ATX 350W PSU
Please be careful with ATX: These have mostly the 12V rail as their main regulation rail, which puts the 5V rail in kind-of-a random regulation status if the 12V rail is nearly-unloaded like it's the case with Amiga computers. You may be lucky with the specific type you have there, but the general rule of thumb is that the delivered power on 5V must be higher than on the 12V rail, and that's not the case for most of today's ATX PSUs.
There may be two reasons why this does not work; either the memory where the disks are being unpacked to is not OK, or the flash contents got corrupted somehow. If you believe that memory is not OK, then launch the full 3.1 config and run a memory test (there are numerous tests on Aminet).
For the flash contents, I have created a repair archive here: http://wiki.icomp.de/wiki/ACA5…are#OS_installation_files
That said, the plan has always been to make a DC-DC converter, so a modern switchmode PSU with high efficiency can be used. Just like the Commodore power supplies that we have in the shop now, we can have these made by a larger company and get all the required certifications and paperwork for our product liability insurance. On the low-voltage DC side, there is only CE/FCC to observe, and that's daily business for us.
The acceptance of such a unit would still be limited, as you'd have a two-box-setup: A laptop-like AC-DC power supply and another box that contains the DC-DC converter. Further, there's another weak part inside the Amiga, and that's the current-compensated coil at the power-entry point. Here's where another voltage drop happens, so yet another connector would have to be introduced, likely reducing the acceptance of the solution even more.
Then there's the problem of getting the 5-pin square DIN male connectors. None of the connectors that I have found are RoHS-compliant, because they are "new old stock" from the 1990s. Since the sources that I have talked to a few years ago are depleted now, my guess is that these non-RoHS parts have now changed hands to those who are offering those "new" power supplies. If we wanted to get that fully legal, we'd have to make new toolings to have new, RoHS-compliant 5-pin square DIN connectors.
However, I'd still be hesitating to spend that kind of money on tooling, because the 5-pin female square DIN connector on the computer side is rated for 2A of current only. Commodore gave it a stretch from the very beginning with it's 2.5A rated PSUs for the first A500 computers and the C128 (which is also 2.5A rated on the 5V rail), and then continued to stretch it up to 4.5A, bearing the danger of the connector burning with a high-powered setup that loads the 5V rail excessively (for example with a B1260 or Apollo 1240, which is even worse).
So if you look at this from an engineering point of view, there is only one way to get things absolutely right, and that's to make a circuit board that replaces the 5-pin DIN connector AND the current-compensated coil that's located right behind it on the A500/A600/A1200 boards. This would probably have yet lower customer acceptance, as it would be an expert-only-install.
However, it would eliminate the weak point of the connector, the voltage drop on the input coil would not have to be accounted for, and the outer appearance would be even nicer, as a new laptop-like PSU is much smaller than the old bricks. The DC-side cable would be thinner because it only has one core plus GND/Shield, and I bet the overall efficiency would be higher than with the old bricks, so you'd even save money on your electricity bill.
Still working on that flash issue, but we're making progress. My next step is to investigate the Indivision ECS V2 flash in an A600 setup. Already spoke to a customer on the phone, so I can replicate his setup exactly like his and hope that I can really brick the unit, so I can see what's going on. That'll be after easter, though. The A500 has now moved to the table where the new Samsung monitor is, so it's closer to where I will do the ghosting test. Getting there!
I'd use one of the old 4.5A units from an A500 and re-cap them.
Thanks for the report! We'll look into this and develop a fix. For a test case, this is a bit annoying to load, but we may be able to isolate it into a separate routine. Not having watched the demo yet, I assume that this screen uses dual-playfield and sprites, which may be a seldomly-used combination.
The image contains copyrighted software, especially the OS isntall disks, which are tied to the Buddha. We can't make that available for download, and especially if you have purchased from a reseller, we can't do direct service (as already mentioned via e-Mail). With the solution you've mentioned in your eMail, let's just wait until you get the chance to try that out.
The X-Surf-100 does not make use of the Zorro memory space, but it does use 64k Z2 autoconfig space at $e8.0000 (plus a bit). Where did you hear about problems? Just asking, because I don't know of any trobule with mem/HD controllers.
