Announcement

Collapse
No announcement yet.

Interesting Power Supply Issue...

Collapse
X
  • Filter
  • Time
  • Show
Clear All
new posts

  • #46
    Tom,

    My testing points to the TDA7402 module as the culprit.

    I tested two TPA311X 2.1 amps with three different power supplies. One amp is an TPA3118 Sanwu 2 x 50 W + 60 W TPA3118 (pictured, $11 ebay). The other is the new PE DTA 2.1 BT 2 x 50 W + 100 W amp. The PE amp was removed from its enclosure (ongoing reverse engineering project sponsored by Kevin K).

    The three power supplies: (1) The DTA's 24 V 4 A PS with a huge ferrite bead and 3 prong plug; (2) A PE 24 V 5 A brick, 2 prong, no bead; and (3) An 15 year-old Dell 19 V 3.3 A, 2 prong, bead was removed to put on a 5.5 plug.

    The DTA, at maximum volume and nothing connected to the RCA inputs was dead quiet with all three PS's.

    The Sanwu, at maximum volume and nothing connected had some pink noise. The same with all three PS's. I had to put the 4" full range right to my ear to hear it (but I'm half deaf). With a the RCAs plugged into my laptop and no signal, it became dead quiet.

    I think the TDA7402 board is picking up emissions from the bricks switching due to a high gain setting on the chip - either via the air or, more likely, via the PS + rail on the board. Could be just a poor design. I believe the TDA7492 module has higher gain than the 26 dB typically used for TPA311x modules. On my other forum, it's noted that the TPA311Xs on higher gains are noisy.

    I would test the TDA with a low impedance source connected (head phone output from a laptop/PC or phone). If the noise is acceptable with that, I think the faceplate has a abnormally high output impedance. That was indicated when you tried 1.3K? resistors to ground but couldn't get the volume up. In that case, I might try different value resistors on the inputs to find a happy medium between noise and volume from the faceplate (I can mail various values from my stock for you to try).

    Mike

    2.1 Amp Red.png

    Comment


    • #47
      Hey Mike, thanks for doing all this.
      Oh, I tried the radio in several locations in the house, several different 'zones' or breakers on the service panel.....they were all noisy.

      Well, I just got in the DC-DC regulated convertor, hopefully the ground isolator for the audio input will be here by Tuesday or so. I'll try again what that comes in. I'm hopeful.

      I also ordered another of the TDA7402 amps, the original ones that had noise with nothing hooked to them.... I remember messing with these to see if they could be used in single channel mode like the product description says... I think you'll remember that post... I tried to hook it up a few different ways, and I'm wondering if it's possible that I goofed it up somehow.... anyway, I have another one coming so I can be sure.

      I've been doing some woodwork/veneering etc. while I wait for parts to come in so at least I'm trying to be productive!

      BTW, I think it's great that you're taking a peek under the hood of PE's new 2.1 amp, that really looks promising if the Lowpass filter can be brought down to a reasonable range from the 180 Hz or whatever it is. You're really have something there for the Bluetooth users, the phone could be the remote control in a way with that.

      I searched page after page of 2.1 amps/preamps, etc. on the 'AliX' and 'Zon' sites today... there are dozens and dozens of various implementations of that arrangement. Someone is making a lot of little amplifiers in this world!

      TomZ
      *Veneering curves, seams, using heat-lock iron on method *Trimming veneer & tips *Curved Sides glue-up video
      *Part 2 *Gluing multiple curved laminations of HDF *Cello's Speaker Project Page

      *Building the "Micro-B 2.1 Plate Amplifier -- Part 1 * Part 2 * Part 3 * Part 4 * * Part 5 'Review' * -- Assembly Instructions PDF

      Comment


      • #48
        Neal, just got the little DC/DC converter, I didn't realize it was going to be so small!
        Here it is next to the little linear regulator and a quarter for size comparison.
        Click image for larger version

