ST Modifications underway

[1024MAK]

Having received a HD FDD circuit, a MPU Overclock circuit and a LASCAN circuit from Ralcool (again,thanks Ralcool ), as the Atari ST that I am fitting them to is a different version main board to Ralcool's I am documenting the modifications here .
The PCB part number of the Atari STFM that I am modifying is C103414-001 REV 1.1 (a 520STFM which already has had 1.4 TOS fitted and a 4MB RAM upgrade fitted).

PCB.jpg

Anyway, after reading up and examining the PCB, doing lots of tracing of the PCB tracks and testing with a meter (this took quite some time with some head scratching ), I worked out where I am going to connect the many wires that need to tap into the existing circuitry .
Next up, make up the "wiring loom" with the connector sockets. Job done - See picture.

wiring loom.jpg

Now the HD FDD modification circuit has provision for an external floppy disk drive to be a HD type, but an extra connection is needed . The documentation for the HD FDD modification suggests using pin 7 of the external FDD connector (J4) so I had a close look at the 14 way DIN connector and the PCB tracks in the area of the connector. Pin 7 is one of two 0V / ground terminals.
There not being much space , I decided not to cut the pin of of the connector, but instead I decided to unsolder the connector to see if I could modify the PCB tracks instead. As with a lot of things in life, easier said than done . Well after some effort (with a solder sucker and desoldering braid), I have now freed the connector from the board and have started to modify the tracks so that pin 7 will no longer be connected to ground. Then I can connect a wire from the HD FDD circuit board to pin 7.

Ext Floppy PCB top side.jpg

Ext Floppy PCB underside.jpg

That's all for now, more updates will follow

[spiny]

yours looks a bit more full on then mine was good luck, hope it all works first time

[ralcool]

The overclock and FDD mod are quite straight forward.... The Lascan is a slightly tricky beast as a display signal is replaced/augmented by the mod- and involves CUTTING motherboard traces.

Good luck Mak.. My thoughts are with you.

(That is quite a harness you've made there!)



Sean

PS, I'll attach the program I modified so you can run 8 or 16mhz mode 'autoswitching'. (Midi ACIA Pin5 to the mods' lifted 74LS74 pin 4)

16MHZ.zip

Place 16mhz_1.prg in the auto folder to enable the mod while booting. Run 16mhz_0.prg anytime to switch it back off. (Run either anytime- doesn't matter)
If you set the jumper to disable... it will not switch on at all.

Sean

[ralcool]

Looking at the overclock harness.. your supplied mod has 7pins, but you've installed wiring for 8.... (Future versions of this board have 8pins for a new RTS input)
That would be the brown wire of that connector.. it will be unusable / no header pin to find contact with.

The mod overlay diagram I sent you has this mentioned/shown.

Looking at the harness... 5v and ground appear go to all the correct places. And the 16mhz line linked to the HD mod. (Why didn't you do the same with the 8mhz?)


Are you planning to integrate this late feature into the header wiring- then hand wired to the 74LS74? You could use the brown wire as the RTS control line.....
Sean

ps, You went to some effort to utilise Pin7 of the ext FDD connector ... I just used a small 2.5mm headphone jack on the rear of the machine. (Which when 'unjacked' grounds the pin)
You'll have to inspect the external drive cable and modify it to suit as well. And HOPE it didn't get cheap and not use just a single ground wire. Unless of course its a home made cable that is easily adjusted.
I read about this spare pin in my 'Elco HD kit' instructions And thought I'd add it- since it sounded useful... if tricky. It not compulsory to configure external HD drives... HDD1 pin of the mod can remain unconnected and it (the mod) will operate an existing external DD drive perfectly. (One of the mods' resisters pulls that pin low- if unused)

[ralcool]

This should help some.... From the original Overscan manual.

In German from here...http://dev-docs.atariforge.org/files/Ov ... ndbuch.pdf

lascan_STrev1.jpg

[1024MAK]

Part Two (a)
So after desoldering and removing the external floppy drive connector, its track cutting time!

