amperage requirements?

[Superdad]

On 11/3/2018 at 7:19 PM, the_doc735 said:

 

If you really want an accurate measure of current draw, then cut the wire and put a fast multimeter/ammeter in series with the circuit. But when you cut one wire to put your meter in series, the land mostly just shifts to the remaining wires where there is less resistance (than the meter) to pass through, that is why this method does not work.”

 

Now you have spliced together things I said separately. I don’t appreciate that. Oh, and “land” was a typo on my part as I of course meant “load.”

 

On 11/3/2018 at 7:19 PM, the_doc735 said:

So, are you now saying that it is impossible to measure the amperage of a 24 pin ATX mobo input @ 3.3, 5, 12v ?

 

Not at all! Clearly you are doing something incorrectly, but without being there we can not say what.

[the_doc735]

1 hour ago, jabbr said:

 

Are you having fun?

 

Would you get satisfaction from figuring this out?

nice to hear from you Jab! - so kind!

Yes I am quite absorbed in this, in my spare time! (LOL).

I've got "the bit between my teeth" now!   ...And I'm not letting go any time soon! (LOL).

Tried lots of wiring configs. that don't work! (lol) ~ 

So, now I'm back to square one (again)   ...never mind!

 

Yes, I'd love to get a definitive answer for measuring amps on a 24 pin ATX mobo, and yes I would get lots of satisfaction from figuring this out! (LOL).

But it seems that even the experts on various forums are dumb founded at giving a definitive answer to this question? (LOL). Sadly paul hynes doesn't have the time to educate me in this process. (shit! - LOL).

The shunt resistor technique intrigues me, but I don't fully understand which resistor(s) to purchase, or where to place them in the circuit? ....OR where to place the probes in "the lane"?

I thought of this:

....then I got stuck! (LOL)   ...no confirmation from PH! (sad!)

Hope you are well and enjoying your Hi-Fi?

Take Care.

Many thanks for your reply!

[the_doc735]

8 minutes ago, Superdad said:

 

Now you have spliced together things I said separately. I don’t appreciate that. Oh, and “land” was a typo on my part as I of course meant “load.”

 

 

Not at all! Clearly you are doing something incorrectly, but without being there we can not say what.

lovely to hear from you again! Apology for the "splice", I am just trying to figure things out at my end (LOL!) - Sorry no offence intended!

I hope you are enjoying your Hi-Fi/listening?

Yes, it's a real shame, but there is no way you can be here matey! 

Also, please see my response 'above' to 'Jab'.

Many thanks for your reply, much appreciated!

Do take care!

[jabbr]

Can’t say because without seeing what you are doing but consider that you have a loose connection.

[paulhynes]

Doc735,

 

Under normal circumstances I like to be as helpful as I can with power supply issues, and whether the questions come from customers, or not, does not affect this.

 

When you asked me how to measure the current requirements of the ATX supply rails I stated that breaking each rail (3v3, 5v and 12v) and inserting a low value resistor would allow measurement of the voltage drop across the resistor under load. This was a general statement and I was not specific about how to do this. Indeed, I told you I was currently unable to help due to pressure of work as I was in the middle of preparations for transferring the workshop and office over to new premises. When I have all the work benches, office desks, equipment and stock transferred and installed, I then have to get on with the work schedule and this will have to take precedence over development projects and any other distractions until the schedule stabilises.

 

What I did not say was that for personal requirements I have in development (currently on hold while I deal with the above mentioned business re-location) a test jig to allow the process of ATX rail current measurement.

 

I can say that the design of the jig allows for having a low value resistor in each individual lead of each high power rail of the ATX loom. I am not a motherboard designer and prefer not to second-guess specialists in fields where I have limited knowledge. Whilst it is probable that all the leads for each rail are connected together at both ends, there is also the possibility that this may not be the case with some motherboard designs either now or in the future if the trend for supply rail separation continues.

 

Having a resistor in each rail lead would maintain balanced current in the paralleled leads so an accurate assessment of the overall rail current would be possible. It would also allow accurate rail current assessment if the leads are powering different sectors of the motherboard without a parallel connection on the motherboard and it would also highlight a broken wire or connection fault in the loom. In other words the jig would be versatile enough to cope with future power supply experimentation requirements.

[gstew]

I haven't read through this entire thread yet, so I may be repeating info already posted... sorry if I'm duplicating info already posted or assuming something is useful and it doesn't apply to the current content.

 

First, for measuring the current of each rail, consider getting extender adapters. I found these online for a reasonable price for both the ATX-24 and P4 connectors. That way you can mutilate the wires on the adapter and leave your current power supply wires alone.

