Gauging interest in Atari based Eprom Burner

[Dropcheck]

what is the biggest size eprom that this will burn? Thanks

 

I have successfully burned a 27C512 (512kbit 64KBX8) Eprom chip

[Dropcheck]

It looks great, but as an Atari 800 user, what I'd really like to see is a purchasable re-implementation of the hard-to-find product known as Aprom:

 

http://www.atarimagazines.com/v3n4/productreviews.html

 

Looks like the burner part is somewhat limited. But the other features would be cool to have. Of course it's only compatible with an 800.

[Mathy]

Hello Lenore

 

IIRC It's "Ralf" David, not "Ralif" David.

 

I seem to remember that the 6820's and 6521's don't use the same characters (A, B) for the same frequency versions. As in, for instance, a 2MHZ 6521 has a different character at the end of it's name than a 2MHZ 6820.

 

BTW Western Design Center is still selling them, but not for the prices mentioned above. :-(

 

Sincerely

 

Mathy

[GroovyBee]

If acquiring legacy parts is a concern why not implement their functionality in a CPLD instead.

[dl7ukk]

Hi,

 

this is a good idea

please 2 piece for me

greetings
andreas

[Sleepy]

Hi,

 

nice idea. Please put down one for me.

 

Sleeπ

Edited July 2, 2016 by Sleepy

[Dropcheck]

Hello Lenore

 

IIRC It's "Ralf" David, not "Ralif" David.

 

I seem to remember that the 6820's and 6521's don't use the same characters (A, B) for the same frequency versions. As in, for instance, a 2MHZ 6521 has a different character at the end of it's name than a 2MHZ 6820.

 

BTW Western Design Center is still selling them, but not for the prices mentioned above. :-(

 

Sincerely

 

Mathy

 

 

Name correction noted.

 

My proto type is using 2 R6520P. According to the datasheet that means it is the 1 Mhz version. So assuming the same pinout a MC6821P, R65C21C1 or R65C21J1 should work. Right?

[Dropcheck]

If acquiring legacy parts is a concern why not implement their functionality in a CPLD instead.

 

 

Have no experience doing CPLD. Has someone actually duplicated the R6520P in a cpld yet?

[IndusGT]

 

I have successfully burned a 27C512 (512kbit 64KBX8) Eprom chip

Great! Thanks

[1050]

My proto type is using 2 R6520P. According to the datasheet that means it is the 1 Mhz version.

Maybe, maybe not. Depends very much from which line you got that number from. AND

the trademarked logo as a clue to who made the chip. Then what their datasheet say.

 

 

 

Both examples are made by rockwell so only a rockwell datasheet can be followed as

to what suffix characters mean. Top chip carries full rockwell number of R6520AP

in the first line which is the only valid chip number line.

Which makes it a 2 MHz. Bottom chip only carries Atari numbers but mixed in there

is one that can be mistaken for a valid Rockwell number and it just is NOT. I'm

speaking to R6520-26 -- it's an in house Atari number and you can't go by it. You

also can't go by second line of top chip with R6520-13 as it's also an 'in house'

number only this time Rockwell and we have no idea what dash13 means which sums up

the issue with in house numbers.

 

Rule of thumb is first line only please and the rest are formally 'in house' stuff

that can still be very useful. I would suspect the top chip was made in Mexico for

example, beyond that I would need in house documents to look up the meaning with

which generally isn't public information.

 

So assuming the same pinout a MC6821P, R65C21C1 or R65C21J1 should work. Right?

Depends on who made by their logo and if these numbers offered have been lazyily

modified by omission. MC68B21P would be fine if it carries the Hitachi or Motorola logo and that number

is from the first line. MC is a traditional Motorola prefix just as LM is a traditional

National number prefix. So finding a MC number is almost a guarantee that it has

a Motorola pin for pin equivalent and most IC makers will never cross this line to

produce a rouge part that isn't the exact drop in replacement for the MC number.

 

They still make these new. There is no issue of we can't get them anymore.

$1.55 each for 50 count includes shipping at eBay. 250 available from this one

source and I do believe that if you bought every single one today that tomorrow

they would magically have another 900 for sale.

http://www.ebay.com/itm/50pcs-MC68B21P-high-quality-Peripheral-Interface-Adapter-DIP40-/181326029267

I think CPLD just got priced out of the picture entirely.

[GroovyBee]

Have no experience doing CPLD. Has someone actually duplicated the R6520P in a cpld yet?

I assume most of the PIA port bits that are used are set to output for address and control lines and maybe a programming pulse of some kind. In that case you could replace that part with latches and ignore the DDR part of that chip. Knowing more about what the software does would help to find some better solutions and future proof the design too.

