Yikes.... I bought an EII+

[HideawayStudio]

Well - I couldn't resist.   I've always wanted one and now I'm assembling a team of people to assist in developing my virtual floppy drive it's the ideal opportunity to use the real thing to help develop the concept. 

It's got to be shipped from the US to the UK - which is costing a fortune! 

The ultimate goal is to have the entire official EII sample archive residing on one  MicroSD card in the virtual floppy unit sitting in place of one of the drives.  The virtual floppy name being displayed on a LED backlit 2X20 char LCD to give the name of the bank using INC and DEC buttons.

[JMP]

Congrats Dan. Good luck with the shipping and getting it through Customs unnoticed... 

A nice addition to your existing Emu collection eh?   

[Elmbeatz]

YEAH!!!!!!!!!

[HideawayStudio]

It looks like my EII has had an interesting owner much like my EMAX II.  The previous owner was none other than Allen Adkins, the founder of OMI!  A bit of an irony really considering what I want to use the EII for....

....and just in case you were wondering about my EMAX II kb it was owned by Jason Young having been handed down years ago by DM's Alan Wilder in a slightly poorly state.  What I suspect noone will ever answer for me is whether it was used by DM or in Alan's studio.      I fully restored the beast and put it in pride of place in the studio now with internal HD some time ago now.

I'm a little concerned that my keyboard stands are not going to take the 70+ pounds weight of an EII!

Is there anyone in the UK who can provide/swap Emu EII sample disks??  I'm in the SW of the UK.    These can be on either 5.25 or 3.5" as the first thing I'm going to do is convert one of the drives to 3.5" so I can experiment imaging them for data characterisation.

[JMP]

Finding original OMI CD's and the drive will be extremely difficult unless someone here or on the Yahoo Users group will give them up. Otherwise only realistic way of getting OMI sounds in the EII is via Sound Designer and an old Mac.

I'm in NW UK and have an EII with dual 5.25" drives. They work but can be tempermental so copying floppies and them working in yours could be hit and miss.... but can still give it a go.

[esynthesist]

Nice to hear you don't give up the virtual floppy project !!
I have to say I was a bit surprised to read you didn't have an EII yet 

These can be on either 5.25 or 3.5" as the first thing I'm going to do is convert one of the drives to 3.5" so I can experiment imaging them for data characterisation.

What exactly do you mean by "data characterisation" ? The only thing not demystified today is the physical disk layout (sectorsize, filler bytes, etc.) A few years ago I was hoping that using 3.5 EII disks instead of 5.25 ones would increase the chance to find out these disk parameters. But it didn't... a PC with floppy disk recognition software such as OmniFlop/OmniDisk can not handle the 3.5 disks either - same problem as 5.25 ones, which is normal in fact, because most probably the EII simply uses sectorsizes which are not supported by the PC/Mac diskcontrollers/BIOS.

Anyway, for the sake of the floppy replacement project, I don't think you need to buy/swap EII disks at all. Having one disk with the OS on it is sufficient for that project. The soundbanks on these disks have the exact same layout as the SoundDesigner for EII files and are saved as one big 485888 block of bytes on successive sectors/tracks of the disk (before or after the OS - that's unclear). SD4EII files can be transferred via the internet, so you'll not need physical disks for that.

///E-Synthesist
 

[HideawayStudio]

Nice to hear you don't give up the virtual floppy project !!
I have to say I was a bit surprised to read you didn't have an EII yet 

These can be on either 5.25 or 3.5" as the first thing I'm going to do is convert one of the drives to 3.5" so I can experiment imaging them for data characterisation.

What exactly do you mean by "data characterisation" ? The only thing not demystified today is the physical disk layout (sectorsize, filler bytes, etc.) A few years ago I was hoping that using 3.5 EII disks instead of 5.25 ones would increase the chance to find out these disk parameters. But it didn't... a PC with floppy disk recognition software such as OmniFlop/OmniDisk can not handle the 3.5 disks either - same problem as 5.25 ones, which is normal in fact, because most probably the EII simply uses sectorsizes which are not supported by the PC/Mac diskcontrollers/BIOS.

Anyway, for the sake of the floppy replacement project, I don't think you need to buy/swap EII disks at all. Having one disk with the OS on it is sufficient for that project. The soundbanks on these disks have the exact same layout as the SoundDesigner for EII files and are saved as one big 485888 block of bytes on successive sectors/tracks of the disk (before or after the OS - that's unclear). SD4EII files can be transferred via the internet, so you'll not need physical disks for that.

///E-Synthesist
 

Hi,

             By "data characterisation" what I mean is looking at the nature of the raw digital data signal coming off an EII disk when it is spinning in a drive on a storage scope and/or a logic analyser.   Due to the nature of the device, as said before, I'm not interested in the data format.  That said, it is vital that I can oversample and playback each track exactly as it came off the disk and time aligned to the spindle sync.   In order to determine the maximum rate I should oversample I need to carefully analyse the data to determine minimum pulse width, typical jitter, and raw data rate.  This will also determine the amount of scratchspace buffer required to store one track as an oversampled burst of data and just how tight the timing preservation has to be.  Fortunately some of these parameters are specified in the Shugart standard.   General thoughts by myself and a couple of friends is that we will need to sample at 1-2MHz - this will produce disk images of up to 8Mbytes / disk.  This is inefficient but remember on a modern memory stick this is still several hundred floppy disks!

This project is all about managing the time domain.  The data is serial and binary but very much asynchronous as the clock is effectively embedded in the data which the drive itself is unable to understand - thus what matters is that each bit is sampled and delivered at the right time.   What you have to remember though is this is exactly the same as a normal mechanical Shugart drive - the floppy drive has no concept of format - ie. it blindly plays back data originally set down by the disk controller.   The disk controller is where the data is pulled apart and decoded - thankfully I don't have to worry about that provided my virtual drive doesn't dish up complete gibberish.

