Why did Atari not follow Percom Standard?

[+mytek]

I just came across this little tidbit while searching the web.

 

Quoted from article at: www.it1me.com/learn?s=Atari_DOS

 

"Percom standard
In 1978 Percom established a double-density layout standard which all other manufacturers of Atari-compatible disk drives such as Indus, Amdek, and Rana--except Atari itself--followed. A configuration block of 12 bytes defines the disk layout."

 

Does anyone know why Atari made the decision to not be compatible with a standard that four other 3rd party companies were following in their disk drive products? Other than plain foolishness or the "not invented here syndrome", was there a valid reason to do their own thing independent of everyone else?

 

- Michael

 

[flashjazzcat]

They did comply once they brought out a DD drive (the XF551). Stock 1050 doesn't need it.

[+mytek]

They did comply once they brought out a DD drive (the XF551). Stock 1050 doesn't need it.

 

Too funny, just shows you how much I have used Floppy Drives lately (as in NOT, been at least 20+ years). When I got back into the Atari scene, I was quickly won over by things like the SIO2PC-USB and SIDE2. No real use for a floppy anymore when I have these devices to use.

 

Thanks for the quick response to my question

 

- Michael

Edited August 23, 2016 by mytekcontrols

[+CharlieChaplin]

So it remains the question why the 810 or the 1050 was not DD right from the start ?!? Imagine the 810 drive would have been DD, then we would all have DD drives and all A8 games would be min. on a DD disk (or image). The only change or upgrade would have been DSDD then... and no need for SD/90k and ED/130k formats...

Edited August 24, 2016 by CharlieChaplin

[+Stephen]

So it remains the question why the 810 or the 1050 was not DD right from the start ?!? Imagine the 810 drive would have been DD, then we would all have DD drives and all A8 games would be min. on a DD disk (or image). The only change or upgrade would have been DSDD then... and no need for SD/90k and ED/130k formats...

Cheapness. It takes 256-bytes to hold a DD sector, and more RAM needed for code, etc.

[ijor]

Well, not making the 810 is understandable. And it's not just an issue of RAM. The floppy controller on the 810 is FM only.

But on the 1050, the ED was really nitpicking (and let's not even mention DOS 3, jaja).

 

OTOH, if the 1050 would have been already DD, already with a customizable firmware as the 810 rev E, that would have spoiled all the fan of getting enhancements, wouldn't be ?

 

[ricortes]

I think that 1978 date may be misleading or not quite tell the whole story. I think Percom's did not come out for Atari until 1983

 

http://www.atarimagazines.com/creative/v9n6/114_An_alternative_to_the_Ata.php

 

Lot's of double density systems including IBM @160k and 180k SSDD and CP/M systems by then. Atari had its 815 on the drawing board and the 1050 wasn't designed in house. If I have it correctly, it was designed at Tandon and didn't become an Atari product until after the failure of the 815. So the 1050 was designed to compete with the 810 by Tandon so even the modest increase in capacity with 26 sectors/track was good'nuff. Still a less then optimum design given the technology available for several reasons. The voltage doubler circuit so it can use a 9 VAC supply, picking a 128 byte RAM. Certainly by 1982-3 the 1k & 2k static RAMs i.e. 6116s, were available and probably about the same price.

[thorfdbg]

 

So it remains the question why the 810 or the 1050 was not DD right from the start ?!?

That's due to the layout of the hardware. The 1050 (and likely the 810) consists of a 6502 derivate as CPU, and two standard 6502 peripheral chips (IIRC, two RIOTs - RAM, IO, Timer). There are two of them in the 6502, each of which comes with whooping 128 bytes of RAM. One of the chips serves as zero-page and stack for the 6502 (due to incomplete addressing), the other is the sector buffer. 256 byte sectors would have required *either* a lot of engieering skill (and hence time and costs) *or* an additional 128 bytes of RAM, and hence also increased costs.

 

Atari wanted to keep the disk drive as cheap as possible.

[tschak909]

I was just pissed that for years, they kept the 810 without any decent data seperation (not until the Grass Valley analog boards made in 1982)... Irony: Percom was just down the road from me, so it was _really_ easy to walk up to them and get a data seperator.

 

-Thom

[+kheller2]

All the 1050 needed was another RAM chip and ROM upgrade. Oh yeah: US Doubler. But I don't think anything really needed DD back in 83 and Atari seemed to have enough issues with bad drive mechs from Tandon. (I'm not sure sure the 1050 was designed outside of Atari). All software released to date was SD and it made no sense for a SW company to produce a DD disk that required a 1050 to run.

 

Having said that, Atari was designing DD drives right after the 1050.

[+slx]

I would have loved to have DD in '83 as it would have halved my expenditure for blank floppies (with all floppies sold being DD anyway because of the C= 1541 and no price advantage for SD's) and added a bit of convenience, i.e. less swapping.

