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

Blogs & guides and tales of woo by forum members.
User avatar
crashman
Posts: 47
Joined: 25 Jan 2018 14:55

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

Post by crashman »

If you need help on troubleshooting, I am willing to help.
Even if it's just supporting your work helping to pay those first pcb's

BR
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 »

crashman wrote: 04 Mar 2024 07:47 If you need help on troubleshooting, I am willing to help.
Even if it's just supporting your work helping to pay those first pcb's

BR
cool, will get back to you then :-)
Until now this still fits into my hobby budget :)
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: 04 Mar 2024 06:00 Since it is the first prototype, I am quite satisfied
Although I am not that happy with the overall routing…
I am not 100% happy with the routing of the H5 even though it was totally manually routed. Unfortunately if you wait until everything was 100% perfect, then nothing would ever get released. As I think the saying goes, the last 10% of a project takes 90% of the time.

Also why the idea of the H4/H5 series was to have future add-ons. If I had put every little thing people wanted onto the motherboard right from the start, it would have taken considerably longer and have been considerably more expensive, ultimately the project would have just taken too long and would likely have simply been abandoned.

Indeed it took weeks if not months to route the H4/H5 motherboards manually. Though I really struggled to complete these projects because of my RSI. So in effect some compromises had to be made just to get the project finished.
I have routed some crucial parts manually, then let the autorouter do its work, and then changed a lot manually again.
indeed some things should be manually routed like the video circuit and the audio circuit etc. If you let the autorouter do the work then you could end up with RGB signals wrapped 20 times around the motherboard.
Depending on how much I have to change during troubleshooting, I am considering doing the routing again. Manually then…
Let’s see
As long as you are using a solid ground plane, then you are going to mitigate a lot of the original issues already. I don't know how many layers you ended up with ? I struggle to route the H5 on 3 layers. 1 was dedicated gnd and power, but I had to use the 5V layer for a lot as well. But that is a lot less critical than taking chunks out of the gnd layer. I think the autorouter would even struggle to do that on 6 layers with power planes.
BTW: for this first prototype, I decided to let JLCPCB only assemble the „small SMD stuff“
I am already having most of the 74 logic chips available (for hand soldering), but for the final run most of them can also be assembled by JLC
It makes sense, though I gave up with JLC for full assembly. I already wrote a lot of issues about all that. It is becomes too much hassle if they don't already stock the parts. And their customer support is always like a bunch of teenagers on the first day of their job. It is becomes incredibly frustrating to deal with them ultimately. I use them for small PCBs and small "birdseed" boards but nothing more now.
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 »

exxos wrote: 04 Mar 2024 14:23 As long as you are using a solid ground plane, then you are going to mitigate a lot of the original issues already. I don't know how many layers you ended up with ? I struggle to route the H5 on 3 layers. 1 was dedicated gnd and power, but I had to use the 5V layer for a lot as well. But that is a lot less critical than taking chunks out of the gnd layer. I think the autorouter would even struggle to do that on 6 layers with power planes.
It is a 4 layer board
I managed to really have a solid ground plane and also a quite solid VCC plane
All signals are on the top and bottom layers
The VCC layer only has a few wider routes for +/-12V around the edge of the PCB

The drawback is, that some of the signals take a longer route with a few vias back and forth - let’s see, how this works out :-)

I am really impressed, that you did all the routing manually! Esp. with your RSI - chapeau!
If you sit in front of the huge birdsnest, it is quite overwhelming :lol:
…where to start… WTF

So, I can imagine that it took several months… :D

Your comment on „the bunch of teenagers“ in China left me with a big grin on my face…
I am working for a German company having a Chinese Shareholder - so I exactly know what you mean
Communicating with them is a nightmare sometimes…
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: 04 Mar 2024 16:54 I am really impressed, that you did all the routing manually! Esp. with your RSI - chapeau!
If you sit in front of the huge birdsnest, it is quite overwhelming :lol:
…where to start… WTF
Yes absolutely.
So, I can imagine that it took several months… :D
I likely documented progress on the forum. But one of the main annoyances is when you realise that if you move something or rotate slightly, routing ends up being a lot better.. But then it also means you have to rip up 100's of tracks and re-route them all again :roll:

I think if I did a new H5 based board again (which I really don't plan on doing) I would reorganise how some of the layers used. Though it is difficult to predict how things will turn out until it is actually finished and you can see everything as a whole and then realise some things will be a lot better if done on different layers. But also I wish I had put the GLUE on the right side of the board. I had the address and databus running all over the place anyway, and most of the chipselects are more towards the right side of the board. So I ended up with having to route chipselect lines right across the board which could have been avoided. Though really I wanted to punch together the main chips anyway for future use of a plug-in FPGA board. This would have been a lot more problematic if the GLUE was on the other side of the motherboard.. Design compromises suck..
Your comment on „the bunch of teenagers“ in China left me with a big grin on my face…
I am working for a German company having a Chinese Shareholder - so I exactly know what you mean
Communicating with them is a nightmare sometimes…
8-) They are always trying to help but there becomes a point where it just becomes way too much effort. I used PCBWAY previously for example. While there was some mistakes on my part, They also made mistakes. Don't get me wrong their quality of work was excellent. But when they repeatedly ask the same questions like 30 times over.. Don't properly check the parts which I send which then cost me a fortune in shipping packages to them.. It was just one thing after another all the time . Overall communicating with them was extremely difficult in the end. Everything took like 50x longer than it should have. I just need companies who actually pay attention to whats going on and are not a huge effort to deal with all the time. Anyway, I've already wrote volumes about problems of PCB supplies over the years..
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 »

1st prototypes arrived
IMG_0525.jpeg
IMG_0535.jpeg
IMG_0521.jpeg

Checking footprints, distances, silkscreen, etc. in detail first - will take a while…

Already identified some minor quirks - and one major problem: picked the wrong footprint for the Mini-DIN ports for Mouse&Keyboard *sic*

At least it fits perfectly into the ATX case and the two ISA ports are placed exactly where they should be :D
You do not have the required permissions to view the files attached to this post.
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: 11 Mar 2024 19:11 1st prototypes arrived

:yay:

Already identified some minor quirks - and one major problem: picked the wrong footprint for the Mini-DIN ports for Mouse&Keyboard *sic*
Don't feel bad. I screwed up every IO port on the original remake Alpha :lol: :roll:
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 »

exxos wrote: 11 Mar 2024 19:39
Don't feel bad. I screwed up every IO port on the original remake Alpha :lol: :roll:
:mrgreen:
Thanks

Not unexpected, that things will not fit
Let’s see, how the rest will look like
User avatar
stephen_usher
Site sponsor
Site sponsor
Posts: 7376
Joined: 13 Nov 2017 19:19
Location: Oxford, UK.

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

Post by stephen_usher »

Good luck! :-)
Intro retro computers since before they were retro...
ZX81->Spectrum->Memotech MTX->Sinclair QL->520STM->BBC Micro->TT030->PCs & Sun Workstations.
Added code to the MiNT kernel (still there the last time I checked) + put together MiNTOS.
Collection now with added Macs, Amigas, Suns and Acorns.
User avatar
PhilC
Moderator
Moderator
Posts: 7441
Joined: 23 Mar 2018 20:22

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

Post by PhilC »

I remember those alpha days ,,@exxos, are you sure it was just the ports? :lol:

It is good to see an actual PCB, good luck with getting it all up and running.
If it ain't broke, test it to Destruction.

Return to “MEMBER BLOGS”

Who is online

Users browsing this forum: Baidu [Spider], ClaudeBot, petal [bot], rubber_jonnie and 14 guests