Loading ...
Sorry, an error occurred while loading the content.

Re: How to fix a noisy ground?

Expand Messages
  • nelsonj_sce
    Hmm...I am going to have to get the schems from Kawai - I will also do some research on how to filter a ground. If the analog ground is noisy the op-amps will
    Message 1 of 41 , Oct 3, 2004
    • 0 Attachment
      Hmm...I am going to have to get the schems from Kawai - I will also do
      some research on how to filter a ground. If the analog ground is
      noisy the op-amps will by amping noise. If the digital ground is
      noisy, it is possible that the digital chips are actually producing
      noise but they "think" they are just producing an accurate digital
      signal.

      BTW, last night I replaced IC4 and IC1 (both 064D op-amps) with
      Burr-Brown OPA4228. These are super, super clean op-amps. They did
      improve things, but there is still a lot of noise even with these
      great op-amps. I have designed a headphone amp that uses an OPA4228
      and I plan on replacing the LM386 nock off, but I have not built it
      yet, I might get to it tonight. But again I only expect a bit of
      improvement because that ground noise will still be there.

      Anyway, I have a nice blue backlight, I will soon have a nearly all
      Burr-Brown amplification path (at least to my headphones). If we can
      just figure out a way to get the ground noise out, my K5 will look and
      sound killer.

      Must eliminate ground noise, must eliminate ground noise, must...


