jamesbeat wrote:

It looks as though an ATX connector is too difficult to remove with the tools I have, or I would have tried it out on this spare board to see what happens.

O.I.C. You are contemplating removing the connector to NOT modify the board. 😕 😕
Pardon my confusion. 😕 😕 😕

I was contemplating the feasibility of doing it - I do not need to modify any of my boards, because (at least for now) all of my Dell boards have working PSUs.

This is a spare board that I don't need, and it has an undesirable PPGA socket, so I was toying with the idea of attempting to modify it just to see if it would work.

As I assume you read above, I have already decided against it, because I don't have the equipment to desolder the connector.

Those adapters might not be available in a few years time.
They appear to be pretty thin on the ground as it is, and I expect the ones that are available are old stock.

Maybe this idea will help a future collector.

I'm not confused about what you were doing but you seem to be

jamesbeat wrote:

so I was toying with the idea of attempting to modify it just to see if it would work.

But then you say you aren't working out a mod.

jamesbeat wrote:

I was contemplating the feasibility of doing it

Of course it's feasible, but as I tried to tell you before it's not as simple as moving some wires around on this one.
You need to check what is going on with the POL (Point of Load) voltage regulators.

Dell moved 3x 3.3v wires to the 6-pin and added 2x 5v wires to the 20-pin.
Each pin (if the standard part, probably is for this) in the connector is rated for 9 amps.
Dell presumably had a reason to add 18 amps of capability to the 5v rail.
Presumably something other than the power connections is also different between the Dell and non-Dell versions.
.

If you want to desolder the Dell-ATX connector for cheap, buy some desoldering braid. It's like a thin ribbon made out of copper, and you place it between the iron and the joint and use it to draw solder away from the joint. It's not as clean as a desoldering vacuum gun and there's a bit of technique to it, but I've used it many times to desolder multi-legged components. The size of the ATX connector would make it a bit easier than with other components like ICs.

Of course the real question is, is it really worth it? I see one big drawback. Someone comes along in the future and sees your Dell board with a real ATX connector, but the power supply has gone. This person knows it's a Dell board which needs a special PSU, so they go to the trouble of hunting down a Dell PSU (or getting an adapter), plugging it in to your board and then bang, it blows up because you have the one Dell in the world that uses a non-standard connector (a real ATX instead of the Dell-ATX that all the other boards use.) I think the risk of this outweighs the benefits.

Not to mention that as time goes on, often it's the 'curiosities' that get a lot of interest. Things like VESA Local Bus and its forerunners, or bus mice that look like they have PS/2 connectors but they plug in to a custom card. You might find that as years go on, these unusual Dells have a lot of interest because they are a bit weird, but if you've converted them to be just like every other retro motherboard then that uniqueness is lost!

Personally I don't think there's anything inelegant about an adapter. If nothing else it points out that you're looking at something a little out of the ordinary. Let's also not forget that even if adapters are hard to come by, you can always rewire the ATX connector on the power supply to turn it onto a Dell-ATX one as well. You can get tools to remove the pins from the plastic ATX housing which makes the process pretty straightforward (also very useful for rewiring cheap ATX extension cables to make new adapters.)

I had the board sitting on my desk at work, and noticed the unpopulated pads next to the connector.

I idly wondered if it would be possible to convert the board to ATX, and briefly considered trying it out on this spare board just to satisfy my curiosity.

I discovered that I can't do it with the tools I have, so I dropped the idea.

I don't understand why you're having trouble wrapping your head around that.

You may be right about the voltage differences being significant, though I can't think of any reason why Dell would have legitimately needed to do this.
These machines don't have anything unusual connected to them that might require a modified power system.

I strongly suspect that the reason Dell 'needed' to do this was because they wanted to sell their proprietary power supplies and motherboards.

Judging by what I have read in old articles, that is what everybody else at the time suspected too.

Malvineous wrote:

If you want to desolder the Dell-ATX connector for cheap, buy some desoldering braid. It's like a thin ribbon made out of coppe […]
Show full quote

If you want to desolder the Dell-ATX connector for cheap, buy some desoldering braid. It's like a thin ribbon made out of copper, and you place it between the iron and the joint and use it to draw solder away from the joint. It's not as clean as a desoldering vacuum gun and there's a bit of technique to it, but I've used it many times to desolder multi-legged components. The size of the ATX connector would make it a bit easier than with other components like ICs.

