Tio, I got this link from Salas
Transistor matching
I ordered 40 PF5102 and 15 was under 7.5mA for me.
Transistor matching
I ordered 40 PF5102 and 15 was under 7.5mA for me.
Its not a make or break requirement but its the right thing to do. As for how, do as rowli has linked
I have got an extra BiB +IRF kit and board to make a LPSU for my Raspberry Pi3 to give it around 5V 1.5A which is the max current that 1.3 bib board can supply as per the posts here. Am I correct in my assumption?
Also I have got the small kit for 5V from Tea-Bag to be used for the +IRF board. Apart from this kit do I need to make any more changes from the kit provided by Tea-Bag to get 5v 1.5A max current? My Pi does not have any external USB drive so its just Moode OS that is running with LAN port and the Amanero i2s DAC which gets powered from this PSU via the Audioquest Jitter bug. What should be the ideal resistor value for R1 with the CCS current @ 1.5A? Also what load resistor should I choose to test the board on the output.
Thanks
Also I have got the small kit for 5V from Tea-Bag to be used for the +IRF board. Apart from this kit do I need to make any more changes from the kit provided by Tea-Bag to get 5v 1.5A max current? My Pi does not have any external USB drive so its just Moode OS that is running with LAN port and the Amanero i2s DAC which gets powered from this PSU via the Audioquest Jitter bug. What should be the ideal resistor value for R1 with the CCS current @ 1.5A? Also what load resistor should I choose to test the board on the output.
Thanks
That's a difficult region for readily found values because it takes 0.36 Ohm (3W) R1 for 1.61A (allowing 110mA spare current not to bring current limit at 1.5A sharp and the Pi3 will reset). Maybe a combination of resistors like two 0.33 Ohm 2W in series.
With good sinking and >100mA Pi3 idle draw the standard M1 type could possibly suffice but an IRF9630 should be a safer bet if you will see a problem. R9 value should also be upped to 1.5K for guarding against oscillation better when the Pi3 will be running idle or low tasking and huge spare current (thus heat also) will be wasted on M2 affecting its characteristics.
In any case better measure what your Pi3 will be drawing as peak in your use. So not to set too much CC i.e. don't waste too much in heat if not really needed.
For dummy load testing a 3.3 Ohm 20W resistor should simulate 1.5A draw at 5V rail and take the heat without smoking away but it will certainly get very hot so beware not to let it rest on anything other than glass, ceramic, or metal.
With good sinking and >100mA Pi3 idle draw the standard M1 type could possibly suffice but an IRF9630 should be a safer bet if you will see a problem. R9 value should also be upped to 1.5K for guarding against oscillation better when the Pi3 will be running idle or low tasking and huge spare current (thus heat also) will be wasted on M2 affecting its characteristics.
In any case better measure what your Pi3 will be drawing as peak in your use. So not to set too much CC i.e. don't waste too much in heat if not really needed.
For dummy load testing a 3.3 Ohm 20W resistor should simulate 1.5A draw at 5V rail and take the heat without smoking away but it will certainly get very hot so beware not to let it rest on anything other than glass, ceramic, or metal.
Thanks for your help SalasYou understood well. Your whole post's logic is following the instructions and suggestions correctly
I have found a R-Core transformer at 2*16v AC. I am to have +/-15 on dc side. Would there be any issues in "burning" 6,2 V (16*1,414-1,4=21,2V) at the voltage regulator or would that be on the high side and pushing the UltraBib?
Thanks in advance
Can you give a hint as to the priority of this?
My phono preamp went to heaven, and I have been in the marker for a replacement for a few months.
Every time I go to drop the dime on the PassDIY boards, I end up needing the cash for something else.
-Josh
@Salas,
I have a couple of questions on transformers.
The first pertains to the Allo Katana DAC.
Amazon Drive
Near the bottom is a list of requirements.
The 5V requirements for the Katana board, and the Controller board are straight forward.
If I am reading the table right, the onboard DC-DC converter requires 5V 180mA.
This converter will be bypassed, and the opamp board will be powered with +-15V.
It calls for 5V 800mA, so I figure +-15V has 6 times the potential, and thus the opamp board should get 800/6 ~= 130mA.
So we want 600/230 = 2.6R for R1.
Allo seldom replies to questions (but they do seem to read them).
Am I crazy?
The next is what are your thoughts about split bobbin "semi-toroidal" transformers. In theory they should be much better about common mode EMI. My house has dirty grounds, so it is something I always worry about.
Hammond and Signal both make a 10V 48VA that should work for the 5V rails. I like the size of them, and they could be put right up near the PS.