[GoodByteXL]

 

My proto type is using 2 R6520P. According to the datasheet that means it is the 1 Mhz version. So assuming the same pinout a MC6821P, R65C21C1 or R65C21J1 should work. Right?

Maybe, maybe not. Depends very much from which line you got that number from. AND

the trademarked logo as a clue to who made the chip. Then what their datasheet say.

 

Despite many people seem to be fond of replacing "old" chips with whatever emulated programmable thing they favor, in my view this is no advantage for the users as nearly everybody could replace a broken PIA acquired from whatever source. You will be stuck with any programmed thing (remember how difficult it was to obtain e.g. a MMU not long ago).

 

And how many fans or users are capable of fixing such things without the means to reproduce it when the originator is gone and nobody has the data. And when done in SMD style it will be even more difficult as fewer users will have that kind of means for their hobby.

 

I got stuck lately in repairing the IDE Plus V2 because of that. If the SMD buffer chip would have been a standard DIL on a socket, no frills, but SMD soldering is out of my sight. So I needed to acquire professional help.

 

So no, I am not voting for a CPLD or whatever if it comes to remake things from the past which do the job well in their original design.

 

In the notes with the original design sheets of the burner it is mentioned that the left PIA should be a 2 MHz version.

 

On the other hand my burner came with the 2 MHz PIA in the right socket and is doing fine for more than 2 decades.

 

I wonder if it makes any difference at all, so may be testing with 1 MHz PIAs only can tell. Also any pin compatible PIA that was found to be working in a XL/XE should do. My guess is that they used what was found in the drawer back then.

Edited July 3, 2016 by GoodByteXL

[Pmetzen]

I am interested in one

[Dropcheck]

I assume most of the PIA port bits that are used are set to output for address and control lines and maybe a programming pulse of some kind. In that case you could replace that part with latches and ignore the DDR part of that chip. Knowing more about what the software does would help to find some better solutions and future proof the design too.

 

 

So from your answer I assume that no one yet has done a R6520 in cpld. I like the idea of reducing chip count and 'future proofing' something, but it is way beyond my experience and skill to recreate a R6520 on a programmable chip. Maybe a couple of years down the road after I've spent dollars and time learning to do advanced cpld programming I can.

[Dropcheck]

Maybe, maybe not. Depends very much from which line you got that number from. AND

the trademarked logo as a clue to who made the chip. Then what their datasheet say.

 

index.jpg

 

 

Both examples are made by rockwell so only a rockwell datasheet can be followed as

to what suffix characters mean. Top chip carries full rockwell number of R6520AP

in the first line which is the only valid chip number line.

Which makes it a 2 MHz. Bottom chip only carries Atari numbers but mixed in there

is one that can be mistaken for a valid Rockwell number and it just is NOT. I'm

speaking to R6520-26 -- it's an in house Atari number and you can't go by it. You

also can't go by second line of top chip with R6520-13 as it's also an 'in house'

number only this time Rockwell and we have no idea what dash13 means which sums up

the issue with in house numbers.

 

Rule of thumb is first line only please and the rest are formally 'in house' stuff

that can still be very useful. I would suspect the top chip was made in Mexico for

example, beyond that I would need in house documents to look up the meaning with

which generally isn't public information.

 

 

Depends on who made by their logo and if these numbers offered have been lazyily

modified by omission. MC68B21P would be fine if it carries the Hitachi or Motorola logo and that number

is from the first line. MC is a traditional Motorola prefix just as LM is a traditional

National number prefix. So finding a MC number is almost a guarantee that it has

a Motorola pin for pin equivalent and most IC makers will never cross this line to

produce a rouge part that isn't the exact drop in replacement for the MC number.

 

They still make these new. There is no issue of we can't get them anymore.

$1.55 each for 50 count includes shipping at eBay. 250 available from this one

source and I do believe that if you bought every single one today that tomorrow

they would magically have another 900 for sale.

http://www.ebay.com/itm/50pcs-MC68B21P-high-quality-Peripheral-Interface-Adapter-DIP40-/181326029267

I think CPLD just got priced out of the picture entirely.

 

I think I was pretty specific on the part numbers when I asked the question. I had pulled up the datasheets and from what I read these exact part numbers were matches. I just wanted to see if I was missing something about them that others might know.

 

Thank you for the link. I have a smaller order coming in the next couple of weeks to see if the MC68B21P chip will work in the prototype. If it does then hopefully the link will still be good to place a bigger order if needed. That would definitely lift the restriction on fully populating the burner.