Of course the real question is, is it really worth it? I see one big drawback. Someone comes along in the future and sees your Dell board with a real ATX connector, but the power supply has gone. This person knows it's a Dell board which needs a special PSU, so they go to the trouble of hunting down a Dell PSU (or getting an adapter), plugging it in to your board and then bang, it blows up because you have the one Dell in the world that uses a non-standard connector (a real ATX instead of the Dell-ATX that all the other boards use.) I think the risk of this outweighs the benefits.

Not to mention that as time goes on, often it's the 'curiosities' that get a lot of interest. Things like VESA Local Bus and its forerunners, or bus mice that look like they have PS/2 connectors but they plug in to a custom card. You might find that as years go on, these unusual Dells have a lot of interest because they are a bit weird, but if you've converted them to be just like every other retro motherboard then that uniqueness is lost!

Personally I don't think there's anything inelegant about an adapter. If nothing else it points out that you're looking at something a little out of the ordinary. Let's also not forget that even if adapters are hard to come by, you can always rewire the ATX connector on the power supply to turn it onto a Dell-ATX one as well. You can get tools to remove the pins from the plastic ATX housing which makes the process pretty straightforward (also very useful for rewiring cheap ATX extension cables to make new adapters.)

Yes, I have desoldering braid, but I'm still reluctant to attempt it.

These are through-hole connections that appear to be through-hole plated, so the solder goes all the way through the connection.

I just don't think braid would wick enough out, and there are so many pins that would all have to be desoldered perfectly before I could remove the connector.

This is all made worse by the fact that there are some extremely fine traces going to the pins - it's a recipe for disaster unless you have proper equipment.

I did try to snap the pins off a scrap board that I have, thinking maybe I would be able to pry off the plastic block and desolder the pins one by one, but it was still stuck fast.

That was the point where I pretty much lost interest, but it wpuld be cool if someone else gave it a try.

You are right about the prospect of a future owner trying to use a Dell PSU - the board would have to be clearly marked to show that the modification had been made.

I'm exploring feasibility and practicality. I could care less if you're going to actually try it or not.
I don't understand why you're having trouble wrapping your head around that.
.

You're probably right.

I might solder wires directly to the pins and the board to test it out one day, but I'm in the middle of restoring five machines and building a sixth, so it's not high on tthe priority list.

Re the voltages, perhaps Dell had a surplus of PSUs designed for the earlier +5V heavy machines, and some engineer was convinced that newer machines would need extra +5V power?

I forgot to mention it's a really interesting find that 1) it's not a completely custom connector, it's just shifted a bit, and 2) that there's an Intel board that supports it at the trace level. I've seen Intel boards with unpopulated 6-pin headers but I was always confused because they didn't seem to match the 6-pin connector that briefly appeared on some ATX supplies. I guess now I know, it was probably there for Dell!

The trick to desoldering these types of things with the braid is that you need a way to move the pin as it's cooling down. The braid generally wicks all the solder away, and you're just left with a thin coating on the pin and PCB which sticks together. But if you can wiggle the pin while it's cooling down (or even hold it in the middle of the PCB hole if it's large enough) then you can get the solder to harden without making contact with the pin. If you can do this for each pin one by one then the connector will slide out once you get the last pin. I'm not saying it's easy, it's still very frustrating, but it's doable.

I do have a vacuum desoldering gun and while it is much quicker to use than desoldering braid, you still end up with the same problem when you're trying to desolder a multi-pin component (as in the residual solder sticks when it cools.) It also doesn't work well unless you can put the nozzle completely over the PCB hole and suck air through it, to suck all the solder out. If you don't get a good seal it won't suck out the solder in the PCB hole. The hot-air gun is useful for smaller components, but for larger ones it's too easy to start burning the PCB before the pins get hot enough to all melt at the same time, and it makes the component you're desoldering get *really* hot (so probably not great if you want to reuse it in something else.)

Only basic soldering gear is needed for this but the right basic tools do matter.

Solder braid comes in different sizes. (widths)
For something like this you want a wide type so the braid isn't completely full before solder is removed.
Add flux to the braid. The pre-fluxed types generally don't have enough flux so you end up adding anyway.
Also the iron needs to be one with enough grunt to heat all the way through the board. Say 60-80 watt.