Ext Floppy PCB top side after pin 7 pad drilled out.jpg

Ext Floppy PCB underside after track cutting.jpg

As you can see, I drilled out the PCB copper pad on the top side to isolate the pin of the socket when refitted. I used a 'veroboard' cutting tool (really just a 3.5mm drill bit with a plastic molded handle) and then a 10mm drill bit which I worked by hand.
On the underside I isolated the track for pin 7 but cutting it with a cutting disc in my 12V 'PCB' drill. You may wonder why you cannot see any golden brown copper showing. Well when I disturb a track by cutting it, I like to tin the exposed copper and then coat it with PCB protecting lacquer (I have a very handy pen like applicator which is great for repairs, rework or small PCB's ).
Next I routed the HDD1 wire via a hole that I desoldered (provided for the screening can that Atari never fitted - bit of a pig to clear due to the large amount of copper ground plane sucking heat from my iron , nearly had to get the big iron out). Then soldered it in place.

Wiring feeds #4 (feeds to HD FDD).jpg

Ext Floppy PCB underside after resoldering and connection of HD wire.jpg

Next was the HDD0 wire which I soldered to pin two of the internal FDD header on the underside of the PCB routing via the same hole described above.

Connection to pin 2 of internal FDD header.jpg

Getting in the swing of things, the other connections followed .
DS1 feed to HD FDD

Wiring feeds #2 (DS1 feed to HD FDD).jpg

8MHz feed to HD FDD

Wiring feeds #7 (8MHz feed to HD FDD).jpg

[1024MAK]

Part Two (b)
CLK (clock out) feed from HD FDD to WD1772 controller

Wiring feeds #3 (CLK from HD FDD to WD1772).jpg

Note that I have looped the +5V and the 0V / ground connections from the LASCAN board and the 16MHz clock from the Overclock board.
So on to the next one, the LASCAN connections
First up, the +5V, the 0V / ground, the H Sync and the Reset connections

Wiring feeds #6a (feeds to LASCAN).jpg

Then the V Sync and the DE in connections

Wiring feeds #1 (feeds to LASCAN).jpg

The DE out connection

Wiring feeds #12 (DE out from LASCAN).jpg

And the RTS 'control' signal from U54 pin 5 (keyboard ACIA). At the same time I also connected a wire to U52 pin 5 (MIDI ACIA) for future use.

's).jpg

Now time for a test fit...

Getting there, LASCAN & HD FDD in place.jpg

PCB 66% of wiring done.jpg

[1024MAK]

Part Two (c)
Now for the MPU Overclock circuit
First the 16MHz feed from the Shifter. As this machine has a socket half soldered on to the shifter (not all pins are soldered) I connected the wire to the leg of capacitor C108 nearest to the shifter.

Wiring feeds #9 (16MHz feed from shifter).jpg

Next some power supply feeds (+5V and 0V)

Wiring feeds #10 (+5V & 0V for Overclock).jpg

And the BG and AS signals from the 68000 MPU (U64), plus the new 8 / 16MHz clock out to the MPU

Wiring feeds #11 (feeds for Overclock).jpg

And of course an 8MHz feed to complete this part

Overclock connector & loom fitted.jpg

Now the LASCAN circuit needs the 'DE' signal line track (trace) cutting. 'DE' is produced by U65 (GLUE) and then goes to U43 (68901 MFP), U56 (MMU) and U49 (Shifter). But the LASCAN circuit needs to send a modified DE signal to the MMU and the Shifter, but the 68901 MFP still needs the orignal DE signal from U65 (GLUE). On this version of the main (mother) board I found that I could cut just one track to achieve this, the cut is made on the underside at the 'front' near the left hand side of the MPU right under the area of resistor R94.

' line for LASCAN.jpg

Now depending on how you modify your machine, you may need some additional track cuts. I found a dead handy section of the 8MHz line between U15 (WD1772) and U31 (DMA) which I could cut, this saved having to cut and lift pin 18 of U15 (WD1772). It is on the underside between U15 (WD1772) and U31 (DMA)

Track cut for CLK line to WD1772.jpg

I also used a track cut to isolate U64 (68000 MPU) pin 15.

[1024MAK]

Part Two (d)
Nearly done now, wiring now 100% complete
Just time for a couple more pictures...
You can also see that one 'ROM' has a different label to the others, that's because Ralcool kindly supplied a EPROM with a modified version of TOS 1.4 (UK) so that the step rate is 6ms. This allows the ST to boot from either DD 720kB or HD 1.44MB floppy disks

PCB 100% of wiring done.jpg

PCB 100% of wiring done (2).jpg

Modules ready for fitting.jpg

Now for some sleep before I dare to plug and pray
Reassembly and testing can wait till tomorrow...zzzzzzz
PS if anyone wants more info, ask me or Ralcool. All the pictures I have shown you are available as full size if needed.

