer10]

We did try a spare encoder which is known to work. Remember that the core also has the problem of not seeing the memory too so I wont rule the core out as the problem.

These two tiny problems, the encoders and the memory not working are doing my head in!! :}

[stryd_one]

Uhm I know it would be a pain in the backside, but have you considered switching cores with your SID?

There is a lesson in this - always test everything *before* you case it up and then again afterwards :)

[dcer10]

Hiya,

I did test it all on the bench before putting it in and it was all ok, but that wasnt with the encoders just the boards. I did consider swapping the core with my sid, but its the only thing I have left that works and its all sealed up in a difficult to get into case, so Im reluctant, but I know my memory board issue will be gone in one second if I did it. Its just way too much work. Rather than sacrifice the SID i will be swapping over the core with a V4 one from smash as soon as it gets here.

One thing occured to me today...

How many people have V2 cores, from mikes shop or otherwise, which work with the 8 bank memory board? Its odd as I have 4 v2 boards and 1 v3 board. All v2 boards worked with the memory to begin with, then the memory chips blow {I think as I had to swap them out to use in my SID} and no matter what you do the memory never works again, my V3 board from mike works with the memory without any problems. Strange odds?? A board fault?

Id be interested to hear how many people may have the same problem? Or people with V2 cores who might be able to test the 8 memory bank board on theirs?

As far as the rest of the sequencer, theres a lot of pain in front of me but im going to get there somehow.. Cant wait for my order from smash to get here!

Also I just quickly uploaded midio128_v2_1c again to see if the encoders still registered, this is the results:

Encoder 10 - when turned left cc33 val 127 occasionally cc32 val 127 randomly when turned right exactly the same.

Encoder 14 - when turned left cc34 val 127 occasionally cc35 val 127 randomly when turned right exactly the same.

Bear in mind, these are the encoders which in the seq app do not do anything when moved, but randomly generate their own values sometimes usually G8 note.

Thanks again,

John

[dcer10]

Just to let everyone know, thanks for the help so far. Ive now got a 100% working front panel!

What I did with the encoders is I moved them to another SR after the normally used ones and with TK's help changed the setup file for the application. Once uploaded encoder 14 worked instantly. Encoder 10 stuffed me about for a while as I could only get it to work in one direction, I tested another encoder in its place, same thing, I tested another wire, same thing, finally just before tearing my hair out I tried another cable cut from a different spool of ribbon and it worked instantly too! There was connection on both ends of the wire for all three pins, but perhaps some sort of intermittant fault in the wire??? Odd but a new one did fix it up!

So now I can use the sequencer but only with one pattern as I still have the problem with the memory not working as described before. Also I am trying to sort out the sync output

http://www.midibox.org/forum/index.php?topic=7245.msg53957#msg53957

And also I am going to connect up the Aout board soon once my power supply is ready, and then Im going to add 2 or 4 IIC boards, the new pic, and the V3 software.

If anyone has any thoughts on the memory problem {core wont see the memory} please let me know.

Thanks again,

John

[TK.]

There is no special application available to test the memory (you mean: the BankStick), because it's normaly not required. If MBSEQ detects it, the connections are ok, if it doesn't detect it, you can be sure that there is a hardware problem.

Maybe you missed to solder the connection between J4:SC and Pin #28 of the PIC?

Best Regards, Thorsten.

[dcer10]

Hiya TK,

Thanks for the reply. I did solder the jumper wire on as shown in the diagram. As I have said before, all of my V2 boards from mike have had this problem, and it lands up killing the memory chips which are plugged into it. Ive got about 30 seemingly dead 24LC256 chips :}

Usually the V2 boards read the memory for a while, then dont do it any more, or ever again.

I register the earth connection, the input voltage, no shorts on the board I can see {no connection between the input pins} and for all other purposes each V2 core works fine. My V3 boards dont do this.

If I cant solve this its not too bad as I have some V4 boards on the way from smash to replace, and I can use the V2 for making a midi monitor box or similar. Im having enough fun for now with one pattern only :}

Another question about the seq, when I upgrade from V2 to V3 and add some IIC boards, does this mean the midi will not work from the core anymore and the only midi should be from the IIC boards? If so what effect would this have on using the LTC board on the seq if any? Currently I have 1 in 2 out and 1 thru from the core and LTC module, and 4 holes drilled into the case for midi, I dont mind using only the IIC as 2 boards would supply 2 in 2 out I guess and I wouldnt need more holes :} Is this all correct? I know there are other posts on the V3 seq, but it was not clear to me about this point. Is there going to be a way to change between the 16 tracks of the V3 using the track 1-4 buttons from a V2 seq? I noticed you said that you will use your existing panel, which is handy for me as mine is the same!

Also when using the Aout module, Im not certain about the gates. It says in the Aout page that in the seq RC0 (CORE::J6::RC) and RC1 (CORE::J6::SC) of the PIC are the extra gates, does this mean there are 8 cv outs and 4 gate outs? In this case you could control 4 synths, and 4 extra cv's for filters etc on a modular??

Thank you so much for you help once again!!

John

PS, this sequencer is totally awesome!!

[TK.]

