Expansion board for the CD32

[kipper2k]

Originally Posted by plasmab

How do you spec that in eagle? That seems like a 3D option?

I found the option on pcbway board options, select "gold fingers" and then enable 30 degree bevel. I think all board houses have this option, mabbe worded differently, not sure.

[spudje]

Originally Posted by supaduper

My second TF328 now built and luckily working straight off the bat so a 24 loop test..and it will then be going to its new home in France...JOTD
Attachment 54479

Oh, do you want me to contribute a bit. I think JOTD deserves some thank you for all his work on CD32Load!

Originally Posted by kipper2k

nice! would be nice to get whdload games working with full iso images etc

Well, the biggest challenge will be the CDaudio! Jotd's CD32Load already offers some CD32 tailored way to run WHDLoad slaves from CDROM. Now with this board I hope he will further improve his CD32Load run from HDD functionality. However I think supporting CD audio, or let's say wave files from the CDAudio will most likely not work, I don't think there is a way to directly stream these audio files from HDD in the same as from the CDROM drive. But that is maybe something to discuss here in the near future: http://eab.abime.net/showthread.php?t=81114

[Amiga1992]

Originally Posted by kipper2k

if you want the KB picchip 16f628a as well it will be an extra $4

Sorry to ask again, but can you provide instructions on how to put that on the older boards like mine? And do I have to buy this PIC off you then?

Originally Posted by supaduper

and it will then be going to its new home in France...JOTD

You're a star!!

[supaduper]

Just to say guys with this second tf328 I am having it crash with Defender of the Crown2 and Mad professor Mariarti , with games like frontier left on loop it does not miss a beat..strange
its very similar to the Timing problem on some A1200 when using the indivision accelerators and we had to remove to a couple of components to cure it.

They only difference between the two cards is that the memory on the 1st board was from mouser , that Steve linked to and ran out of stock, this board runs great no probs no crashes with the above games.

The second board I used these ram chips,https://www.utsource.net/itm/p/453595.html, that Steve was going to try, so do we think it could be these that are causing the crashes with timing errors ?, I ask because I am no expert on these things .

Also I must tell you guys again that the cheapo £3 CF/IDE adapters are still no good when using with the tf328, so i used another TRU IDE adapter and its fine with that.

[mattymuc]

Originally Posted by plasmab

I have some cheap 60ns jobs from UT source.

I have 2 stable working TF328 with mouser RAM and one TF328 with these UT source 60ns RAMs (HY5118160BTC-60). The UT Source TF328 is very unstable. Every whdload game/demo hangs/resets after some seconds.

Systest is running between 1 and 15+ rounds before I get RAM errors.

So I think these RAMs are not 100% compatible with the first firmware. Perhaps we have some timing issues here?

Edit: Oh, same problem as supaduper has?

[Stedy]

Originally Posted by supaduper

Also I must tell you guys again that the cheapo £3 CF/IDE adapters are still no good when using with the tf328, so i used another TRU IDE adapter and its fine with that.

I think this is down to a lack of buffering on some of the PATA (IDE) signals. It's not limited to the TF328, my A600 also exhibited issues. I connected an IDEFIX 97 to the system and it worked fine. I have an idea what causes the incompatibility, when time allows I'll investigate further. I added some inital ramblings here:
http://eab.abime.net/showthread.php?t=88096

Ian

[mattymuc]

Buffering is important when using long cables. My adapters are connected directly to the TF328/TF530 so there shouldn't be a problem with buffering. I have some very sensitive CF cards that don't work stable but they don't work stable with an IDE buffer also.

I'm going to post the "bad" CF-cards in your thread.

[supaduper]

Just a thought the CD32 is basically an A1200 and has the same signal set up regarding the Alice chip, would removing the E123C, E125C, help with the timing issue regarding these other Ram chips ? as the condition it displays is exactly as the one you get with Indivision accelerators

[Amiga1992]

Originally Posted by supaduper

Just a thought the CD32 is basically an A1200

I thought it was an A4000-based motherboard.

[Jope]

Originally Posted by kipper2k

@jope, i wonder if wrapping the wires in tin foil and using heatshrink will clean it up a bit ?

I was thinking perhaps a ferrite balun might eat some of the noise coming down from the PSU, but never got around to trying it.