 

As there were only few 810 drives around in Europe before the 1040 was introduced, Europe would probably have accepted DD as the Atari standard had Atari sold the 1050 with true double density. (For the US the installed base of 810s was probably too large to make software manufacturers change to DD.)

 

And there is a 1050 upgrade that allows DD without extra RAM by using software that does not use subroutines. Read more here.

[ricortes]

All the 1050 needed was another RAM chip and ROM upgrade. Oh yeah: US Doubler. But I don't think anything really needed DD back in 83 and Atari seemed to have enough issues with bad drive mechs from Tandon. (I'm not sure sure the 1050 was designed outside of Atari). All software released to date was SD and it made no sense for a SW company to produce a DD disk that required a 1050 to run.

 

Having said that, Atari was designing DD drives right after the 1050.

I've repaired a number of 1050s that had 'Copyright Tandon' stickers on the ROM.

 

The original 810s were a bit of a mess too. I have one and it has worked perfectly for me but subsequent models had an extra board<analog?> to make them work right. Double density was no sure thing back then. One of the reasons Percom was in the business and successful is they knew how to do it right.

 

BTW: I always get into semantic fights over density. I have history and manufacturers' data sheets to back me up. Single density according to the manufacturers is FM while double density is MFM. This is because the original meaning of density was ~bits/inch. For that matter you could have 4k byte/sector FM and 128 byte/sector MFM. Of course the 4k bytes/sector may have only one sector/track.<sic> All academic now, like debating if a T-Rex could win against a pack of Raptors.

[+kheller2]

Not sure where this is going but 1050s were manufactured in two countries, and included Tandon or WST mechs and ROMs/EPROMs. The keyword "designed" is what threw me, and unless someone can correct me, I didn't think the design was outsourced. Even the 810 had different mechs and motherboard versions.

[Mathy]

Hello slx

 

As there were only few 810 drives around in Europe before the 1040 was introduced, Europe would probably have accepted DD as the Atari standard had Atari sold the 1050 with true double density. (For the US the installed base of 810s was probably too large to make software manufacturers change to DD.)

 

If the 1050 would have been able to do DD, it probably have been able to read SD too. So software manufacturers didn't have to switch to DD if they didn't want to.

 

Sincerely

 

Mathy

[+mytek]

Hello slx

 

 

If the 1050 would have been able to do DD, it probably have been able to read SD too. So software manufacturers didn't have to switch to DD if they didn't want to.

 

Sincerely

 

Mathy

 

Good point. And please excuse my ignorance, but doesn't the XF551 do this? What would it have taken to have the 1050 also do this as well? Just curious

 

- Michael

[+kheller2]

See above. A RAM chip and ROM updates. Again, there was probably very little cost justification to do a DD drive as the only benefit, at the time, would be for the end user. Software companies would be silly to produce a DD game.

 

The XF551 seems to be a newer version of the 1053.

Edited August 26, 2016 by kheller2

[ijor]

Not sure where this is going but 1050s were manufactured in two countries, and included Tandon or WST mechs and ROMs/EPROMs. The keyword "designed" is what threw me, and unless someone can correct me, I didn't think the design was outsourced. Even the 810 had different mechs and motherboard versions.

 

I tend to agree with ricortes. Not that I know for sure, but I believe Tandon did wrote the firmware.

 

One thing is to provide mechanism, or even the whole hardware, another is to claim copyright on the firmware. And if the firmware for one hardware version was copyright Tandon, then it was Tandon copyright in all the 1050s because the firmware was the same, disregarding the mechanism or the manufacturing country.

 

I also seem to remember references to "Tandon code" somewhere, don't remember where I saw that.

 

Now, writing the firmware doesn't necessarily means that Tandon designed the 1050, or the ED concept. But it is very possible they did. Do we know who wrote DOS 3? They should be related.

[tschak909]

Bill Wilkinson wrote DOS 3 based on specifications from Atari.

 

-Thom

[BillC]

I've repaired a number of 1050s that had 'Copyright Tandon' stickers on the ROM.

I believe the sticker just indicates that a 2732 EPROM was used rather than a ROM.

This is why there are 4 jumpers to configure for the differences between the 2 pin-outs(2332/2532 versus 2732).

[+hunmanik]

Percom drives for the Atari, the first alternatives to the 810, came out in spring 1982. Antic announcement; InfoWorld review.

 

 

I think that 1978 date may be misleading or not quite tell the whole story. I think Percom's did not come out for Atari until 1983

http://www.atarimagazines.com/creative/v9n6/114_An_alternative_to_the_Ata.php

Lot's of double density systems including IBM @160k and 180k SSDD and CP/M systems by then. Atari had its 815 on the drawing board and the 1050 wasn't designed in house. If I have it correctly, it was designed at Tandon and didn't become an Atari product until after the failure of the 815. So the 1050 was designed to compete with the 810 by Tandon so even the modest increase in capacity with 26 sectors/track was good'nuff. Still a less then optimum design given the technology available for several reasons. The voltage doubler circuit so it can use a 9 VAC supply, picking a 128 byte RAM. Certainly by 1982-3 the 1k & 2k static RAMs i.e. 6116s, were available and probably about the same price.

