TF4060 Beta Program

[terriblefire]

Code: Select all

[315ms] WARNING: Detected supply voltage drop to 2628mV. Verify power supply is adequate.
[606ms] WARNING: Detected supply voltage drop to 2573mV. Verify power supply is adequate.
[3536ms] WARNING: Detected supply voltage drop to 2521mV. Verify power supply is adequate.

Could we eliminate this please before going further?

[kludge]

New fw for the 4060 for those able to update.

Significant fixes.

IDE lockups sorted out.
SCSI issues fixed included weird DMA to chipram issue.
IDE speed at max pio.

tf4060r5_20611_2024_07_24_4252a06_alpha.zip

Had IDE lockup almost instantly before, now it's solid! Tried a few different cpuspeeds as well; 60 (freeze), 70 (violent crash), 75 – 100 (seems stable), 101 (freeze). Didn't try the different speeds for very long, just enough to get SysInfo going :)

AmigaOS 3.2.2.1, rev 6 060, Cybervision 64/3D (no graphics library installed yet, though), FastLane SCSI Z3 (w/ BlueSCSI networking).

Uploaded my heatsink holder to GitHub if anyone wants to tweak it for their own heatsink: https://github.com/andersmilton/TF4060-Heatsink-holder

[terriblefire]

• locks up immediately during - I guess - partition lookup as it happens before first DF0: access

Any updates on this. I wouldnt have expected it to take you two days to plug power into the ZuluSCSI and retest.

I have tested here and with the 4000CR 030 you can get away with the ZuluSCSI and not plugging the power in. You cannot with the 4060 running.

[Chritoph]

• locks up immediately during - I guess - partition lookup as it happens before first DF0: access

Any updates on this.

Other things on my table as well, sorry, currently helping my daughter to prepare for an exam.
Will try later today with power added or a real SCSI HDD.

[Chritoph]

So this is how it looks like with ZuluSCSI getting power delivered via Berg connector:

Code: Select all

[10ms] Platform: ZuluSCSI RP2040
[10ms] FW Version: 23.05.25-release May 25 2023 19:01:02
[11ms] DIP switch settings: debug log 0, termination 1
[11ms] SCSI termination is enabled
[12ms] Flash chip size: 4096 kB
[14ms] SCSI target/disk mode selected by DIP switch, acting as a SCSI disk
[98ms] SD card detected, FAT64 volume size: 121910 MB
[98ms] SD MID: 0x03, OID: 0x53 0x44
[99ms] SD Name: SN128
[99ms] SD Date: 2/2023
[99ms] SD Serial: 0x5AB38CCE
[100ms] Reading configuration from zuluscsi.ini
[101ms] Active configuration:
[102ms] -- SelectionDelay = 255
[102ms] -- EnableUnitAttention = No
[103ms] -- EnableSCSI2 = Yes
[103ms] -- EnableSelLatch = No
[104ms] -- MapLunsToIDs = No
[104ms] -- EnableParity = Yes
[113ms] -- Opening A3000T-32GB.img for id:0, specified in zuluscsi.ini
[114ms] ---- Read prefetch enabled: 4096 bytes
[188ms] Finding HDD images in directory /:
[193ms] -- Platform supports ROM drive up to 3740 kB
[193ms] ---- ROM drive image not detected
[194ms] SCSI ID:0 BlockSize:512 Type:0 Quirks:0 ImageSize:33800760kB
[294ms] Initialization complete!
[304ms] WARNING: Detected supply voltage drop to 2578mV. Verify power supply is adequate.
[3847ms] WARNING: Detected supply voltage drop to 2523mV. Verify power supply is adequate.

Not much of a difference IMHO, issue also persists.
Digging out old SCSI HDD now...

With old SCSI Disk, looks like this:

• SCSI disk LED permanently lid,

• No floppy drive clicks,

• No floppy access when inserting one

[terriblefire]

It's still reporting a bad PSU. will see if Fred gets the same issue on his 3000T.

it has been reported working before so a little bug confused.

