16MHz BOOSTER 2021 edition

[exxos]

pcb.jpg

cpu.jpg

results.jpg

The successor to the V2.2 ST BOOSTER

This 2020 /2021 (rev 1.4) edition uses a smaller board and 16MHz DIP CPU instead of the PLCC68 HC variant. The reason for this move is to greatly decrease the size and number of layers on the PCB. The problem with the V2.2 was the fairly large PCB coupled with a six layer board was just simply too expensive to produce in low quantities, so ultimately the project was abandoned.

Mostly the booster when this way with the thought of adapting to 32MHz in the future, full clock switching on the ST series to 32MHz has been deemed as not possible. The SEC BOOSTER (not yet available so don't ask!) uses a totally different topology which I have had running up to 75MHz with a 68000 SEC CPU. As I currently do not have any ST accelerators in stock,at this booster will at least be something in production.

The same PCB is also used for the TOS206 decoder board. The booster also has the TOS206 decoding on board. You will need a suitable DUALTOS board. Currently I only produce one for the C070789 or MEGA ST boards. If you have a H4/H5 series remake board, then they already have a DULATOS ROM socket, so you can just plug in the booster and connect up.

This board was developed on the H5 series motherboard. There is no reason why this booster cannot be used on original ST/F/M machines, however all the mandatory fixes must be done prior to installation. As there are so many issues with original machines and there could likely be other problems, particularly across the vast number of motherboard revisions, I am not officially supporting this booster with original machines anymore.

As with other boosters, there is a auxiliary input to disable the booster and run at stock 8 MHz speeds. There is also a solder pad to permanently disable the booster if required (TOS206 decoding will still function even if the booster is disabled).

STORE LINK

https://www.exxosforum.co.uk/atari/store2/#0212

[exxos]

FITTING INSTRUCTIONS

The booster will have wires already soldered, these are for testing the module before they go to my store. You can re-use the wires, shorten them, or change them etc.

As mentioned above, if your fitting on a original machine, all the mandatory fixes must be done prior to installation. As there are so many issues with original machines and there could likely be other problems, particularly across the vast number of motherboard revisions, I am not officially supporting this booster with original machines anymore.

pcb.jpg

ROM_CS WIRE
H4/5 Right pin of JP13.

On 70789 style boards with the DUALTOS board like below..

UBESTFM.jpg

The solder jumper needs to be changed from illustrated above, TO the top and middle pads soldered. ROM_CS from the booster connects to the via above CE SELECT text on the PCB.

16MHz INPUT
Wire connects to SHIFTER pin 39. Cable should be short as possible.

8MHz ONLY JUMPER PAD
Forces 8MHz mode.

8MHz ON WIRE
Connect to a switch and when connected to 0V, Forces 8MHz mode.

[exxos]

CPU clock modification

Basically I found that some DIP 68K CPUs did not play nice with some boosters and was unstable. This was down to a 100R resistor in the CPU clock line. Once the resistor was removed (and solder blobbed over it) the problems went away. It gives the CPU a stronger clock drive.


The highlighted resistor below , value should be changed to 47R 0602.

R5_MOD.PNG

[exxos]

PRELIMINARY FIXES
(The current "stable" state of my booster)


BR modification

While testing some CPUs I ran into a weird situation on my H5 where the CPU would work fine in the booster, but be totally unstable in the ZIF socket. It had been working fine for several CPUs the night before ! It is possible this problem may show up in other hardware combinations.. I really don't know..

IMG_4220.JPG

TL;DR
It seems to be a issue with /BR on the CPU. There's no BR logic in the firmware, its a direct connection to the MB. Adding 50pF as shown crudely below seems to solve the problem (47pF also OK)

fix.PNG

Scrape some of the top solder resist off (yellow arrow grey square) and solder one leg of the 50pF there, and the other on the SOT23 package as shown by the other yellow arrow.

H5 BGACK FIX - Likely needed on original STs as well.
See this topic viewtopic.php?p=95672#p95672


REMOVE THESE HIGHLIGHTED CAPACITORS

caps.PNG

FIRMWARE UPDATE

ORIGINAL RELEASE FIRMWARE

206_V14_16MHZA.zip

2026 BETA FIRMWARE - Various fixes.

2026.zip

I use this programmer..

IMG_4219.JPG

Video HERE on how I use the software.

There may be other programmers to program ATF1502 but I am not currently aware of any (let me know? )



Current benchmark with blitter ON. Using TOS206 decoding.

IMG_4223.JPG

[exxos]

BG_SYS modification.

There is a missing pullup on BG_SYS. This is needed as BG goes via the PLD and the PLD release firmware mistakenly tristates the pin.

To fix this issue, solder a 2.2K pullup between the 2 highlighted red circle points. I suggest a 1/8w resistor as its tricky to solder with the 68K socket in place.

bg_fix.PNG

resmod.jpg

UPDATED FIRMWARE

2026_ALT_RAM_B.zip

The new firmware fixes the tristate BG_SYS issue.

It also switches into 16mhz for accessing alt-ram (8MB block starting from the 4MB mark in the memory map). You must have a 16mhz capable alt-ram board such as flashyclock (H5 platform only) to use this firmware.