--- > MANDATORY FIXES < ---

The latest tweaks and fixes to keep your machine running in tip-top shape.
ijor
Posts: 171
Joined: Fri Nov 30, 2018 8:45 pm

Re: --- > MANDATORY FIXES < ---

Post by ijor » Fri Apr 22, 2022 8:30 pm

v836 wrote:
Wed Apr 20, 2022 9:44 pm
I found as well on my 520ST that my WD1772-PH 00-02 failed sometime during a ~250hr burn in with 10k pullup resistors on DAL0 through 7. Possibly it was heat maybe this is just too much current. I'm a little nervous about applying a pullup again, if I do so I might add a heatsink. It's a shame to break parts that are getting increasingly rare.
I reckon that some of the other fixes with much stronger pull-ups might be a bit controversial; but the extra current generated in this case, with 10K pull-ups is not really significant.
0.5mA isn't a lot, but it's something.
0.5 ma is about the input leakage current of TTL chips. To give you an idea. The ACSI port in the 520ST is directly connected to the same data bus. That means that connecting an older Hard Disk will probably be more significant to the FDC than these weak pull-ups. Because even when the FDC outputs don't target the ACSI port, they will still sink current from anything connected there when driving the data bus low.
http://github.com/ijor/fx68k 68000 cycle exact FPGA core
FX CAST Cycle Accurate Atari ST core
http://pasti.fxatari.com

v836
Posts: 5
Joined: Fri Apr 01, 2022 7:24 pm

Re: --- > MANDATORY FIXES < ---

Post by v836 » Sat Apr 23, 2022 12:23 am

rubber_jonnie wrote:
Fri Apr 22, 2022 4:55 pm
v836 wrote:
Wed Apr 20, 2022 9:44 pm
I found as well on my 520ST that my WD1772-PH 00-02 failed sometime during a ~250hr burn in with 10k pullup resistors on DAL0 through 7. Possibly it was heat maybe this is just too much current. I'm a little nervous about applying a pullup again, if I do so I might add a heatsink. It's a shame to break parts that are getting increasingly rare.
A word of warning, if I seem to be a bit blunt here, sorry, I've had a bit of a day.

My first question would be "Why on earth are are you doing a 250 hour burn in on an ST?" For what reason would you ever need one to run for that long?

These machines are home computers and 30 odd years old, and even back in the day, they would probably never run continuously for much more than a 10th of that. I expect the Atari burn in was "does it switch on and get to the desktop and switch off again afterwards".

The ICs are going to get hot when run for that long without cooling of any sort, and they were never designed to do that, they are a home computer, not an industrial machine, built to a budget and now suffering deterioration due to age.
...
First, I hope you have a better weekend. :)
Industrial computing is one of the applications for our software, so yeah I got a kick out of you saying these aren't that. :) 250 hours is barely scratching the surface but enough to settle in to whatever steady state it should find itself in. In my world a machine that has inherently limited uptime just for thermal characteristics is a broken machine, and I'll fix it till it isn't broken.
At any rate I suspect you're right about the cause. I've added passive cooling to the new chip just as a precaution, the heatsink warms right up so the heat transfer seems to be effective. I'm just using thermal adhesive (not tape which is near useless and possibly counterproductive). Thanks as well to the other poster for the reminder to look at it in context of whatever else is on the bus, actually that gave me some other ideas.

User avatar
rubber_jonnie
Moderator
Moderator
Posts: 6531
Joined: Thu Aug 17, 2017 7:40 pm
Location: Essex
Contact:

Re: --- > MANDATORY FIXES < ---

Post by rubber_jonnie » Sat Apr 23, 2022 10:08 am

v836 wrote:
Sat Apr 23, 2022 12:23 am

First, I hope you have a better weekend. :)
Industrial computing is one of the applications for our software, so yeah I got a kick out of you saying these aren't that. :) 250 hours is barely scratching the surface but enough to settle in to whatever steady state it should find itself in. In my world a machine that has inherently limited uptime just for thermal characteristics is a broken machine, and I'll fix it till it isn't broken.
At any rate I suspect you're right about the cause. I've added passive cooling to the new chip just as a precaution, the heatsink warms right up so the heat transfer seems to be effective. I'm just using thermal adhesive (not tape which is near useless and possibly counterproductive). Thanks as well to the other poster for the reminder to look at it in context of whatever else is on the bus, actually that gave me some other ideas.
Thanks, a little better today.

Machines from this era are pretty bad thermally, I just fixed a VIC-20 and every single IC is hot to the touch within a minute of powering on. The VIC chip was marginal, and that was hottest of all. My 8 bit Ataris are much the same, but seem to have more thermally resistant chips. Commodore did use a lot of MOS chips, which are generally regarded as quite unreliable.

I suspect you had a similar scenario with the WD1772. Unfortunately these machines were almost disposable, and built to a low price, so I doubt thermals were even a consideration, so whilst I get what you're saying about steady state, it might be a bit of a stretch to have an ST manage it unassisted!

