GVP Impact series II for A500 not working

[rmdelzen]

Hi all, A few weeks ago the GVP SCSI interface for A500 stopped working. At cold boot it simply gave a white screen with LED (HD access) multiple blinks, then the white screen disappeared and the requester appeared asking for the WB disk. I searched for answers to this and any seemingly related problems with the GVP, and after a few dozen reads eventually ordered and soldered in a replacement SCSI interface chip (the replacement was AM33C93A-16P, not the "original" WD chip). Now on cold boot the A500 gives a white screen and no further indication of activity, and the HD LED stays dark (although in "game" mode I still have access to the RAM onboard the GVP Impact). SCSI_Inquire doesn't show any indications of gvpscsi.device. GVPinfo (on the FaaastPrep GVP install disk) displays the GVP Impact board parameters and the onboard RAM, but the GVP Install FaaastPrep193 also doesn't see gvpscsi.device.

Any suggestions/comments welcome. Thanks to one & all.

-- Rich

[rmdelzen]

Oops ... I meant ROM for my GVP Impact II is 3.07 (and the board is Rev. 3)

[alexh]

Dead HDD?

[rmdelzen]

Hi, I tried two HD, Quantum ProDrive LPS, Rodime Type E9. Both had been working before, now the A500 can't access them when either is connected. The software (GVP FAAASTPrep) running on the A500 sees the controller but can't see gvpscsi.device and nothing connected to the GVP Impact. Not sure whether both drives could have gone bad at the same time, I suppose it's possible tho.

[alexh]

Originally Posted by rmdelzen

Not sure whether both drives could have gone bad at the same time, I suppose it's possible tho.

Depends. If both have been inactive for the same amount of time. Stored in the same location.

Try soft-loading (binddriver) the SCSI driver? Just in case it is the onboard EPROM?

Might sound stupid, but the disable switch is not in the wrong position?

[Dale423]

Sounds like could be a bad autoboot ROM.

On a side note, if you find that it is something more serious and this unit is dead, please let me know if you would be interested in selling the plastic top cover . I am in need of one.

[rmdelzen]

Disable switch is set properly. Haven't checked the ROM (not really sure exactly how I can check the ROM).
Not selling anything (yet...)

[Matt_H]

Might be worth re-seating the ROM. Between age, heat, minor oxidation, it might have lost full electrical contact with the socket and just needs to be popped back into place.

[rmdelzen]

Good suggestion! Tried reseating the ROM in the GV Pboard after doing ISP wipe of pins, but sadly no change.

[alexh]

Flip the switch to non-autoboot. Boot from a floppy disk containing Workbench and the binddrivers file in the SYS:Expansion dir.

You can get the binddrivers version of the driver from HERE

http://babel.de/download/gvpscsi4.tar.bz2

This should automatically softload the ROM. See if you can access your HDD?

I'm not an expert in this area, you may need to mount your drive after using binddrivers .

A real Amiga Workbench disk *should* have binddrivers in the startup-sequence.

[rmdelzen]

I downloaded the driver for the GVP Impact per your suggestion (for which, alexh, many thanks!) and copied the gvpscsi file to ENV: of my Workbench disk. After a cold boot, same problems: Yellow LED signifying HD access remains dark, no HD access is possible. Doing "version gvpscsi" gets me 4.15, as appropriate for the newer version. I'm suspecting a hardware glitch but haven't the slightest idea where to start looking. I'll go to my soldering job on the AM33C93 and look for cold (or no) solder on the connections.

[rmdelzen]

Resoldered a few suspect pins on the Am33C93 but no improvement.

[Dale423]

That is strange. I am not sure what the issue with it is.

[alexh]

See if you can confirm the SCSI disks as working. I once had several break at once because I foolishly put them next to a magnet source (speaker)

[rmdelzen]

Haven't found a kludge to verify that my SCSI HD are OK, still looking at the GVP SCSI interface and am ordering a replacement for the AM33C93A in case I fried it while soldering it in. Maddening since GVP software can see the board and the onboard RAM is accessible...

[Dale423]

Your Bio says St Charles, so not sure if that is Illinois or Missouri maybe? I so I am in Northwest Arkansas and would be willing to help you test the drives. I have several Amiga's with SCSI.

[rmdelzen]

Hi Dale423, I could mail you one of the scsi drives if you'd be willing to test it on your system. Please let me know. Thx in advance!

[thebajaguy]

Troubleshooting 101:
1. The GVP HD8+ is an AutoConfig device. Any RAM in it will produce 2/4/8MB RAM expansion. Since the SCSI controller and memory are all integrated on one chip you should see, with the autoboot ROM disabled (no short on pins) the expansion board(s) with a tool that shows expansion boards. The GVPInfo tool on the support disk does this, as does SysInfo and ShowConfig. If you are not getting AutoConfig of 1-2 boards, there may be edge connector damage (HD8 or Amiga), or the HD8 controller is dead. You can find the disk on gvp-m.com
2. Short the J2 jumper (switch connects here) to enable the ROM. This will produce a longer boot delay at power on/reset before the floppy is accessed to boot if there is no HD that is accessible/ready on the SCSI bus. If you know #1 is valid, and #2 is not producing the added boot delay, the ROM could be bad. Rare.
3. If you get an extended boot delay, the HDD LED may be faintly blipping if you look closely. This points to a dead HD, or could mean a bad 12V line on the HD8 PSU. No power to a spinning disk's 12V (but has 5V for the GVP-HD8 and HD) will easily produce this. Check the input voltages on the HD's 4-point power connector. 5V and 12V are on the outsides. Ground is the inner pair. The HD8 fan is usually 12V - if it's properly connected and not spinning, that may also indicate a PSU issue.
4. If all power is good, and the HD LED has to those tiny blips, the HD is suspect failed.

[rmdelzen]

Hi bajaguy, Many thanks for the tips! I noticed in the process of replacing the AM33A93C I damaged a trace (or two, can't tell...). Any chance of finding a circuit schematic so I could verify all paths between pins and where they're supposed to go? Thanks again for your help!
Rich

[thebajaguy]

As I replied in 1:1, the 33C93A is mostly mapped directly back to the DPRC - almost every Series II card will be wired the same.

I'm on a work trip for the next 2 weeks, but I can try to trace out the suspect pins against another A500-HD8 I have. Even someone with an A2000-HC8 can do the same for a few pins with a continuity meter. Post the suspect pins. Pin 1 on the surface mount 33C93A is designated by the dot in the center of the top side above the part identification text.