[ralcool]

Man you really commit.

That harness looks excellent. And well thought out.

The first boot will be the litmus test. Doing all three mods at once raises the error stakes.

Bloody well done.

[1024MAK]

yours looks a bit more full on then mine was good luck, hope it all works first time

Thanks Spiny

Good luck Mak.. My thoughts are with you.

PS, I'll attach the program I modified so you can run 8 or 16mhz mode 'autoswitching'. (Midi ACIA Pin5 to the mods' lifted 74LS74 pin 4)

And thanks Sean (Ralcool)

[ralcool]

I'd like to make some recommendations for the first powerup..

Set the Overclock to disabled- We don't want a forced 16mhz cold boot problem like I experienced... Lets see if the ST runs at 8mhz through the mod- and will fire up with all the other cuts, etc.

Check the ACIAS for continuity between the pin 5s.. My 'Autoswitch-Overscan' manual warns that some early STs had these linked together for no good reason.

Test the FDD drive for HD signalling on its pin 2. Some drives don't do this- and will force the HD mod hard into DD or HD mode and not autoswitch correctly.
Also check again after being connected to the floppy cable- the pull down resisters might affect the signalling results too. (I had to use a 1meg resister for my external drive- and wire direct to the switch)

Consider making some loop back test headers out of single row header pin strips to be able to 'reenable/link' sections of the loom headers if a mod fails... (eg DE out to- oldDE, or 8mhz to FDD CLK)

Have a stiff drink. Say a Prayer.

Sean

ps. Have the Lascan boot disk prepared to go, but ensure the other important mods work first.

[spiny]

whoa ! thats all a bit full on ! I'm not sure I have the steady hands needed to cut actual motherboard traces

[wongck]

Wow you did those wiring in 2 days?
It took me 2 days to just solder my NVram battery mod

[1024MAK]

Wow you did those wiring in 2 days?

Yep!
But a week earlier (actually more like 10 days earlier) I studied the information, disassembled the Atari and worked out how I was going to fit the extra circuits including tracing the tracks (both by eye and with a little meter that I have). This part took a day and a half. So what you see in the pictures is the result of just over three days work
'
Not had much time to do anything today (I was looking after my niece for most of the day, and this evening had a union meeting to attend).
Did find a little time to take the lid off the two external 3½ floppy drives that I have. I have previously had the cover off one of them before, but wanted to check to see if the signal connector was how my memory believed it to be

External FDDs with case top removed.jpg

Anyway, I will have to check and test, but one of them has a replacement Sony drive, so may be a HD drive
Looking at the wiring, moving the dark green wire (which goes to pin 7 of the Atari 14 way DIN plug) from the 34way IDC connector pin 1 to pin 2 should connect through the HD signal (assuming that the drive actually generates said signal! ). The pin 1 connection on the Sony is not ground anyway. Note that the green and yellow wires just below the connector are nothing to do with the signal connector and you cannot see much of the dark green wire.

External FDD (Sony drive) signal connector.jpg

If the existing Sony drive is not HD, then I have another Sony FDD which is a HD type standing ready .
'
The Chinon drive signal connector.

External FDD (Chinon drive) signal connector.jpg

The dark green wire looks more like a off colour blue in this picture (something to do with the light I was using I presume)

You may also be interested in my 14 way DIN to 34 way IDC floppy adapter that I made up many years ago. I have not yet modified it to made it HD compatible, but will do in due course.

My DIN to IDC floppy adaptor.jpg

[wongck]

That DIN to floppy adapter looks interesting.... mmmm may make one so that I can have both floppy + floppy emulator.

[1024MAK]

That DIN to floppy adapter looks interesting.... mmmm may make one so that I can have both floppy + floppy emulator.

I've put some more details at http://www.atari-forum.com/viewtopic.ph ... 95#p170395

[1024MAK]

The Atari ST lives!
Well, at least with the three mods disabled or in non active states...
First I connected up the existing floppy drive, switched on and with a disk in, got a normal desktop. Good!
'
So on with the testing....

