1050 repair help needed

[+kheller2]

Got me here a 1050 with Tandon mech and Tandon Rev K EPROM. The unit powers up, even spins the floppy the appropriate few seconds when you put a floppy in. However, when connected to a computer, all I get is a one second blast of bzzzz from the computer and then the Ready prompt.

 

Where on earth do I start looking for the issue? No obvious signs of problems on the board. Even with the mech unplugged, same symptoms.

 

Karl

[+Guitarman]

Look at the SIO connectors. Make sure there are no cold/broken joints. That is the type of behavior the XF551's had when the flimsy SIO connections broke. Also make sure the drive is set to Drive 1.

Edited April 21, 2007 by Guitarman

[+kheller2]

Nope.. not solder joints.. and the drive is set to 1. I don't get the standard, um, fart for lack of a better term that happens when you turn on the system w/o a drive. It definitely knows something is there just sounds horrible when probing.

[mimo]

Nope.. not solder joints.. and the drive is set to 1. I don't get the standard, um, fart for lack of a better term that happens when you turn on the system w/o a drive. It definitely knows something is there just sounds horrible when probing.

try a different power supply if you can. I had problems with my 1050 tonight, and it turned out to be the psu

 

also try swapping the sio lead, I also had a problem where a lead works fine with sio2pc but not at all with a 1050

Edited April 21, 2007 by mimo

[puppetmark]

First, make sure you have 5V and 12V from the internal power supply, not the power brick but inside the 1050. The Drive uses linear voltage regulators that can burn out. On the 1050 board, you will see metal posts sticking out of the board, these are test points hence the TP marking. Check TP13 it should read 5VDC check TP14 it should read 12VDC. Also, make sure the caps in the power supply are not drying out, the tops will be peaking up like little tents if they are.

 

If your voltages are good, then the problem is probably U1. It is the chip near the SIO port. It is a transistor array, and I have replaced many of them in the 1050s. My bet is that this is the problem. you can replace the chip with a LM3086, its the same thing. You can get them from Jameco.com or Best electronics. If you do order the chip from Best, do yourself a favor and get a new drive belt too.

 

Let me know how you make out. I have attached a copy of the schematic. If you want, I would be happy to look at it for you sometime. you are only an hour or so away from me.

 

Edited April 21, 2007 by puppetmark

[+kheller2]

puppetmark, that was my guess too after running through the 1050 Field Service Manual. I haven't checked the voltages yet, but it does pass the power on tests. It comes down to the SIO path through U1 to the PIA U7. U1 is some sort of SIO "amp" is how Atari worded it I think. I'll swap out the 6532PIA from another drive first (I'm guessing a 6520 won't work).

 

You should visit the #atari IRC channel.... it was Atari Hardware Radio tonight.. and I got just about exactly to where your post is. Wish I had seen it earlier.. but the folks on #atari seemed to have some sick liking to my ramblings.

 

Karl

Edited April 21, 2007 by kheller2

[puppetmark]

Looks you are on the right track. I think I have a couple LM3086 chips. I can send you one if you need it.

 

I have been meaning to get onto the atari IRC Channel. I am not sure I understand how. guess I have to try it to understand it.

Edited April 21, 2007 by puppetmark

[SpicyChronos]

That chip was the problem with my Super Archiver II drive I won off of eBay a long time ago, even though the person said it didn't talk to the computer, I guessed that was the problem. So I promptly got some IC sockets and about a dozen of those chips (never know when your gonna need spares), and viola! it worked. So I hope for your problem it is the CA3086 chip at U1. Best of luck to you!

 

-Marshall

Edited April 21, 2007 by SpicyChronos

[re-atari]

That chip was the problem with my Super Archiver II drive I won off of eBay a long time ago, even though the person said it didn't talk to the computer, I guessed that was the problem. So I promptly got some IC sockets and about a dozen of those chips (never know when your gonna need spares), and viola! it worked. So I hope for your problem it is the CA3086 chip at U1. Best of luck to you!

 

-Marshall

I agree. Had the same thing here with a 1050. A not-so-competent previous owner had soldered in a write protect switch, but hadn't noticed he had bridged one of the pins of the jack with an adjacent via with a large blob of solder. I still ask myself how you can not notice that.

Result was the same symptom as with your 1050. I don't even remember what was bridged, but removing the solder didn't get the drive back on its feet. It just would not communicate with the XL. Replacing the 3086 transistor array fixed it, chances are it will be the culprit in your drive. It cost me just 75 cents (including the socket) and a few hours of soldering to get my 1050 up and running again.

