Problems burning Kickstart

[hit]

Originally Posted by Jope

WaitIDE adds delay to the 3.0 bootup.

ah, ok. thought so too, in the end
but is it just some sort of a "number" which has to decrease to have a shorter delay in the scsi.device of 3.1?

[yaqube]

Originally Posted by orange

for some reason it still won't work in my A500 (rev.6a). I get red screen and ,after several programming 'passes', yellow. (I did several passes thinking that might improve response time)
it seems something is wrong with A500 because it wouldn't work with original 3.1 ROM (not EPROM) either. (strange, happens with two A500)

I had burned an EPROM (AMD AM27C400) with Kickstart 3.1 image for my A500+ but when I replaced the original ROM I got a red screen with flashing power LED. Very similar situation to the one described by orange.

The kickstart was the 40.063 version intended for A500/A600/A2000. I had verified the CRC-32 checksum before burning it. The image wasn't corrupted or modified. I had checked it with the SKick. Everything had worked fine.

With the CPU replaced by the turbo board (030 CPU) the kickstart booted correctly (I'm sure it wasn't a 020+ image because it had been soft-kicked earlier and had worked with the stock 68000 CPU, the EPROM was also verified if it contained the programmed data).

Strangely when my Action Replay III card was connected the kickstart booted correctly but required disabling of the cartridge (actually it was A2000 version of the Action Replay III connected through a bus board). I have burned another EPROM with the same data but the story repeats.

Then I started to suspect that something strange must be happening on the CPU bus. I noticed that the Action Replay card had pull-down resistros on the data bus but my rev 8A motherboard was missing data bus pull-ups. I populated the missing resistor packs (RP108 and RP110) but nothing happened. I tried both pull up and pull down configurations without success.

Then I started to measure signals at the CPU socket with my scope and found out that when I touched with a probe some low order address lines the power LED flashing stopped. The system hung. It was deterministic behaviour.

I realized that also low order address lines (A16-A1) were missing their pull-up resistor packs (RP106 and RP107) so I removed data lines pull-ups and moved them to low order address lines. Since that moment everything works flawlessly.

You can locate aforementioned missing resistor packs next to the CPU socket on the A500+ rev 8A motherboard.

I don't really know why it initially worked with the original mask ROM but didn't with the CMOS EPROM. One possible explanation is an input leakage current of the ROM/EPROM chip. The 68000 CPU was manufactured using NMOS technology with relatively weak sourcing current of high output state (0.4mA). According to the datasheet of the ROM chip its input leakage current is ten times larger than the one of the EPROM.

I hope someone will find my description helpful.

[hit]

Originally Posted by yaqube

I hope someone will find my description helpful.

i'm sure that

just to see if i understood this. you installed two resistor packs at RP106 and RP107 to make the "AMD AM27C400" eprom working. what are the types/specs of these resistor packs?

[yaqube]

Originally Posted by hit

just to see if i understood this. you installed two resistor packs at RP106 and RP107 to make the "AMD AM27C400" eprom working. what are the types/specs of these resistor packs?

Yes, I installed two resistor packs (aka resistor networks or resistor ladders) as specified in the rev 8A schematics: 9 x 4k7 in SIP-10 package.

One of the suitable components is BOURNS 4610X-101-472LF. You can buy them at Farnell.

[hit]

ah, ok. thanks for the additional explaination

Originally Posted by rkauer

Yes, there is two advantages: less memory consume (few kb) and less boot time.

Also, by doing this (removing unneeded libs) gives one more room in kickrom for newer libraries from BB#2, and some of them are a lot larger than original ones from 1993. Bear in mind, no need to overkill the kickrom, since it is slower than most memory on accelerator boards. Thus its a good idea to do kickstart remaping into fast mem, using tools such as QuickROM (or others that may come on your accelerator install disk).

[rkauer]

Oh, another (excellent) advantage is support for (very) large HDs right on boot, if you burn the ROM with the fixed SCSI.device from our friend EAB moderator.

[spotUP]

so, how are your kickstarts turning out boys?
doobrey, any news from you?

[rkauer]

Mine is the 3.9 + BB1 + BB2 + patched SCSI.device in ROM. Runs perfectly and no crashes at all.

[spotUP]

where can i get this patched scsi device?

[Calgor]

@spoUP

I am asuming he is talking about the patch which includes fix for LBA48:
http://eab.abime.net/showthread.php?t=35801

Post 20 in that thread shows where to get it off aminet.

[orange]

@yaqube
it works great! I just put in RP106 and RP107 like you said and A500 booted KS3.1 fine. thanks

[spotUP]

@calgor

thanx!

[aperez]

I just scored, big-time...511 pcs of M27C4002 in a 44-pin PLCC package, $53. Obviously not pin-compatible but not hard to fix. And yes, they are One-Time-Programmable, so I'll be testing it before I burn

http://cgi.ebay.com/ws/eBayISAPI.dll...m=260401091574

[Eclipse]

Originally Posted by rkauer

Mine is the 3.9 + BB1 + BB2 + patched SCSI.device in ROM. Runs perfectly and no crashes at all.

How much?

[rkauer]

I've paid USD37 for the chips and postage. I have original OS3.9 CD & original 3.1 ROM, of course.

[Eclipse]

Can a 1.3 rom from an A500 be put into an A600?
I mean take it from one machine and plonked into another?

[DoogUK]

A500/600/2000 roms are interchangeable between each other...no other amigas are

[Calgor]

I thought that the Toshiba TC574200D and the M27C400 were both 512KB with identical pinouts. However, the Wellon programmer comparison list (http://www.weilei.com/per.htm) says that the TC574200D is only supported by higher model with full 48pin driver rather than cheaper VP280 (I have) and VP380 (Toni has) that have 48pin "popular" driver. Is there any reason for this?

Toni, any comments if you are reading? I wanted to use the TC574200D for my 1MB ROM A1200. Will I destroy the ROM by programming it as an M27C400? (The device ID feature of the programmer did not work for some reason, and is not documented in the help section that I could find).

[Toni Wilen]

Originally Posted by Calgor

I thought that the Toshiba TC574200D and the M27C400 were both 512KB with identical pinouts. However, the Wellon programmer comparison list (http://www.weilei.com/per.htm) says that the TC574200D is only supported by higher model with full 48pin driver rather than cheaper VP280 (I have) and VP380 (Toni has) that have 48pin "popular" driver. Is there any reason for this?

Toni, any comments if you are reading? I wanted to use the TC574200D for my 1MB ROM A1200. Will I destroy the ROM by programming it as an M27C400? (The device ID feature of the programmer did not work for some reason, and is not documented in the help section that I could find).

I compared datasheets I found and didn't find anything interesting. TC574200D seems to be 100% identical to 27C400..