 

I did a DIY'd ATX-24/P4 supply for my computer music player of the time, a cMP/cPlay setup, back in 2010. I've attached pictures of the unit and measuring the current on one of the rails after it was running. For that measurement, I just un-soldered one of the output connections and used a meter that had a 10A setting. Paul's suggestions above would of course be better.

 

I initially used monolithic Linear Technology LT1083 regulators, which was a significant upgrade above a linear-powered PicoPSU, but later moved to Belleson providing a further upgrade. I STRONGLY suspect that a unit supplied by Paul Hynes would be  well above what I did and likely the best available .

 

It took some fiddling to get the target motherboards to fire up with the unit and after trying several, I found that the requirements varied from board model to board model. I'll look for my notes on power-up for the boards I tried and update here if / when I find them.

 

Greg in Mississippi

 

P.S. The form-factor of the unit looks a bit strange, but I configured it to fit inside of the HTPC case I was using replacing the existing ATX PSU without taking up too much additional space.

 

P.P.S. I continued working to optimize the power setup of that system and eventually had additional linear power supplies for the SSD, monitor, keyboard, and mouse. The output was via I2S from a HIGHLY modified Juli@ soundcard's digital portion with a small DAC board (ES9022-based) mounted directly on the Juli@ and fed via it's own set of linear supplies. I was planning a DAC card upgrade and separately powering the video card along with modifying the motherboards for additional power filtering and linear regulation for the memory when some alternative computer music playback setups became available and I moved away from this system.

 

P.P.P.S. BTW, this setup had an EXTREME-MINIMIZED WinXP Op Sys and player setup that was only 15MB or so big!

 

  

[Superdad]

2 hours ago, gstew said:

 

Greg: Don't you know you frighten people with photos like that. Post those pics on your door at Halloween and the children will run away! (Especially like those 6 AC plugs going into one duplex socket in the background.)

[the_doc735]

6 hours ago, jabbr said:

Can’t say because without seeing what you are doing but consider that you have a loose connection.

yes, could be so. I may not have pushed the 24 pin connector in properly because when I removed it a few hours ago, it came out/off far too easily! (LOL).

Cheers!

[the_doc735]

6 hours ago, paulhynes said:

Doc735,

 

Under normal circumstances I like to be as helpful as I can with power supply issues, and whether the questions come from customers, or not, does not affect this.

 

When you asked me how to measure the current requirements of the ATX supply rails I stated that breaking each rail (3v3, 5v and 12v) and inserting a low value resistor would allow measurement of the voltage drop across the resistor under load. This was a general statement and I was not specific about how to do this. Indeed, I told you I was currently unable to help due to pressure of work as I was in the middle of preparations for transferring the workshop and office over to new premises. When I have all the work benches, office desks, equipment and stock transferred and installed, I then have to get on with the work schedule and this will have to take precedence over development projects and any other distractions until the schedule stabilises.

 

What I did not say was that for personal requirements I have in development (currently on hold while I deal with the above mentioned business re-location) a test jig to allow the process of ATX rail current measurement.

 

I can say that the design of the jig allows for having a low value resistor in each individual lead of each high power rail of the ATX loom. I am not a motherboard designer and prefer not to second-guess specialists in fields where I have limited knowledge. Whilst it is probable that all the leads for each rail are connected together at both ends, there is also the possibility that this may not be the case with some motherboard designs either now or in the future if the trend for supply rail separation continues.

 

Having a resistor in each rail lead would maintain balanced current in the paralleled leads so an accurate assessment of the overall rail current would be possible. It would also allow accurate rail current assessment if the leads are powering different sectors of the motherboard without a parallel connection on the motherboard and it would also highlight a broken wire or connection fault in the loom. In other words the jig would be versatile enough to cope with future power supply experimentation requirements.

many thanks for this, I am humbled. Most grateful.

[the_doc735]

5 hours ago, gstew said:

I haven't read through this entire thread yet, so I may be repeating info already posted... sorry if I'm duplicating info already posted or assuming something is useful and it doesn't apply to the current content.

 

First, for measuring the current of each rail, consider getting extender adapters. I found these online for a reasonable price for both the ATX-24 and P4 connectors. That way you can mutilate the wires on the adapter and leave your current power supply wires alone.

 

I did a DIY'd ATX-24/P4 supply for my computer music player of the time, a cMP/cPlay setup, back in 2010. I've attached pictures of the unit and measuring the current on one of the rails after it was running. For that measurement, I just un-soldered one of the output connections and used a meter that had a 10A setting. Paul's suggestions above would of course be better.

 

I initially used monolithic Linear Technology LT1083 regulators, which was a significant upgrade above a linear-powered PicoPSU, but later moved to Belleson providing a further upgrade. I STRONGLY suspect that a unit supplied by Paul Hynes would be  well above what I did and likely the best available .

 

It took some fiddling to get the target motherboards to fire up with the unit and after trying several, I found that the requirements varied from board model to board model. I'll look for my notes on power-up for the boards I tried and update here if / when I find them.

