quadra 630 owners - question re: serial port performance

[Cory5412]

Well, the tech note that mentions SerialDMA drivers v8 and v9 is dated 1996. This thing you've just posted is dated 1998, so I think this is probably a significantly newer file. It wouldn't at all be surprising if this newer file was in fact an update for the 5400 and its brethren rather than the 5200 and its.

I don't believe that the serialdma driver you want was ever published on its own. It is likely part of a newer version of system 7 than what was included on the system you have. That means 7.5.3 at absolute minimum. 7.5.3 with no footnotes was released in January 1996, so it's more likely a revision, or that 7.5.5 would have it.

To be honest, I just recommend moving to 7.6.1 as that gaurantees the newer serial driver, perhaps even a newer one than that, and 7.6.1 + speeddoubler 8 will likely run better overall.

This is presuming you're not already there, I don't know if you've mentioned.

7.5.1 is the preload version of the OS for the 5200 and it's worth noting that Apple has a moderately notable history of preload versions of Mac OS doing very poorly on its hardware, and errors being fixed with newer versions of the OS.

 

[supernova777]

my interpretation was that the article was blatantly spelling it out that the "serialDMA 2" file originates from the time period specified in the article (which was July 1993 through to November 1994)

the macOs versions that were current at that time period were: 
System 7 (May 1991) 7.1.1 (Oct 1993), 7.1.2 (Mar 1994), 7.5 (Sept 1994)

so yes, i think its logical to assume that an important fix would be included in later mass distributions of the OS.

so i guess my next assumption would be that this serialDMA driver is probably already on my peforma mac 

the versions that followed that time period: 

7.5.1 (Mar 1995), 7.5.3(Jan 1996), 7.5.5(Sept 1996), 7.6.1(Apr 1997)

im already running 7.5.5 now... so, my current version is from 1-2 years past that time period

i only updated to 7.5.5 to be able to get my damn networking working properly, installing openTransport 1.1.2 etc to be able to get

the chooser to allow specifying an IP address (appletalk over ip? ie: AFP://x.x.x.x) 

going forward updating to 7.6.1 is not what im looking to do,

like i said the whole reason for thaving the computer itself was:

for me to be able to have a vintage system, running 68k cpu,

which can run some music programs that dont seem to run on newer more modern ppc macs like g3/g4/g5 systems.

 

[supernova777]

http://web.archive.org/web/20011128081309/http://www.info.apple.com/support/oldersoftwarelist.html

Macintosh/Networking-Communications/Other_N-C/SerialDMA/SerialDMA_2.0.2.sea.bin

Macintosh/Networking-Communications/Other_N-C/SerialDMA/SerialDMA_2.0.2.txt
Macintosh/Networking-Communications/Other_N-C/SerialDMA/SerialDMA_2.1_Update.smi.bin
Macintosh/Networking-Communications/Other_N-C/SerialDMA/SerialDMA_2.1_Update_Info.txt  

wow I actually found the files i think!

but the 2.0.2.txt says:

Name: SerialDMA
Version: 2.0.2
Released: November 15, 1995
Description: Included in 7.5.2 Printing Update. Fixes problems while
printing on serial printers and improves overall performance
of the serial port on AV and PowerPC based Macintosh computers,
except 5200, 5300, 6200, and 6300 series. Requires System 7.5
or later. For complete information, please see the Read Me
included in this archive.

so there u go - there must be a real hardware reason as to why. its logical to assume that wouldnt u think?

this is just very clear further evidence to support the fact that the ports are somehow physically unique + different from the others.

 

[Cory5412]

It's not at all logical to presume this is a hardware problem. This is borne out in the experience of people who have successfully used faster modems with these machines, successfully participated in full-speed localtalk operation, and in the developer notes of the machines themselves, which show all the same hardware in the 5200/6200 as in the 630 and in the other AV Macs and PowerMacs.

It's good that you've done the research on SerialDMA. It's interesting that they say the fix is "not for" the 5200/6200. Perhaps this is because serial wasn't "broken" on those systems, or perhaps because the next-generation driver for those systems was already bundled into the OS.

Just to put it out here:

630 dev note https://cdn.preterhuman.net/texts/computing/apple_hardware_devnotes/Mac%20LC-Quadra%20630.pdf 

5200 dev note https://cdn.preterhuman.net/texts/computing/apple_hardware_devnotes/PowerMac%205200-6200.pdf

 - worth noting: on page 24, the serial port pin-out shows the hardware handshaking pins.

You should look at Figure 2-1 on both dev notes. You'll see that both systems use the same F108 memory controller chip to control memory, serial, and scsi.

This differs from a machine like the 605 (dev note: https://cdn.preterhuman.net/texts/computing/apple_hardware_devnotes/Mac%20LC%20475-Quadra%20605.pdf) where the SCC is a separate chip.

There's no indication that the SCC in the F108 is really any different from the 8530 used on previous Macs.

However, it isn't literally the exact same physical chip, so if literally nothing other than the 8530 will do, then you should buy a Mac that has it: Not the 630.

There's something different about the ports, but Apple considers them to be functionally identical to serial ports powered by the 8530 chip.

