Internal->External PSU mod

Discuss Yocto construction and related issues
  • Something doesn't sound right to me in this scenario. I have installed a large 17V 1A non-toroidal transformer inside of my unit. It sits about a cm from the board and there is no hum.
  • mine isn't toroidal, as you can see in the pic
    i'm not familiar with transformers too much, but i think i read somewhere that toroidal transformers are supposedly better insulated or something
    also, are you with a metal case and are you using the "loop breaker"?
  • The only purpose of a groundloop breaker is to avoid ground loop hum when you interconnect different audio devices together with all of them connected to mains.

    [resize=800]http://www.prosoundweb.com/images/uploa ... HumRFI.jpg[/resize]

    So, if you say you have no hum with the groundloop breaker disconnected, that means that the loop breaker circuit has a fault.
  • jeroenbvo wrote:The only purpose of a groundloop breaker is to avoid ground loop hum when you interconnect different audio devices together with all of them connected to mains.
    i'm interconnecting the yocto together with two x0xb0xes (on AC-AC adapters), a korg ES-1 (AC-DC adapter without earth connection), the computer (pretty normal chunky old PC), and the mini-HIFI (no earth connection)
    So, if you say you have no hum with the groundloop breaker disconnected, that means that the loop breaker circuit has a fault.
    well, no idea, i also didn't expect that
    btw, i'm talking about buzz and hum which is quite quiet, but it pretty quickly becomes a huge issue when i want to push the gain up and distort the drums

    for the actual hum/buzz tests i did, i did not plug the yocto into the normal setup, but only into the HIFI's microphone input (which has a gain knob) ... the HIFI itself was still connected to the computer (which itself was connected to the rest of the setup)

    so that is some extreme amplification gain, surely, but there definately is buzz (with the loop-breaker) and hum (picked up electromagnetically)
    when i exclude the loop-breaker and move the "green board" 15-or-so centimeters away from the rest of the stuff - all that is left is the noise floor of the yocto (which is pretty clean)
  • This mod is easy.

    If you follow Jeroenbvo...

    viewtopic.php?f=16&t=331&start=50#p6702
  • as i said in the first post, i already have that mod, my yocto came with it (sort of)
    now whether it is implemented the same way - i can't tell
    with this thread i am talking about taking that mod and moving it "outside" the yocto case

    i read some stuff about mains voltage, earthing, safety, etc
    things appear so messy
    in my country it appears we use the so called "TN-C" standard, newer buildings (should) have 3 wires going to each power outlet, but all of the older ones typically only have two - Live and Neutral, and the earth pin on the "schuko" socket is typically connected to the Neutral (which as far as i understand is the least safe option)

    now, i sort of understand that the metal case should be earthed if there is a chance for the 220V Live to touch it by accident, which may be very possible if there is 220V going into the device
    sadly in practice in my situation, that "earth" will end up being equivalent to the Neutral, which as far as i understand could be bad for the buzz/hum situation
    i also read that with audio electronics devices the buzz and hum are very hard to deal with pretty much all over the globe no matter which mains standards scheme is used

    i'll have to stirr this information some more to really get closer to some clearer answers to some of my dillemas...
  • I think the easiest way for you to go is to go back to the original design by Vincent. Use the external power brick.
    You will be perfectly isolated from the mains. And probably have no hum/buzz.

    Reason why some people go for build in power supplies is reliability. You don't want that flimsy 16V plug fall out of the instrument during a gig in front of 500 people.. :mrgreen:
  • update: my latest iteration:
    Image

    Image

    so, the PCB is actually two boards that can be chopped off (or not)
    - mains voltage goes on the bottom, into the transformer, rectifier, and some smoothing capacitors
    - the result goes to connector K1, rectified 15VAC, which should actually be more like 21VDC
    - from K1 it goes to K2 on the smaller board, where there's a voltage inverting regulator which would generate the negative voltage
    - the final result comes at the connector at the top:
    --- negative rail (-17V should be enough to drive the 7915 regulator on the yocto IO board)
    --- positive rail (even more filtered version of the rectified 15VAC, to drive the 7815)
    --- a current-limited version of the positive rail, which can be used to drive the 5V regulator
    - the smaller board also includes the optional loop breaker

    the two boards can be chopped off and connected via a cable with XLR3 connectors, the smaller one inside the yocto, while the bigger one outside (that's what i intend to do)
    or they could be kept as a single board and stuffed inside the yocto.. but you'd get chances for electromagnetic interference from the transformer

    important: the transformer i picked is rated 6VA, and might be too small for the job
    my yocto certainly draws less current than a "normal" yocto because i used bigger current-limiting resistors for the LEDs (3.3k) than what was according to the BOM (220ohms iirc)

    i measured the current consumption from the power rails on my yocto as it is currently, but my measurements might have been non-ideal (and thus wrong) because i measured between the rectifier and the first reservoir capacitors, so at that point in the circuit there would be violent currents to begin with
    still, what i measured was roughly:
    > 96mA on the positive rail
    > 85mA on the negative rail
    > 7.7 to 12.5mA on the GND rail depending on how many LEDs are lit and whether the sequencer is running or not
    so i expect more consumption on the positive rail in a "normal" yocto due to the LEDs
    i'll try to re-do the measurement, i'll add additional reservoir caps after the rectifier and measure the current after them

    initially i was going for a transformer with a 14VA rating because that should certainly do the job, and i found a 18VA transformer from ANG
    sadly it's huge, it's roughly the same size as the blue transformer, and that would make the bigger PCB yet even bigger
    i also considered the option to use a chassis-mounted transformer, that could make the PCB very small
  • i finally got time to revisit this, and i modified the circuit and PCB slightly since last time
    i fixed a few little issues, removed some components, made the PCB a tiny bit smaller

    Image
    Image
    Image

    splitting the two boards was easier than i thought.. i just ran a knife on the top to cut the copper tracks, then i pushed the weak spot onto an edge..

    i made a small adapter plate for the metal case, to allow me to mount the XLR3 socket onto the panel hole for the mains socket
    ..in librecad, got it fab'ed locally

    Image

    the bigger board is now sitting sealed inside a plastic box, up-side-down (it's mounted onto the lid)
    it's see-thru, so i can see the power LED

    Image

    on the inside, some of the cables are still kinda "temporary", i'll get better connectors later

    Image

    the schematic:

    Image

    i was worried that the transformer i chose might not be strong enough to supply the yocto
    ..it's actually good enough
    the big board gave about 26VDC with no load, and falls down to 20VDC with the yocto running
    my local electronics shop only had an SMD version of the switching regulator chip, i put it on a SOIC-to-DIP8 adapter, but it works fine
    you might have noticed that there are some fusible resistors, i couldn't get them so i used just a jumper wire..
    there should be 22ohm between the transformer and the rectifier, it would limit the charge currents i think (saw it in the tr-808 schematic)
    the regulator generates -16.9VDC when the trimpot is centered, just like i expected

    so..
    even the way it is now, i finally don't have mains voltage in the yocto, and i can close it and i don't have any hum, or buzz in the audio, just an innocent noise floor
    the whole case is connected to protective earth

    interesting observation - the loop-breaker seems to add mains buzz, i left it disconnected