tep392

Ha Ha. I just noticed that you photoshopped the power supplier and power light on the first photo. You are a strange character johannesmutlu.

I guess I have a high annoyance threshold.

Of course I’m sure. Download the file I posted in the other thread and run in emulation if you want to verify.

It's really not that hard. I took apart my 400 when I was 15 and had never seen the inside of a computer before. I didn't have any instructions or Youtube videos to guide me. If you can handle a screw driver then you should be able to do it. Just follow any Youtube video that goes thru the process. Have some confidence in yourself.

The FDC doesn't provide any "enable" signal for the stepper. Just a step pulse and a direction signal. These signals are sent to the PC floppy drive which has additional logic and drivers that generate voltage sequence for the A and B phases of the stepper motor to advance it. The Tandon uses flip-flops that turn A and B on/off with the appropriate phase to step in the desired direction. This simple logic always leaves power applied to the stepper coils. Some additional logic would be needed to decide when both stepper coils can be de-energized or re-energized. Maybe the spindle motor on/off signal or drive enable? The Atari 1050 drives the stepper using the CPU and RIOT chip, so there's more smarts available to make the energize/de-energize decision. The phasing of A/B is controlled by software, rather than logic gates.

Then just accept that you will never play Mario Bros. on your 400.

The MPI that came with mine was spinning the disk but wasn't clamping down hard enough so the disk was slipping. The cause was the metal lever inside the mechanism that pushed down on the clamp was bent. When I tried to bend it back,it broke. A cheap casting that failed from fatigue. There are two broken mechanical parts in the MPI, so it is now a parts drive.

The Tandons look nearly identical to the MPI drive that was originally used. Problem is that the PC configured Tandons don't power down the stepper motor when idle, so they are generating a bunch of heat. With two Tandon drives, the unit still warms up substantially over time, and pulling warm disks out of the drive is a pet peeve of mine. These disks are old enough that I don't want any further degradation from running hot. The TEAC's I'm using now don't have that problem. I decided to do a quick check on my 1050 drive with Tandon mech, and it looks like Atari was smart enough to de-energize the stepper when idle. These are the TEAC drives. TEAC's are plentiful on Ebay, but I had to look a bit to find them in black. Most are beige.

Here is the .car file that will work on a 400/800 with 32K. Any flash cart should be compatible with your 400, but main thing to check is that cart is not too tall to close the lid. My Ultimate Cart fits fine, but I don't know if they are available anymore. Maybe someone can reply with some currently available carts. I am actually looking into building a few carts of this version, if there is any interest. MarioBrosXE_400_32k.car

Tried it but it doesn't work. I'm not surprised either because the Glen 5200 conversion is an XEX file. The entire thing (21K) is loaded into RAM. But then I ask why you wouldn't just run the hacked XE version that I just tested on a 32K machine and works. You could run it from any of the various multi-carts that are available. (edit: not Atarimax Maxflash cart. doesn't support .car format)

Very interesting. I just hacked the Mario Bros XE image to replace all the 9FFC 00's with 01's. It seems to be running just fine now on my 800 with 32k of RAM installed. I'm using the cart image in my Ultimate Cart. Sadly, I don't think Johannesmutlu will ever be brave enough to open up his 400 and install a 32K RAM card, so he is going to miss out on running Mario Bros XE on his 400. edit: FYI for Johannesmutlu. 32K RAM cards can be installed in the 400 without any soldering required. Just pull out the 16K card and replace with 32K card.

It was a necessity in order to run two drives in that enclosure. I played around with a small internal fan to force airflow, but it was noisy and not very effective. For my final configuration I replaced the tandon drives with two TEAC FD55-BR double sided, double density drives. I just detest heat in my floppy drives, and the two Tandons still generated a fair amount of heat, mostly from the stepper motors which are always drawing power. Newer drives like the TEAC are very energy efficient, generating little heat and are quite reliable. My ebay Tandon purchases all needed repair to get running. So I don't waste the Tandons, I installed one in an older PC, along with a Quad density TEAC drive, so that I can create floppies of various formats for my retro computers.

I don't care either way. I just posted the question because of what was said on Facebook to try and clarify the situation.

Yes, I saw the post, which is why I asked about the photo's on Atarimania. People assumed a scam because the photo's are also on Atarimania. That's why the mod locked the post.

@www.atarimania.com These same pictures are on your site. Can you confirm if they were just added and came from the poster of this topic?

This might be my last update on this drive. I've been using the printer interface on the AT88-S1PD to control my 825 printer and it has been working well. I had to make an adapter cable to connect the printer's 34 pin connector to the Atari CX86 printer cable. I thought I would post how I did it for future reference. I purchased a DA15 male/female adapters from Amazon that used a ribbon cable and cut the male end off. I then made a few specific cuts to the ribbon cable and connected it to a 40 pin connector that plugs into the AT88 interface. I discovered one difference between the Percom interface and the 850 that had to be accounted for. The Fault signal is active high on the Percom and active low on the 850. The CX86 cable connects this signal to printer pin28, Logic 1 5V. I could have just jumpered the Fault pin on the interface side to one of the unused signal grounds but I decided to rewire the printer end connector to move this signal to pin 23 which is 0V when the printer is powered on. I did this so that the fault logic would activate if there is no printer connected or the printer is powered off.

You are still quite confused. First off, stop associating the size of banks with the size of RAM. The two things have nothing to do with each other. There's no benefit to making bank size equal RAM size. Also, bank switching is a hardware mechanism. There is no such thing as software bank switching. The entire 64K XE Mario Brothers is not copied into XEGS RAM. There is certainly a portion of it that runs directly out of ROM. If you think about the 7800 which has games that were released up to 144k and that console only has 4k of RAM. There is very little of that code that get's copied to RAM. By the way, did you know the 5200 can map 32K of ROM without bank switching? The entire 32K is visible to the CPU at all times. That is actually an advantage over the computer line which only maps 16K to the address space.

Why do you think that everything needs to be loaded into RAM? The CPU can run code out of ROM or RAM. How do you think it’s possible for a system like the 7800 to run 128k ROM carts with only 4k of RAM? You seem to not understand some fundamental concepts of computer architecture.

5200 version is not 16k. It’s a 32k cart with 20k of data.

Boards for a bare bones Altair mounted in the backplane. I still need to order power supplies and get everything mounted in the case. I'm getting close!

I bought a NOS 825 printer, which ended up needing repair because the ribbon drive motor would go backward. After designing and printing a 3D part to fix it, it's now working, even with the original 40 year old ribbon. I don't know if you still need this, but I printed out all the characters in every font and scanned at 1200 dpi. I also did a set of 10 CPI with underline to show the 9th pin in use. It's a bit hefty at 42mb.

Eleven years newer than I am.

Great read! I got thru most of the IMSAI prop stuff but ran out of time to continue reading. Interesting that the IMSAI was just doing the flashy lights and the monitor/keyboard was connected to a different computer. I feel like I read about that fact a while back, but my memory isn't what it used to be.

I didn't say that I don't want to do it for the 2600. I'm just guessing it might be more difficult to fit header data with the space constraints of 2600 carts. I don't even know if the 2600 has an existing header format. If not, it might not have to be as lengthy as the 7800 is. If 2600 programmer are willing to develop a header and can fit it in their games, then they should go for it.

My suggestion is keep the stella.pro process for existing carts, future 2600, and add the option to check the 7800 dump for the header data so that it can configure the emulator without need of the stella.pro database. I would guess that initial updates will be limited to the database, but if there is a larger update to the 7800 emulator, to fix emulation issues, then that would be the time to add A78 header support.