When I connect the 1.3 PSU outputs to the Soekris power input side both the positive and negative boards are not working like the LEDs are not glowing. But when I remove the Soekris board they lit and output voltage is stable +/-12v. I am using a small connections to extend the wiring purposes. I followed the Soekris DAC board power supply connections from the hifiduino site from the below pic. The only difference from the hifiduino pic below is that the 0 from each of the positive and negative goes independently to the Soekris 0,0. Do I need to tie up these 0's from both the positive and negative boards and then take 2 wires to the Soekris 0's?
Thanks
Thanks
Last edited:
First double check the Mosfet tabs are well insulated to sinks/chassis. Use the DMM's continuity mode beeper sound for that. Do the test before you wire their outputs to their load.
The zeroes of the pos & neg UBiB outputs must meet either with a link locally at their output connectors or at the load's zero power node. In any case the DMM must show almost zero Ohm between UBiBs 0 & 0 when everything is wired up.
The zeroes of the pos & neg UBiB outputs must meet either with a link locally at their output connectors or at the load's zero power node. In any case the DMM must show almost zero Ohm between UBiBs 0 & 0 when everything is wired up.
Hi Salas
I checked all the shorting with the mosfets and they seem to be fine as I have mounted them on the bottom plate of my cabinet with nylon screws with the washer properly insulated.
Seems that the negative board is the culprit with the Soekris DAC board connected both the positive and negative boards LEDs are switched off but without the Soekris connected they work fine. Looks like the Soekris is pulling more than what the negative board is providing which was my initial issue of not able to set to -12v with the 62R 22W resistor with the max voltage coming to only -9.5v. Because of this I am presuming that both the PSU boards are going to protection mode and not powering up. When I tried the old 1.1 negative board with the new 1.3 positive board with the same combination it works fine the Soekris DAC board. How do I resolve the negative board voltage issue?
Thanks
I checked all the shorting with the mosfets and they seem to be fine as I have mounted them on the bottom plate of my cabinet with nylon screws with the washer properly insulated.
Seems that the negative board is the culprit with the Soekris DAC board connected both the positive and negative boards LEDs are switched off but without the Soekris connected they work fine. Looks like the Soekris is pulling more than what the negative board is providing which was my initial issue of not able to set to -12v with the 62R 22W resistor with the max voltage coming to only -9.5v. Because of this I am presuming that both the PSU boards are going to protection mode and not powering up. When I tried the old 1.1 negative board with the new 1.3 positive board with the same combination it works fine the Soekris DAC board. How do I resolve the negative board voltage issue?
Thanks
Last edited:
12V/62R=193mA load. 0.6V/3.9R is 153mA limit setting. Not to can serve that dummy load demand is expected. If your Soekris revision model pulls more than 150mA on its negative rail is another thing and better confirm from its spec that it should or not.
If it should, first measure voltage drop between R1 legs on the negative reg to make sure the relative semiconductors are good and its near 0.6V, then add a parallel resistor to it for making their total value smaller thus allowing a higher current limit that serves the load's negative rail spec with 100mA headroom. The formula is R1=0.6V/A.
Example, say you want to set 200mA current limit in UBiB for 100mA load + 100mA headroom, then R1=0.6V/0.2A=3 Ohm.
If it should, first measure voltage drop between R1 legs on the negative reg to make sure the relative semiconductors are good and its near 0.6V, then add a parallel resistor to it for making their total value smaller thus allowing a higher current limit that serves the load's negative rail spec with 100mA headroom. The formula is R1=0.6V/A.
Example, say you want to set 200mA current limit in UBiB for 100mA load + 100mA headroom, then R1=0.6V/0.2A=3 Ohm.
R1 resistor type/brand
Hi there
Sorry if this already have been discussed, but what type of resistor would you guys recommend for the R1 to set CCmA?
Would it be fine to use a Vishay PR03 series, 3w 250 PPM Metal Film Resistor, like this one:
PR03000203008JAC00 Vishay / BC Components | Mouser Denmark
I am to use a 2.7, 3 and 3.3 ohm resistor in 2-3w config...
Tea-Bag from the GP also mentioned a wire wound resistor as a possibility but he wasn't sure if it would have any impact at this location in the circuit.
Have a good weekend out there
Hi there
Sorry if this already have been discussed, but what type of resistor would you guys recommend for the R1 to set CCmA?
Would it be fine to use a Vishay PR03 series, 3w 250 PPM Metal Film Resistor, like this one:
PR03000203008JAC00 Vishay / BC Components | Mouser Denmark
I am to use a 2.7, 3 and 3.3 ohm resistor in 2-3w config...