 

Greg in Mississippi

 

P.S. The form-factor of the unit looks a bit strange, but I configured it to fit inside of the HTPC case I was using replacing the existing ATX PSU without taking up too much additional space.

 

P.P.S. I continued working to optimize the power setup of that system and eventually had additional linear power supplies for the SSD, monitor, keyboard, and mouse. The output was via I2S from a HIGHLY modified Juli@ soundcard's digital portion with a small DAC board (ES9022-based) mounted directly on the Juli@ and fed via it's own set of linear supplies. I was planning a DAC card upgrade and separately powering the video card along with modifying the motherboards for additional power filtering and linear regulation for the memory when some alternative computer music playback setups became available and I moved away from this system.

 

P.P.P.S. BTW, this setup had an EXTREME-MINIMIZED WinXP Op Sys and player setup that was only 15MB or so big!

 

  

very interesting comments, thank you. Looking forward to your update!

[the_doc735]

2 hours ago, Superdad said:

 

Greg: Don't you know you frighten people with photos like that. Post those pics on your door at Halloween and the children will run away! (Especially like those 6 AC plugs going into one duplex socket in the background.)

I thought "maybe it's a time machine or something?" (LOL). No offence to greg intended. Looks very complex! ?

[the_doc735]

16 hours ago, Superdad said:

 Clearly you are doing something incorrectly, but without being there we can not say what.

I think you have already said what is going wrong in this statement(?):

"Most ATX connectors are running a number of duplicates of each voltage because otherwise the 18AWG wires used are not heavy enough for the current carried.  In other words, several 12V lines are joined together at the motherboard end.  So when you cut one wire to put your meter in series, the load mostly just shifts to the remaining wires where there is less resistance (than the meter) to pass through.  That is why your method does not work."

I think that is what is going wrong?

 

But please remember prior to this you said:

"If you really want an accurate measure of current draw, then cut the wire and put a fast multimeter/ammeter in series with the circuit…..". But when I did this based on your advice, the result was: "the load mostly just shifts to the remaining wires". "That is why this method does not work."

 

This is why I am confused! Because it looks as though you have said: put DMM in series for an accurate measurement, but this won't work because the load will mostly shift to the remaining lanes (of the same colour).

 

Do you see my dilemma? Do you see my problem here? No offence intended!

I really do need your clarification please?

Many thanks!

Take Care.

[Superdad]

1 hour ago, the_doc735 said:

This is why I am confused! Because it looks as though you have said: put DMM in series for an accurate measurement, but this won't work because the load will mostly shift to the remaining lanes (of the same colour).

 

Well my mind was on one proper way to measure current.  It was not until you came up with very low readings that I thought about it and realized what was happening with several wires of the same voltage running.

And Paul is correct: there are likely some 12V and 5V runs that are discrete, so bundling and spicing for the measurement could cause you some problems.

As before, I think you should just plan on buying several nice beefy power supplies and being done with it.  Or consider getting/building a lower draw PC server or just buy a NUC or some streamer/renderer.  I don't recall that there was anything particularly special about the big PC build you are wanting to power. 

Best of luck,

--Alex C.

[the_doc735]

1 hour ago, Superdad said:

And Paul is correct: there are likely some 12V and 5V runs that are discrete, so bundling and spicing for the measurement could cause you some problems.

 

consider getting/building a lower draw PC server I don't recall that there was anything particularly special about the big PC build you are wanting to power?

 

Best of luck,

--Alex C.

so this won't always necessarily work then? (click link)?

 

Why is mine a BIG PC? For example it's only a mini-ITX board with a 8700T (35W) CPU, fairly small surely? To me it seems tiny because in the past all my PC's have been huge FULL ATX TOWERS, like for gaming with 2 GPU's etc. But by small you maybe mean like a raspberry? iPhone size device? No it's not anything special really! (only dedicated bit being the PF i2s PCIe card). Size wise: "it's all relative", depends what you are comparing mine to! i.e. mine to a raspberry is big, but to a huge gaming rig it is small. (And fanless!).

Cheers!

[the_doc735]

15 hours ago, paulhynes said:

Having a resistor in each rail lead would maintain balanced current in the paralleled leads so an accurate assessment of the overall rail current would be possible. It would also allow accurate rail current assessment if the leads are powering different sectors of the motherboard without a parallel connection on the motherboard and it would also highlight a broken wire or connection fault in the loom.

...I know you are busy with your premises move, and your current workload/schedule, and that you can't take on more projects until next year like you said in your private emails to me. I do respect this situation Paul.

But! ~ I just wondered if I could trouble you for the values of the resistors to place in each of the 11 rails (yellow (x2), red (x5), & orange (x4)?

