ATX PSU in PEB mod

[Sinphaltimus]

 

Well, I completed the installation of an ATX power supply in my PEB, and I started experiencing strange issues. First, DSK2 stopped being accessed, only DSK1 and DSK3. Then the power supply wouldn't even start if the pcode card or the HDX modified RS232 card was in the PEB. The IDE card worked fine.

The Geneve wouldn't even start up.

I have the 5V and +12V lines both connected to the PS 12V, and the -12V line to the PS -12V. Ground to ground as well.

It seems no one else has had these issues. What gives?

I switched back to a backup PEB with the original PS and everything was back to normal...

 

Hrm. I'm assuming you've verified the PSU through and through. Maybe pull your cards and check end to end traces from PSU to car slots?

 

What's the rating of the PSU you chose?

[+Vorticon]

400W. However, upon further inspection, it's not outputting the full +/-12V. I get with no load 11.8V and -10.9V. That does not sound quite right, but I'm not sure that it's enough to cause the issues I'm experiencing...

[Tursi]

Well, I completed the installation of an ATX power supply in my PEB, and I started experiencing strange issues. First, DSK2 stopped being accessed, only DSK1 and DSK3. Then the power supply wouldn't even start if the pcode card or the HDX modified RS232 card was in the PEB. The IDE card worked fine.

The Geneve wouldn't even start up.

I have the 5V and +12V lines both connected to the PS 12V, and the -12V line to the PS -12V. Ground to ground as well.

It seems no one else has had these issues. What gives?

I switched back to a backup PEB with the original PS and everything was back to normal...

I guess I'd check the 12v line when it's in the failure mode -- maybe it's drawing too much? Are you running the drives off the drive power leads? Strange symptoms though. :/

[Stuart]

400W. However, upon further inspection, it's not outputting the full +/-12V. I get with no load 11.8V and -10.9V. That does not sound quite right, but I'm not sure that it's enough to cause the issues I'm experiencing...

 

Apparently the +/-12V outputs of a PC switched mode supply normally only approach +/-12V when under a reasonable load. With no load then they'll be less than +/-12V as you have seen. One solution apparently is to wire in a car brake light or two to give a bit of load, or a couple of high power resistors on heatsinks.

[+Schmitzi]

So, an AT-power supply (instead of ATX) maybe would be better (as this is "unregulated" ?) ?

Or does this not deliver the needed voltages ?

[Stuart]

So, an AT-power supply (instead of ATX) maybe would be better (as this is "unregulated" ?) ?

Or does this not deliver the needed voltages ?

 

I don't think an AT supply is "unregulated". As far as I know they work in pretty much the same way, but the form factor (size) and connectors are different.

[Sinphaltimus]

It must be that the gotek display, temp gauges and probes provide just enough load so I don't experience this by chance. I am powering 2 floppy drives, a gotek with display and 3 cards (ram,controller,interface).

[Tursi]

My system works fine with an ATX supply and just the 32k, RS232, and two floppy drives. I doubt it's a case of too little load.

 

I'm more wondering if he's on one of the stand-by lines by accident, but rather than guessing I was hoping to get some data first.

[+Vorticon]

My system works fine with an ATX supply and just the 32k, RS232, and two floppy drives. I doubt it's a case of too little load.

 

I'm more wondering if he's on one of the stand-by lines by accident, but rather than guessing I was hoping to get some data first.

 

What do you mean by the standby line?

Aside from measuring the voltages and making sure they are being supplied to the right pins, I'm not sure what else to test for short of trying another ATX supply. The one I have looks pretty cheap...

[Sinphaltimus]

 

What do you mean by the standby line?

Aside from measuring the voltages and making sure they are being supplied to the right pins, I'm not sure what else to test for short of trying another ATX supply. The one I have looks pretty cheap...

Check THIS out. Seems legit.

 

"An ATX PSU requires what is called a Power Supply or PSU On signal (Active low) to actually turn on. It gets this signal

from the motherboard via the green wire connected to pin 14"

 

EDIT:

From above link.