It's amazing that any of these machines run at all at 30-40 years old, but I'm glad they do :)
Collector of many retro things!
800XL and 65XE both with Ultimate1MB,VBXL/XE & PokeyMax, SIDE3, SDrive Max, 2x 1010 cassette, 2x 1050 one with Happy mod, 3x 2600 Jr, 7800 and Lynx II
Approx 20 STs, including a 520 STM, 520 STFMs, 3x Mega ST, MSTE & 2x 32 Mhz boosted STEs
Plus the rest, totalling around 50 machines including a QL, 3x BBC Model B, Electron, Spectrums, ZX81 etc...

Steve
Posts: 1875
Joined: Fri Sep 15, 2017 11:49 am

Re: --- > MANDATORY FIXES < ---

Post by Steve » Sat Apr 23, 2022 10:54 am

When you touch a hot chip, this is experiencing the true 'power without the price'tm

v836
Posts: 5
Joined: Fri Apr 01, 2022 7:24 pm

Re: --- > MANDATORY FIXES < ---

Post by v836 » Sat Apr 23, 2022 3:09 pm

rubber_jonnie wrote:
Sat Apr 23, 2022 10:08 am
v836 wrote:
Sat Apr 23, 2022 12:23 am

First, I hope you have a better weekend. :)
Industrial computing is one of the applications for our software, so yeah I got a kick out of you saying these aren't that. :) 250 hours is barely scratching the surface but enough to settle in to whatever steady state it should find itself in. In my world a machine that has inherently limited uptime just for thermal characteristics is a broken machine, and I'll fix it till it isn't broken.
At any rate I suspect you're right about the cause. I've added passive cooling to the new chip just as a precaution, the heatsink warms right up so the heat transfer seems to be effective. I'm just using thermal adhesive (not tape which is near useless and possibly counterproductive). Thanks as well to the other poster for the reminder to look at it in context of whatever else is on the bus, actually that gave me some other ideas.
Thanks, a little better today.

Machines from this era are pretty bad thermally, I just fixed a VIC-20 and every single IC is hot to the touch within a minute of powering on. The VIC chip was marginal, and that was hottest of all. My 8 bit Ataris are much the same, but seem to have more thermally resistant chips. Commodore did use a lot of MOS chips, which are generally regarded as quite unreliable.

I suspect you had a similar scenario with the WD1772. Unfortunately these machines were almost disposable, and built to a low price, so I doubt thermals were even a consideration, so whilst I get what you're saying about steady state, it might be a bit of a stretch to have an ST manage it unassisted!

It's amazing that any of these machines run at all at 30-40 years old, but I'm glad they do :)
For what it's worth, I'm measuring internal ambient temperatures at the moment with a thermocouple and ambient temps are stabilizing out to 29C after an hour or so (measured near the 1772) which is about 9C over room ambient, and far below the 70C ambient that the WD1772 data sheet says it is rated to, so there is a fair bit of headroom for passive cooling for any chip that is of particular concern. Basically the case ventilation seems to be allowing an adequate airflow, it's not just heating up to higher and higher temps over time. I should mention that there is no internal RF shield on this 520ST, that went away many decades ago (I'm the original owner, this was one of the first wave of retail machines I had preordered it before the initial release of the ST). Basically I don't think crazy runaway overheating is an issue at all here. Maybe things might be different for the versions that include power supplies, floppy drives, etc. inside the case so YMMV.
I'm also very glad to have this machine running all this time later, and appreciate the support and information of exxos it really is a huge help.

User avatar
exxos
Site Admin
Site Admin
Posts: 18041
Joined: Wed Aug 16, 2017 11:19 pm
Location: UK
Contact:

Re: --- > MANDATORY FIXES < ---

Post by exxos » Sat Apr 23, 2022 5:03 pm

v836 wrote:
Sat Apr 23, 2022 3:09 pm
Basically I don't think crazy runaway overheating is an issue at all here. Maybe things might be different for the versions that include power supplies, floppy drives, etc. inside the case so YMMV.
As another suggestion , I wouldn't rule out the PSU either. Just because its branded doesnt mean its reliable. I would scope it out cold and after a very long run and measure its regulation and temperatures. I don't think anyone has done any long term testing if the meanwell PSUs etc a lot of people use them, but I don't think anyone has done any long term tests past "it works".

I've seen PSUs as they warm up the voltage can drift a fair bit. The DMA / 1772 circuit seems to suffer first when the voltage is getting towards 4.5v YMMV. Also it could be increasing to 5.5v+ and if the thing has huge voltage spikes, it can kill chips over time.

But also as explained earlier, "external" things like hard drives or floppy drives which plug into the ST could also be a cause of problems. Also the motherboard could be bending as things heat up and cool down. The soldering on the things isn't great to start with. So you may have to resolder the entire thing to rule that out. I had to resolder a virgin board a few years back as the soldering was a total mess, the machine couldn't even boot. The list of problems is pretty endless. Maybe time for a H5 ;)

Post Reply

Return to “TWEAKS & FIXES”