Given the pinouts in here (System Overview 1-9 to 1-12)

I'm guessing something like this is what you'd need

https://www.atxpowersupplies.com/Dell-24-to-2 … -16-adapter.php

* it lists the 410 at least in the compatibility table at the foot of the page

Ahh, thank you. Nice find there. $19.99 + $36.99 shipping to Canada. I may just make my own with one of those sub $3 20-pin ATX to 24-pin ATX connectors, though that will require a lot of cutting and soldering. Is it easy to get the cable out of the connector without breaking the tabs? If so, I'd only need to 3 cables at most.

Get the ATX connector extractor tool, From UK.

Cheers,

Thanks. Apparently you can use two staples to remove the ATX pins. https://www.youtube.com/watch?v=GKh_gUBQwvw

Good info. I just tried using staples on one of my ATX extenders, the micro clips were slightly different and could only get one very small staple in due to lack of space inside the male end. I think it would be easier on the female side though have not tried yet.

Experience tells me it is best to have a properly made tool. I had to carefully file down my spare tweezers just right and tips (strong and thick, lightly beveled to slide into between metal connector easier and push over the hooks without digging in, shaped just right then things works properly. Tricky if you are using home made extractor.

Also strength of the tips is essential. Staples are too soft.

Each tine of the tool It is slightly thicker 1mm or so and about 1.5 mm wide and nearly 2.5 cm long.

The ones from UK are properly made just like mine, the others that sells tool made of round tube and slit to make two tines is not good made.

Insert tool first all the way in, then push the wire and tool together into plastic connector this will be felt when wire move into little further and that helps tool to seat fully, by few wiggles, then tug wire out.

Cheers,

I finally received my $2 ATX-20 (female) to ATX-24 (male) adapter from eBay. The female end plugs into the non-Dell PSU, while the male end plugs into the Dell Precision Workstation 410 or 610 motherboard. I was able to use two staples to very easily remove all the ATX pins on the adapter male end. No special tool required. Perhaps it was so easy because the $2 adapter cable is a cheap piece of junk, implying that the spring force on the retainers is rather weak.

At any rate, the pinout as supplied in the Dell Precision Workstation 410 and 610 Service Manual is incorrect. Not one, but three assignments are wrong. I was shocked and a little applauded that Dell never bothered to fix this in their electronic version PDF file. It is a good thing that I have an actual Dell PSU that came with a Precision Workstation 410 to verify this mistake. The mistakes are shown below:

After figuring out the location of the motherboard's header for speaker and power_on, I installed a Xeon PII 450, a terminator card, the CPU case fan, single stick of 256 MB ECC REG PC100 memory, floppy drive, and an AGP Diamond Fire Pro 1K (came with the Dell 410 Workstation). I am using a 300W AOPEN PSU attached to the re-wired ATX adapter, attached to the Dell MB.

The specs of the 300W AOPEN are:
3.3V = 20A
5V = 30A
12V = 13A
-5V = 0.3A
-12V = 0.8A
5Vsb = 2A
Combined +3.3V & +5V = 175W

The specs of the 330W Dell NPS-300GB B Rev 01 are:
3.3V = 18A
5V = 35A
12V = 14A
-5V = 0.3A
-12V = 0.3A
5Vsb = 1.2A
Combined +3.3V & +5V = 230W

Using the Aopen PSU, I powered on the Dell 610 motherboard, but received a halt error at POST; and it won't let me into the BIOS. The Dell PSU worked fine, so the motherboard works. The error was:

1Alert! Power supply fan failure detected.
2System halted!

The AOPEN power supply doesn't have a pin assignment called TFSC, which stands for "thermal fan-speed control", so I figured that this must be the cause of the system halt. Since my Dell PSU comes from my cased Dell Precision Workstation 410, I wanted to have a benchtop CPU for this Dell 610 Xeon motherboard and not have to remove the PSU from my case. So I started investigating ways to trick the Dell 610 motherboard into seeing the signal it wants to see so that the motherboard would boot.

I decided to run the Dell PSU on the benchtop without the motherboard connected to it. I used a solid core 18 AWG wire to short the PS_ON pin to GND on the DELL PSU's open cable. I also connected a floppy drive to the PSU, otherwise the PSU detects no load and turns off after 10 seconds. I then connected my oscilloscope to the TFSC pin to see what kind of output the PSU is generating on the TFSC pin. Note that the manual calls the TFSC a "power-supply input signal used to control power-supply fan speed in special applications." I find this curious because this pin is definitely outputting a square wave. Oscilloscope screenshots are shown below. The top half shows the waveform for the TFSC pin without the motherboard connected, while the bottom shows it with the motherboard connected and powered-on.

Measurements on the scope identify the waveform as a square wave having a 50% duty cycle, with frequency of around 57 Hz, and a peak Vpp of 900 mV. I decided I would use a common 555 timer IC in astable mode to mimic this waveform. While there are dozens of ways to create simple oscillators, I wanted to pick the method which required the least soldering, least head cracking, and to use only components I had on hand.

