XF 551 - Troubleshooting

[Mathy]

Hello Robert

 

So many things have been upgraded the last few years, that it shouldn't be impossible to upgrade the Speedy XF too. There must be someone who's willing to take the challenge.

 

Mathy

[Dropcheck]

Hi,

 

Does anyone have schematics for the XF551 motherboard? I have the ones from here http://www.dereatari.republika.pl/schematy.htm

[Mathy]

Hello Drophead

 

If you're a member of the ABBUC, you can download it from our website www.abbuc.de

 

Mathy

[Dropcheck]

Hello Drophead

 

If you're a member of the ABBUC, you can download it from our website www.abbuc.de

 

Mathy

 

Well I knew my mother did something to me when I was little but dropped on my head. No, don't remember that.

 

I signed up two days ago and so far have not heard anything. No email with user name password, no denial, nada. So I've contacted Bradley at Best Electronics and purchased the schematics he has. Some of the components on the schematics here are missing values and I'm not entirely sure it's trustworthy. Assmann's schematics do not include power components and values. So critical info is missing between the power jack and the first vcc for the ICs and the 12V power for the floppy mech. Hopefully Bradley's schematics will show that info.

 

On the cost of the boards.... I'm running into what I was afraid of and one of the reasons this hasn't been done before, cost. A dropin replacement (say 152 x 127 mm) just a little shorter in length is going to cost around $75 dollars and up. That's without parts. Remember we are looking at a small run, probably no more than 15 or 20.

 

Perhaps with some fancy redesign of the the power circuit with modern parts I can cut another 10/20 mm off the length. I don't fancy that as I am not an engineer. I will look into that though.

 

In the meantime I am working on the schematic in Eagle to finalize a board.

Edited May 31, 2012 by Dropcheck

[BillC]

On the cost of the boards.... I'm running into what I was afraid of and one of the reasons this hasn't been done before, cost. A dropin replacement (say 152 x 127 mm) just a little shorter in length is going to cost around $75 dollars and up. That's without parts. Remember we are looking at a small run, probably no more than 15 or 20.

 

Perhaps with some fancy redesign of the the power circuit with modern parts I can cut another 10/20 mm off the length. I don't fancy that as I am not an engineer. I will look into that though.

 

In the meantime I am working on the schematic in Eagle to finalize a board.

I haven't used this company but saw them recommended in another thread, you can get 5 prototypes of a board smaller than 20cm X 20cm for only $38.

http://iteadstudio.c...dex&cPath=19_20

For some reason the link always takes me to the larger than 20cm x 20cm maximum board option no matter what URL I post, just click on the PCB Prototyping Service link to go to the page with the different sizes offered.

[31336haxx0r]

Off the shelf wall wart PSUs with +5V and +12V supply should do the trick. They just need to supply enough current. Like at least an Amp per rail.

[Dropcheck]

On the cost of the boards.... I'm running into what I was afraid of and one of the reasons this hasn't been done before, cost. A dropin replacement (say 152 x 127 mm) just a little shorter in length is going to cost around $75 dollars and up. That's without parts. Remember we are looking at a small run, probably no more than 15 or 20.

 

Perhaps with some fancy redesign of the the power circuit with modern parts I can cut another 10/20 mm off the length. I don't fancy that as I am not an engineer. I will look into that though.

 

In the meantime I am working on the schematic in Eagle to finalize a board.

I haven't used this company but saw them recommended in another thread, you can get 5 prototypes of a board smaller than 20cm X 20cm for only $38.

http://iteadstudio.c...dex&cPath=19_20

For some reason the link always takes me to the larger than 20cm x 20cm maximum board option no matter what URL I post, just click on the PCB Prototyping Service link to go to the page with the different sizes offered.

 

Due to the poor english used I'd suspect China or other Far East source. Which isn't bad. What is giving me a headache is when you try to nail down the cost no two quotes match. They seem to be set sizes too, either too big or just slightly too small for my needs. Still a possible option that I'll keep in mind.

 

Thanks

[Dropcheck]

Off the shelf wall wart PSUs with +5V and +12V supply should do the trick. They just need to supply enough current. Like at least an Amp per rail.