      --- In k5synth@yahoogroups.com, "antdes45dragonmaster" <antdes45@n...>
      wrote:
      > Digital Ground is plugged to one bridge's negative pole and analog to
      > 0v line of the transformer(Black wire). I know it's a noisy ground
      > because when I transfer MIDI between K5 and computer, I hear some low
      > volume beeps in the noise, and my intermittent inverter whines(Is on
      > DGround) in the headphones when it wants. I looked at the schemes, and
      > every different PCB's DGrounds are plugged together on the PSU
      > mainboard, but are going to it thru different wires.(Same for
      > AGround). Problem is I don't remember how to filter ground. Something
      > to try. Also, when I unplug the brown socketed black bus wire going to
      > outputs board, there is no signal from the headphones, but still
      > noise. So it's not coming from the opamps(Headphones noise). Might be
      > that the LM386's ground has no filter. While for the rear outputs
      > noise, the opamps might be too noisy(-45dB with some TL072 instead of
      > some TL062, which might have not changed anything noise talking, using
      > some OPA????, don't remember which, will have to re-find the name on
      > DIYAudio, might be a lot less noisy, because a signal passing 4 times
      > in pretty noisy OpAmps in series might not be really good).
      >
      > --- In k5synth@yahoogroups.com, "nelsonj_sce" <nelsonj_sce@y...> wrote:
      > > It looks like you have already changed out the op-amps and found that
      > > the noise is still there. Have you tried anythink to fix the noisy
      > > ground? Is it that the digital and analog all share the same ground?
      > > Is there a way to "split" the ground or add a new ground plane and
      > > reroute say the digital chips to the new ground plane? Are all the
      > > analog parts on a single board so that we could just cut the ground to
      > > that board and then wire up a new, "independent" ground? Are there
      > > just a few key digital parts that are creating the noise that can be
      > > rerouted to a seperate ground plane?
      > >
      > > Has anyone at Kawai told you how to solve this problem? Even if I can
      > > get a hold of the schematics, if Kawai or others can identify a few
      > > key components it should not be that hard to create a seperate ground
      > > for those parts.
      > >
      > > Any thoughts?
      > >
      > > --- In k5synth@yahoogroups.com, Antoine Deschênes <antdes45@n...>
      wrote:
      > > > The LM386 doesn't produce any noise. Even if you unplug the signal
      > > there
      > > > will still be some. Sctually the problem is a noisy ground. It's
      > > sharing
      > > > the same ground for all the board(s).
      > > >
      > > > nelsonj_sce@y... a écrit:
      > > >
      > > > > Here is the latest, it is good news.
      > > > >
      > > > > Today I got my parts for DigiKey. As below I replaced my
      backlight
      > > > > with a blue backlight from nightlaunch. It worked with my
      stock K5
      > > > > inverter. However, my stock inverter was very flakey, even
      after I
      > > > > did the resolder job so I decided to replace the stock
      inverter with
      > > > > the DigiKey part. (Digikey part #289-1032-ND). The good news
      > is that
      > > > > inverter works (even though it is only about 95V @ about
      550Hz). It
      > > > > is a slightly different color than before, but it still looks
      great.
      > > > > Also- THERE IS VIRTUALLY NO HUM!
      > > > >
      > > > > The only warning is the Digikey part is not the exact same size
      > as the
      > > > > original inverter - it is a bit smaller in footprint and a lot
      > > > > shorter. The stock inverter has only three connections (in -
      out -
      > > > > ground) and the new inverter has a fourth pin for mounting.
      Because
      > > > > the part is smaller, I used the original hole for the "out", but
      > I had
      > > > > to wire up jumpers to the "ground" and the "in" - not a big
      deal but
      > > > > it is some extra work - and you don't want to see pictures of
      > my ugly
      > > > > job. :). But so far so good.
      > > > >
      > > > > Also, I did the 147K resistor hack and it seems to have gone
      > well. My
      > > > > K5 is a louder, but the headphones have even more noise now.
      I have
      > > > > been looking things over and I think the headphone noise is
      because
      > > > > they use a cheap LM386 op-amp and a bad overall design for the
      > > > > headphone amp. I am trying to see if I can replace this with a
      > better
      > > > > op-amp, otherwise I think the only good solution is to just bypass
      > > > > their circuit completely and make a better headphone amp and
      wire it
      > > > > up to the K5.
      > > > >
      > > > > Anyway, again the good news is that for less than $30
      delivered you
      > > > > can get nightlaunch backlight and a Digikey inverter and have
      a new
      > > > > backlight that does not hum! :)
      > > > >
      > > > > Now for the headphone amp...
      > > > >
      > > > >
      > > > >
      > > > >
      > > > >
      > > > >
      > > > >
      > > > >
      > > > > --- In k5synth@yahoogroups.com, "nelsonj_sce" <nelsonj_sce@y...>
      > > wrote:
      > > > > > Update:
      > > > > > Well today I ordered a new inverter from www.DigiKey.com. I
      > > found an
      > > > > > inverter that looks like it will do the job - for those
      interested
      > > > > it is:
      > > > > >
      > > > > > Digikey part # 289-1032-ND and it only cost $11.90 - not
      > bad.
      > > > > >
      > > > > > I am not sure if it will fit exactly in the same slots as the