Name:	IMG_0633 (Small).JPG
Views:	1
Size:	86.3 KB
ID:	1358961


        I just wanted to check with you on wiring it up. I'm assuming that I just connect the voltage in + and - as well as the voltage out + and - and I'm good? I can just ignore the other leads? Looking at the data sheet, it isn't clear what the other leads do. Here is the data sheet for it: https://media.digikey.com/pdf/Data%2..._Series_Ds.pdf

        Also, do I need to be careful about soldering to this? My plan is to solder some wires to the leads I need, and heat-shrink those and the remaining leads to keep the terminals from contacting anything, with probably a few wraps of electrical tape around the area where the terminals come out of to give it some strength against bending the leads. Will this thing get hot enough to need to be isolated from wires, etc.? Questions, questions, right? I should be paying you guys college tuition money!

        I should get the line-in input device Tuesday hopefully so I'll give everything a run then.

        TomZ
        *Veneering curves, seams, using heat-lock iron on method *Trimming veneer & tips *Curved Sides glue-up video
        *Part 2 *Gluing multiple curved laminations of HDF *Cello's Speaker Project Page

        *Building the "Micro-B 2.1 Plate Amplifier -- Part 1 * Part 2 * Part 3 * Part 4 * * Part 5 'Review' * -- Assembly Instructions PDF

        Comment


        • #49
          Originally posted by tomzarbo View Post
          Neal, just got the little DC/DC converter, I didn't realize it was going to be so small!

          Nah--that's a big honking component in my world. I'm used to SMD stuff that requires magnifying lenses Anything that can fall on the carpet and not get lost is big!


          I just wanted to check with you on wiring it up. I'm assuming that I just connect the voltage in + and - as well as the voltage out + and - and I'm good? I can just ignore the other leads?

          ​The N.C. pins are "no connect", so don't worry them--no heatshrink tubing required. The R.C. I assume is "reserved connection", probably a pin used for testing. It should be isolated.


          Also, do I need to be careful about soldering to this?

          ​I don't think it requires extreme care since the housing is an epoxy that is heat resistant. I always use a temperature controlled iron, and I would recommend one of those. But a simple 25W soldering pencil or other electronics-oriented soldering equipment should be fine.
          Will this thing get hot enough to need to be isolated from wires, etc.
          Nope. With 24V input it is 83% efficient. It the load is 150ma, the convertor will only dissipate only 0.3W (12 x .15 x .17).

          ​For this type of component, I usually make the solder connections and then hot melt glue the part to the chassis or some other device. Cover the terminals with hot melt glue if you like to give the wires and connections some extra strength. You can even use Velcro to mount the device, as long as the wires aren't too heavy.

          ​I hope this works out for you. Grounding is always more difficult than you would expect, and unfortunately, it is not taught in college. I'm assuming the problem is due to unwanted current flowing in the grounds rather than radiated interference, but this will at least help isolate the problem.
          Free Passive Speaker Designer Lite (PSD-Lite) -- http://www.audiodevelopers.com/Softw...Lite/setup.exe

          Comment


          • #50
            Originally posted by neildavis View Post
            [LEFT]
            I'm assuming the problem is due to unwanted current flowing in the grounds rather than radiated interference, but this will at least help isolate the problem.
            Can't be 100% sure, but I tend to agree (and mentioned this near the start of the thread). Have worked on many mixed digital and analog HW designs for work and have never seen radiated (or conducted, it doesn't have to be radiated) interference remotely near the levels that would be required for the issues Tom is seeing. If it is interference, it would probably be conducted vs radiated as noise level depends on what is plugged in

            Comment


            • #51
              Originally posted by DDF View Post

              Can't be 100% sure, but I tend to agree (and mentioned this near the start of the thread). Have worked on many mixed digital and analog HW designs for work and have never seen radiated (or conducted, it doesn't have to be radiated) interference remotely near the levels that would be required for the issues Tom is seeing. If it is interference, it would probably be conducted vs radiated as noise level depends on what is plugged in
              First, these are both DC based units. That implies that signal ground between them is the same ground when connected to the same DC supply; the amp having decoupling caps on the inputs to negate any DC bias. I do agree that the noise, if from the PS, is conducted. But radiated noise can be a problem. IIRC, Neil has written about issues of radiated noise from the digital output filters due to the chips switching frequency.

