LarryL’s blog - creating an ATX version of the Phoenix H5

Blogs & guides and tales of woo by forum members.
User avatar
exxos
Site Admin
Site Admin
Posts: 28344
Joined: 16 Aug 2017 23:19
Location: UK

Re: LarryL’s blog - creating an ATX version of the Phoenix H5

Post by exxos »

LarryL wrote: 13 Feb 2024 13:47
  • all supporting caps are 1uF - usually I always see 100nF. Any specific reason to go for 1uF?
100nF caps were used in "the old days" because of the physical constraints of the package. These days we can push higher values in much smaller packages, so why not make use of ?
  • The MMU bus isolators & buffers are specced as 74AC373 and 74ACT244 - in my H5 I see LS types, so I think both is OK, depending on availability
That was due to parts shortages at the time of some production runs. LS types are preferable.

  • the H5 design shows a separate 16MHz OSC for the HD GAL, but there is an option to use the existing 16MHz from the Shifter. Is the latter tested and proven, or shall I go for a separate OSC?
This was done because often I found the shifter 16MHz clock caused problems when driving the 1772 on original machines. But because the H5 design was updated, it may not be a problem now. Unfortunately I don't believe anyone has done significant testing to prove this either way.

I have personally used the OSC a lot during daily usage for a long time and not noticed any problems. Maybe just wire to the shifter 16MHz feed via a resistor of course, and just have a option for a OSC if something causes a problem...
  • for the FDD, the Drive1 signal only goes to the external DIN14, not to the internal IDC. Since an ATX housing has enough space, I would also wire the Drive 1 signal to the IDC. Any reason not to do this? Of course, then internal and external Floppy should not be used in parallel. I could add another jumper to select...
That would be up to you if you wanted to put both drivers on the same cable. But of course any changes will undoubtedly result in future consequences" at some point . I don't believe anyone has even attempted it on a original machine. Though I would imagine for a ATX it would seem to make sense in having A & B on the same cable.
  • in the H5 reset circuit, the reset signal is driven from 3V3. Do I need to change things, if I drive this from 5V?
It can be 5V. It was changed on later boards because other items on the bus may use the reset line and be 3.3V devices. So it just made sense to run the reset signal from 3.3V. But as you are not using the 3.3V bus, you can stick to normal 5V levels.
  • I have not seen any polarized caps (electrolytic, tantalum) in the H5 design (and also not on my H5 board) Is this observation correct, or did I miss something?
This is also the same thing as your first question of why to use 1uF caps. Because there are dedicated power planes I don't believe that bulk capacitance is necessary. Of course with all the 1uF caps you are effectively having bulk capacitance anyway. I did not see any difference between having them and not on the H4.

I never use tantalum because they tend to catch fire, and I also tried to avoid using electrolytic for long-term reliability.
  • the original H4 audio circuit is using a BC337, only available as THT. Is a BC817-40 (like in the video circuit) the right equivalent?
You could pretty much use any generic transistor, however you would have to run a circuit simulation to see if any of the biasing voltages end up different. I had to do some small changes IIRC when I changed to the BC337 because of that.
  • the IDE circuit uses a 2N3906 to drive the reset line. I checked for a SMD equivalent and found BC857B. Good choice?
I don't know what circuit you are referring to because the H5 doesn't have IDE on-board.
  • for Keyboard&Mouse I use the Eiffel schematics. Do you see a need to add a RJ12 socket for e.g. a Mega ST keyboard?
That would be up to you.
  • for all resistors and capacitors I will go for 0805 footprint (hand soldering possible). Only for the few 75R resistors, JLCPCB has only 0603 in their "basic parts list". Going for smaller footprints for all resistors and caps would save some space, but (at least for me) makes it quite impossible to hand solder. What do you think?
You would have to re-evaluate all the power requirements of all the resistors to see if you could actually do that or not. I personally wouldn't solder a bunch of packages smaller than 0805 anyway.

Even so, extended parts sourcing generally costs very little anyway. So no need to limit to basic parts. I normally use more extended parts than basic as the basic range kinda sucks.
LarryL
Posts: 241
Joined: 20 Nov 2022 14:42
Location: Germany

Re: LarryL’s blog - creating an ATX version of the Phoenix H5

Post by LarryL »

Hi Chris,

thanks for the detailed answers - very much appreciated!

Caps: OK, I will stick to the 1u caps and will not use any polarized/tantalum ones - wanted to be own the safe side.
And yes, I also do not want to set the board on fire :lol:

LS types: good, I will see what is available...

16 MHz OSC: have to check, if space is sufficient for a separate OSC

FDD: I will go for having Drive0&Drive1 on the internal cable. But I will also equip the DIN14 for external Drive 1. Will add some additional jumpers to select

Reset with 5V: good!

BC337 in audio circuit: I think I will stick to the BC337, to be on the safe side

IDE circuit: I will share some schematics later - then it will be clearer. But maybe also here I can stick to the 2N3906 - doesn't need so much space

Eiffel or RJ12: I think some folks prefer to use a Mega ST keyboard instead of PS/2.
Since I anyway plan for a "2nd level PCB" in the ATX-Port-Area, the RJ12 can fit there

You are right re the "extended vs basic parts" at JLCPCB: they do a surcharge, but not that much

OK, in the meantime I did most of the component placing and did a first routing attempt using the KiCAD autoroute plugin "Freerouting" - not that bad.
BUT: Some optimization needs to be done, definitely!
LarryL
Posts: 241
Joined: 20 Nov 2022 14:42
Location: Germany

Re: LarryL’s blog - creating an ATX version of the Phoenix H5

Post by LarryL »

OK, that's the IDE & TOS 2.06 decoding schematic from the ATX design
IDE_2.06decode.png
It is basically the ppera design from here using the very same GAL code

TOS 2.06 decoding is designed for 2 Chips (e.g. 27C020)
But I will use the H5 ROM design with the AT27V4096, so I think this will do the trick:
TOS_select.png
You do not have the required permissions to view the files attached to this post.
User avatar
PhilC
Moderator
Moderator
Posts: 7440
Joined: 23 Mar 2018 20:22

Re: LarryL’s blog - creating an ATX version of the Phoenix H5

Post by PhilC »

Just wanted to say that I’m watching your progress with a lot of interest.