Very grateful for any help you can bestow upon me!

Good luck with the business move, I hope it all goes smoothly for you!

[paulhynes]

Hi Doc,

 

The resistor values need to be chosen so that they do not drop enough voltage to take the voltage regulation at the load out of the prescribed ATX voltage regulation tolerance.

 

For the individual 12v wires each resistor should be no more than 0R05 (0.05 Ohm) and the power rating of the resistor should be 11W or more.

 

For the individual 5v wires each resistor should be no more than 0R02 (0.02 Ohm) and the power rating of the resistor should be 5W or more.

 

For the individual 3v3 wires each resistor should be no more than 0R01 (0.01 Ohm) and the power rating of the resistor should be 3W or more.

 

There are a number of options available for low resistance value resistors at RS components, Farnell, Mouser etc

 

When you have taken a voltmeter reading across one of the resistors, dividing the voltage (in volts) by the resistance in Ohms will give you the current passing through the resistor in Amps. A peak reading multi-meter with a peak hold function will be useful as this should then give you the maximum peak current passing through the resistor.

 

For example measuring on a 12v wire with a 0R05 resistor :-

 

0v3 (0.3V) / 0R05 = 6A

 

Regards

Paul

[the_doc735]

On 11/6/2018 at 12:57 AM, Superdad said:

I don't recall that there was anything particularly special about the big PC build you are wanting to power. 

Best of luck,

--Alex C.

No it's not anything special really! (only dedicated bit being the PF i2s PCIe card). Size wise: "it's all relative", depends what you are comparing mine to! i.e. mine to a raspberry is big, but to a huge gaming rig it is small. (And fanless!). Mini-ITX mobo.

Also it only has the O/S, plus audio music apps like fidelizer and optimizer, graphic EQ 32 auto levelling, foobar, VBAUDIO, ASIO, etc. Don't forget it only has the bare minimum hardware too, nothing excess to case noise with, and (35W 'T' CPU).

So. it's not too bad really! (LOL).

Cheers!

[the_doc735]

....at mouser (uk) found these: can anyone confirm they are correct, as described by paul hynes in the above post: "Posted Tuesday at 01:15 PM"? For obtaining voltage drop on 24 pin ATX mobo input?

 

https://www.mouser.co.uk/ProductDetail/Vishay-Dale/RH010R0500FE02?qs=sGAEpiMZZMtbXrIkmrvidDNaDpN5VXc5pbnZdRmhvU8%3d

 

https://www.mouser.co.uk/ProductDetail/Welwyn-Components-TT-Electronics/LOB5R020JLF?qs=sGAEpiMZZMtlubZbdhIBIAPP8GYSuulIGAs3qcxRbmQ%3d

 

https://www.mouser.co.uk/ProductDetail/Welwyn-Components-TT-Electronics/LOB3R010JLF?qs=sGAEpiMZZMtlubZbdhIBIEpWRY6HDroi7HyDnPbs3NU%3d

 

many thanks!

[Superdad]

7 minutes ago, the_doc735 said:

....at mouser (uk) found these: can anyone confirm they are correct, as described by paul hynes in the above post:

 

Yes, those match what Paul advised.

[the_doc735]

1 hour ago, Superdad said:

 

Yes, those match what Paul advised.

"Super!"....?

[the_doc735]

On 11/6/2018 at 1:15 PM, paulhynes said:

The resistor values need to be chosen so that they do not drop enough voltage to take the voltage regulation at the load out of the prescribed ATX voltage regulation tolerance............

Regards

Paul

B.T.W. do you still sell the SR4 please?

Cheers!

[paulhynes]

Yes

[the_doc735]

3 hours ago, paulhynes said:

Yes

many thanks!

[the_doc735]

On 11/3/2018 at 11:58 PM, the_doc735 said:

based on this:  (click link)

 

I did this to measure amperage @ 3.3, 5, 12v.

 

....but! ~ it won't boot, keeps resetting every 3--4 seconds?

The other cables are all connected as standard/default. i.e. green, blue, purple, grey, white, black.....

also this thread   (click link)

 

Any ideas why this doesn't work?

@jabbr    ...erm!   it was a loose connection after all. The 24 pin to mobo wasn't driven all the way home! .....ooophs!!

And the readings set with peak hold function are:

12v   ...3.88A

5v   ...5.20A

3.3v   ...3.85A

 

...does that seems accurate now?

Cheers!

[jabbr]

4 minutes ago, the_doc735 said:

@jabbr    ...erm!   it was a loose connection after all. The 24 pin to mobo wasn't driven all the way home! .....ooophs!!

And the readings set with peak hold function are:

12v   ...3.88A

5v   ...5.20A

3.3v   ...3.85A

 

...does that seems accurate now?

Cheers!

Yes.

Consider rounding up to 5/6/5 A