              Gent.'s ... for this problem I'm forming a different opinion than ground issues.

              In this module, the TDA7492 chip's gain is set fairly high. That tends to make the module susceptible to noise. Having a low impedance source helps. That was evidenced when Tom tied 1.3K resistors to ground on the inputs thus reducing the noise somewhat. But the high gain in a single stage is always going to be problematic. The amp was noisy without the face plate connected. And, the face plate seems to be a high impedance source as evidenced by the lack of volume with resistors to ground on the amp's inputs. There's a well known YouTube video that measures an amp's noise reduction by replacing a 50k input pot with a 10K pot on an amp. That swap essentially insures a maximum of 10K input resistance to ground

              Based on info from following other forums and having tested the other amps that performed well, several things are evident. First the chips I tested, although different, are set to a lower gain (26 dB compared to 31.3 or 33.6 dB in the TDA7492 module). That's significant. 26 dB is ~ 20x gain while 33.6 dB is ~ 48x gain. To get the overall gain up, the modules have op-amp stages feeding the chips. Those stages are also lower gain, perhaps in the neighborhood of 12 dB, So there isn't a single stage with excessive gain that is sensitive to induced noise (conducted, radiated, take your pick). And the op-amps provide a low impedance source to the chip's inputs further resisting induced noise. On the $11 ebay Sanwu module, there was slight noise with unconnected inputs but it was dead quiet with the laptop connected. The TPA311x series has been out a while now. And every current design sets the gain at 26 dB (or lower) based on early experience with higher gain settings and noise.

              If it wasn't 0o degrees out, I'd suggest taking the amp outside and tap a 12 V car battery. If it's still noisy, then the chip's switching is inducing noise into the module. If not, I suspect the system would be quiet with both the amp and the face-plate connected to the car battery.

              I would like the amp tested on the PS with a known low impedance source - headphone output from a phone or laptop (on batteries). If still noisy, it's the amp's high gain susceptible to the PS noise. And shorted inputs would determine the basic noise floor for the module.


              Side Note: Tom I didn't mean to steer you to 2.1 amps. I just happen to have those two units on hand to do the PS tests in my previous post.
              Last edited by Millstonemike; 01-07-2018, 01:23 PM.

              Comment


              • #52
                I finally got to mess with this a little more this morning. I've been pretty sick for a few weeks, plus I had a small surgery a few days ago...I just haven't felt like doing anything that required me to use my brain.

                The isolated DC-DC convertor seems to have done the trick. I took power right from the amp board went to the isolated convertor, and out to the faceplate/preamp. It was quiet as could be. I tried it on the newer Tripath amp with three different power supplies -- all silent, I also tried it with the original 50 watt per channel amp with a few different ones, and that setup was quiet too.

                Also, I put the wireless thermometer on the isolated convertor after a few minutes of play time, and it was hardly any warmer than ambient with the bluetooth working on the faceplate... nice! I never knew they even existed; it pays to have smart friends.

                Of course, this stuff all worked on the bench. Shoving everything into the confines of the radio could change that I suppose, but I'm hopeful that this will fix the issue. I tried this first as I had the parts to do it and it was pretty easy to do. Those little Isolated DC-DC converters are like $9 or something, but if it really ends up working, I don't mind the few extra bucks. I plan to try Mikes suggestion to connect a lower impedance input, such as my laptop on battery just to see what that does to the process.

                Well, I'm not getting my hopes up too much until this thing is singing noise-free all put together, but things are looking up, and I just want to thank all of you guys so much for helping me get the gremlins evicted from this project.

                I'll report back once I get it all put together again.

                TomZ
                *Veneering curves, seams, using heat-lock iron on method *Trimming veneer & tips *Curved Sides glue-up video
                *Part 2 *Gluing multiple curved laminations of HDF *Cello's Speaker Project Page

                *Building the "Micro-B 2.1 Plate Amplifier -- Part 1 * Part 2 * Part 3 * Part 4 * * Part 5 'Review' * -- Assembly Instructions PDF

                Comment


                • #53
                  Hooray...this is sounding more promising.