"It is important to note that the + and - 12V supplies are normally not considered by the PSU before it asserts a

Power Good\OK signal! Understand that just because you have a Power Good\OK signal, does not guarantee +- 12Vis ok, although with a well designed PSU, there is a high probability that most of the PSU circuitry for these voltages is OK.

Edited December 9, 2016 by Sinphaltimus

[towmater]

Could you add a well heatsunk nine-volt regulator inside the PSU?

[+InsaneMultitasker]

 

Check THIS out. Seems legit.

 

"An ATX PSU requires what is called a Power Supply or PSU On signal (Active low) to actually turn on. It gets this signal

from the motherboard via the green wire connected to pin 14"

 

EDIT:

From above link.

"It is important to note that the + and - 12V supplies are normally not considered by the PSU before it asserts a

Power Good\OK signal! Understand that just because you have a Power Good\OK signal, does not guarantee +- 12Vis ok, although with a well designed PSU, there is a high probability that most of the PSU circuitry for these voltages is OK.

 

I recall having to do something similar when I converted my system to a PC power supply. I say similar because I am not able to confirm it at this time. The wire was green... and without it connected, the power switch did not activate the supply.

[Stuart]

I think Vorticon has got that bit sussed out already otherwise the PSU wouldn't be working at all.

[Tursi]

Yes, you need that power-on line, but he gets power, so it's not that... presumably.

 

Just put a voltmeter on it while it's faulting and see if it's still delivering 12v, that's the first step.

[+InsaneMultitasker]

Yea, I misread an earlier post while using my phone to browse the history.

[+Vorticon]

Yes, you need that power-on line, but he gets power, so it's not that... presumably.

 

Just put a voltmeter on it while it's faulting and see if it's still delivering 12v, that's the first step.

 

Yup, I am quite familiar with the power-on line. No that's not it...

I'm particularly curious as to why DSK2 would not get accessed though. Does not make any sense.

I'll do some further testing this weekend and see what I can tease out. One thing I did note was that the cards causing the issues (pcode and RS232) have original regulators in, whereas all my other cards except the PEB interface have newer ones (IDE, SID Master, SAMS, TI disk controller)...

[+Vorticon]

Well it looks like the drives were not getting enough power to run both of them. DSK3 was working because it was external and had its own supply... When I powered them from an external PS, everything worked fine. The ATX PS I'm using looked really cheap and I've had it for a number of years sitting around in a box. The best part, as I was running further tests, the ATX sparked and tripped one of the house's circuit breakers. Upon closer inspection, both of the large capacitors appeared to be leaking. So much for that

Time for a quality supply to enter the picture...

[Sinphaltimus]

Glad you found the issue.

[wyerd]

You might want to look at my PSU replacement as described here. http://atariage.com/forums/topic/234391-how-to-replace-peb-psu-with-new-atlpx-psu/

[+arcadeshopper]

Well it looks like the drives were not getting enough power to run both of them. DSK3 was working because it was external and had its own supply... When I powered them from an external PS, everything worked fine. The ATX PS I'm using looked really cheap and I've had it for a number of years sitting around in a box. The best part, as I was running further tests, the ATX sparked and tripped one of the house's circuit breakers. Upon closer inspection, both of the large capacitors appeared to be leaking. So much for that

Time for a quality supply to enter the picture...

 

[+Vorticon]

Exactly!

[+Vorticon]

Well good news. I replaced the PSU with a better quality one and everything now works as it should including the pcode card. But here's the twist: I did not take the PSU apart but installed it whole inside the PEB because it actually fits just fine in there. A little added safety The only thing I had to do was cut off with a Dremel the transformer stands which were riveted to the bottom of the PEB so that the PSU will be flush with the floor then I just Epoxied it there. The height of the power plug is exactly similar to the original plug hole. The one potential issue is that the fan faces the front of the PEB, but there is quite a bit of clearance inside the enclosure to create decent airflow nonetheless and I left a good 1 cm space between the PSU and the fan grille in the PEB. We'll see...