I was going to do a Mega ST conversion but had to give up on it due to a lack of time.
If it ain't broke, test it to Destruction.
LarryL
Posts: 241
Joined: 20 Nov 2022 14:42
Location: Germany

Re: LarryL’s blog - creating an ATX version of the Phoenix H5

Post by LarryL »

PhilC wrote: 21 Feb 2024 16:02 Just wanted to say that I’m watching your progress with a lot of interest.

I was going to do a Mega ST conversion but had to give up on it due to a lack of time.
cool :D

it's fun, so far - although risky, since I am not an expert at all :lol:
LarryL
Posts: 241
Joined: 20 Nov 2022 14:42
Location: Germany

Re: LarryL’s blog - creating an ATX version of the Phoenix H5

Post by LarryL »

OK, next question...

Since I like to be able to switch between Drive1 on internal FDD-Header and external DIN14, I was wondering how to do this...
Is it sufficient, to simply only switch the drive1 signal to internal/external drive?
What about the Motor_On signal?

Lets assume, I have 2 internal drives on the internal cable (where Drive0 and Drive1 is connected)
If I connect an external drive to the DIN14 (with jumper for Drive1 now only connecting to the DIN14), will the internal 2nd drive still react on the Motor-On signal, even when the Drive1 signal is not connected to the internal drive?
:?
LarryL
Posts: 241
Joined: 20 Nov 2022 14:42
Location: Germany

Re: LarryL’s blog - creating an ATX version of the Phoenix H5

Post by LarryL »

LarryL wrote: 21 Feb 2024 16:16 OK, next question...

Since I like to be able to switch between Drive1 on internal FDD-Header and external DIN14, I was wondering how to do this...
Is it sufficient, to simply only switch the drive1 signal to internal/external drive?
What about the Motor_On signal?

Lets assume, I have 2 internal drives on the internal cable (where Drive0 and Drive1 is connected)
If I connect an external drive to the DIN14 (with jumper for Drive1 now only connecting to the DIN14), will the internal 2nd drive still react on the Motor-On signal, even when the Drive1 signal is not connected to the internal drive?
:?
if above is sufficient, this should work...
DRV1_int_ext_1.png
DRV1_int_ext_2.png
You do not have the required permissions to view the files attached to this post.
User avatar
JezC
Posts: 2782
Joined: 28 Aug 2017 23:44

Re: LarryL’s blog - creating an ATX version of the Phoenix H5

Post by JezC »

@LarryL Not sure if you have already looked at how exxos has done it on the H4/H5 boards but that works well in practice.

Only complication is your desire for two internal floppy connections...

I too am following with great interest.

:thumbup:
LarryL
Posts: 241
Joined: 20 Nov 2022 14:42
Location: Germany

Re: LarryL’s blog - creating an ATX version of the Phoenix H5

Post by LarryL »

JezC wrote: 21 Feb 2024 18:06 @LarryL Not sure if you have already looked at how exxos has done it on the H4/H5 boards but that works well in practice.

Only complication is your desire for two internal floppy connections...

I too am following with great interest.

:thumbup:
Hi,
Thanks for following :-)

Since I am mostly copying the H4/H5 architecture, the drive-A-B swap is already incorporated.
As you spotted right - my additional requirement is, having two internal drives and potentially one external drive - of course not constantly

So, if I attach an additional external one (to the DIN14), I need to disable one internal drive

The H4/5 only has one internal and one external drive - so a simple swap of Drive0/1 is sufficient

Don’t ask, if I need three drives (of course not) - but it would be nice to have the option :D

But your comment made me think again :-)
Even with only two attached drives, the only signal which is different at both drives, is indeed drive0/1
All the other signals are connected to both drives anyway
So, it should be sufficient, to switch the two drive select signals between my 3 drives

Regards
Michael
User avatar
alexh
Site sponsor
Site sponsor
Posts: 1335
Joined: 17 Oct 2017 16:51
Location: Oxfordshire

Re: LarryL’s blog - creating an ATX version of the Phoenix H5

Post by alexh »

Thanks for posting. This looks very exciting. I am watching with interest.

I don't know if it is practical but if you can route digital Red Green Blue + Hsync + Vsync +5v +gnd to a header (approx 16 pins) you could easily add HDMI output to the ATX back-plate (which would look like a gfx card) using a SiPeed Tang Nano 9k which are about $10

The logical design for an equivalent of an RGB2HDMI using a Tang Nano 9k has already been written. As well as outputting over HDMI it can also output via MIPI camera connector.

Why would you want that?

It can connect to a PiStorm for automatic switching between Native gfx and VDI gfx card.
Senior Principal ASIC Engineer - SystemVerilog, VHDL
Thalion Webshrine - http://thalion.atari.org
ST,STf,STfm,STe,MegaST,MegaSTe,Falcon060
A500+,A600,A4000/060,CD32,CDTV

Return to “MEMBER BLOGS”

Who is online

Users browsing this forum: Bing [Bot], ClaudeBot and 12 guests