                  It's always difficult to resolve grounding problems because they often don't make much sense and they are difficult to locate. The most intuitive discussion that I've found as to why they happen is in the datasheet for the LM3886 (page 21). The problem is that there is no such thing as a perfect ground--all ground paths have some resistance. That means that the ideal amp circuit in the first figure below often ends up like the second figure, where R1 and R2 are different parts of the same ground wire. That circuit provides positive feedback, and you can get oscillation. And when you have multiple amps sharing the same ground paths, those shared currents get amplified and can create lots of unwanted noise. The "fix" is to use separate ground returns for the inputs and outputs, as shown in the final circuit. Also, try to separate signal ground from power ground and only allow them to connect at a single point. That is the "star" ground approach. However, sometimes it isn't possible to have separate ground paths (especially with "modules"), and in those cases the isolated DC-DC converter can help separate the shared currents.

                  Back in the old analog days, I made many LM3886 amps that "hummed", and it was always due to bad grounding practices. But separate out all those ground paths and return them to a single point, and voila--silence.



                  Free Passive Speaker Designer Lite (PSD-Lite) -- http://www.audiodevelopers.com/Softw...Lite/setup.exe

                  Comment


                  • #54
                    I'm about ready to try and re-work the radio in a few minutes, but I just wanted to report that hooking the old original amp to my unplugged laptop's 3.5mm output resulted in silence as well.

                    Once I unplugged the audio jack from the computer, it started making noise again. But when the cable was plugged in -- even when the laptop was paused, it was quiet.
                    Sounds like Mike was onto something.

                    I'm starting to feel about electronics the way I used to feel about girls in high school...... "I like 'em, but I sure don't understand 'em!"

                    Off to solder an isolated DC-DC convertor in my original radio.... Please work, Please work, please work, please work....

                    TomZ
                    *Veneering curves, seams, using heat-lock iron on method *Trimming veneer & tips *Curved Sides glue-up video
                    *Part 2 *Gluing multiple curved laminations of HDF *Cello's Speaker Project Page

                    *Building the "Micro-B 2.1 Plate Amplifier -- Part 1 * Part 2 * Part 3 * Part 4 * * Part 5 'Review' * -- Assembly Instructions PDF

                    Comment


                    • #55
                      IT'S WORKING!

                      I've been listening to the radio for a while now and it's totally quiet, just like it should be. I hot glued the convertor over the contacts and glued it in the faceplate area which had a little more room and less wires to contend with.

                      I'll run it for a few days just to make sure everything is working okay, then it will finally go to it's intended recipient.

                      Thanks, you guys for the help... I really am not sure what my next move would have been had this not worked out.

                      TomZ
                      *Veneering curves, seams, using heat-lock iron on method *Trimming veneer & tips *Curved Sides glue-up video
                      *Part 2 *Gluing multiple curved laminations of HDF *Cello's Speaker Project Page

                      *Building the "Micro-B 2.1 Plate Amplifier -- Part 1 * Part 2 * Part 3 * Part 4 * * Part 5 'Review' * -- Assembly Instructions PDF

                      Comment


                      • #56
                        Congrats, nice learning experience...
                        John H

                        Synergy Horn, SLS-85, BMR-3L, Mini-TL, BR-2, Titan OB, B452, Udique, Vultus, Latus1, Seriatim, Aperivox,Pencil Tower

                        Comment


                        • #57
                          Originally posted by tomzarbo View Post
                          I'm about ready to try and re-work the radio in a few minutes, but I just wanted to report that hooking the old original amp to my unplugged laptop's 3.5mm output resulted in silence as well.

                          Once I unplugged the audio jack from the computer, it started making noise again. But when the cable was plugged in -- even when the laptop was paused, it was quiet.
                          Sounds like Mike was onto something.

                          I'm starting to feel about electronics the way I used to feel about girls in high school...... "I like 'em, but I sure don't understand 'em!"

                          Off to solder an isolated DC-DC convertor in my original radio.... Please work, Please work, please work, please work....