[Chritoph]

Spinning hard disk is pretty unimpressed by anything PSU I guess 8-) .
I can use either, SCSI HDD or Zulu for booting with other accelerators (e.g. CyberStorm MK2 040/40MHz) or with onboard 030/882,
so I guess the suppy voltage message is kind of a smoke grenade just distracting from something else.
Also: if Zulu would be in trouble or crashing, how did it write the message to SD Card? ;)

This is now from a pretty successful boot with the CS MK2 mentioned above:

Code: Select all

[10ms] Platform: ZuluSCSI RP2040
[10ms] FW Version: 23.05.25-release May 25 2023 19:01:02
[11ms] DIP switch settings: debug log 0, termination 1
[11ms] SCSI termination is enabled
[12ms] Flash chip size: 4096 kB
[14ms] SCSI target/disk mode selected by DIP switch, acting as a SCSI disk
[97ms] SD card detected, FAT64 volume size: 121910 MB
[98ms] SD MID: 0x03, OID: 0x53 0x44
[98ms] SD Name: SN128
[99ms] SD Date: 2/2023
[99ms] SD Serial: 0x5AB38CCE
[100ms] Reading configuration from zuluscsi.ini
[101ms] Active configuration:
[101ms] -- SelectionDelay = 255
[102ms] -- EnableUnitAttention = No
[102ms] -- EnableSCSI2 = Yes
[103ms] -- EnableSelLatch = No
[103ms] -- MapLunsToIDs = No
[104ms] -- EnableParity = Yes
[113ms] -- Opening A3000T-32GB.img for id:0, specified in zuluscsi.ini
[114ms] ---- Read prefetch enabled: 4096 bytes
[188ms] Finding HDD images in directory /:
[193ms] -- Platform supports ROM drive up to 3740 kB
[193ms] ---- ROM drive image not detected
[194ms] SCSI ID:0 BlockSize:512 Type:0 Quirks:0 ImageSize:33800760kB
[294ms] Initialization complete!
[304ms] WARNING: Detected supply voltage drop to 2569mV. Verify power supply is adequate.
[4490ms] WARNING: Detected supply voltage drop to 2517mV. Verify power supply is adequate.
[9605ms] SCSI ID 0 negotiated synchronous mode 5 MB/s (period 4x53 ns, offset 12 bytes)
[18785ms] SCSI ID 0 negotiated synchronous mode 5 MB/s (period 4x53 ns, offset 12 bytes)

The box also works with both, Zulu + HDD together on CS MK2 + onboard 030/882, so - what should I say?
Next best thing I could do is write an oscilloscope printout of the voltages and search for drops, multimeter says 5.05 on mainboard connector.

[chucky]

I am thinking if Termination is fine?

reason is I installed a 4060 in a friends 3000T somewhat more then 1 week ago without issues.
and the 3000T is crazy picky about termination.

Also the 3000T I installed it on have the INT2 patch done. maybe it is required?

[Chritoph]

Why would termination change when changing CPU card?

My usual SCSI chain is:
External active terminator --- A3000T Controller, not terminated --- Plextor CD-ROM, not terminated --- ZuluSCSI, terminated,

in video above:
External active terminator --- A3000T Controller, not terminated --- Plextor CD-ROM, not terminated --- IBM HDD, terminated.

Sounds correct so far. With CS MK2 I can also run it w/o the external terminator, does not seem to matter much.
If you think it helps I can try the TF again w/o active external terminator or try a shorter cable with IBM HDD alone,
but I do not see how the CPU board would have any influence there.

I also do not believe the PSU is a major issue.
CPLDIcy in top-most Zorro slot reports 4.83V, multimeter telling 5.05V measured on mainboard power connector sounds fine.

I use an cable from a socket between U350 / CIA and mainboard, running a wire from pin 21 to the int2 header of the TF,
that should do the job I guess. I would not wonder if the signal quality is not great, just because of the connectors in between.
Another suspicion would be DMAC 2 + RAMSEY 4 + WD 04 and their infamous tolerances.
Do all other A3000Ts in the test group use onboard SCSI?
With which DMAC + RAMSEY + WD?

[chucky]

termination shold not change. but as SCSI and termination is damn close to magic you never kow :-D

My friends A3000T have Dmac2, ramsey 4 and the 00-04PROTO SCSI Chip. and booted fine from a mechinacal drive.

motherboard termination and to that disk.. WIDE SCSI with Active termination on all bits.

he is sadly "out of town" now but I asked him to test newest firmware when he gets back home.