1050 tandon vs the other mech

[rockdoc2010]

It seems that I have a Tandon mech that works just fine with a happy upgrade.

but it sticks out like 1/8 inch in front.

 

any thoughts?

 

Douglas

 

 

[StickJock]

Maybe it is not sitting on the posts properly?

[rockdoc2010]

nope the posts are solid

[rockdoc2010]

was there a difference from tandon tech to the others one?

 

[rockdoc2010]

I am totally happy that i have a happy..

 

 

[StickJock]

The faceplate & the top piece of the case have a few differences.  IIRC, one of the hold-down posts on the top shell is in a different place, so it presses on a different part of the drive mech.  (I've got a drive where someone put the wrong top & face on it).

 

This post shows the difference between the Tandon & WST drives.

 

Edited June 25, 2020 by StickJock

[rockdoc2010]

it seems the tandon is a bit off of the others

i am getting parts to redo that motherboard

so we will see

 

[rockdoc2010]

it is a matter of posts  

[rockdoc2010]

sorry the piano has been drinking

[_The Doctor__]

7 hours ago, rockdoc2010 said:

sorry the piano has been drinking

I knew it!

[+Nezgar]

There's two "main" types of 1050's:

 

"Made in Singapore" - Tandon Mech

"Made in Hong Kong" - World Storage Technologies mech

 

The cases, bottom and top posts are in slightly different places as you have found, so you'll have to match up the right components. The WST drives also use a different ROM with different stepper motor phase encodings and step delay, which do not work with a Tandon mech.

 

Using a happy board with a WST mech "downgrades" the stepper speed to that used by the Tandon drives. There is an alternate USDoubler ROM that takes the stepper phase encodings and track delay from the WST ROM which might be better suited for that drive until such time someone creates a happy rom "optimized" for a WST mech. Heck, I used a happy board in a WST 1050 all my 90's life not knowing the stock ROM actually stepped faster!!

[Kyle22]

What about using a WST stepper motor in a Tandon mech?

 

[rockdoc2010]

Thank You Very Much!

 

I was chasing issues with three 1050's giving BOOT ERROR  and swapping mechs and MB

 

Wound up being a bad power supply.

 

I DL instructions for what I thought was a dead happy and whooldathunkit... It works with a tandon mech and installation instructions.

 

 

[cwilbar]

14 hours ago, Nezgar said:

There's two "main" types of 1050's:

 

"Made in Singapore" - Tandon Mech

"Made in Hong Kong" - World Storage Technologies mech

 

The cases, bottom and top posts are in slightly different places as you have found, so you'll have to match up the right components. The WST drives also use a different ROM with different stepper motor phase encodings and step delay, which do not work with a Tandon mech.

 

Using a happy board with a WST mech "downgrades" the stepper speed to that used by the Tandon drives. There is an alternate USDoubler ROM that takes the stepper phase encodings and track delay from the WST ROM which might be better suited for that drive until such time someone creates a happy rom "optimized" for a WST mech. Heck, I used a happy board in a WST 1050 all my 90's life not knowing the stock ROM actually stepped faster!!

I have a 1050 here with a Tandon mech, but the WST ROM.  works nicely.... quicker.

 

I can't recall I where, but I read somewhere that Tandon used a couple of different stepper motors.... maybe only one of them can handle the WST ROM ????  or maybe the one I have is pushed beyond spec, but I got a mid week build :-).

 

 

[+Nezgar]

1 minute ago, cwilbar said:

I have a 1050 here with a Tandon mech, but the WST ROM.  works nicely.... quicker.

 

I can't recall I where, but I read somewhere that Tandon used a couple of different stepper motors.... maybe only one of them can handle the WST ROM ????  or maybe the one I have is pushed beyond spec, but I got a mid week build :-).

I think you're lucky if it's working for you - pretty sure I tried and it didn't step properly with at least one mech. Yeah I've participated in at least one thread here on AA comparing various steppers... When I was trying to figure out if a corrupt US Doubler ROM was not working because of stepper type... turned out the phase encoding bytes were messed up in the bad ROM and that was a red herring...