[Sinphaltimus]

Interesting. The thought never even occurred to me. I would put a temp gauge in the psu since that's going to be your primary heat source I think. Pumping hot seit in to the peb and trying to push it out with same fan my create hot pockets (Not the snack food).

 

Maybe consider a case fan if you still have space for it?

[+Vorticon]

I might consider installing a small fan on the side of the PEB at the level of the PSU. But I'm still not sure if heat is going to be excessive at this point. More testing is needed.

[CC Clarke]

After reviewing this interesting thread, a couple of observations come to mind:

 

1. The TI peripheral clamshells are not effective heat sinks; they serve to protect the encased circuit boards. As such, they actually retain a certain amount of heat. The (heat-generating) voltage regulators are usually mounted to the bottom - near the edge connectors; a more efficient design would place them near the top with venting provided. It is what it is.

 

2. Using a fan to circulate air through the PEB enclosure basically turns it into a mini, desktop convection oven. Sure, positively-displaced air will vent through any exterior openings, but the OEM fan is designed to extract air from the interior and exhaust through the provided vent on the rear of the PEB as the most efficient air path. Nearly all power supply cooling follows this tried-and-true, established practice. Heat, --being the sworn enemy of electronics, should be removed as completely as possible. For the sake of extended reliability, I would re-visit the fan mounting methodology.

 

3. There's no such thing as sort-of pregnant. Likewise, the same applies to the proper output of regulated power supplies. They either are or they aren't. Regulated power supply output voltages aren't under-rated as far as the load is concerned. If a 12 volt supply not under load, isn't --then it's got issues. If a 12 volt supply is rated as such, that's what it's supposed to be, regardless of the load, except. . .

 

Increasing the load will have the effect of making the power supply current increase. At some point, the current will exceed the rating of the components and you'll end up with a tripped breaker, blown fuse, smoked component, or melted wire/connector - depending on the protection scheme employed.

 

4. From the pics I see, it looks like electrical tape was used to insulate the heavier gauge wiring. (Correct me if I'm wrong.) If so, electrical tape is for electricians, --not electronic assemblies. Heat and age cause the adhesive to become loose and gummy. I recommend using the proper diameter and voltage-rated heat shrink (preferably polyolefin) for any splicing or wire termination.

 

5. Heat measurement of an interior volume can be subjective, and sometimes deceptive, which is why thermistors are placed on the heat-sensitive CPUs, etc. as well as the interior of a system. They both have their place, but with a properly-operating fan drawing in fresh air, the surface of the more sensitive components may or may not be greatly affected -depending on location and mounting scheme to the PCB.

 

6. Securing any components that are involved with lethal (or non-lethal) voltages with epoxy is not a recommended practice, even if their wiring is insulated. General-use epoxy is easily broken down by heat over time. Using hot glue guns are not recommended either. High-heat epoxy exists, but this isn't the proper way to secure assemblies. Instead, a better method is to use fasteners and insulated stand-offs. It's more work, but it's the right way to ensure a reliable, long-lasting, and trouble-free mod.

 

If it isn't obvious, I'm a stickler for safety, --and for good reason, after experiences I've been involved with or witnessed. The safety-related tips I'm passing on are important in that someone who is unfamiliar with wiring things up can easily and unknowingly make a fatal mistake or damage hardware if their reach exceeds their grasp.

 

I assist a lot of folks on different types of online forums, and the over-riding concept I always make a point to highlight in procedures or advise I put out has to do with safety. One simply doesn't know what another's technical ability is when the (helpful) things we post can literally be viewed by others all over the world. While the ability to be able to reach a far-ranging audience is a pretty cool concept, it carries a certain level of responsibility by the people trying to assist others.

 

Safety first!

 

CC

Edited December 11, 2016 by CC Clarke