[+kheller2]

 

I tend to agree with ricortes. Not that I know for sure, but I believe Tandon did wrote the firmware.

 

One thing is to provide mechanism, or even the whole hardware, another is to claim copyright on the firmware. And if the firmware for one hardware version was copyright Tandon, then it was Tandon copyright in all the 1050s because the firmware was the same, disregarding the mechanism or the manufacturing country.

 

I also seem to remember references to "Tandon code" somewhere, don't remember where I saw that.

 

Now, writing the firmware doesn't necessarily means that Tandon designed the 1050, or the ED concept. But it is very possible they did. Do we know who wrote DOS 3? They should be related.

 

 

Actually, the firmware is different. But then again, that depends on what you mean by the "same" or "different"... a few bytes here or there can be ignored or scrutinized. You can't take a WST mech and run it in a Tandon based unit and vice versa. The ROMs are slightly different, and I'm not sure if any of the POTS are calibrated differently -- I wouldn't think they would be, but its not something I've looked at.

 

Usually, but not always: Singapore = Tandon, Hong Kong = WST. I've seen both Tandon and WST EPROMS, but only ever masked ROMs by Tandon. The plastics are mentioned to be slightly different as well (but I've never bothered to confirm that). The WST boards are screwed down and the connectors are all facing the proper way, unlike in Tandon units where the drive mech connectors are flipped on some of the wire groupings. This of course means nothing, but thought it would be interesting to provide. Also, I remember reading but can't find the notes at the moment, about the problems Atari was having getting good drive samples from Tandon. That might have been related to the 1450, but then again I don't think Tandon was sourced to design the floppy board.

 

This also leads me into asking where people think the 810 was designed? It is certainly a more elaborate(?) device given that WD didn't have (or didn't have an affordable) FDC at the time for data seperation.

 

As a side note, the Indus GT drives used two different mechs as well, and were connected slightly differently, but the ROMs were the same.

 

This is an interesting discussion, totally unrelated to the topic, and steers more towards the emotional/political side of what constitutes "designed by" vs "manufactured by" vs a joint effort.

Edited August 26, 2016 by kheller2

[ricortes]

I should have included the Tandon stickers on the ROMs were ~Copyright Tandon to make it clearer. Doesn't/wouldn't make sense for that to be anything other then the software for sure and probably the drive was done at Tandon. I mean if they were just manufacturing the drives, they would have just put an Copyright Atari sticker on them. I also accept Curt as pretty much a definitive and accurate source for most things Atari.

http://www.atarimuseum.com/computers/8bits/xl/xlperipherals/atari1050.html

 

"The disk drive electronics and its mechanism were done by Tandon, the case design was done by Tom Palecki, formerly of Atari's Industrial Design group."

 

IIRC: World Technology<?> is just another incarnation of Tandon.

[thorfdbg]

Single density according to the manufacturers is FM while double density is MFM. This is because the original meaning of density was ~bits/inch. For that matter you could have 4k byte/sector FM and 128 byte/sector MFM. Of course the 4k bytes/sector may have only one sector/track.<sic> All academic now, like debating if a T-Rex could win against a pack of Raptors.

Indeed. The ED you have on 1050 is MFM coded, so in that sense it is a DD format. However, it does not hold twice the capacity of the SD format because the overhead is larger. If I remember correctly, the floppy controller on the 1050 can even do MFM at 256 bytes/sector, but this feature wasn't used because there was not enough RAM for a sector buffer.

[Bryan]

And there is a 1050 upgrade that allows DD without extra RAM by using software that does not use subroutines. Read more here.

Does anyone have this code? I couldn't find it.

[+kheller2]

I should have included the Tandon stickers on the ROMs were ~Copyright Tandon to make it clearer. Doesn't/wouldn't make sense for that to be anything other then the software for sure and probably the drive was done at Tandon. I mean if they were just manufacturing the drives, they would have just put an Copyright Atari sticker on them. I also accept Curt as pretty much a definitive and accurate source for most things Atari.http://www.atarimuseum.com/computers/8bits/xl/xlperipherals/atari1050.html"The disk drive electronics and its mechanism were done by Tandon, the case design was done by Tom Palecki, formerly of Atari's Industrial Design group."IIRC: World Technology<?> is just another incarnation of Tandon.

I agree Curt is the final word. The PROM/ROM and mech are clearly Tandon devices. WST was probably just another source for the mechs ... Just in case. It's always odd for me to see Tandon units because my very first 1050 was a WST and so were the next few. It wasn't until I started collecting that I thought the Tandon was the odd ball. And there aren't any WST technical guides...