 

Consider a US Doubler upgrade recently patched for "Speedy" like stepping speed with proper tandon stepper phase encoding. (Plus DD, UltraSpeed, etc hehe). Currently assembling 2x6810 PCB's after waiting eons (2 months!) for more pin headers to show up from China...

[ijor]

20 hours ago, Nezgar said:

The WST drives also use a different ROM with different stepper motor phase encodings and step delay, which do not work with a Tandon mech.

 

After Jeff used the patched USD ROM, I think we know now that the faster step rates do work in a Tandon mech. At least in some of them, probably in all later ones. Phase encoding should be the same, otherwise 1050 enhancements (USD, Happy) wouldn't work on WST mechs. What might be incompatible, is the extra step that the Tandon ROM performs and the WST doesn't. Somebody should try a stock WST ROM on a Tandon mech and we will know.

 

[_The Doctor__]

when we say try, that means filling disks in all densities and seeing them work on both a wst and tandon afterwards... many things appear to work and don't get fully tested... later we pay the price...

[cwilbar]

2 hours ago, ijor said:

 

After Jeff used the patched USD ROM, I think we know now that the faster step rates do work in a Tandon mech. At least in some of them, probably in all later ones. Phase encoding should be the same, otherwise 1050 enhancements (USD, Happy) wouldn't work on WST mechs. What might be incompatible, is the extra step that the Tandon ROM performs and the WST doesn't. Somebody should try a stock WST ROM on a Tandon mech and we will know.

 

As I noted previously, I have one drive in that configuration, and it appears to work fine.  I have not done extensive testing, but booted a number of disks (copy protected and not) w/o issue.  This does not mean (as @_The Doctor__ points out above) that everything would be right.  I did not test ED, I did not test write behaviors, etc.  I thought the only difference between the ROMs was basically the step rate.  Wonder what other mech specific changes could be in there ?  Anyone ever concretely determine the 1050 ROM versions and document the differences ?

 

I'm sure it would be better to use a Tandon ROM patched for faster stepping (I presume something like that exists), or step up to the patched USD ROM with faster stepping and get DD out of it at the same time.

 

Can't seem to locate at the moment the thread on the PCB that eliminates the need to stack RAM ?

 

[ijor]

4 hours ago, cwilbar said:

As I noted previously, I have one drive in that configuration, and it appears to work fine.  I have not done extensive testing, but booted a number of disks (copy protected and not) w/o issue.  This does not mean (as @_The Doctor__ points out above) that everything would be right.  I did not test ED, I did not test write behaviors, etc. 

(Re: WST ROM on a Tandon mechanism)

 

Interesting, good to know. I doubt ED would make much of a difference, although of course, we need more exhaustive testing to be sure it works reliably. But the main issue that might be difficult to test is that it might depend on the specific stepper. It seems reasonable that perhaps earlier drives with older stepper would not work, and newer ones wouldn't have a problem.

 

Quote

I thought the only difference between the ROMs was basically the step rate.  Wonder what other mech specific changes could be in there ?  Anyone ever concretely determine the 1050 ROM versions and document the differences ? I'm sure it would be better to use a Tandon ROM patched for faster stepping (I presume something like that exists),

 

A Tandon ROM patched for faster stepping would be identical to a WST ROM. Below is a link to thread about all the 1050 ROM revisions.
 

 

 

 

[+Nezgar]

On 6/26/2020 at 12:48 PM, ijor said:

A Tandon ROM patched for faster stepping would be identical to a WST ROM.

"Functionally & Apparently" yes... but "technically" the phase encodings *are* different between Tandon and WST ROM's as supplied by Atari...

 

Phase encodings are in the first four bytes of Atari's 1050 ROM's:

• Tandon Rev H: FB F7 EF DF
• Tandon Rev J: FB F7 EF DF
• Tandon Rev K: FB F7 EF DF
• Tandon Rev L: FB F7 EF DF
• WST Rev 5: F3 E7 CF DB

Even with the stepper rates patched, I think we should try preserve the Forward/back step for tandon mechs, and alternate phase encodings for WST mechs and clearly identify the target mech for any new ROM patches...