That said, I really haven't heard anywhere else that these machines are specifically problematic for MIDI, so it's pretty difficult to tell what exactly is going on.

 

[Cory5412]

So, there is a difference because around 1994 Apple switched to a different serial controller, but the 840/660, 6100/7100/8100, and also every subsequent PowerPC mac used either the new SCC design or something else, so I think there's something else going on - because there has been no widespread pandemonium about the serial performance of those systems.

If you have money to throw at this for troubleshooting purposes, I recommend that you buy a Performa 630 or variant and try your MIDI work on it. If it works, then there's something "wrong" with the 5200, but the thing that's wrong is not the serial controller, because that entire part of the block diagram of the two systems is literally identical. If it doesn't work on the 630, then it's possible the new serial controller is to blame.

Then, go buy a 68k Mac that uses the 8530 serial controller, something like a 475/605, a 610/650/800, a 700/900/950, or anything older than those. (Mac II series, Performa 600.) If the task works there but not on the 630, then we can presume that either the "new SCC" in the F108 is to blame, or something else on the system or how your machine was set up is to blame.

Also of interest could be a system like the 7500/8500, which were much more "for" tasks like this than their contemporaries, the 5200 and 6200. The 7500 and 8500 have the same SCC that the 5200/6200 have.

 

[supernova777]

that block diagram is in no way conclusive -- just because it says "SCC" on the block diagram does not mean that its the same "SCC" etc etc 

the F108 is not a memory controller but rather an ASIC (Application Specific Integrated Circuit) and you arent shedding any new light or information in stating that as IT WAS LITERALLY SAID IN THE ARTICLE I POSTED because the quadra 630 is from within the time period that the article mentioned and has been discussed over and over again.

From the time of the Centris 660av (July 1993) through to the Power Macintosh 8100/110 (Nov 1994) the SCC was integrated into an Apple ASIC with a DMA  controller supported by a SerialDMA driver. However the first (driver file version?) design was flawed and it is essential to ensure that SerialDMA 2 is used. Later Power Macintoshes (from 1995???) up to the beige G3s (introduced in Nov 1997) have an improved ASIC. For the full story lookup Technote 1018 on Apple's web site 

re: "There's no indication that the SCC in the F108 is really any different from the 8530 used on previous Macs."
the indication, history + premise was provided in the articles i posted:
https://www.fenestrated.net/mac/mirrors/Apple%20Technotes%20(As%20of%202002)/tn/tn1018.html

http://web.archive.org/web/19990505184713/http://www.hinton.demon.co.uk/mac/macmidi.html

it seems that most of the things you are posting originate in the taylordesign.net article:

http://www.taylordesign.net/classic-macintosh/the-mythical-road-apple/

like i said before; 

"if u are challenging that the article is false, thats your deal. im not debating that, im exploring the possibility that its true."  i said that in response to seeing the taylordesign.net article -- this is the internet, alot of people say alot of things out of their ass that is simply not true; or backed by correct logic/real hard evidence. 

this is kind of going in circles...

 

[supernova777]

wow its really annoying how this site concatenates links 

https://www.fenestra.../tn/tn1018.html

http://web.archive.o...ac/macmidi.html

are u the 68kmla.org owner/operator/admin?

 

[Cory5412]

I am in fact suggesting that the articles you posted are either incomplete, misinformed, or are straight-up sales pitches. Almost every article about these machines on LEM is based on totally wild guesses made by people who appear to not have looked at technical documentation about the machine when they launched, and much of that was initially published 1997-1998, so anything that refers to those articles is at best misinformed and at worst, meant specifically to rip on these machines for no particularly good reason.

• The problem with the "Mac MIDI" article is that it is basically a sales pitch, and sales pitches often don't rely on particularly accurate information about alternatives to the product being presented. Plus, the way it's presented talks not (as you say) about multi-port MIDI adapters, but about multi-port Mac serial multiplexers. A product that somebody using a Mac in the '90s might have used to connect a modem, PDA, printer, digital camera, MIDI devices, and perhaps other things to their computer all at once.
• Calling on TN1018 is of questionable value at this point, because as you discovered earlier, it looks like these particular SerialDMA drivers pertain to machines other than the 5200.

It is going in circles, so what I'll say is, if you want to try the 630, nothing I can do will stop you, but it is literally the same machine as a 5200, but with a much slower CPU. Especially if your 5200 does not have a modem installed, almost any problem you have with the 5200, you're likely to experience with the 630. The only reason this wasn't considered a problem in the '90s was because 630s were cheaper and older and because nobody was buying either of these systems to do demanding professional work.

As you say, there appears to be no hard evidence that physical problems with the 5200's architecture are causing the problems you're having, and there's no hard evidence that switching to a 630 will fix them.

One more note: The F108 is veritably the memory controller. Apple literally has this text in the developer note: the F108 memory controller and I/O support IC. It's on page 30 of the document, in the list of custom chips.

