A short update about this topic:

Quote:

The reason for this is that NMOS and HMOS circuits do not have push/pull FETs on their output pads, but an always-on-pullup-FET.

On a German forum I posted measurements and a die photo which shows that this is not true. And that a quite large current is needed to pull down A15 w/o using DMA.

Jens told that under the condition that this current is only needed for a very short time per cycle it is very unlikely that something gets damaged. Finally he came to the conclusion "better safe than sorry" and added an idea how the concept from my paper could be further improved.

All in all this means that the content of the paper is valid and correct. It may not be the most simple way to implement it, but it is the safest so far because all components are used as "written in the books".

Jens: Yes I'm going to do it with a small CPLD this time, because unlike the EasyFlash this does not mainly aim at DIY-people. And it saves development time There are some aditional ideas already which will also take some macro cells but I'll think about the idea with the address latch nevertheless.
_________________

---
skoe.de/easyflash

Any updates? I hope this is not dead.

I was thinking about this the other day myself.. I sure hope it's not. This would really come in handy..
_________________

It's still under development. Unfortunately the first batch of PCBs contained a serious design mistake, that's why I had to order again This time I was more careful with the checks.

I decided (and asked Jim about his opinion) to merge the KERNAL cartridge and the EasyFlash 3.

As some people know I planned to develop a successor for the non-released EasyFlash 2 for long. The parts for the EasyFlash 2 were too expensive (about 40 or 50 Euro all in all) even still without USB. Then, a few month ago, Stingray told me about his idea with his AlienFlash, which looked quite impressive. So I thought there was no real reason to make an EF3. However, the KERNAL project and the experiences from the EF2 convinced me to go on with it.

I think the PCBs will arrive next week. Most of the parts are here already. As soon as something can be seen I'll post it here.

Edit: Ah, there is something. At least virtually

It fits into both types of cartridge shells which are available currently. Buttons and USB connector are on left-hand and right-hand side.
_________________

So, what's planned for EF3, multiple E.F. slots like the AlienFlash?

Yes, there will be many similarities with AF. Stingray and I realized quite a long time ago already that we had similar ideas. No surprise, some of them were born here in this forum

I don't want to talk about too much about features until they actually work. In case I botched something again...
_________________

Great news! Thanks for the continued work

Excellent, I better start saving my money to buy one when it comes out. Much appreciated for the update.
_________________

to skoe

Skoe,

Are you on target to be finished by mid next month and meet the bounty deadline?

thanks

DLH

I think so, yes.
_________________

I guess I'll have to save up as well, bundled as EF3 this is going to be something in the $100-150 area... maybe.

I guess we will when we buy the cart instead.

My response was quite short yesterday, I was quite tired already. Here is the more detailed state:

As I told already that the first batch of PCBs had a design mistake, the second one is still not here. In Germany we have a public holiday today and one one two weeks ago, so the 10 working days lead time for the PCBs are two days more than I though. I hope they'll arrive tomorrow. All of the parts I need are here already.

Our (2nd) baby who was born last fall and the design mistake in the first revision delayed the whole project a bit, but I think it's still doable on schedule.

In the meantime I took an EF2 and modified the hardware, mainly be adding a 25 MHz oscillator. This was a good starting point for the VHDL code. I wrote it from scratch instead of taking the EF2 code as base to get a more efficient (and hopefully readable) design.

The first version I release before the bounty deadline will only contain the KERNAL and the EasyFlash mode. The latter one will be used to flash the KERNAL images. A few prototypes are reserved for people who can verify the functionality. If I didn't make a mistake with the design, it will be possible for them to update the CPLD over USB later with a firmware version which has more features.

e5frog wrote:

I guess I'll have to save up as well, bundled as EF3 this is going to be something in the $100-150 area... maybe.

I will not be the one who sells them later, that's why I'm not able to say a price here. But I am quite sure that your guess is much too high. The parts were chosen carefully to provide an inexpensive BOM and a small PCB, so it's more in the spirit (and league) of EasyFlash than of 1541U etc.

Quote:

I guess we will when we buy the cart instead.

I will not get money from the carts being sold (I think - and that's okay for me), as the whole thing will be open source and somebody else will sell them. But indirectly you are right: The main bounty sponsor will hopefully earn more than the money he spent on the bounty, which could motivate him to sponsor future projects...

For me it's enough to get back some money for the cost of the parts and PCBs for the prototypes (including the screwed up 1st design). This is the case with this bounty. I don't expect to get money for the development time - in this case some zeros are missing at the bounty value anyway It's just a hobby for me.
_________________

---
skoe.de/easyflash

There it is. After I tested the basic functions with the first prototype, I soldered a few more yesterday.







The first two pics show how it fits into the icomp cartridge shells (although I will make an improvement to make it fit even better) and the cartridge shells from UK. The last one is one with shorter buttons, that one is thought not to be put into a housing.

The jumpers are only needed to update the CPLD firmware over USB. They do not need not to be touched for normal use.

There's still much to do in the next two weeks. Best I go on now.

Thomas
_________________

Looks nice Thomas!

Nice progress!

Excellent progress, Can't wait to add one or two or.... of these to my Cartridge collection.
_________________

hmmm... looking amazing
can't wait...
_________________

Looking forward to the final outcome. God Speed.
_________________

Will this be able to have multiple kernels in it?
_________________

love your work, as always Skoe! This is going to be another great device

Thanks guys for your encouragements.

In the last days (or even weeks?) I spent lots of time to test and modify the KERNAL implementation. It looked like it run stable already, but after having tested it on different board revisions - and even on different devices with the same revision - I came to the conclusion that the DMA line influences the 6510 in very weird ways sometimes, when being used for the fraction of a cycle. This may happen because DMA does not only control AEC, but also RDY.

So I decided to discard the full-blown design and to use a much simpler one. In this solution A14 is pulled down for 80 ns without having the bus in tri-state. This is a kind of a hack, but turned out to run stable on different boards. This way was also the one mentioned by Wiesel, so I think it's okay to do it like this

moiree ported and improved a tool which makes it possible to update the CPLD over USB with different firmwares. This is really useful, so more features may be added later and bugs can be fixed. It can be found here: https://bitbucket.org/skoe/easp - Thank you, moiree!

Can somebody compile it for Win32 and write a short manual how to install the needed libraries?

I didn't send the prototype boards to others yet, because I wanted to be sure that no hardware changes are needed before I do this. But today I'll send out another one.

Now I have to complete the software and the VHDL design to make it to a simple to use device with good software support in the spirit of EasyFlash.

By the way: The simplified concept should also be possible to be implemented on the first hardware I made, the one with the design mistakes. That one is a real low cost design, but has no other features than three flashable KERNALs. I'll attach a picture to this post. But currently I work on the larger version only (EasyFlash 3 + KERNAL).


Remark: This picture shows the first, small design.
_________________