Posts by Jens

In that case, please make more tests after re-capping the original PSU. We did have negative reports with Gotek drives ever since the ACA500 came out in 2013, and I never had the chance to investigate that. If you have some electronics knowledge, it may be worth a try to insert an inductor into the supply line of the Gotek drive, so pulsing power consumption does not fire back into the host system. That's just a shot in the dark, but the best "educated guess" I can provide.

Please also understand that official product support ends here - you have successfully shown that our products work fine with the A500. Only if an extra component by a third vendor is added, you have trouble. I'd like to hear what the vendor of the Gotek drive has to say, especially about power consumption and power filtering.

Ideally, you follow the IPC reflow profile: Roughly 3 minutes of pre-heating at 150 deg C, then ramp up to 260 deg. air temperature and make sure that you don't expose the board to this temperature for more than two minutes. I just keep lifting on the chip with tweezers while I apply hot air, and I move the hot air away as soon as I have the chip off.

Don't force cooling, but let it cool in free air. You should solder the new chip on with a normal soldering iron, not with hot air, as that means less stress for the materials.

Comos wrote:

Does the board design allow to have a option for this in the future fw release?

Firmware cannot change the behaviour of a hardwired circuit.

The first C64 baords had the reset line of the IEC wired directly to the IEC circuit of the C64. This made reset-buttons on the IEC port possible. However, with many devices connected to the IEC at the same time (up to four floppies, a printer and a plotter), the reset line was pulled up my each of these devices and the computer's internal reset circuit had trouble dealing with the load.

Starting with the second revision of the C64 main board, Commodore introduced an open-collector driver "from computer to IEC" which was strong enough to deal with a higher number of external devices.

No Commodore computer has ever had an open reset line on the IEC port. What you are asking for is something that was hacked by crackers back in the day - you may have forgotten about it after all these years  

The C64 Reloaded MK2 of course also implements "open collector driver on IEC", and if you want to apply the hack of "reset C64 but not IEC devices", you'll have to remove resistor R80, which is the base current-limiting resistor for transistor Q13 that pulls low on the IEC-reset line.

Kazz wrote:

Anything special I should be concerned about while changing that chip?

Remeber that the board is produced RoHS compliant, so the melting point of the solder is higher than you may be used to from other hardware that has been produced pre-2006. It may take awfully long to melt the solder with hot air, but that should not make you impatient - it's tempting to increase hot air temperature, but that may have negative effects on the glue that holds the copper on the FR4 material (i.e.: pads rip off if you pull on the chip too hard).

In short: Take your time, don't apply force. Tiny tools, very little force.

KKR75 wrote:

I am using an RT-50B

That's no good for the Amiga. I am aware that there are commercial sellers that believe this is OK, but we've had multiple support cases where this PSU chassis has been found to be the root source of the problem.

KKR75 wrote:

If I remove the ACA500plus completely, I get a full green screen on boot. If I remove the external gotek it all works fine.

Looks like the power consumption of that thing is already a problem for the MeanWell chassis.

KKR75 wrote:

Could the ACE2 be the source of any issues, since a green boot is an Agnus or RAM related issue usually ?

I doubt that - since your computer works normally without Gotek but with ACA500plus, you know that ACE2 is not the problem. Trust me, the MeanWell PSU is your problem - please retire that.

Ah - so not a user port cartridge, but an expansion port cartridge. This makes Peter's suggestion for exchanging the HCT139 (U15) more likely. That's a standard part, just in SO16 package. I guess if you can find a shop that still sells TTLs, they may even have the SMD version in stock.

These diagnostic carts may be wrong - as long as they don't show precisely what went wrong, it might as well be a software error.

Yes, a schematic would help - I found the Eagle schematic file, but I dont have Eagle installed on any machine (I don't use paint programs, I prefer real CAD software). Can someone convert it to PDF?

Kazz wrote:

untested link232

That might give more hints - was it hot-inserted? Did you do any repairs on that so it works now? Whatever you repaired on that unit might give us a hint about what happened to the CIA chip.

link232 is a user port device, right? Which signals does it use?

gully666UK wrote:

Any mass storage hints would be fantastic.

If you can wait until Xmas, we'll have a new top-of-the-line accelerator for the A1200 that also works on the ACA500plus - and it comes with a micro-SD card slot that you can boot from. Speed will be comparable to CF cards on the ACA500plus.

Other than that, there is no "clean" solution. There's the new Checkmate 1500 case, but that has lots of issues with it's slot-adapter board: Neither ACA500plus, nor known-good Zorro cards work reliably. I have offered help; see other thread in this forum. If I'm successful debugging that board, this may be a solution for you, although you still would not get the "exchange data with FAT-formatted CF cards" back. A Buddha may work in that (well, if I find the fault, but I'm confident I will).

For data exchange, I'd go for the X-Surf-500 - that's more convenient anyway. Mounting network drives on the Amiga just rocks I fear that you have lost the "data exchange with FAT-formatted CF card" for good.

Luckily, all updates are provided as ADF images, so even without the CF card slots, you can execute all updates by unpacking the UPD file to a floppy drive or floppy emulator.

BTW, dry food for that cat for the next six weeks

Solder pads are ripped off - I guess the card will come up normaly after making sure that there are no shorts on data lines.