 

re-atari

Edited April 21, 2007 by re-atari

[re-atari]

puppetmark, that was my guess too after running through the 1050 Field Service Manual. I haven't checked the voltages yet, but it does pass the power on tests. It comes down to the SIO path through U1 to the PIA U7. U1 is some sort of SIO "amp" is how Atari worded it I think. I'll swap out the 6532PIA from another drive first (I'm guessing a 6520 won't work).

 

You should visit the #atari IRC channel.... it was Atari Hardware Radio tonight.. and I got just about exactly to where your post is. Wish I had seen it earlier.. but the folks on #atari seemed to have some sick liking to my ramblings.

 

Karl

A 6520 will definitely not work, a 6532 has more things onboard, like 128 bytes of RAM.

 

re-atari

[+kheller2]

Ok it is not the PIA.. next stop, the 3086.

[+kheller2]

1050 Fixed. Puppetmark was kind enough to send me a 3086 and the socket for it. Note to folks.. Q6 is insulated from the heatsink ground... via a piece of plastic tape and rubber washer through the regulator. If you don't put that back right.. it won't pass the power on self test and you'll get upset.

[puppetmark]

1050 Fixed. Puppetmark was kind enough to send me a 3086 and the socket for it. Note to folks.. Q6 is insulated from the heatsink ground... via a piece of plastic tape and rubber washer through the regulator. If you don't put that back right.. it won't pass the power on self test and you'll get upset.

 

Nice Work! Its nice to see 8-bit stuff fixed instead of the alternative.

[+kheller2]

OK.. I got another 1050, looking for advice again.

 

This is a World Storage Technologies drive where every chip is socketed!

 

When powered on: Power light comes on, motor spins then stop, does not seek to track 0. Any other power cycle and the motor won't spin -- it will only spin once after being unplugged for several hours. I ran through the Tandon manual and verified the test points had the correct voltages, so I don't think its power related. The mech works fine, as I swapped it into another unit and tested it.

 

Before I start really digging into this, I was wondering if anyone has a pointer to a possible cause.

[WizWor]

Looks you are on the right track. I think I have a couple LM3086 chips. I can send you one if you need it.

 

I have been meaning to get onto the atari IRC Channel. I am not sure I understand how. guess I have to try it to understand it.

Look here for some help with this.

[+kheller2]

Um.. IRC? Yes, I'll ask there too.

[puppetmark]

Its a shame all the chips are soldered. That always complicates repair. If you leave the drive on for a few minutes, I would check for an over heating component.

 

If nothing is obvious then I would start with the OS Chip then the 6532 chip.

 

Problems like this often require an oscilloscope and a good service manual, otherwise you just have to guess and replace parts till it works, and of coarse, with everything socketed you have your work cut out for yourself.

 

I think I have some spare 1050 chips, I will look and see what I have.

[+kheller2]

Every chip *is* in a socket, even the LS chips. I haven't had time to poke at it yet.

[+kheller2]

It was the WDC 2793.. or is it a 2797.. now I need to reopen it to remember what I need to order.

[re-atari]

It was the WDC 2793.. or is it a 2797.. now I need to reopen it to remember what I need to order.

Order a 2793, most 1050's used them. A 2797 will work, but needs to have 1 pin disconnected (snipped or bent away out of the socket). 2793's will be easier to come by, if you can find one. Check with Best; the price might scare you, though...

 

re-atari

[ijor]

It was the WDC 2793.. or is it a 2797.. now I need to reopen it to remember what I need to order.

Order a 2793, most 1050's used them. A 2797 will work, but needs to have 1 pin disconnected (snipped or bent away out of the socket).

 

In addition, not every ROM works with the 2797. IIRC only the latter (don't remember the letter code right now) 1050 ROM revision was updated for the 2797. You could also have troubles with 1050 enhancements. The USD standard ROM needs a 2793, there was a different version for the 2797.

 

(Catching up slowly after a break from Atari)

[re-atari]

In addition, not every ROM works with the 2797. IIRC only the latter (don't remember the letter code right now) 1050 ROM revision was updated for the 2797. You could also have troubles with 1050 enhancements. The USD standard ROM needs a 2793, there was a different version for the 2797.

That's correct. I recall a member of an A8 computerclub I occasionally visited, who had bought a Happy-ish clone that didn't work in his 1050. Turned out it had a 2797 installed. After swapping it with a 2793 (no additional adjustments of any pots required) the speeder worked like a charm.

BTW, the pin that's snipped/bent out on a 2797 in the 1050 is pin 25. Its use is side select when using double-sided drive mechanisms.