 

With that said, I used a happy board in my WST 1050 for most of my 80's/90's life without issue, which used the Tandon phase encodings and slower stepper rate. I had never realized the stock ROM actually had a faster stepper rate since it's time was very short-lived before being upgraded.

[+Nezgar]

So I did some reading on phase encodings... The 1050 uses a 4-phase stepper, and 4 of the bits from these encoding bytes (5,4,3,2 - inverted?) are output by the PIA as the signals to the four coils of the stepper.

 

Tandon ROMs:

FB = 11111011

F7 = 11110111

EF = 11101111

DF = 11011111

 

WST ROM:

F3 = 11110011

E7 = 11100111

CF = 11001111

DB = 11011011

 

We can see only 1 bit shifting in the Tandon ROM, and a pair in the WST ROM. According to wikipedia, the Tandon ROM single bit shift method is called "Wave drive (one phase on)", and the WST ROM double-bit shift method is called "Full-step drive (two phases on)."

• "Wave drive (one phase on) has the same number of steps as the full-step drive, but the motor will have significantly less torque than rated. It is rarely used."
• "Full-step drive (two phases on) is the usual method for full-step driving the motor. Two phases are always on so the motor will provide its maximum rated torque. As soon as one phase is turned off, another one is turned on. Wave drive and single phase full step are both one and the same, with same number of steps but difference in torque."

Interesting that the method employed by the Tandon ROM uses the method that "is rarely used" and has "significantly less torque." It would appear the encoding used by the WST ROM would be superior for all drives. This may improve reliability/performance when using shorter stepper delays too - something to consider for future ROMs and patches.

 

Edit: I found probably the main reason to make sure the encodings are matched to the mech - according to this page: https://www.motioncontroltips.com/faq-what-are-stepper-drives-and-how-do-they-work/

 

"Engineers rarely use wave driving: it is inefficient and provides little torque, because only one phase of the motor engages at a time. Two-phase-on driving has its name because two phases are on at a time. If the drive energizes both A and B poles as south poles (shown in green), then the rotor’s north pole attracts to both equally and aligns in the middle of the two. As the energizing sequence continues on like this, the rotor continuously ends up aligning in-between two poles."

 

That last sentence suggests that mismatching the Tandon/WST ROMs and mechs equates to a permanent 1/4 track misalignment in a 1050 since it takes 2 steps to complete a track step. Factory alignment would have accounted for the intended ROM's phase encoding. (This could explain some of my own reliability issues reading marginal disks in my WST 1050 that had a happy board equipped, or disks formatted by that drive used in other drives...)

[cwilbar]

Interesting finding.  I suppose the best way to determine if it would be 1/4 track misaligned would be to take a Tandon 1050 and using an alignment disk, make sure it is in alignment, swap out to a WST ROM, then check the alignment.  It should be plainly evident if the different encodings for the stepper throw off the alignment.  If it operates as described then it does indeed sound like it would be off.

 

To step one track, does it simply move one step, or does it run through a number of steps to move the heads one disk track ?

 

[_The Doctor__]

it becomes cumulative... so it may be here and there across the disk.

[+Nezgar]

1 hour ago, cwilbar said:

It should be plainly evident if the different encodings for the stepper throw off the alignment.

Yeah, in theory you'd see the amplitude of the target track decreased somewhat.

 

1 hour ago, cwilbar said:

To step one track, does it simply move one step, or does it run through a number of steps to move the heads one disk track ?

The ROM has to send two steps to the stepper to for a complete track step in a 40-track drive. 1 step per track for an 80 track drive. The stock ROM in a Tandon 1050 actually performs an extra step forward/back at the end of a forward seek (Half physical track), but not when stepping backwards. Atari's WST stock ROM does not do this.

 

17 minutes ago, _The Doctor__ said:

it becomes cumulative... so it may be here and there across the disk.

It would be a consistent 1/4 track offset on a 40 track disk across the entire disk.

[_The Doctor__]

Think about that as the head moves back and forth essentially skipping a beat on each reversal (the correction back step going forward but none on going back)

Edited June 29, 2020 by _The Doctor__