 

[supernova777]

just to make a record, and to answer more clearly + illustratively Cory's earlier questions:
heres a few images to show what the error is when my interface refuses to function properly..... everythign else is properly configured trust me..

swapping it out for the opcode studio 3 yields better results but it doesnt work 100% either.. it (the studio 3 opcode midi interface) can "bug out" and need to be power cycled etc

but these images are of the 64xtc - which detects properly.. i can "view the rom version" in OMS setup.. but then when i try to actually do any MIDI data recording or playing back 

it all goes KAPUT and this is what i get.. all the lights light up + blink (four times?).. not sure if its an actual error code they are blinking but then the system says the interface isnt responding and

i have to basically power cycle either the 64xtc by itself .. or the 64xtc as well as the mac.. depending on luck

 

[Cory5412]

Out of curiosity, do you have this device working on another computer, either a newer or older Mac or a PC?

EDIT: And, one more thought: Does your 5200 have a modem in the comm slot? If so, you need to remove that modem, because it's using the serial port.

 

[Cory5412]

One more random follow-up thing. I found out that the SerialDMA drivers being discussed earlier aren't "for" the 5200/6200 because they are for computers with GeoPort. GeoPort came out on the 660/840. The 630 and thus the 5200/6200 did not have GeoPort. The 6400 gained it, though.

 

[supernova777]

yes the 64XTC works fine on a windows computer!!!!!!! recently hooked it up to a pII 266Mhz slot 1 440bx board running windows 3.1 works 100% running windows 95 works 100%

back to the 5200CD...

for the record there is no Modem in the commslot 

i just pulled out the motherboard to check... theres no commslot modem installed
only a Farallon PDS network card that i ordered from Ebay
the modem interface is connected on modem serial port
the printer port has nothing connected
the scsi is attached to an external ZIP100 plus which works fine
its connected to my local network via 10BASE-T  port on the PDS card
 
>> Calling on TN1018 is of questionable value at this point, because as you discovered earlier, it looks like these particular SerialDMA drivers pertain to machines other than the 5200.
 
fact: the textfile for the v2.0.2 serialDMA extension does not mention the performa 5200
BUT! the text file included with v2.0.2 describes the information clearly:

What is the SerialDMA 2.0.2 extension?
Version 2.0.2 of this extension is simply version 2.0 of this extension plus a couple of additional fixes and support for
the new PCI Power Macintosh computers (7200, 7500, 8500, and 9500). 
Version 2.0 of this extension was a complete rewrite of the DMA serial driver which originally accompanied the:
Centris 660AV, Quadra 660AV/840AV, Power Macintosh 6100/60, 6100/66, 7100/66, 7100/80, 8100/80, 8100/100, 8100/110. 
It also supports the various Performa and Apple Workgroup Server products derived from the models listed. 
The SerialDMA driver is a replacement for the standard set of Macintosh serial device drivers referred to by their driver names:
�.AIn,� �.AOut,� �.BIn,� and �.BOut.�
SerialDMA is applicable only to the set of Macintosh models which incorporate DMA channels servicing the standard Z8530 SCC(Serial Communications Controller)
 
The original serial driver for these products suffered from occasional system hangs, communications timeouts, poor communications performance, and various other glitches.
The new SerialDMA 2.0.2 driver corrects these problems and provides the targeted products with a serial driver that should exceed the performance and reliability standards for serial communications on the Macintosh platform.

lets get our proper logical thinking caps on
notice it says "the NEW 7200, 7500, 8500 + 9500" so one can dedeuce this file obviously originates from 1995!!
fact: the 2.02 serialDMA extension says it was created on October 23rd 1995
 
so; the next logical question would be to confirm: does the 5200CD use the Z8530 controller? 
well lets refer back to the taylor design article: http://www.taylordesign.net/classic-macintosh/the-mythical-road-apple/
 

"An 8530 SCC serial port controller was embedded in the F108 ASIC. The developer note documents the handshake pins and specifically notes that the GPi pin for each port is connected to the controller and available for use. The GPi pin is what’s missing from earlier entry level Macs that limits their transfer speeds."

so in this article the writer is saying that his 6300 performa can handle a 56k modem, and that it also contains the F108 ASIC which according to him (?) also contains the 8530 SCC. 
but can we logically assume that if its present on the 6300 should it not be present on the 5200?
i would guess NO you cant logically assume that because the 5300/6300 + 5400/6400 series were different in many ways from the 5200/6200 series.. right??
so we need to confirm somehow that this is a docmented fact that the performa 5200CD does include this F108 ASIC, and that it is true whether or not it contains the 8530 SCC.
 
 
http://www.gla.ac.uk/~gwm1h/macos7.6/aboutappleremoteaccess.html this document says:

In order to use the serial port at 115200 bps on one of these computers, you need a new serial extension called SerialDMA. System 7.5.3 or later incorporates this driver into the operating system.
 

that being said. i have just now installed the serialDMA extension and i thought at first that it was crashing my system; but no.. its fine.. and ive got the extension activated
im just trying to get a proper MIDI app installed to do some realworld testing!
fact: when i first tried to install this version of the serialDMA extension, my performa froze on startup screen but its working now?? strange
 
this page looks pretty relevant:
http://wikivisually.com/wiki/Macintosh_Performa_5200CD