      > > original
      > > > > > inverter so it might require some additional wiring.
      > > > > >
      > > > > > Anyway, I had my K5 open to look at the power supply to plan
      > the new
      > > > > > inverter install when my K5 started "humming", and guess what,
      > I now
      > > > > > have a beautiful blue backlight. It looks great, a very cool
      > blue.
      > > > > > So it appears my inverter was not bad, but it just has bad
      solder
      > > > > > joints. It is working for now but tomorrow I will likely take
      > > out the
      > > > > > board and redo all the solder joints.
      > > > > >
      > > > > > Just some additional info I learned in researching
      backlights and
      > > > > > inverters. The "brightness" of the backlight is a function
      of the
      > > > > > voltage and frequency applied to the backlight. It appears one
      > > has a
      > > > > > lot of flexibility in selecting the inverter output voltage and
      > > freq.
      > > > > > For example, nightlauch says you can "test" the backlight by
      > > plugging
      > > > > > it directly in to a wall socket (120V AC @ 60 Hz here in the
      > > US)! The
      > > > > > higher the voltage and freq the brighter the backlight.
      > > However, the
      > > > > > backlight will "burn out" quicker the higher the freq. So there
      > > is a
      > > > > > trade off, high freq = bright but shorter life.
      > > > > >
      > > > > > Right now my stock inverter is putting out about 102V @ 380
      Hz. I
      > > > > > think the inverter I ordered will put out about 95V @ 540 Hz and
      > > > > > should work with a 3 - 6 V DC input.
      > > > > >
      > > > > > Next step, 147k resistors (also ordered from Digikey today).
      > > > > >
      > > > > >
      > > > > >
      > > > > > --- In k5synth@yahoogroups.com, "nelsonj_sce" <nelsonj_sce@y...>
      > > wrote:
      > > > > > > Well I replaced my backlight with one I ordered from
      > > > > > www.nightlaunch.com.
      > > > > > >
      > > > > > > It took less than two weeks from when I ordered and the
      > parts look
      > > > > > good.
      > > > > > >
      > > > > > > Good news/bad news though on my backlight.
      > > > > > >
      > > > > > > The good news: I had no problems removing the LCD, taking out
      > > the old
      > > > > > > backlight and sliding in the new backlight. The new light has
      > > some
      > > > > > > wires and a plastic "tab" that kind of gets in the way, so I
      > > soldered
      > > > > > > the backlight wires to where the wires connect to the LCD
      board.
      > > > > > > Hint:You have to cut the new backlight to size, so make the
      > > backlight
      > > > > > > a little longer (say 1/4 inch) than the original to make
      things
      > > > > easier.
      > > > > > >
      > > > > > > Bad news: My backlight does not work but I think my inverter
      > > is the
      > > > > > > problem. I am only getting about 1/2 Volt at the output of my
      > > > > > > inverter, and I am supposed to get 120V @400Hz if I
      > understand the
      > > > > > > part correctly. I tested my inverter before I did the
      backlight
      > > > > > > install and saw it had low voltage, but I was hoping that my
      > > existing
      > > > > > > "dead" backlight was causing the problem (i.e. it was
      > shorted.) I
      > > > > > > guess not because even with the new backlight I get very low
      > > voltage.
      > > > > > >
      > > > > > > Anyway, it looks like I need to get an inverter. I can get an
      > > > > > > inverter from nightlaunch for $17.99+s/h but it needs 9V
      DC. I
      > > > > > > understand the inverter in the K5 uses 5V DC.
      > > > > > >
      > > > > > > Q1: Does the K5 inverter use 5V DC?
      > > > > > > Q2: If I measure the voltage on the two wires to the left
      of the
      > > > > > > inverter, should I get 120V if everything is working? (My
      > > Fluke says
      > > > > > > it is good <=1kHz)
      > > > > > > Q2: Does anyone know where to get a compatible K5 inverter
      > and how
      > > > > > > much it will cost?
      > > > > > >
      > > > > > > Thanks again for all the help. This group is great.
      > > > >
      > > > >
      > > > > Yahoo! Groups Sponsor
      > > > > ADVERTISEMENT
      > > > >
      > >
      >
      <http://us.ard.yahoo.com/SIG=129jk20o5/M=273541.4247512.5496808.2248467/D=g=
      > =
      > >
      >
      rplch/S=1705032144:HM/EXP=1095918198/A=2061551/R=0/SIG=11sr5b9n1/*http://la=
      > u=
      > > nch.yahoo.com/artist/videos.asp?artistID=1001301>
      > >
      > > > >
      > > > >
      > > > >
      > > > >
      > >
      ------------------------------------------------------------------------
      > > > > Yahoo! Groups Links
      > > > >
      > > > > * To visit your group on the web, go to:
      > > > > http://groups.yahoo.com/group/k5synth/
      > > > >
      > > > > * To unsubscribe from this group, send an email to:
      > > > > k5synth-unsubscribe@yahoogroups.com
      > > > >
      > <mailto:k5synth-unsubscribe@yahoogroups.com?subject=Unsubscribe>
      > > > >
      > > > > * Your use of Yahoo! Groups is subject to the Yahoo! Terms of
      > > > > Service <http://docs.yahoo.com/info/terms/>.
      > > > >
      > > > >
      > > >
      > > > --
      > > > antdes45@n...
    • antdes45dragonmaster
      Also, I get some -47 when the volume isn t even at max. So I ll give a try to these OPA4228, and the 2 chan version in some preamps and DACs. ... the line in
      Message 41 of 41 , Nov 10, 2004
      • 0 Attachment
        Also, I get some -47 when the volume isn't even at max. So I'll give a
        try to these OPA4228, and the 2 chan version in some preamps and DACs.

