Storm Column

Well shit. Phoebe ordering window was open what, 5 hours? And you’ve managed to fill the list up.

This time I cut it a bit short, I’d rather have a few spares than explain to dozen people why I have nothing to sell to them. But still, this is way more devices than I planned to build, I figured I will take it slow and the rest will just trickle out. Now I have to hurry with everything so that previous estimate of 4 weeks? Make it 6.

I’ll start replying to emails tomorrow. I don’t want to look at my inbox now.

EDIT: I get asked this a lot so: Yes, there will be another batch. Hard to say when because this is an unprecedented interest in Saturn ODEs. Had you guys (any gals?) been here last time I made Phoebes I would have ordered more PCBs and parts now. And planned things differently.

Also, replying to so many emails takes time, I can’t do this all day you know. Have patience. I can already tell there are some duplicates so let me remind you the one device per batch per person rule is still in effect. On the bright side there should be a bunch of leftovers.

Don’t Think Twice

Last 2-3 weeks were quite hectic for me but situation has somewhat improved now. Hopefully it’ll get back to normal in a few days.

So, I’ll open Phoebe orders this weekend. Saturn devices are not in such high demand as GDEMU is – which means I can manage assembly even if I don’t have a lot of free time. I expect to finish that run in about 4 weeks and then I’ll start with GDEMU again.

Note that Phoebes are made in batches by type so your order might get processed out of order if I have (or don’t have) what you need at the moment.

As for Rhea, I will try to make some during GDEMU run, assuming it doesn’t turn into a disaster like the previous one did. I’ll announce when/if I have something for sale.

Ain’t No Rest For The Wicked

You keep asking so here’s PCE status: In progress.

I know it doesn’t look like it from the outside but I am working on it. Most of the Saturn code changes I’ve introduced lately are actually meant for PCE. It’s quite difficult to do a proper low-level optical drive emulation. For example dumps usually don’t have track 1 pregap or leadout data in them – it has to be spoofed. Missing sector headers, EDC/CRC data has to be regenerated on-the-fly. Copy protection has to be spoofed. Not all dumps have subs so basically the whole system must be able to do without them, whereas it should actually be based on subs only. Then there are hardware limitations, I want to avoid PLD/FPGA devices as these make things more expensive and difficult to upgrade – so I have to resort to some tricks that aren’t always easy to implement or timed perfectly. And don’t get me started on hardware bugs, both in the stuff I use and in the consoles.

Case in point: After some tinkering Rhea has only one code path now that handles both audio and data sectors, which actually requires more processing but the result is closer to what a real drive does. This, coupled with some changes in the internal logic and status reporting made Bug! work. New FW will be available once I test things a bit more.

For PCE I’m going to go even further and limit supported image types, subs will be required. Now that I have one less thing to worry about I can focus more on hardware design again. I will post some photos but only once I have something booting at least a few games.

Speaking of Rhea, I’ve sold all the leftovers – there really weren’t many so it all happened pretty quickly. I’m planning to start Phoebe run next, probably near the end of this month. Then Rhea or GDEMU, depending which will look like less work. As usual watch the blog for announcements.

More Australian Winter

GDEMU list is finally done. Not everything is shipped yet but I have all devices ready. I will now go through the backup list with any leftovers, that shouldn’t take more than a week.

As for the next odering window – it’s still way too hot to do any serious assembly so I can’t quite give you any dates right now but as soon as the weather gets more reasonable I’ll make a new post here. In the meantime I have a few unsold Rheas it seems so I’ll briefly open ordering this weekend to find them a new home.

Don’t be surprised to see a newer FW in your device than what’s available on the download page. There isn’t any new functionality (well, officially :P) but there was a considerable code refactoring/cleanup done. I also moved to CMSIS 5.0 and will now try gcc-arm 5.4 so there might be another “empty” version bump or two. I’m only saying this in case there are regressions, otherwise there’s nothing you need to worry about.

Australian Winter

GDEMU orders are 50% done and the rest is at least half-assembled. But it’s getting too hot to work in the afternoons as much as I was until now so… still 2-3 weeks, at best, before it’s all moped up. And then I’m probably getting a break.

From now on I will again change the shipping procedures a bit, I have a box of GDEMUs by now that just sit there and wait for payment time-out so that I can sell them to the next person. I’ve said this many times but I will repeat it again: this far in the queue it becomes the number 1 issue that causes further delays. Plus it really kills any motivation to make even more devices. So now I will ask for payment before your device is actually ready (but will be in few days) in order to clear the stock. This approach has a downside – if people actually pay quickly I will end up short until I can catch up so there might be up to a week of waiting between your payment and actual shipping. Keep in mind though that I process the payments in the order of receiving them so if you pay right away chances are you will get next day shipping.

Too Quiet

The good news: I’m going to finish assembly of all the outstanding Rheas this week, though actual shipping might happen after Sunday. Testing, packing, processing payments – these things take time too. Quite a lot actually.

The bad news: GDEMUs are only about 25% done. It’ll take weeks to wrap it up – and this is exactly why I dislike doing long runs, but the whole gmail blunder left me with little choice. I patiently await the next disaster to strike, this would be the most opportune moment to make this mess even worse.

Also, as of now I’m going to silently ignore any questions about the devices being ready yet, next planned batches, and your order status. If it feels like me being a passive-aggressive jerk here then it’s probably the case but I’m tired of having to answer those same questions again and again on top of all the work I have to do. If you want to respond in kind and rage quit / cancel your order, please do. This will actually shorten the queue and speed things up so you’re welcome.

One more thing, since a few months ago some PayPal payments are coming without shipping address and I have to ask for it via email. Seems like it happens mostly when you use the mobile app so if possible please try the usual web interface instead. It’s not a big deal now but if your PP email is on hotmail I might not have an easy way of contacting you if that happens.

Double Cross

You won’t like this but unfortunately I can already tell there will be delays, especially on GDEMU orders. Basically there’s so many that shipping takes a lot more of my free time. So much in fact that I can’t do any assembly on the same day I’m readying the envelopes – which means I’m going to fall behind. Also, very sorry but the one-week-to-pay rule is now very ruthlessly enforced, and I won’t be making any exceptions this time unless you email me first and explain your situation properly. It’s your job to check your email/PayPal to make sure you won’t miss my payment request.

In other news, I got a couple of FPGAs from a different supplier that offered a better price:

Fake FPGA

Those were supposed to be Cyclone II devices, except as you can see the marking clearly says Cyclone III. Except the chip name is EP2C5. Also, and this is not so obvious unless you take a hi-res photo at a certain angle, the laser-made markings are not very uniform and look like were done in two different steps. The cyclone logo is off too, twice to thick compared to original, although again you can’t tell unless you have the real Cyclone next to it to compare.

So, clearly a counterfeit chip. I do wonder if it’s even actual “chip” or just a piece of plastic – I’m going to experiment a bit on those. Well, so much for sourcing cheaper parts🙂