 

Thanks. It was something I considered. Still may do. Was just trying to reuse as much as I could and keep the board as original as possible, ie dropin and go while using the Atari PS.

[BillC]

On the cost of the boards.... I'm running into what I was afraid of and one of the reasons this hasn't been done before, cost. A dropin replacement (say 152 x 127 mm) just a little shorter in length is going to cost around $75 dollars and up. That's without parts. Remember we are looking at a small run, probably no more than 15 or 20.

 

Perhaps with some fancy redesign of the the power circuit with modern parts I can cut another 10/20 mm off the length. I don't fancy that as I am not an engineer. I will look into that though.

 

In the meantime I am working on the schematic in Eagle to finalize a board.

I haven't used this company but saw them recommended in another thread, you can get 5 prototypes of a board smaller than 20cm X 20cm for only $38.

http://iteadstudio.c...dex&cPath=19_20

For some reason the link always takes me to the larger than 20cm x 20cm maximum board option no matter what URL I post, just click on the PCB Prototyping Service link to go to the page with the different sizes offered.

 

Due to the poor english used I'd suspect China or other Far East source. Which isn't bad. What is giving me a headache is when you try to nail down the cost no two quotes match. They seem to be set sizes too, either too big or just slightly too small for my needs. Still a possible option that I'll keep in mind.

 

Thanks

From what I see that is the maximum size of board for the price listed, your design doesn't necessarily have to be that size. If you could change your layout to under 10cm x 20cm it's possible you could get 2 boards per 20cm x 20cm panel, that would be 10 boards for $38.

[+Larry]

Here is a picture that you may find interesting -- it is a 720K version "Hyper-XF" work-alike board from World of Atari in '98, IIFC.

 

Redesigned with the same major IC's (with Hyper-XF rom OS), but with other components replaced with SMD's. The board actual board space used for devices is about 3" X 4".

 

This was an "intermediate stop" for the developer -- he wanted to make a totally new floppy using an Atmel or similar controller, but I don't think he ever finished it.

 

-Larry

[guus.assmann]

Hello Dropcheck,

 

There's a gap in the communication or your understanding of things and my explanation.

And that's just what the forum should be helpfull for. But sometimes it takes me some time to see.....

 

This is the case: The XF uses a power supply that provides an AC voltage.

And that's converted to 12V and 5V DC. This is done by some standard stuff on the XF board.

 

The design I made, totally lacks the power supply.

I used a supply from another device to power the floppy mechanics with 12V for the motors and so on.

All logic, on both the controller and the floppy unit requires 5V.

This is also supplied by the same (from seperate former SCSI device) power supply.

As I used the casing of this SCSI device, I also didn't include the SIO-connectors on the board.

 

In conclusion, my design had a somewhat different purpose.

 

BR/

Guus

[Dropcheck]

Hello Dropcheck,

 

There's a gap in the communication or your understanding of things and my explanation.

And that's just what the forum should be helpfull for. But sometimes it takes me some time to see.....

 

This is the case: The XF uses a power supply that provides an AC voltage.

And that's converted to 12V and 5V DC. This is done by some standard stuff on the XF board.

 

The design I made, totally lacks the power supply.

I used a supply from another device to power the floppy mechanics with 12V for the motors and so on.

All logic, on both the controller and the floppy unit requires 5V.

This is also supplied by the same (from seperate former SCSI device) power supply.

As I used the casing of this SCSI device, I also didn't include the SIO-connectors on the board.

 

In conclusion, my design had a somewhat different purpose.

 

BR/

Guus

 

Hi,

 

I didn't mean to imply your design was incomplete or lacking in any way. I realized early on that you were assuming power off the board. If it seemed I was casting negatory comments on your work I heartly apologize. That was not my intention.

 

I was just commenting that what schematics I had available to me were missing infomation to replicate the power circuits for the XF551. Actually I was bemoaning that. I still am. I just got the schematic from Best Electronics and guess what? It's a preliminary and there's no parts list and no value for the transformer/inductor or the bridge rectifier parts.

