IDE+2 Questions

[+Larry]

Is it possible to change partitions of D1: on the fly on the IDE+2 (not ATR's)?  I'm pretty sure that this question came up a long time ago, and may have been asked about the earlier IDE+.  Drive swapping has always been a handy feature of the MIO and Black Box.  I'm going to get both those out and reacquaint myself with their operation.  The IDE+2 is my "daily driver" due to its speed and compact size, but it sure would be nice to swap partitions on the fly.

 

My goal is to have two 720-sector partitions and be able to boot from one partition then "flip my disk" (partition) to the other.  Only one needs to be bootable.  I've read the configuration manual about "Boot Drive" but it doesn't appear that I can do the swapping that I want. (?)

 

[Rybags]

I don't think so (I've not used mine in a fair while and am probably multiple firmware releases behind).

 

You can change which one to boot from though.

And thanks to the OS versatililty that substitutes the IDE IO method almost transparently with SIO, you can have in effect a floppy image as a drive, it's not really the preferred way to do things.

[drac030]

36 minutes ago, Larry said:

Is it possible to change partitions of D1: on the fly on the IDE+2 (not ATR's)?  (...) Drive swapping has always been a handy feature of the MIO and Black Box.  I'm going to get both those out and reacquaint myself with their operation.

Currently nothing like that is implemented. It is of course pretty easy to add a swap table, but there always is a question remaining on how the user has to access the swap feature, i.e. what "on the fly" really means. Through the IRQ button like ATRs, or entering the menu with Start/Reset? In any case this means that the configuration menu should be expanded/reworked, and this is exactly what I am procrastinating

[+Larry]

Thanks, Konrad & Rybags. I think that one would have to be able to activate a non-destructive menu, and then return to the same point as you left.  Yes, it sounds like a lot of work! 

[Roydea6]

@Larry if this is a MyDos application then use the R function to set D2: and use H function to set the environment for future use.

if it is SpartaDos the type swap 1,2.

 

[+Larry]

Hi Roy-

It's for boot games -- bootable is on Side A; data on Side B.  SIO works great, but it is relatively slow.

[Rybags]

Something that would be cool would be the ability to have D1: mapping by a hotkey to allow disk swaps.

Maybe a large file container that held multiple ATR images.

But then it'd need to be guaranteed contiguous.  And the firmware would need to be asssured all the mapping possibilities fall within the bounds of the file.

Gets complicated fairly quickly.

 

Or maybe just have a particular partition type that is just assumed to contain lots of small disk images in groups of 4 or something?

[drac030]

Actually I think the easiest way to what Larry wants is to have the multidisk game stored as ATRs on a SDFS-formatted HDD partition, mount them all from the loader level to consecutive disk units (like D1, D2, D3, D4), boot, then use the IRQ button to rotate the ATRs. However, Larry probably does not have an SDFS-formatted partition, and MyDOS format does not allow to access disk images (such as ATRs) in a random manner.

 

4 hours ago, Rybags said:

Or maybe just have a particular partition type that is just assumed to contain lots of small disk images in groups of 4 or something?

This reminds me that the APT partition type $02 is just this: the partition split internally into data chunks of fixed size. So maybe it is time to use this bit of the APT specification and implement this. Booting from such a partition could then enable the IRQ button to rotate the chunks.

[+Larry]

Actually I have a couple of cards with SDFS on them. But would that be faster (loading) than APE with a low pokey divisor? I'll probably attempt it, though. Lots of ways to "skin this cat," just depends on how much hassle they are.

[drac030]

23 minutes ago, Larry said:

But would that be faster (loading) than APE with a low pokey divisor?

Reading speed from a "mounted" ATR in IDE+ should be about 25 KB/s. The writing is much slower, only about 7 KB/s.

[foft]

While Larry is asking IDE2+ questions, I thought I could ask another one!

 

A year or so ago Larry sent me his old IDE2+ revision E to test on the EclaireXL (Many thanks Larry!).

 

I've since found it quite useful and it seems to work ok, except my Atari 600XL only boots when I hold option if it is connected! I think this is because some settings are corrupted...

 

Larry told me the board has an issue with the settings getting lost and had previously added a big capacitor by the external power supply (which I'm not using) to alleviate this. I guess this is to smooth the power off curve on shutdown? He told me that had worked for a bit, but then stopped working. I changed the CR2032 recently since it seemed a bit flat and it seemed somewhat better.

 

Anyway I was wondering if anyone knows a proper fix that I can do to the Rev E to make it store settings - and more importantly to allow it to boot with basic enabled? Is the schematic for the various revisions available somewhere? 

[drac030]

The IDE+ settings (let me say the obvious at the outset) are stored in the NV-RAM of the onboard hardware clock. While I am not competent to advise you what to do if there is something wrong with the hardware, there exists a common cause of setting corruption in IDE+ even if the hardware itself works as designed: it is hitting Reset or switching the power off while the system is accessing the clock to fetch the current time.

 

The reason is that the V3021 clock used in rev. A, B, C, D and E does not really have NV-RAM separate from the time memory. So, to read the clock, the firmware has to do the following:

 

a) read out the settings and store them in the internal RAM

b) read the time

c) restore the settings

 

If you hit Reset or switch off the power while the firmware is doing b) and has not yet managed to complete c), the settings are lost.

 

This has finally been fixed in IDE+ rev. F which is rev. E with DS1307 clock, and this one has real NV-RAM, so this problem is gone.

 

Now I have no idea if it is possible to replace the V3021 with DS1307 without also replacing the CPLD core, @Simius could probably tell you more.