[+davidcalgary29]

Please add me to the list

[1050]

I think I was pretty specific on the part numbers when I asked the question. I had pulled up the datasheets and from what I read these exact part numbers were matches. I just wanted to see if I was missing something about them that others might know.

Fair enough, omission issue is obvious where I can't know if MC6821P is the valid full number or not.

In this case it is full by your confirmation and No, that one is 1 MHz and should not be purchased for this project. Your other

queries are likewise 1MHz devices and should not be considered at any price.

 

A lot of 1 MHz units will work for 2Mhz, but it's not guaranteed by the chip maker,

and also no guarantee against data loss, just FYI there. This can explain why your prototype works

while using the 1 MHz version, but I would change them out asap.

 

For all three devices asked about you apparently didn't see the ordering information table box where frequency

rating is deciphered and/or specified by part number suffix. Page one for MC6821P and page two for both others

which are using the same datasheet.

R65C21C2, R65C21P2 or R65C21J2 would work speed wise. But J package is 44pin PLCC quasi surface mount so not gonna drop right in. But it would make a nice fit for a ProBurner II version, eh watt?

 

Thank you for the link. I have a smaller order coming in the next couple of weeks to see if the MC68B21P chip will work in the prototype. If it does then hopefully the link will still be good to place a bigger order if needed. That would definitely lift the restriction on fully populating the burner.

Most welcome there, I certainly did not want you buying expensive chips when there is no need. If these

guys aren't around when you need them, then there will be somebody else there.

Edited July 4, 2016 by 1050

[GroovyBee]

So from your answer I assume that no one yet has done a R6520 in cpld. I like the idea of reducing chip count and 'future proofing' something, but it is way beyond my experience and skill to recreate a R6520 on a programmable chip. Maybe a couple of years down the road after I've spent dollars and time learning to do advanced cpld programming I can.

It doesn't have to be an exact copy of a 6520. For example, if the ports are always used as outputs e.g. to latch address lines then the DDR part of those ports can be removed from the design and a parallel latch used for the port instead. From the schematic it should be easy to identify bidirectional, input only and output only signals. Even if you didn't go down the CPLD route, its possible that some functionality could be replaced with readily available 74 series latch devices.

[freetz]

I'd be interested in a fully populated version (only), thanks to SantosP for arranging the distribution within the EU! How should one prove one's ABBUC membership?

 

Cheers!

[+tf_hh]

From what I have read, 6821 and 6521 should be interchangeable. However, I have no first hand experience.

 

The 68B21 and 6520 are the same IC. They can interchange without issues. The BitWriter original PCB uses the 6821 (mostly A version, 1 MHz, the B version could be used with 2 MHz clock), but I used for 3/4 of the BitWriter Replica boards the 6521 - works fine.

 

Jurgen

[Defender II]

Is there a projected availability date yet?

[Dropcheck]

Is there a projected availability date yet?

 

Final testing with a production board should begin this next week. Just got notice from Itead that they shipped my production board order.

 

By the way I think I have developed an adapter board for burning 2732 eproms. This burner does not natively support it, but should be able to with this adapter.

 

Also design on the 3d printed case is progressing. Translation of the manual from German to English is commencing as well.

[Mathy]

Hello Lenore

 

I'll take one. From the stack Panos is putting together. Maybe he could send them to one of us Eastern-Atlantic-ABBUC-members and we can pick them up at the Fujiama meetings (mid august) or the annual meeting (late october).

 

Sincerely

 

Mathy

[dl7ukk]

Hello,

 

I have a change.

Please 5 pieces for myself and others ABBUC User.

thank you

 

andreas

(translated by G)

[Dropcheck]

Final testing with a production board should begin this next week. Just got notice from Itead that they shipped my production board order.

What I thought was the final production board order turned out to be incorrect. Errors in the ZIF footprint are going to force a redo. That's okay, because when I went to test the MC6821 chip to see how compatible it would be, it wasn't. At the same switch settings that gave me 12.5 volts for Vpp with the R6520, I was getting 21.5 volts with the MC6821. At the switch setting for 21.5 volts I was getting 12.5. Now I have to figure out if it is a software issue or hardware.

 

If it's hardware I might be able to include a jumper on the board to easily use one or the other.

 

If it's software, then it will take a rewrite and even then I don't know if it can be done.

 

Just double checked and the part is actually a MC68"B"21P not MC6821P. Wonder if the 2Mhz is too fast?. Ordered the correct 1Mhz from ebay. Hopefully it won't take too long to get here. Checking the 2732 adapter now.

Edited July 14, 2016 by Dropcheck