First melt the solder all the way through (preheat) THEN bring in the wick before it cools too much in the middle.
If not then you may end up with 1/2 empty holes because the solder inside the board is too cold.
-- Wick works by capillary action so it's not going to pull anything from a 1/2 full hole.
-- Preheating is not usually necessary for component leads or smaller pins but it helps there too.
If a hole doesn't come clean then add some solder to that one and start over. (Adding also helps with lead-free solder.)
With practice you can clean the solder out completely leaving a tinned hole.
When you go to pull the connector some pins may be stuck to the sides of holes. A sideways push with a hot iron will free them up.
https://www.youtube.com/watch?v=htrcZuK_ZsY
https://www.youtube.com/watch?v=Z38WsZFmq8E
.

Something like this would work https://www.ebay.com/itm/Athena-Power-CABLE-D … in/160976241947?

oeuvre wrote:

Something like this would work https://www.ebay.com/itm/Athena-Power-CABLE-D … in/160976241947?

The theory is that you might be able to modify the power connector on the board itself, thus removing the need for those adapters.

@ jamesbeat
Do you know the part number for the stock Dell PSU?
I want to look at the rail ratings to see if all those extra pins are even necessary.

I wouldn’t have thought that connector would be too difficult to remove with a desoldering pump iron and/or braid (I usually prefer the former for multi pin components such as ICs and connectors like this)

Sometimes with a very stubborn pin on a ground plane I use the desoldering pump with another iron on the same joint but you can easily lift a pad that way on a fragile board.

Failing that you could always redneck it and Dremel the old connector off and then remove the pins one by one but one tiny slip and you could end up with a whole lot more work or even knacker the board.

Perhaps a soldering iron with an old, wide tip could be used to remove the plastic part instead, with a bit of persuasion from some pliers, then remove the pins individually, clean up the pads and Bob’s your uncle. I like the idea of replacing the connector instead of using an adapter personally, much cleaner.

PCBONEZ wrote:

@ jamesbeat
Do you know the part number for the stock Dell PSU?
I want to look at the rail ratings to see if all those extra pins are even necessary.

The PSU for that board is in my office at work, so it will have to wait until tomorrow.

Dell NPS-200PB-123A

Dell PS-5201-1D

oeuvre wrote:

Dell NPS-200PB-123A

Dell PS-5201-1D

Mine's different:

HP-P1457F3

If it helps, the part # is 8765D

Thank you everyone.

The pins typically come in 9a, 11a, 13a but some smaller manufacturers do make other sizes.
At some point the ATX PSU Standard started specifying 13a pins but much of the industry ignores it.
Since there is no way to tell which pins got used the safe bet is to assume 9a pins.

This next is the number of power pins needed to handle full PSU output using 9a pins.
The other voltages are all good with 1 pin.

PS-5201-1D [200w]
+12v = 6a => 1 pin
+5v = 22a => 3 pins
+3.3v = 14a => 2 pins

NPS-200PB-123A [200w]
+12v = 10a => 2 pins (Allot 1a+ to peripheral (HDD/ODD) then 1p is still fine in the main power connect.)
+5v = 22a => 3 pins
+3.3v = 14a => 2 pins

HP-P1457F3 / 8765D [145w]
+12v = 3.5a => 1 pin
+5v = 14a => 2 pins
+3.3v = 8a => 1 pin

Hope that helps.
.

Thanks for investigating this. I've noticed the shift in the ATX connector on these boards before, but was too lazy to confirm the matchup of the pinouts. The pinout charts you posted show that it really is that simple.
Dell did this altered pinout thing back in the slot-1 and PPGA370 (Mendocino Celeron) days. At that time it was normal for PSUs and motherboards to be +5V centric. Perhaps that was a valid motivation for them making this change, but with these boards I agree it doesn't seem the change was really needed.
By the time everybody started making more use of the +12V rail, Dell had long since switched back to a standard ATX connector (they did that during the P3 days).

Bonez, thanks also for the info on the amp capacity of the pins.

I have a few of this era of Dell boards and at least one, maybe 2 Intel versions. I don't know if I have any exact matching models in both Dell and Intel versions, but if I do, maybe I should take a close look at them and see if I can find any other differences in the voltage regulation circuitry.
I've thought about using one of the Dell boards for a build, but the point for me would be to utilize a matching Dell case and PSU so keeping it Dell would actually be more convenient in my case.