Attached is my simple schematic, to be spliced into the ATX20-ATX24 adapter cable. Values for R1, R2, and C1 affect the duty cycle and free running frequency and were selecting accordingly. There are several ways to play with these values, even using entirely different values to reach the same waveform, but since I wanted to use what was on-hand, this is how I arrived at my starting point. From the theoretical calculation shown on my diagram, values of R1 = 499 ohms, R2 = 32.4 Kohms, and C1 = 0.33 uF should result in a 5V square wave with a frequency of 67 Hz and a 50% duty cycle. To drop the waveform's voltage from 5 V to around 900 mV, I am using a simple voltage divider circuit with Ra = 10 Kohm and Rb = 2 Kohm.

Here is a photo of the components soldered onto the NE555 package:

I tested this on the benchtop out of motherboard first and the resultant waveform as expected. So I soldered it onto the ATX20-ATX24 adapter cable, being careful not to sever any strand of wire with a mid-cable splice.

Also shown in motherboard:

Below are images from the oscilloscope. The top row of images show the waveform with just the PSU powered solo (no motherboard), while the bottom row shows it with the motherboard connected and powered up. The in-board 555 timer waveform is running at 57 Hz, just as was the case with the Dell PSU. The waveform voltage from the Dell PSU in-board has about 60 mV more of a DC offset compared to the 555 timer in-board. The Dell PSU's in-board waveform has a peak of about 900 mV while that of the 555 timer in-board has 700 mV. This difference did not effect operation. To ensure a sufficient safety margin, I swapped Rb for 1 Kohm instead of 2 Kohm, which brought the peak voltage down to 400 mV and it still did not affect operation, thus I left it at 2 Kohm.

The ATX adapter with spliced in 555 timer has solved the issue with the system halting due to a detected power supply fan failure, however in comparison to using the Dell PSU, there is another error:
Alert! Previous voltage failure.

This error does not prevent the system from booting, but I am wondering why the error is there. When the system is powered up, I measured all the voltages on the ATX-24 cable and they are all within 2% of nominal. One thing I did notice, though, was that the cables from this cheap ATX adapter get warm when the system is powered on. A good amount warmer than the cables on the Dell power supply. I can tell that the manufacturer skimped on this adapter as some wires are even 20 AWG instead of 18 AWG. Perhaps the Dell motherboard is detecting a drop in voltage over time as the cable heats up? Maybe I should have bought the $10 cable, but shipping to Canada was too high. Anyone else have any idea other idea on this 'voltage failure' alert? Is the Aopen PSU too weak?

That link provided by PC Hoarder Patrol shows a photo of a Standard ATX-20 to DELL ATX-24 adapter, however I do not see in the photo any kind of timing circuit on the TFSC pin. I wonder what they did, if anything, to solve this issue?

Some photos of the Aopen and Dell power supplies.

feipoa wrote on 2020-08-31, 08:37:

One thing I did notice, though, was that the cables from this cheap ATX adapter get warm when the system is powered on. A good amount warmer than the cables on the Dell power supply.

This is your problem!
the write quality of your adapter is bad so when your computer is starting up, the voltage is drop way to much which cause the voltage error.
I think you should find some local computer shop that do wire modding and asked them to build the adapter for you or buy the crim tool and build the adapter yourself.

Thanks for your input. I agree, the crummy wire quality is probably the issue. The wire appears to be pure copper at least. I didn't think the motherboard would measure the transient voltages, but rather the steady-state ones. I guess it measures transient as well. The simplest remedy would be to get a higher quality ATX20-ATX24 adapter cable and mod it. I do have one, but I want a 2nd for the modification. This adapter is intended for the testbed, not to go into a system.

EDIT: I wonder if soldering the crimp will help at all. Most of the wires are 18 AWG, but some of the black ones are 20 AWG. 18 AWG is what most of these adapters use. I haven't seen one with 16 AWG.

I don't think the quality of cable matters that much for such a short wire run. I am not sure its measurable when we are talking about a couple inches.

Cover removed error sounds a lot like anti tamper switch. some cases have, and Id be willing to bet this board has a header for one.

Hard disk thermal probe, it couldn't possibly know that, only thing that I can think of is there is a header for that too. Or it had a card that had that on it that the bios could read.

Is there a hardware error log in the bios? try clearing it.

Yes, these Dell workstation cases came with a switch on the cover, which is normally closed. To get rid of the case error, I just need to short two header pins. These cases also came with a thermal probe on the HDD enclosure. These two errors I don't care about because they are expected. The voltage error is the one I'd like to fix.

There is no hardware error log in the BIOS, but using the DELL PSU makes the voltage error disappear, so the issue is related to using a non-Dell PSU and an adapter. I made sure there aren't any bulging caps in the Aopen PSU. I also tried using a cheapo 400W supply and the voltage error is still there.