PCB with modules fitted and working.jpg

So the first thing to try is a HD capable floppy drive.
But the Atari is not happy , keeps attempting to read the nice new HD floppy but boots into desktop without reading the disk Strange, I know this floppy reads double density disks without problems and it is a HD floppy drive...

Sony MPF920-E FDD label.jpg

Okay, so what voltage am I getting on pin 2 of the floppy IDC connector. Bugger, can't get to it . So next best thing, get to the HDD0 pin on the HD FDD board . Damm, I'm getting 0V .
So off with it's head, sorry case (in this case the bottom section).
Oh look, a unconnected pin 2. Explains why no voltage . But look here, a "pad" that can be soldered across. Best check that the feed from the chip does actually produce a high logic level (5V or thereabouts) with a HD floppy in the drive. Testing with a meter confirms this , so power down and out with the soldering iron and hey presto, a logic high feed on pin 2 of the IDC connector . When measured on the HDD0 pin on the HD FDD board I get 3.49V. Okay, a little lower than I would have liked, but still (as far as TTL logic chips are concerned) a logic high. BUT I later found out that this part of the floppy drive circuit is a HD disk select INPUT not an output as it does not change to a low level when there is a DD disk in the drive, nor when there is no disk in the drive! To get the HD output signal, solder a wire from "SW3" "HD IN" switch (the pin next to the "SW3" marking) to pin 2 of the floppy connector (switch "SW3" is in the bottom right of this picture, better picture coming soon!). To enable the Atari to sense when a disk is changed you also have to solder a wire across the two pins of "SW2" "CT IN" (top right of picture).

Sony MPF920-E FDD PCB.jpg

Sony MPF920-E FDD PCB close up.jpg

Pic showing meters.jpg

In goes the HD floppy, reset the Atari and the friendly sound of an Atari booting from floppy is the result
And bonus, the Atari boots the Lascan driver (slight thought of panic as I see strange coloured lines on the screen before my memory kicks in again- phew!)
Whoo-hoo!! It works

Testing in progress.jpg

LASCAN in action.jpg

[1024MAK]

Modules fitted and working.jpg

Okay, so the Lascan settings need tweeking a bit, no matter, two out of three modules tested okay so far...

Now for the 16 MHz Overclock (the one that caused Ralcool so many problems).
Well mine does not like a cold boot either. Plus when reset, desktop comes up okay, but does not like reading program files from disk, program crashes just after loading

But switching to 16 MHz and then loading the program (in this case GEMbench) works
Anyway, will have to do the extra mod for this later.
Attached is the results from GEMbench

[ralcool]

Great stuff.. You speed test results aren't far from mine... TOS2.06 probably helps some GEM/VDI areas.. But the CPU & ROM access reports are within 1%.

How good is LASCAN?! Borderless computing like the PCs.

Be interesting to see how the reliability possibly improves using the software switch. At least you can then leave it permanently jumpered 'on'

You've posted some great pics.

You had me slightly nervous doing everything at once!

How good is installing mods with a soldering iron. Brings back the good old days of 8bit Z80s and good old fashioned hackery.

Cheers

Sean

ps, You've aparently got a newer version of Gembench than I... v3.28 vs 3.03... want to attach it?

[wongck]

That DIN to floppy adapter looks interesting.... mmmm may make one so that I can have both floppy + floppy emulator.

I've put some more details at http://www.atari-forum.com/viewtopic.ph ... 95#p170395

Thanks !

[wongck]

ps, You've aparently got a newer version of Gembench than I... v3.28 vs 3.03... want to attach it?

Actually you can get v4 from Umich Atari Archives.

[ralcool]

Wow, Lots of stuff there... and found Gembench v4.03. Ta.

[simbo]

worth an addition to this topic
frank_b and myself tried a 16mhz 68010 mpu inplace on a 8mhz 68000
as 68010 has some added ins and op's
ste.. booted to gem however hddriver breaks with a 2 bomb bus error
so perhaps a catch is needed to ident other mpu used where a 68000 lives normaly

not sure how he idents the mpu in code???

its a hddriver issue other may work

ill ask uwe,...

[1024MAK]

The attached word-processor file (in OpenOffice.org and Micro$oft formats) gives details of the wiring, including the colour coding of the wires I used and details of the connection points.