Help with internal 1050 to PC via USB connection

[Dropcheck]

I am trying to design an internal 1050toPC circuit using the FT232RL chip and USB connector. I am planning to incorporate it into either a XF551 or 1050. I would like to use either the 1050 to PC or Pro System software on the PC at the flick of a switch. Also I think I need to plan for the worst case of having the drive connected to both Atari and PC at the same. My memory isn't what it used to be and I have a bad habit of leaving things on/connected.

 

I have come up with a basic schematic, but I am thinking there needs to be some more protection for the devices involved.

 

[ricortes]

I believe you need add 5V to READY<?> line or the drive will ignore everything.

[Dropcheck]

I've added both a on/off switch and forward biased voltage on the SIO pin 10 for a 1050. The XF551 doesn't use SIO pin 10 but the 1050 does. Weird.

 

Power for the circuit comes from the USB port ie PC.

Edited March 3, 2015 by Dropcheck

[Dropcheck]

Added a second diode to the previous circuit. Also setup this circuit for a XF551 which only uses SIO pins 3,4,5,6 and 7.

 

[foft]

I think in 1050-PC mode you need a pull-up on the sio input (computer sio in, drive sio out).

Edited March 4, 2015 by foft

[Dropcheck]

I think in 1050-PC mode you need a pull-up on the sio input (computer sio in, drive sio out).

 

Not following you.

[1050]

I've added both a on/off switch and forward biased voltage on the SIO pin 10 for a 1050. The XF551 doesn't use SIO pin 10 but the 1050 does. Weird.

That's what I noticed when I looked but haven't confirmed with an actual XF551 board just yet. Sobola's schematics are sometimes wrong at this level of detail so conformation with real hardware is best. If so then you can shelve pin 10 issues by applying only this workaround to 1050 drives installed with this circuit internally. Run the drive's +5 volt power thru a 10K resistor directly to pin 6 of the 3086 and this will cause the drive to operate as if the Atari was on, even when it isn't and most importantly you can't back feed the turned off Atari an alternate source of power this way. Diodes aren't needed and won't ever fail if they are not there in the first place, save some money. 2nd switch then just for 1050s also not needed.

 

As to agreed weird part, Atari itself seems to be the biggest offender out there in NOT complying with their own specifications and published standards. I'll assume they just did a why bother and didn't put that weak 'Atari devices for Atari computers only' interlock system in the XF551.

 

 

Not following you though.

 

I still don't see the PC's serial port to this device's connections, so I'm left wondering if 1050 to PC software does USB? News to me if it does and the source to most of my confusion. Really need to stay in the loop better.

 

As to original device protection issue, all TTL outputs are designed for fanout where several devices are listening to the data/address buss, they are made to take one giving high and another low at the same time so no real issue other than data scrambling which is solved with the command line switch.

 

'you mean like this?'

Absolutely not, that's just bleeding data in to data out line. A pullup resistor goes to +5 and since we are dealing with data in, data out, and command line only I have no clue where he wants to put the pullup at in the first place, best to ignore until clarified so it can even be considered first? A reference to this need would be a handy thing to link to as well, I'm thinking. Hasn't been done on the 1050, XF551, this chip or the Atari so...?

[Dropcheck]

That's what I noticed when I looked but haven't confirmed with an actual XF551 board just yet. Sobola's schematics are sometimes wrong at this level of detail so conformation with real hardware is best. If so then you can shelve pin 10 issues by applying only this workaround to 1050 drives installed with this circuit internally. Run the drive's +5 volt power thru a 10K resistor directly to pin 6 of the 3086 and this will cause the drive to operate as if the Atari was on, even when it isn't and most importantly you can't back feed the turned off Atari an alternate source of power this way. Diodes aren't needed and won't ever fail if they are not there in the first place, save some money. 2nd switch then just for 1050s also not needed.

 

As to agreed weird part, Atari itself seems to be the biggest offender out there in NOT complying with their own specifications and published standards. I'll assume they just did a why bother and didn't put that weak 'Atari devices for Atari computers only' interlock system in the XF551.

 

 

Not following you though.

 

I still don't see the PC's serial port to this device's connections, so I'm left wondering if 1050 to PC software does USB? News to me if it does and the source to most of my confusion. Really need to stay in the loop better.

 

As to original device protection issue, all TTL outputs are designed for fanout where several devices are listening to the data/address buss, they are made to take one giving high and another low at the same time so no real issue other than data scrambling which is solved with the command line switch.

 

'you mean like this?'

Absolutely not, that's just bleeding data in to data out line. A pullup resistor goes to +5 and since we are dealing with data in, data out, and command line only I have no clue where he wants to put the pullup at in the first place, best to ignore until clarified so it can even be considered first? A reference to this need would be a handy thing to link to as well, I'm thinking. Hasn't been done on the 1050, XF551, this chip or the Atari so...?