 

PS. And the system is constantly reading time when one has TD-line on or the clock display is enabled in the lower right corner of the right panel in Sparta Commander...

 

PS.2. And the same stuff happens when one hits Reset or switches the power off while being in the IDE+ configuration menu, since this one also is reading the time constantly.

Edited October 28, 2020 by drac030

[+Larry]

 

So long as it was powered, the settings did not get corrupted/lost.  Otherwise, I had to hold Start at power-up to reconfigure the settings. It lost the settings from the time that I got it from Simius, and he suggested the capacitor.  I've had 2 rev C, 2 rev D, one E and now an F.  Only E had this problem, and it worked fine for a couple of years after I added the cap.  I asked Simius, and he said (essentially) "buy an F."  I did that, and so far, no issues. I don't recall the BASIC/Option issue on my NTSC Antonia 4MB 600XL.

[drac030]

So it seems there was a hw problem, nevertheless my comments on V3021 still apply and settings can be lost in circumstances described, even if the hardware is working perfectly.

 

F definitely is the best revision, or rather would be, if it allowed to attach a real HDD to itself, and not only a CF card.

[drac030]

1 hour ago, foft said:

except my Atari 600XL only boots when I hold option if it is connected!

Try this IDE+BIOS and see if it gets any better:

 

 

[foft]

Well this big cap could allow it enough time to finish ‘restore the settings’. 
 

Looking at the data sheets it seems the ds1307 is i2c and v3021 is something else, so not a drop in replacement. I guess if required a new board revision to change pins + cold core. I wonder if I can change cpld core + solder some wires...

 

[foft]

Will give that bios a spin tomorrow night thanks.

[foft]

@Larry, did you always run it with an external psu then?

[+Larry]

19 minutes ago, foft said:

@Larry, did you always run it with an external psu then?

No, only for testing.  I just went back and used a rev D (and then got the F).  D is larger with a lot of discrete components, but it is my favorite.  It was the first version to work with the 65816 cpu.  Antonia + the IDE+2 is a really nice combination.  And add drac's 65816 OS. Great stuff!

[+Larry]

Test #1 -- The Black Box

Since I couldn't use the IDE+2 to swap partitions, I got out a BB and set up two 720-sector Single Density partitions. (The BB allows 99 partitions.) It wasn't too difficult, but with all the screens for the BB configuration, it took some time to figure it out. (The docs are pretty good.)  Then I copied the two sides of a graphic adventure (Adventureland) to the two floppy partitions with MyDos.  The first disk side booted way faster than an SIO drive. Then thanks to the BB configuration menu, I could swap the game image (partition) to Side B. Worked flawlessly. But here is the rub -- the computer program takes quite a bit longer than the loading from the hard disk. But the whole process is so much faster than I remember from many years ago, playing these games from a 1X floppy.  Even using a Pokey divisor of 8 (Ultra Speed) with APE, seems very acceptable. 

 

Next test -- I'll attempt to mount two ATR's onto the IDE+2 in a SDFS partition.

[drac030]

7 hours ago, Larry said:

the computer program takes quite a bit longer than the loading from the hard disk

An accelerator should solve this problem

 

8 hours ago, Larry said:

Even using a Pokey divisor of 8 (Ultra Speed) with APE, seems very acceptable

On IDE+ the ATR loading, as I wrote above, should be twice the Pokey divisor 0.

 

[foft]

On 10/28/2020 at 10:00 PM, drac030 said:

Try this IDE+BIOS and see if it gets any better:

 

 

Yes, much better. Many thanks!

 

Now it works with spartados enabled without holding option. So on the old bios, black screen with spartados and basic not disabled is normal?

 

So to avoid settings getting trashed I should avoid turning it off either in the start menu. I was probably doing that all the time without realising. What is the TD line though, is that something I'd know if I was using? (edit: found its an app to display date/time in startados...)

Edited October 30, 2020 by foft
Added thanks!

[foft]

Just linking the post with the buttons, since I couldn't find what they were without a lot of looking:-)

 

[drac030]

46 minutes ago, foft said:

So on the old bios, black screen with spartados and basic not disabled is normal?

I guess, on XL without the RD5 line connected to IDE+, when you do not hold Option, you end up with both BASIC and SDX cart present at the same time in the same place. I personally have an XE, so this problem does not occur and I was not quite aware of it. Once it was confirmed in that topic by @MrFSL that holding Option down makes things better, I made a change in the BIOS code which "holds the Option", i.e. disables BASIC when SDX is active.

 

In the case of BASIC this helps, but when you insert an external cartridge, the problem may be back. Unless I am mistaken, there is no cure for that except adding the RD5 line to the PBI connector or connect it to IDE+ by other means.

[MrFSL]

Yeah things are working well. Since it is a change in how the unit functions, I will also mention that the ^X and ^B methods of saving the config now work as you would expect.  Before when entering the config from SDX ^X or ^B would result in possibly black or green screen that would hang. Now ^X saves and returns you to SDX while ^B saves and reboots into SDX.

 

Also in prior BIOS, when powering on while holding START (and possibly OPTION), ^X would get me to SDX and ^B would skip SDX and try to boot to whatever the boot drive was set to in CONFIG. Now it works more like I expected. If powering on while holding START both ^X and ^B boot SDX.

 

I never thought much about this functionality being related to the OPTION key press that we are discussing here. @drac030 this really removes many of the quirks I was experiencing leaving the device much more intuitive. Cheers.