I didn't get the interference with my flat panel display, but it was very noticeable when used with a real 1084. :-)

[plasmab]

Originally Posted by Stedy

I think this is down to a lack of buffering on some of the PATA (IDE) signals. It's not limited to the TF328, my A600 also exhibited issues. I connected an IDEFIX 97 to the system and it worked fine. I have an idea what causes the incompatibility, when time allows I'll investigate further. I added some inital ramblings here:

http://eab.abime.net/showthread.php?t=88096



Ian

I could put buffering on but I figured the Amiga community already knew how to work round these issues.


Sent from my iPhone using Tapatalk

[plasmab]

Originally Posted by mattymuc

I have 2 stable working TF328 with mouser RAM and one TF328 with these UT source 60ns RAMs (HY5118160BTC-60). The UT Source TF328 is very unstable. Every whdload game/demo hangs/resets after some seconds.

Systest is running between 1 and 15+ rounds before I get RAM errors.

So I think these RAMs are not 100% compatible with the first firmware. Perhaps we have some timing issues here?

Edit: Oh, same problem as supaduper has?

Weird. I'll have a think about this. I got 100 systest rounds with that RAM.


Sent from my iPhone using Tapatalk

[supaduper]

Originally Posted by plasmab

Weird. I'll have a think about this. I got 100 systest rounds with that RAM.


Sent from my iPhone using Tapatalk

Yes I have exactly the same errors as mattymuc using those same Ram chips,

I had a look on the Cd32 motherboard and both the E123, E125 caps are not present, anyway I did put two 22pf there just for fun and it will not boot so we can rule that out I think

[plasmab]

Originally Posted by supaduper

Yes I have exactly the same errors as mattymuc using those same Ram chips,



I had a look on the Cd32 motherboard and both the E123, E125 caps are not present, anyway I did put two 22pf there just for fun and it will not boot so we can rule that out I think

Very strange. I put them on a rev 0 board and they worked out of the box. There isn't much options with the timing. Its RAS then CAS for an access and CAS then RAS for a refresh. @14mhz the period is 71ns so.. pass. Are you both using buffered or unbuffered clocks?


Sent from my iPhone using Tapatalk

[supaduper]

Originally Posted by plasmab

Very strange. I put them on a rev 0 board and they worked out of the box. There isn't much options with the timing. Its RAS then CAS for an access and CAS then RAS for a refresh. @14mhz the period is 71ns so.. pass. Are you both using buffered or unbuffered clocks?


Sent from my iPhone using Tapatalk

Steve I have my clock set to the two pins furthest away from the connector edge,I have tried a jp1 and it was still the same

I have left systest running and its running away happily no errors ...wierd, makes you wonder if its more software related .I.E it wont play with some adapted games like DOTC2 and MPM.

yes strange as the Mouser rams never miss a beat

I have got one more of tf328 board here and have two of the mouser rams and two of the Utsource ones,do you think I should try a mix and match Steve to see what happens ?

[plasmab]

Originally Posted by supaduper

Steve I have my clock set to the two pins furthest away from the connector edge,I have tried a jp1 and it was still the same

I have left systest running and its running away happily no errors ...wierd, makes you wonder if its more software related .I.E it wont play with some adapted games like DOTC2 and MPM.

yes strange as the Mouser rams never miss a beat

I have got one more of tf328 board here and have two of the mouser rams and two of the Utsource ones,do you think I should try a mix and match Steve to see what happens ?

I mixed them on a rev 0. Should not make a difference. Will investigate more when I get back from vacation


Sent from my iPhone using Tapatalk

[supaduper]

Originally Posted by plasmab

I mixed them on a rev 0. Should not make a difference. Will investigate more when I get back from vacation


Sent from my iPhone using Tapatalk

Ok thanks Steve ,I will try the mix of Ram chips and see if there is any difference this end, as it will give us more options.
Will report back on the result.

[plasmab]

I've just ordered 100 bare boards from pcbway with the bevelled gold finger edge. Let's see if there is interest in them


Sent from my iPhone using Tapatalk

[alenppc]

I also just ordered some from EasyEDA. My first board order was messed up. I did 1.6 mm board without the bevelled edge so right now I have a bunch of useless PCBs.

[Amiga1992]

Originally Posted by alenppc

right now I have a bunch of useless PCBs.

Can't the edge be filed to shape?