 

 

In this case I think Sobola's schematics for the XF551 SIO are acurate. I have the prelimary Atari XF551 schematics and they are showing only those 5 connections as well.

 

The FT232RL of course is a USB to TTL serial converter chip. I am going on this http://forum.arduino.cc/index.php?topic=82793.0 and http://atariage.com/forums/topic/141455-sio-to-usb/ as a indication that it will work with just 4 SIO connections. The second poster of the first link indicates that he is using this with an XF551 drive. Not exactly what I am trying to do, but close.

 

From all I've seen the difference in Pro-System and '1050 to PC' is primarily which sig from the pc it works with, RTS vs DTR. Several posters have also indicated that you need to flip the TX/RX connections based on whether you are going to an Atari computer or to a 1050/XF551 drive from the pc.

 

So installing a 10k resistor to pin 6 of the 3086 on a 1050 isn't going to cause issues with a normal Atari to drive connection? If so that would save a dollar or two. Most importantly one less switch on the drive to worry with.

[MrMartian]

Many drive firmwares (and I've been through most of them!) use this signal to indicate whether or not they should be listening to the SIO bus at all... IE, if the signal is low, the system is off so abort any running IO... I would say it's best overall to provide the READY signal as it should be.

[ricortes]

I've thought about it a lot. The Atari systems are designed such that nothing is unterminated either on chip or discrete. You don't want a POKEY with floating inputs or it may pick up signals from the florescent lights and think it is hooked up to a cassette or disk drive. Ditto for everything that is hooked into it, you wouldn't want a 1050 to get that one in a million shot of thinking it format a disk based on someone in house using a microwave oven.

 

Inherent in the system is you will always get a tug of war between pull ups and pull downs. Every time you add another drive or interface it will change the characteristics. i,e, READY being pulled down by eight drives vs. one, all working against the computer 5V ready. Same hold true for every other signal on the SIO bus. IIRC, Tandon/Atari handled it by using the transistor array on signals with something like a 100k resistance going to the base of the transistors with ~200 hfe, We are lucky we get away with what we do. If we were really careful and had all the time/money in the world, we would isolated everything we do like this with either transistors or op amps. This could have been foreseen with some of the features built into the SIO bus, for example smart peripherals that would go into high impedance state except for when the READY/COMMAND lines were asserted to wake them up.

[lemiel]

I do not know about 1050 and Ready line because do not have one but for example Indus/LDW/CA with Toms Turbo uses Clock line and 1050-2-PC does not work in this mode.

[foft]

Not following you.

Yes because the computer data in line is open collector. So afaik the drives output data by pulling to ground. Normally on the Atari end there is a pull up so when the drives are high impedance it returns high.

Edited March 5, 2015 by foft

[1050]

Now I understand foft's concerns and he is correct after all, I had not looked far enough to find the 4.7K pullup resistors living behind 100 ohm series resistors on the lines of interest inside the Atari. The 1050 has no pull up or pull down resistors on these lines at all. I too now believe the 1050 is only capable of dropping lines of interest low but does not bring them high by itself. So the chip used here should be doing that for the 1050 by similar values on all four lines that Drop wants to use. This may be why some errant data and lock ups are encountered using other designs from time to time? I don't see having two devices (pokey/PIA combo, and this chip) doing this at the same time as a critical problem when the Atari is on. With the Atari off the lines will always stay down without the pull ups placed on this chip.

 

But Drop's placement of the suggested 'like this' resistor is in the wrong place as I outlined before. Just to the right of the chip while the four lines are still on a horizontal run, there should be placed four 4.7K resistors to +5. On the vertical run after the corner, four 100 ohm resistors placed in series on all four lines. And then remove the 'Like this' box and the schematic gets thumbs up from me.

 

A lot of talk about the READY signal line with signal being the part that raises my eyebrows. Does no one else see that this so called signal line is connected to the Atari power brick when powered up? It's VCC, +5 Atari buss, I wouldn't exactly call it a signal. If it weren't for the 10 uh inductor on that line we could use it to power the entire drive, printer, modem and anything else on the SIO chain. Some signal. Only the inductor and the small wires of the SIO cable itself cause it to lack some real horsepower. It never goes low until you turn off the Atari and this is supposed to be a drive that we can use in stand alone mode.