        --- In k5synth@yahoogroups.com, antdes45@n... wrote:
        > My mixer plugged to the line in gives me -55 and my K5 plugged to
        the line in gives me -55 too. I did the 147k hack but with 150k ones,
        I got a problem: the last 1/12 boosts the volume x3, and when I switch
        up and down from the 11/12 to the 12/12 part, I hear a pop between
        them(BTW it doesn't distort, except for crappy low W headphones.
        >
        > "nelsonj_sce" <nelsonj_sce@y...> wrote:
        >
        > >
        > >OK - I am note sure this is an apples-to-apples test but here is what
        > >I did. I have a Yamaha DSP2416 mixer in my computer. I hooked up my
        > >K5 to the mixer and then read the levels from the mixer.
        > >
        > >I did all tests while my K5 was in multi mode, I tested both the "mix"
        > >output and the "1" output and turned the mixer channel on my computer
        > >all the way up. I did not hit any notes, so I was just looking at the
        > >noise level, and I did not run through any FX or noise gates, I just
        > >ran the K5 direct to the mixer.
        > >
        > >Results:
        > >
        > >1) With the K5 volume at minimum I was below -72dB on both "mix" and
        > >"1" outputs. This was the same reading as when the K5 was not even
        > >plugged in, so there was no additional noise.
        > >
        > >2)With the K5 volume at MAX, the "Mix" output was just above -66dB and
        > >the "1" output was just above -72dB. NOTE: I did the 147k resistor
        > >hack, so my MAX output is much hotter than a stock K5.
        > >
        > >I am not sure how my K5 would do if I used the same setup you used,
        > >but this was the best I could do with the equipment I have. I hope
        > >this helps.
        > >
        > >
        > >
        > >
        > >
        > >
        > >--- In k5synth@yahoogroups.com, Antoine Deschênes <antdes45@n...>
        wrote:
        > >> Could you do a little VU meter test, with mix output if you have
        > >changed
        > >> the opamp for it(The same opamp is for headphones) or single if not,
        > >> without a noise gate or anything else at minimum volume? I'd like to
        > >> know how many dBs is it now.(I sat my SBLive VOL adj. at -90dB on
        > >the VU
        > >> meter(Software) without anything plugged in the Live, then
        plugged the
        > >> K5 in the Line In and i got a -47dB.
        > >>
        > >> nelsonj_sce@y... a écrit:
        > >>
        > >> >
        > >> > Hack update:
        > >> > Since my last note I replaced the other two op-amps with
        socketed Burr
        > >> > Brown OPA4228. Replacing these two other op-amps did reduce noise
        > >> > levels noticeably. There is still some noise but with 1)the these
        > >> > four super clean op-amps, 2)using Multi, 3)routing everything to a
        > >> > sinlge output on the multi rather than mix, and 4)using my
        noise gate,
        > >> > noise is really not an issue any more. I would love to try and
        > >> > replace the VCAs also, but I have not been able to track down a
        spec
        > >> > sheet (or any other info on the) on them so I do not have a clue as
        > >> > what might works as an upgraded part.
        > >> >
        > >> > If anyone can help, the VCAs are part M5222L and they are from
        > >> > the same off-brand Japanese JRC company that is used elsewhere
        in the
        > >> > K5. If anyone can help me with details on the M5222Ls I will
        see if I
        > >> > can find a replacement parts.
        > >> >
        > >> > My bottom line: If you are willing to lay-out $30 for the four BB
        > >> > OPA4228 and sockets, and you if you are willing to remove some
        > >> > tough-to-remove parts, I highly recommend this op-amp replacement
        > >> > hack. My K5 has much less noise and just sounds much
        warmer/classier
        > >> > with the new parts. I would love remove more noise, but right
        now it
        > >> > sounds very good and I am not very motivate to continue
        hacking. (I
        > >> > do still plan on replacing the headphone circuit though since I
        > >> > already made the new circuit and all I have to do is wire it up.)
        > >> >
        > >> > An interesting aside: After I did the latest op-amp replacement,
        > >> > everything sounded great in multi, but when I went to single I had
        > >> > serious problems. Various ranges of the keyboard worked fine,
        while
        > >> > other ranges sounds like a fried amp circuit. It turns out
        that while
        > >> > I was removing IC8 I accidentally pulled part of the circuit board
        > >> > trace off that connected pins 1 and 2 on IC8. This trace
        provides the
        > >> > feedback loop between the (-) and output of line G2. Once I
        tracked
        > >> > down the problem I was able to reconnect pins 1 and 2 and then
        > >> > everything in single mode worked great. What I conclude from
        this is
        > >> > that the K5 in single mode is splitting different ranges of the
        > >> > keyboard to different op-amps. I am not sure what K% is doing in
        > >> > multi but multi (ouput 1) DOES NOT use the opamp in IC8 pins 1-2-3.
        > >> >
        > >> > Hack on.
        > >> >
        > >>
        > >>
        > >> --
        > >> antdes45@n...
        > >
        > >
        > >
        > >
        >
        >
        > --
        > Antoine Deschênes
        > aka DragonMaster
        > http://thehotcom.deep-ice.com/
        > Files:http://angelfire.com/ca7/hotcomcenter/
        >
        >
        > __________________________________________________________________
        > Switch to Netscape Internet Service.
        > As low as $9.95 a month -- Sign up today at
        http://isp.netscape.com/register
        >
        > Netscape. Just the Net You Need.
        >
        > New! Netscape Toolbar for Internet Explorer
        > Search from anywhere on the Web and block those annoying pop-ups.
        > Download now at http://channels.netscape.com/ns/search/install.jsp
      Your message has been successfully submitted and would be delivered to recipients shortly.