In order to do this it is very likely programmable logic connected directly to a small SRAM will be required to perform track capture and playback.  The processor will manage disk images via mass storage in FAT32 to an SD card as well as dish up track data, monitor the head stepping and side control lines from the controller along with driving the front panel LCD and reading the buttons.

[esynthesist]

Wow, you clearly know what you're talking about.
I'm pretty sure you and your friends will succeed in this project !!

Maybe I'm too with the following question, but I'm wondering about this "x MB" files: so if I understand you correctly, the image of a 560K EII floppy disk could take about 8 MB on a flash disk ? So that would mean that these files will be very different from the SD for EII files we all know now, right ? Or am I too stupid to understand what you'll be doing  ?
Suppose I'm right... do you think there will be any "non-destructive 1:1" way to derive the original files from the flash disk files...
...you can guess already that this would be very nice for the user community (and of course also for my EMXP stuff...)

Good luck !

///E-Synthesist

[HideawayStudio]

Wow, you clearly know what you're talking about.
I'm pretty sure you and your friends will succeed in this project !!

Maybe I'm too with the following question, but I'm wondering about this "x MB" files: so if I understand you correctly, the image of a 560K EII floppy disk could take about 8 MB on a flash disk ? So that would mean that these files will be very different from the SD for EII files we all know now, right ? Or am I too stupid to understand what you'll be doing  ?
Suppose I'm right... do you think there will be any "non-destructive 1:1" way to derive the original files from the flash disk files...
...you can guess already that this would be very nice for the user community (and of course also for my EMXP stuff...)

Good luck !

///E-Synthesist

Hi E-Synthesist - believe me I don't think you're lacking in the brains dept. in the slightest! - it's just that our methods of approaching this challenge are very different.  My fully tranparent belt and braces method has one flaw as you've touched on and that is the raw image files are not much use directly for sample conversion in the pc.  That said it should be possibe to post process the image files as on the pc you have the luxary of not having to do this in real time.  A util would have to be written to parse the oversampled data, determine the pulse length and reconstruct the disk contents into an image format a pc could operate on.  The tricky bit will be resampling this image back into an oversampled raw file that the virtual drive can operate with.  Again - because this doesn't have to happen in real time it is probably quite doable.

Dan.

[esynthesist]

believe me I don't think you're lacking in the brains dept. in the slightest

OK, I'll use that quote towards all my friends, colleagues and even family members who have a different opinion about that 

Well, from a certain perspective it's great that some specific post-processing algorithm has to be engineered: we're all driven by this kind of challenges, aren't we ?

I hope you'll have your first (rev) engineering results soon !

///E-Synthesist

[Escobar]

Very interesting thread indeed! 

This project is all about managing the time domain.  The data is serial and binary but very much asynchronous as the clock is effectively embedded in the data which the drive itself is unable to understand - thus what matters is that each bit is sampled and delivered at the right time.   What you have to remember though is this is exactly the same as a normal mechanical Shugart drive - the floppy drive has no concept of format - ie. it blindly plays back data originally set down by the disk controller.   The disk controller is where the data is pulled apart and decoded - thankfully I don't have to worry about that provided my virtual drive doesn't dish up complete gibberish.

So you have to 'mimic' the different time 'slices' it takes for the real Shugart drive to fetch (i.e. read/write) data from the disk? Do I underdstand you correct? If you don't do that will it be a problem for the floppy disc controller to handle the data from/to disk?
Have you decided for what kind of storge media you're going to use for you solution? SD or CF?

I hope you succeed!

//Escobar
 

[HideawayStudio]

Very interesting thread indeed! 

This project is all about managing the time domain.  The data is serial and binary but very much asynchronous as the clock is effectively embedded in the data which the drive itself is unable to understand - thus what matters is that each bit is sampled and delivered at the right time.   What you have to remember though is this is exactly the same as a normal mechanical Shugart drive - the floppy drive has no concept of format - ie. it blindly plays back data originally set down by the disk controller.   The disk controller is where the data is pulled apart and decoded - thankfully I don't have to worry about that provided my virtual drive doesn't dish up complete gibberish.

So you have to 'mimic' the different time 'slices' it takes for the real Shugart drive to fetch (i.e. read/write) data from the disk? Do I underdstand you correct? If you don't do that will it be a problem for the floppy disc controller to handle the data from/to disk?
Have you decided for what kind of storge media you're going to use for you solution? SD or CF?

I hope you succeed!

//Escobar
 

The aim was to use MicroSD with a push push memory socket on the front panel.   

I have just revisited the HxC website and have noticed his standalone floppy emulator is coming on quicker than I'd imagined.  Before I go and reinvent the wheel I'm going to discuss the possibilities of teaming up and adding EII support to an already working platform.  If we can make this work then it would be a simple matter of nicely repackaging the design into a 5.25" drive form factor.

I will keep you informed whether there is any mileage in this approach.

Rest assured - We WILL come up with a solution as I know EII is crying out for something like this.

[Elmbeatz]

Rest assured - We WILL come up with a solution as I know EII is crying out for something like this.

YESSS   

[psoul]

please add sp1200 support too!!!

[Escobar]

You know what would be REALLY cool? A controller card with a standard PCMCIA-slot for different PCMCIA cards like blutooth wireless network cards so you can transfer the sounds from/to your EII with ease!
Ok guys, so when is it done? Next week, month?
Ok, ok, I sit here daydreaming/drooling sorry...