 

So unless someone has the exact part # or values, I have these options:

 

1. Dissasemble the motherboard and hope that there are part # or values on the parts (Not wanting to do that on a working unit )

 

2. Use a completely external power supply that will pump 12V and 5V to my board something like what you did.

 

I am leaning towards the second option for a number of reasons:

 

It's self contained

You can get a power supply that would be better able to power dual 3 1/2 and 5 1/4 floppies.

It would be cheaper to buy than to make from scratch.

If it fails assuming it doesn't take the board with it, it's easier to replace.

It allows me to make the board much smaller and therefor reduce cost of the board.

 

So you say why don't you simply use your board. I would and could, but..... (There's always a but here ) I want to incorporate a few mods via simple jumpers depending on the user's desire. Minor changes perhaps, but I hate to solder unless I have to. I'd rather be able to simply drop in what I want and go. Also I wanted it to completely replace the current XF551 motherboard yet fit in as a simple drop in replacement, no additional mods for the user other than replacing the power supply.

 

As it stands it will have the same rear footprint of the XF551 case with SIO connectors and power jack and drive # select, utilize some of the same mounting holes, have power connectors for both 3 1/2 and 5 1/4 floppy drives, and have standard floppy IDC connectors.

 

Would it be a souped up version? To a certain extent. It won't scramble eggs.

 

Is it practical? Probably not. But then our obsession with a 40 year old piece of computer history doesn't have much practical value anyway and yet we still spend money on it and for it.

 

I'm getting really good at spending money and time on it.

 

Dropcheck

[Dropcheck]

Here is a picture that you may find interesting -- it is a 720K version "Hyper-XF" work-alike board from World of Atari in '98, IIFC.

 

Redesigned with the same major IC's (with Hyper-XF rom OS), but with other components replaced with SMD's. The board actual board space used for devices is about 3" X 4".

 

This was an "intermediate stop" for the developer -- he wanted to make a totally new floppy using an Atmel or similar controller, but I don't think he ever finished it.

 

-Larry

 

Interesting....

Edited June 1, 2012 by Dropcheck

[31336haxx0r]

Actually, it's less expensive to use a wall wart PSU. If one would clone the XF's or 1050's power supply, you'd need the following:

 

- transformer with 9VAC, 3 amps

- bridge rectifiers

- at least 3 electrolytic capacitors

- at least 4 ceramic capacitors

- 7805KC and 7812KC ICs (that KC suffix is important)

 

For the cost of all that, you could buy several wall wart PSUs with the added advantage that those don't get nearly as hot and consume much less power.

[guus.assmann]

Hello Dropcheck,

 

Thanks, but I don't feel offended in any way.

And I was also trying to choose my words so that I wouldn't offend you either.

 

The fun with our 40 year old stuff is that it's still understandable and we can learn from it.

When it was new, nobody did dare to touch it, the costly stuff might break.

And now, almost anyone does dare to modify and design all kinds of stuff for it.

 

31336haxx0r is right. One or two wall-wart supplies would do the trick quite nice.

Or 1 wall-wart with 12V and a 7805 for the 5V.

 

BR/

Guus

[BillC]

I found this schematic of the XF551 on the Atarimax website. The only part I noticed without a value is the inductor between the output of the LS38 and Data IN on the SIO port.

 

And also the transistors, but I believe that these should be 3904 for Q1/Q2/Q3, and 3906 for Q4.

 

Edited June 2, 2012 by BillC

[31336haxx0r]

Uuuuuuuhhhhhhhhggggggghhhhhhhhhh, they used that voltage doubler circuit (forgot its name) in a modified form to supply the 7812. It works, but you can't draw that much power. Now that I think of it, 9V * 1.41 = 12,69V, which is too low for a 7812 to work and quite high for a 7805. Substract 0.6 Volts forward voltage drop for each diode in the way and you get even lower.

 

I'd say let's NEVER EVER replicate the stock PS design and use some better stuff instead!

[31336haxx0r]

Double post, please remove.

Edited June 2, 2012 by 31336haxx0r

[BillC]

Uuuuuuuhhhhhhhhggggggghhhhhhhhhh, they used that voltage doubler circuit (forgot its name) in a modified form to supply the 7812. It works, but you can't draw that much power. Now that I think of it, 9V * 1.41 = 12,69V, which is too low for a 7812 to work and quite high for a 7805. Substract 0.6 Volts forward voltage drop for each diode in the way and you get even lower.