 

re-atari

Edited July 4, 2007 by re-atari

[carmel_andrews]

First time i've read or heard that d/s mechanisms can be made to work on a 1050 (unmodded of course and assuming you've got a comaptible upgrade like happy/lazer/speedy etc), and there's me thinking that d/s mech's were the EXCLUSIVE preserve of the XF551

[re-atari]

First time i've read or heard that d/s mechanisms can be made to work on a 1050 (unmodded of course and assuming you've got a comaptible upgrade like happy/lazer/speedy etc), and there's me thinking that d/s mech's were the EXCLUSIVE preserve of the XF551

That's not what you've read Well, it's not what I wrote. I stated that in some 1050's Atari used a 2797 instead of a 2793 (probably due to a supply shortage). In order to get these working in the 1050, pin 25 was snipped. This pin is the only difference between the 2 controllers, it controls side-select on double-sided mechanisms. With this pin cut off, the 2797 can only control 1 side.

 

Using a double-sided mechanism just isn't possible with the 1050's PCB standard layout. A few years ago I already looked into this, as I had a spare PCB without a suitable mechanism laying around. My initial idea was to connect a 360Kb PC-drive to this PCB, but this involved so much of a challenge that I just buried it. Besides, after SIO2SD was released, it didn't make sense anymore either.

To give you an indication of the issues you'll meet when trying to connect a double-sided PC-drive mechanism to the 1050 PCB:

- duplicate the circuitry that decodes/encodes the signals from and to the extra R/W head on a add-on seperate PCB. This is roughly the front part of the 1050's PCB (between the on/off switch and the metal shielding containing the 2793, 6507, 6532, etc.). You could bypass this circuitry by routing the R/W signals for both heads directly from the Shugart bus to the 1050 PCB.

- find a way to control the spindle motor. PC-drives usually have a PLL-controlled direct drive motor, that is controlled from the drives' own logic.

- bypass the index-hole. This is also controlled on the drives' own PCB, and it is used to control the speed of the spindle motor. It doesn't help that the 1050 uses 288 rpm instead of 300 rpm (standard with PC-drives).

- find a way to connect the stepper motor. It will not be too difficult electronically, but PC-drives usually have 80-step motors (1.8 deg/step), the 1050 has a 40-step (3.6 deg/step) one. So, to write on the same physical place as the 1050, a PC-drive will have to do 2 steps at a time. Some drives (not all) have a usefull feature called 'double-stepping' built in, some Teac model even have a jumper for this.

- last (and by no means least) rewrite part of the 1050's BIOS to enable side-selection and reading/writing to the extra head. This requires access to a documented source listing, and the necessary know-how.

 

All in all, it seems easier to design a completely new drive electronics. There are plenty old designs floating around the net, that can give an idea on how to interface a 2797 to the Shugart bus. You'd still have to live with small incompatibilites to the 1050, though. On a PC drive mechanism getting the drive speed down to 288 rpm will prove virtually impossible without modifying the PLL circuitry of the spindle motor. You'll need a schematic of that particular drive mechanism, which are generally NA. Trying to find doc's on the jumpersettings can already prove to be a challenge in its own right.

 

But, again, now SIO2SD is available, why go through all this trouble

 

re-atari

Edited July 5, 2007 by re-atari

[ijor]

I stated that in some 1050's Atari used a 2797 instead of a 2793 (probably due to a supply shortage).

 

Yes, most/many/all/some latest 1050 units had a 2797. I'm not sure why, my guess is that Atari got them a few cents cheaper than the 2793.

 

In order to get these working in the 1050, pin 25 was snipped. This pin is the only difference between the 2 controllers, it controls side-select on double-sided mechanisms. With this pin cut off, the 2797 can only control 1 side.

 

The problem is not exactly controlling one side or two. The problem is that this pin is an input in the 2793, but an output in the 2797. It is connected to VCC (at least on older 1050 boards designed for the 2793), and you don't want to connect VCC to a 2797 output

 

Using a double-sided mechanism just isn't possible with the 1050's PCB standard layout. A few years ago I already looked into this, as I had a spare PCB without a suitable mechanism laying around. My initial idea was to connect a 360Kb PC-drive to this PCB, but this involved so much of a challenge that I just buried it.

 

Connecting a PC drive to a 1050 is indeed not simple. Some of the complications you mention can be easily solved, but still not simple.

 

You don't need to duplicate the read/write logic. That's already built-in in a PC drive. So all you need is to skip it and connect the FDC directly to the drive (as you mentioned yourself). Same with motor control, just connect the RIOT output directly to the drive. You don't need to bypass the index hole in a 360K PC drive. Most 360K drives don't need the index hole, or at least can be configured to do so with a jumper. This is indeed a problem for reading the flippy side on HD drives (most HD drives do need to see the index hole).

 

An 80-tracks drive brings additional problem. The most important problem that can't be solved so easily is inter-operatibility with a 40-tracks drive. The different RPM is a big problem. Overall I agree is still doesn't make much sense. Just use an XF-551 board instead.