Tea-Bag from the GP also mentioned a wire wound resistor as a possibility but he wasn't sure if it would have any impact at this location in the circuit.
Have a good weekend out there
Last edited:
Hi there
Sorry if this already have been discussed, but what type of resistor would you guys recommend for the R1 to set CCmA?
Would it be fine to use a Vishay PR03 series, 3w 250 PPM Metal Film Resistor, like this one:
PR03000203008JAC00 Vishay / BC Components | Mouser Denmark
I am to use a 2.7, 3 and 3.3 ohm resistor in 2-3w config...
Tea-Bag from the GP also mentioned a wire wound resistor as a possibility but he wasn't sure if it would have any impact at this location in the circuit.
Have a good weekend out there
All low ppm metal film or wire wound in low inductance style or thick film power resistors (say Caddock) are good and acceptable for R1
Its up to what the load better needs. When having few cm of wire between the UBiB and its load, say 5cm, it should remain stable no mater how its locally decoupled. In any case do check the rails are still steady straight lines without oscillation wave forms if you got a scope.
Thanks Salas, I checked out again and found that the TO-220 washer on the mosfet tabs were giving shorting to the metal screw. So removed all the metal screws on the mosfets mounting and replaced them with the nylon ones. After this when I connect the Soekris DAC board (and mine is 1021 revision 2 or 3 I presume) the DAC gets powered up but the positive voltage drops from +12v to +4.5v immediately on power up. I can also see that the LEDs on the positive board starts with nice brightness but immediately drops the voltage to +4.6v.
On the positive board for R1 using 2.2R 3W and negative board 3.9R 3W. I have rechecked all the mosfets for shorting and they look fine. Replaced the mica with the new silicon type after cleaning the mounting area for any residual metal pieces. Please find the BOM which dimdim provided of his build and I have used the same with the exception of the brand of caps and resistors maybe but the values are exactly the same.
Yes I just measured the voltage drop across R1 of positive and negative boards and its 595mV. This drop is the same with the Soekris board connected and without both for positive and negative boards.
0.595V/2.2R=270mA. When your positive reg can also put 12V on your 62R dummy load that's almost 200mA output. Done and tested. You should look elsewhere in the chain for the problem. Maybe your wiring is wrong or your Soekris has a short, when its nominal peak consumption rail spec is covered by the reg's setting.
I have double checked the Soekris DAC board and actually redid the mounting today with all nylon spacers and nylon screws. Checked for any shortages to the chassis and could not find any. Here is the photo of the running PSU along with the Soekris. As soon as the power is switched on the LEDs on positive board are bright and then go dim in seconds with the voltage dropping to 4.5v and its stuck.
An externally hosted image should be here but it was not working when we last tested it.
Connect the two UBiBs zero out together with one wire and cut its insulation at the middle, solder there and send one zero cable only to the Soekris. Make a Y ground cable run in other words. So we see if something changes. They are far away and reversely positioned the pos and neg with different cable lengths and maybe something weird happens.
More Voltage Drop with Time than Expected
Hello Salas,
Have built the positive board (from Tea-Bag) and when cold, I get 12.5V out into load. I am running a 500ma load with 100ma headroom (CCS set for 600ma). The voltage slowly drops to 12.07. This takes 20 to 30 minutes. (I was aiming for 12.1V).
No parts substitutions except using the orange LEDs instead of the red LEDs. I used 50ppm (or better) resistors.
Any thoughts on where to start looking?
BTW, this is the lowest noise PS that I have ever built.
bsk
Hello Salas,
Have built the positive board (from Tea-Bag) and when cold, I get 12.5V out into load. I am running a 500ma load with 100ma headroom (CCS set for 600ma). The voltage slowly drops to 12.07. This takes 20 to 30 minutes. (I was aiming for 12.1V).
No parts substitutions except using the orange LEDs instead of the red LEDs. I used 50ppm (or better) resistors.
Any thoughts on where to start looking?
BTW, this is the lowest noise PS that I have ever built.
bsk
Its normal for this reg to lose some Vout until fully warmed up. It has enough negative TC due to using Vbe/R I-V reference. Depends on how hot nearby radiating sources to the TO-92 BJTs can be. You may alternatively use a Zener than the trimmer for Vref but if the settling drift isn't critical for your application, better don't go with the Zener alternative. In any case the guide talks about how to install a Zener and how to decide its Vz vs a Vout target (in the notes section titled VR1 & VRR).
I just did the 'Y' cable tweak and still the positive board drops the voltage to 4.5v and the negative is stable at 12v. I also did the wiring as short as possible and equal length from positive and negative but with the same results. So I think some issue within the positive board itself.
Thanks