 

My main concern is when the Atari is off while still attached and how we don't want to back feed it into a powered state while burning up the inductor. Diodes are great, they will work. I just don't see it as a real need. If you feed drive sourced +5 thru a 10K resistor behind the drive's own ready line isolation resistor (also 10K) then you can only back feed the Atari one quarter of a milliamp and that much is entirely harmless on a power buss. Since this is an internal mod and this only needs doing on drives that are using that 5 volt line as an interlock. We can supply the +5 gingerly from the drive itself and have a stand alone drive that always wants to work. NOT going to hurt my feelings if used with a diode. It's just a valid suggestion that the diode is overkill and not really required. You can use the money saved on the second switch to buy purist diodes with? Both the diode and 10K resistor need to be on the 1050 side of the dual SIO drive jacks so really can't be diagrammed with this schematic.

 

OTOH here is a solution that needs no words. A sharp knife and a bit of hook up wire will suffice.

 

Before ------------------------------------------- After

 

[ricortes]

From somewhat faulty memory, the IN lines on the 1050 from the 8 bit go through ~100k resistors tied to the base of the 3086 transistors. In function, this works exactly like a 100k resistor tied to ground through a diode. 8 drives would load the Atari out lines ~12.5k resistor to ground through a diode with ~1.2 voltage drop. I'm just pointing this out in that it is both a good design from isolation/loading standpoint and something we should aspire to. It ain't nothing but it is there. I aspire to only presenting 100k Ohm loadings!

 

Anything else is terminated/pulled high or low down stream on the devices to keep them from floating/spurious signals and doesn't affect the sio bus.

 

We get away with murder because most of us don't hook up 8 drives and a printer at the same time. A single device like SIO2PC USB will present at least one ttl<or CMOS> load per driven line. If we wanted to be good neighbors on the SIO bus, we would buffer everything. Even though I think it would be a good idea, I am personally against it in hobbyist mode. It would add too much complexity and cost to any design. As ~pointed out, if a device uses the READY line, it would be a good idea to buffer/isolate it by at least running it though a resistor that will keep things from smoking. Seems like an oversight in that the 1200XL used a resistor on the 5V line but apparently left the READY line as is.<?>

[Bikerbob]

So Dropcheck, what happened to this project?

I would be interested in this for my XF551 or 1050.. would be nice to get this internal.

 

James

[_The Doctor__]

After watching and using a multitude of these devices and watching people try to save 10 cents and act like its 20 dollars, and then looking at the mess on the sio bus when fully loaded... I just don't understand why we always seem blow it in the Atari world of hobby hardware.... with the prices charged for many of the devices.... why scrimp to save 10 cents or an entire dollar.... The reason why I balk at so many people and prices is because of these kinds of things....

 

I have supported Lenore's projects in the past and indeed more recently with a nice pile of super sparta both varieties and break out carts.... because they were fairly well designed and implemented with good solid components.

 

I will support this project as well if it is THE definitive interface that stands out apart from the crowd, the loading of the bus with corrections and protections will get my money. I am tired of the sio framing errors farts and fails from the offerings of the past...

 

I will praise the individual that make a differential sio interface that can handle it all...and just to put it over the top... can pump a wav. or other audio source on the device as well... wouldn't it be nice to have the audio with the emulated devices (tape included) (or synthesizer) from a pc or android. .... just the way Atari intended? A true SIO to PC emulate everything device... The closest we have are the critical connection and Atarimax devices. I know we can do better and having seen what Lenore, TF, Pano's, electrotrains, and the myriad of others have done both in the past and recently with input from the great minds here at Atariage. We can truly enjoy such a device and it would stand apart as well as above the others.

 

Money and I do not part easily! But when we do, It's always quality!

[BillC]

From somewhat faulty memory, the IN lines on the 1050 from the 8 bit go through ~100k resistors tied to the base of the 3086 transistors. In function, this works exactly like a 100k resistor tied to ground through a diode. 8 drives would load the Atari out lines ~12.5k resistor to ground through a diode with ~1.2 voltage drop. I'm just pointing this out in that it is both a good design from isolation/loading standpoint and something we should aspire to. It ain't nothing but it is there. I aspire to only presenting 100k Ohm loadings!

 

Anything else is terminated/pulled high or low down stream on the devices to keep them from floating/spurious signals and doesn't affect the sio bus.

 

We get away with murder because most of us don't hook up 8 drives and a printer at the same time. A single device like SIO2PC USB will present at least one ttl<or CMOS> load per driven line. If we wanted to be good neighbors on the SIO bus, we would buffer everything. Even though I think it would be a good idea, I am personally against it in hobbyist mode. It would add too much complexity and cost to any design. As ~pointed out, if a device uses the READY line, it would be a good idea to buffer/isolate it by at least running it though a resistor that will keep things from smoking. Seems like an oversight in that the 1200XL used a resistor on the 5V line but apparently left the READY line as is.<?>

There is no +5V power separate from the READY signal, pin 10 is used for both.