                          TomZ
                          I think you had 2 issues: (1) Noisy amp without a low impedance source, and (2) The faceplate generating noise via ground. From prior tests (1.3K? resistors across amp inputs to ground) the faceplate does not seem to be that low of an impedance - but low enough once the ground issue is fixed. But if the faceplate is going to cause such ground issues, it may not be a "go-to" device. Just another extra part needed to get it right.

                          It was trial and error in high school too - perhaps a bit more fun.

                          Comment


                          • #58
                            Tom, try the amp with a battery, should be minimal noise if any. If there is noise, it's that amplifier. I mentioned in an old post that I did not like this amp due to the high gain creating a high noise floor. Good luck, might need a different board.

                            Comment


                            • #59
                              Thanks John,
                              yeah, I kind of feel like I should smoke a cigar or sip some champagne or something after getting this working!

                              Originally posted by Millstonemike View Post

                              I think you had 2 issues: (1) Noisy amp without a low impedance source, and (2) The faceplate generating noise via ground. From prior tests (1.3K? resistors across amp inputs to ground) the faceplate does not seem to be that low of an impedance - but low enough once the ground issue is fixed. But if the faceplate is going to cause such ground issues, it may not be a "go-to" device. Just another extra part needed to get it right.

                              It was trial and error in high school too - perhaps a bit more fun.
                              Funny. But true.

                              I have one more of the isolated DC-DC convertors so I can get the second radio up and running as well. I do like the way the faceplate works, it's a LOT of stuff for what you pay... I just wish it was less glitchy, and noisy. I'd gladly pay $10 more to have one of these dialed in just right... but I guess I'll have to make do. Little stuff like 'clicks' here and there, or the really loud verbal announcement when the volume is really low, or the un-understandable interface buttons... small stuff really, just not 'perfect' or top line product worthy, if you know what I mean.

                              Originally posted by Jake View Post
                              Tom, try the amp with a battery, should be minimal noise if any. If there is noise, it's that amplifier. I mentioned in an old post that I did not like this amp due to the high gain creating a high noise floor. Good luck, might need a different board.
                              I remember you saying that. Well, I've had this amp work fine for a few projects and be a pain for another few. But, the T-amp that I used suffered from the same issue when connected to the preamp/faceplate unit.

                              Someday I'm going to read up on how the isolated DC/DC convertor makes a ground connection without it really being connected. That seems impossible.

                              TomZ
                              *Veneering curves, seams, using heat-lock iron on method *Trimming veneer & tips *Curved Sides glue-up video
                              *Part 2 *Gluing multiple curved laminations of HDF *Cello's Speaker Project Page

                              *Building the "Micro-B 2.1 Plate Amplifier -- Part 1 * Part 2 * Part 3 * Part 4 * * Part 5 'Review' * -- Assembly Instructions PDF

                              Comment


                              • #60
                                Originally posted by tomzarbo View Post
                                Someday I'm going to read up on how the isolated DC/DC convertor makes a ground connection without it really being connected. That seems impossible.
                                ​It's really simple--the key is the transformer inside the module that isolates the input from the output. The incoming DC is turned on and off at a high rate to feed the primary of the transformer. Because it's switching at a high rate, the transformer can be very small (according to the spec sheet, the switching frequency is 100KHz). Then the output of the transformer is converted back to DC with rectification and regulation. Again, because the frequency is so high, the filter components can be very small. So it's magnetic coupling inside the transformer that provides the DC isolation. This is the same way those power bricks or other switching power supplies work to isolate the DC output from the AC line--there's nothing really different going on inside those modules.

                                By using this separate power source, the supply current doesn't flow through the signal ground. It's those shared ground currents that can cause so many problems.

                                Another "trick" for dealing with ground currents is using differential signals. The TPA3116 amplifier chip has differential inputs, but your board just ties IN- to ground, so you can't take advantage of that feature.
                                Free Passive Speaker Designer Lite (PSD-Lite) -- http://www.audiodevelopers.com/Softw...Lite/setup.exe

                                Comment

                                Working...
                                X