Do you see that the disks are being unpacked? You should see a percentage for each disk, shown on both the screen and the DisMo. If that fails, you may have a corrupted flash.
I haven't tried any of the existing solutions, so I can't confirm if they work or not. Since the frequencies from Indivision ECS V1 are not completely standard, you may need some fiddling with the config tool to get it to work.
That's the ACA500plus exception handler. So the CPU works, but obviously does not get all instructions right.
Try to pull back the card by a fraction of an inch, maybe clean the ACA500plus side with IPA and micro fibre cloth. It may have collected fine dust fromt he environment, and moving the accelerator back and forth on the connector may not be enough to remove that.
You may have one of those motherboards that does not have a chamfered edge on the CPU connector. In that case, pull the card back by 1mm - this will make the connection much more stable.
Is there a chance that with faster accelerator Buddha will start to recognise CF cards?
Maybe yes, maybe no. Once again_ Buddha is an IDE controller, not a CF controller. It implements the IDE bus and protocol, and I suggest to try an IDE harddrive first before you jump to conclusions.
Is there a good place to measure the 5 v om the aca500+ or aca1233n to see if there is enough voltage?
The "bottom left" and "top right" pin of the unpopulated oscillator place for the FPU are a good place for measuring the supply voltage. It should be 5V.
Since your card has been tested successfully here, you may need to dig deeper; you may have a ground loop or some other grounding problem with your computer. Try to remove most of the connections you have with your computer, especially Audio connections, as these are most prone to ground loops.
Is the wiring in your house 2-wire or 3-wire (with separate ground)? And what about the PSU, is that grounded? Do you have EMI sources such as a mobile phone or a wireless router near the computer(s)?
so, is this PSU safe to power an accelerated A500? I would have thought having a cable 1 metre in length wouldn't be an issue regarding voltage drop, but Pavel knows nothing about it which is worrying because it was me who recommended the PSU to my mate.
WIth Pavel attempting to put things "factual", but knowing nothing about electric engineering (just a copy&past post about latch-up can't make anyone believe he's up to the task), I'd say that the PSUs offered on the market are not safe to use from an engineering point of view.
From a user point of view, they may be OK in terms of "computer works for some time with them". If you trust that the supplier will help you if the unit damages your equipment, then feel free to use it, but I won't recommend them. The original Commodore PSUs had a regulation circuit that attempts to compensate for cable drop by increasing the voltage under heavy load, but the MeanWell chassis that those aftermarket PSUs are using do not have such a circuit. So essentially, while the chassis may be up to the task, the whole PSU as a product is not.
I guess it's time we take a look at your config - please save the config and upload it here, so we can attempt to replicate the issue here.
The only true difference between the current and the previous model of the ACA500plus is that I'Ve put a 10k pull-up resistor on the reset line, which cleans up the rising edge of the reset signal - this is necessary for some, but not all A500s. You could try that on your machine.
The difference between the connectors may not be the spring force, but some grease that has built up over the years. Please try to clean the connector with IPA and a an old toothbrush, and also use micro fibre cloth and IPA for cleaning the A500 side.
We can't check from this end - only the sender of a parcel can start an official inquiry.
Customs clearance finally happened. Thanks to DHL Express (errmm. not really) for the.. well.. service. The CPUs are supposedly out with the delivery courier now, scheduled to arrive today. Keep your fingers crossed.
If you come to Revision this weekend, I'll have a few cards with me - even cheaper ones that aren't in the shop, for example ACA1233n-26 with 68ec030 CPU (no MMU).
Further, we just got a shipment from the wave solder company this morning, and I have given instruction to put the 40MHz cards through QC first, so I'll have a few 40MHz cards with me at Revision. Everything that I'm taking back from Revision will of course be put into the shop.
We do NOT take reservations. It's first-come-first-serve.
While the shop does have a notification list that you can just add yourself to by ticking the extra box at the top of every product if you're logged in, I do not use the function due to the extremely low quantities that the ACA1233n is available right now (single-digit).