I never heard before, that the EEPROMs will be killed in such a way (especially not, that this happens with Mike's board - most of my own boards are from him). The only answer I have is, that you either forgot to ground the CS pins of the EEPROM (pin 1-3), so that the written data is corrupt, or that you plugged the EEPROM the wrong way into the J4 jumper - note that V3 boards have a slightly different layout, see SmashTV's page for the differences

Another question about the seq, when I upgrade from V2 to V3 and add some IIC boards, does this mean the midi will not work from the core anymore and the only midi should be from the IIC boards?

please read this wiki page which gives you a basic understanding about the IIC modules in relation to the PIC18F4620: http://www.midibox.org/dokuwiki/doku.php?id=mios_pic18f4620

For MBSEQ, we are using the MIDI Outs only, the MIDI INs of the MBHP_IIC_MIDI modules are not used (and don't need to be stuffed)

The MIDI In of the Core module is used as the only MIDI input

If so what effect would this have on using the LTC board on the seq if any? Currently I have 1 in 2 out and 1 thru from the core and LTC module, and 4 holes drilled into the case for midi, I dont mind using only the IIC as 2 boards would supply 2 in 2 out I guess and I wouldnt need more holes :} Is this all correct?

No. In general you can still use the LTC module if you just want to duplicate MIDI OUTs, if you want to have a MIDI Thru, or if you want to have some flickering LEDs. Just have a look to the MBHP_IIC_MIDI module layout, it provides a MIDI Link port as well.

You are allowed to connect for example the LTC::J1:MO pin to the MBHP_IIC_MIDI::J4:MO pin in order to duplicate this MIDI Out, it would have the same effect like in your current setup.

But the usage of a second MBHP_IIC_MIDI module has the advantage, that you will get a seperate MIDI Output which works independent from the first.

In other words this means: you will get 32 MIDI channels

With 4 MBHP_IIC_MIDI module you will get 64 MIDI channels.

I know there are other posts on the V3 seq, but it was not clear to me about this point. Is there going to be a way to change between the 16 tracks of the V3 using the track 1-4 buttons from a V2 seq? I noticed you said that you will use your existing panel, which is handy for me as mine is the same!

It's clearly stated in the other post with repeating examples to be very exact, what is unclear?

Also when using the Aout module, Im not certain about the gates. It says in the Aout page that in the seq RC0 (CORE::J6::RC) and RC1 (CORE::J6::SC) of the PIC are the extra gates, does this mean there are 8 cv outs and 4 gate outs? In this case you could control 4 synths, and 4 extra cv's for filters etc on a modular??

I will release a new build of MBSEQ V3 which will include the new AOUT handling, I've described the mapping of the AOUT channels and Gate triggers in the CHANGELOG.txt file

PS, this sequencer is totally awesome!!

Like to read this! You will be surprised about the new features of MBSEQ V3 - I for myself would not be happy with a V2 anymore ;)

Best Regards, Thorsten.

[dcer10]

Hiya TK,

Thanks for your reply.

With the memory I am using the 8 chip bankstick board from mike. I have 4 of them and have the problem when using any of them with any V2 board, but my V3 board works fine with them. 

either forgot to ground the CS pins of the EEPROM (pin 1-3), so that the written data is corrupt,

Could this be the case when using the bought PCB?? I will tripple check the connections. Also no data can be written as when the seq loads up it doesnt see the bankstick to format, or read the setup. I have tested for connection for earth and voltage on the pins that are meant to be connected to each other, and also tested for connection between the pins for shorts and it all looks ok.

or that you plugged the EEPROM the wrong way into the J4 jumper - note that V3 boards have a slightly different layout,

I have been very careful to plug it up exactly as shown in the images as I have interpreted them. On the V2 boards it seems to be a 1:1 connection when you orientate the boards so the jumpers are closest to each other? Attached is an image showing how it is connected up.

But the usage of a second MBHP_IIC_MIDI module has the advantage, that you will get a seperate MIDI Output which works independent from the first.

In other words this means: you will get 32 MIDI channels

With 4 MBHP_IIC_MIDI module you will get 64 MIDI channels.

Thats fantastic! I knew that was a feature, but I just had to say that again!! Maybe soon I can not have to use my studio 4 midi router any more? Ideally I would like to run all of my gear from the sequencer and have no pc! My seq and my G3 running logic using the studio 4 router dont agree, especially when it comes to timing. The seq seems to be perfect, but the mac reads it +- 2-5bmp fluctuating so a lot of stuff cant be played back in time after the midi is recorded. Dont know why that happens.

It's clearly stated in the other post with repeating examples to be very exact, what is unclear?

What is unclear to me is how you would actually change between all of the 16 tracks from the tracks 1-4 buttons. I may have been looking at the wrong posts :} There is a lot of things to read. I will look again. Im just trying to figure out how much extra soldering I will need to do to the inside of the seq if any. I was reading the text file with the V3 software and I was not sure on the whole track concept, there was a mention of the various Dout settings to show track usage with the LED's and that looked like mine may need to be reconfigured, is that correct {mine is based on yours when yours was a V2 exactly the same}. Id much prefer not to have to solder to any existing hardware if its possible to avoid.

I will release a new build of MBSEQ V3 which will include the new AOUT handling, I've described the mapping of the AOUT channels and Gate triggers in the CHANGELOG.txt file

My Aout isnt even hooked up yet as I am waiting on a power board from a friend which will provide power for the Aout and the core board at once, maybe its best for me to wait for V3 to be finished before I hook it up??

You will be surprised about the new features of MBSEQ V3 - I for myself would not be happy with a V2 anymore

Im still nearly jumping thru the roof in excitement each time I boot up V2, and its not even working 100% yet, so V3 will really knock my socks off Im sure!!! The changelog has some extremely interesting features mentioned in it!! I used to use a program on windows called seq 303, not sure if you have seen it. For at least 6yrs I looked for a hardware alternative which was stable, user friendly, and able to do several tracks at once with cc's note lengths etc. This is the answer to those prayers, and much much more, and it isnt in a big loud PC case :}

Also I know we have spoken about this before, but the data wheel is still very unstable. It moves in large steps sometimes, and occasionally not in the right direction for a moment. It is usable, but frustrating. I have done all the changes you previously mentioned in the config file, do you have any other suggestions?

Thanks again!!!

John