Also, the low VA Antek's might not have the internal shielding. I know for sure the 10VA ones do not have the purple wire. the 25VA ones do not show it in the diagram. The 50VA ones show it in the diagram, but it is not visible in the image and not in the description. Anything 100VA and up has the wire and shield.
Lastly, I assume it is fine to tie the center 0V rail together at the output and let it pretend to be dual mono up to that point. Do you have a thought about tying that to chassis ground via a rectifier like sometimes seen in Pass Labs designs? For the Katana, it is probably best to just let the 0V rail float, but it might be important in another build.
Thank you,
-Josh
I have a couple of questions on transformers.
The first pertains to the Allo Katana DAC.
Amazon Drive
Near the bottom is a list of requirements.
The 5V requirements for the Katana board, and the Controller board are straight forward.
If I am reading the table right, the onboard DC-DC converter requires 5V 180mA.
This converter will be bypassed, and the opamp board will be powered with +-15V.
It calls for 5V 800mA, so I figure +-15V has 6 times the potential, and thus the opamp board should get 800/6 ~= 130mA.
So we want 600/230 = 2.6R for R1.
Allo seldom replies to questions (but they do seem to read them).
Am I crazy?
The next is what are your thoughts about split bobbin "semi-toroidal" transformers. In theory they should be much better about common mode EMI. My house has dirty grounds, so it is something I always worry about.
Hammond and Signal both make a 10V 48VA that should work for the 5V rails. I like the size of them, and they could be put right up near the PS.
Also, the low VA Antek's might not have the internal shielding. I know for sure the 10VA ones do not have the purple wire. the 25VA ones do not show it in the diagram. The 50VA ones show it in the diagram, but it is not visible in the image and not in the description. Anything 100VA and up has the wire and shield.
Lastly, I assume it is fine to tie the center 0V rail together at the output and let it pretend to be dual mono up to that point. Do you have a thought about tying that to chassis ground via a rectifier like sometimes seen in Pass Labs designs? For the Katana, it is probably best to just let the 0V rail float, but it might be important in another build.
Thank you,
-Josh
After some time listening on the test bench, I got time to put my Dam1021 and the Salas boards into a diy box.
To my surprise ... it did not work! I already checked the connections and everything is ok. My suspicion is the GND that in Dam1021 has a bridge between both (- / + GND).
Disconnecting the power supplies from the Dam1021, the Salas boards are Ok +/- 12V, when I call (or try!) The Salas boards turn off (the leds turn off).
I even thought it was the dac but this is working.
Any idea?
Best regards,
To my surprise ... it did not work! I already checked the connections and everything is ok. My suspicion is the GND that in Dam1021 has a bridge between both (- / + GND).
Disconnecting the power supplies from the Dam1021, the Salas boards are Ok +/- 12V, when I call (or try!) The Salas boards turn off (the leds turn off).
I even thought it was the dac but this is working.
Any idea?
Best regards,
An externally hosted image should be here but it was not working when we last tested it.
The Katana thing has been done one time before by a member back in this thread. He used +/-350mA CC but it could be done with +/-300mA CC also I suspect.
Split bobbin are good and low profile. Having corners are not like R-Core but next best thing in low profile frame type transformers style..
As I see Antek states "And they also have the magnetic shield around the outside to reduce the magnetic leakage" in 50VA too, and a purple wire can bee seen for a static shield when looking from the top in their second picture: AS-0515 - 50VA 15V Transformer - AnTek Products Corp
I don't know specifics about various applications like this or that DAC must float etc. you better ask in their specific threads for grounding styles they best accept. Or look back in the thread here to find the member with the Katana & UBiB to send him a pm?
In general the UltraBiB needs independent secondaries since its got independent bridge per section and their zero output points should be connected a load's zero (common). Like a Y in other words.
Thanks for the reply. I went back through the thread and found all the posts on the Katana. There seemed to be alot of confusion. The person in question went with 350mA, which I am positive is way too much. (350-100)*6 = 1.5A. The entire stack including the RPi draws 1.5A (per Allo specs). Using 2 PS's, and driving just Allo boards calls for 1.1A at 5V (1100/6 + 100 = 284mA). Controller + opamp = 1A (1000/6 + 100 = 267mA).
I suppose there is probably no harm in using ~270mA to calculate R1 in stead of 234mA which I believe is correct. I can toss am ammeter on it and change R1 if I need to.
I sent Allo a message, but I might as well be asking the moon. I have never gotten a reply from them on any question. They make a great product, but they kinda suck at the communication side of things.
I just got word from Tea-Bag that I will be getting the new boards from his GB. Putting together my Mouser order now.
I think I will try one of the Hammond split bobbin 10V 48VA transformers for the 5V boards, and use a 15V 100VA Antek for the +-15V rails.