Repairing this will take hours and cost much more than a new card. If you can make it start and boot from floppy, I can give a hint or two for adding mass-storage by other means.

ninex wrote:

Well i can hook laboratory class regulated power supply and finetune it to match baseline parameters.

That's of course a great opportunity. If you have access to that kind of equipment, you surely also understand that the input coil of the A1200 plays a bug role, as voltage drop on both the GND and 5V lines add to the equation: You should measure voltage behind that input coil to be sure that at least 4,85V reach the accelerator.

ninex wrote:

Well if power is an problem then in theory amiga should be unstable in heavy load.
That test is passed.

The accelerator pulls low on the IOR/IOW lines, which are pulled high by the NMOS drivers of the Gayle chip. This appears to be voltage-sensitive.

ninex wrote:

Well if power is an problem then in theory amiga should be unstable in heavy load.

The Amiga chipset is very forgiving if it comes to over/under voltage. Also, the logic on the ACA1233n runs off a local 3.3V regulator, so the input voltage being a bit off doesn't cause big trouble. It just seems that some CF cards (especially on unbuffered interfaces) don't like the not-completely-0V-level of the fast IO lines if controlled externally by the ACA1233n.

Please use the latest ACAtool, not the old software. When ACAinfo was written, the new cards with IDE-speeders didn't exist, so it's expected that it does not detect the card(s) correctly.

ACAtool comes with a GUI, and it also supports the old ACA1233n cards.

The Catweasel MK4plus was not Amiga hardware; it's a pure PCI device. There are two showstoppers for a new production run:

1) we could not get Windows 64-bit drivers to work, even with the help of companies with full Microsoft support

2) the PCI bridge that I chose back then is not available any more.

Further, the FPGA on the unit is a really old type, an ACEX1k30. These aren't made any more either. A full re-design would be required, and this would require quite a number of orders to make it profitable. I don't believe that PCI cards can sell in the required quantities these days.

I did think of re-making the Catweael MK2, as that's fairly easily doable. I remember that AmigaKit used to make special cables, so the Catweasel MK2 could be used on the IDE port of their PPC boards. This would solve the floppy-part of the equation.

Joysticks and Amiga keyboards should probably be extracted into a separate (USB-) product.

FirstNinja wrote:

Gerber files, a list of components and any firmware

Hardly an option, as this would include rights, which we require for the next incarnation of the Catweasel.

Bad Power supply? We've had multiple cases of "second-source power supply messes things up" already.

Try moving the CIA chips in the sockets sideways while the lever is closed (lever down=close). This will solve contact-problems that the chips may have. Also, please do the same with the CPU, as that's the chip on the receiving end of the IRQ signal: The CIAs may produce the IRQ signal correctly, but if the CPU dies not see it, the cursor will not blink either.

Note that every single board has passed QC here before it left. This includes loading software from disk, initiated with a keyboard. Your board would not have the warranty ID if this did not work.

The 4.5A version is indeed preferred. However, without precise measurements, I can't give a definitive go/no-go for the 3A version. I personally test with a 3A unit (labelled A600, not A300), and that test/development machine uses an Indivision AGA MK2, Randy-ROM, Lyra 2 and a Micromys by default (machine is open all the time). The new versions of the ACA1233n (including the 55MHz version) were developed on that setup.

Re-capping is always a good idea, but please make sure to either meet or exceed the voltage and temperature rating of the caps.

There's no direct connection between a drive on the floppy connector and the CPU/RAM subsystem. The only connection I see is power supply; what kind of PSU are you using?

Rockford wrote:

1. I've bought the card from previous user not from any shop. I asked him about any invoice. He said that the card number is enough to have any warranty help because he had serviced different hardware in the past and the serial number was needed only. If it is not true I'll contact the seller to have any confirmation of purchase.

It's just a matter of having all legal documents - I do know who the card was sold to, but unless you have in writing that the card and it's warranty have been signed over to you, I'd have to refuse any warranty claim unless the original owner is actually claiming it. How else would I know that this card is not stolen? Sorry for the red tape, but with parcels getting lost in DHL's deep dungeons every now and then, I have to assume that there are units on the grey market.

Rockford wrote:

Maybe there is an error in the system?

"system" applies to the whole setup that you have there, and includes the software. The funny WB tool that you have there is obviously programmed by someone who never tested it with larger amounts of memory. Just like Sysinfo: It obviously fails to show the actual size due to the large amount. However, it still displays the correct start- and end address, confirming that all memory is there.

In other words: Look for the faulty program in your installation, but don't look for the fault in the hardware - that's perfectly OK.

Rockford wrote:

The seller said that it is in a warranty period.

For the warranty to apply to you, the seller needs to sign it over - just in case there's really something wrong with it.

Rockford wrote:

and only 11 MB of RAM

Please show where that memory is actually located - it sounds like the Z2 compatibility mode is activated, which only adds a total of 9M fastmem. Together with your 2M of chip ram, this would be 11MBytes - but that's just a shot in the dark. Please use "showconfig" from the original WB, Sysinfo or similar programs that list the available memory nodes.

If AmigaKit does not respond to your support requests, you better return the card for a refund (right of revocation) and choose your source for Amiga stuff more carefully for the next purchase. We're here to help, but when it comes to direct service (=you ship directly back to us), you need to be a direct customer.