 

I'd say let's NEVER EVER replicate the stock PS design and use some better stuff instead!

You mention the voltage doubler circuit, then ignore it when calculating the voltage supply to the 7812.

During one part of the phase current from ground flows through D2 charging C102 to about 8V, when the polarity reverses there will be about 17V(9V+8V) at the C102+ terminal, current then flows through D3 to charge C104/C106 to about 16V, this is subtracting 1V for each diode and not multiplying by 1.41. Once charged C104 contains enough energy to keep the 7812 output steady at the designed current draw. It may not be efficient but it is simple and does the job.

[Dropcheck]

I've kind of decided to go with an external self contained power supply like this eBay Auction -- Item Number: 190391450076

 

I'll install a power jack on the board that would fit it. I'm only concerned that I may have too much amp on the 5V line. It's not real clear if it's 2 amps per voltage or 2amps total.

 

I've got one coming to see what my meter reads.

[Fox-1 / mnx]

There's no such thing as "too many amps". The connected system asks whatever it needs, even if it's more then the power supply can deliver. That's why a power supply always should be able to deliver more than a device asks.

[BillC]

I've kind of decided to go with an external self contained power supply like this eBay Auction -- Item Number: 190391450076

 

I'll install a power jack on the board that would fit it. I'm only concerned that I may have too much amp on the 5V line. It's not real clear if it's 2 amps per voltage or 2amps total.

 

I've got one coming to see what my meter reads.

A power supply will deliver current up to the what the label states safely, and should also be able to handle 1.5 to 2 times that for a short period. Only requiring 1/4 or 1/2 of the rated current of the power supply should not be an issue.

 

A nice thing about about using an external power supply is that removal of the diodes/regulators reduces the amount of heat generated in the enclosure.

[31336haxx0r]

Uuuuuuuhhhhhhhhggggggghhhhhhhhhh, they used that voltage doubler circuit (forgot its name) in a modified form to supply the 7812. It works, but you can't draw that much power. Now that I think of it, 9V * 1.41 = 12,69V, which is too low for a 7812 to work and quite high for a 7805. Substract 0.6 Volts forward voltage drop for each diode in the way and you get even lower.

 

I'd say let's NEVER EVER replicate the stock PS design and use some better stuff instead!

You mention the voltage doubler circuit, then ignore it when calculating the voltage supply to the 7812.

During one part of the phase current from ground flows through D2 charging C102 to about 8V, when the polarity reverses there will be about 17V(9V+8V) at the C102+ terminal, current then flows through D3 to charge C104/C106 to about 16V, this is subtracting 1V for each diode and not multiplying by 1.41. Once charged C104 contains enough energy to keep the 7812 output steady at the designed current draw. It may not be efficient but it is simple and does the job.

 

Why only 8V?

[BillC]

Uuuuuuuhhhhhhhhggggggghhhhhhhhhh, they used that voltage doubler circuit (forgot its name) in a modified form to supply the 7812. It works, but you can't draw that much power. Now that I think of it, 9V * 1.41 = 12,69V, which is too low for a 7812 to work and quite high for a 7805. Substract 0.6 Volts forward voltage drop for each diode in the way and you get even lower.

 

I'd say let's NEVER EVER replicate the stock PS design and use some better stuff instead!

You mention the voltage doubler circuit, then ignore it when calculating the voltage supply to the 7812.

During one part of the phase current from ground flows through D2 charging C102 to about 8V, when the polarity reverses there will be about 17V(9V+8V) at the C102+ terminal, current then flows through D3 to charge C104/C106 to about 16V, this is subtracting 1V for each diode and not multiplying by 1.41. Once charged C104 contains enough energy to keep the 7812 output steady at the designed current draw. It may not be efficient but it is simple and does the job.

 

Why only 8V?

I used 1V for the voltage drop across D2/D3 for convenience, the actual value is about 0.7V. Using this voltage drop C104 charges to 16.6V

[31336haxx0r]

Yes, but it's AC. So peak voltage is square root of 2 multiplied with the mean voltage ov 9 V. Isn't it?