Hi,
I've decided to create a new thread, since my issues seem to be very different from the ones in my last post.
Here's the amp I built:
I'm using it with an 18V transformer, with 6,600 uF per side.
I'm also using a solid state buffer between the inputs of the chassis and the inputs of the amp.
The LM1875's signal and output grounds are connected in a star pattern. The buffer's input grounds aren't connected to the star circuit.
I don't remember (I'll check today) whether the signal and output grounds are connected to the mains earth, or to the chassis. I don't think they are.
The LM1875's are properly insulated from the chassis - the heatsinks are mounted on a plastic board.
Here's the problem.
Whenever I'm using a battery powered source, such as a laptop or an iPod, the sound is simply perfect.
When I'm using a mains-powered source, such as a CDP, or a desktop PC, all I get is some quite powerful hum and noise.
I know the issue has to be with the grounds. It's just that I don't know where to start.
Can you help me please with this issue?
I've decided to create a new thread, since my issues seem to be very different from the ones in my last post.
Here's the amp I built:
I'm using it with an 18V transformer, with 6,600 uF per side.
I'm also using a solid state buffer between the inputs of the chassis and the inputs of the amp.
The LM1875's signal and output grounds are connected in a star pattern. The buffer's input grounds aren't connected to the star circuit.
I don't remember (I'll check today) whether the signal and output grounds are connected to the mains earth, or to the chassis. I don't think they are.
The LM1875's are properly insulated from the chassis - the heatsinks are mounted on a plastic board.
Here's the problem.
Whenever I'm using a battery powered source, such as a laptop or an iPod, the sound is simply perfect.
When I'm using a mains-powered source, such as a CDP, or a desktop PC, all I get is some quite powerful hum and noise.
I know the issue has to be with the grounds. It's just that I don't know where to start.
Can you help me please with this issue?
What is the need for the solid state buffer?
Why the high impedance on the input (50K pot). Large impedance on the input means the potential for electrical noise pickup. Use shielded cable on the input leads inside the chassis if impedance has to be this high.
The 50k pot in series with 10K resistor means gain can be below 10 and manufacturer recommends gain of 10 or better for stability.
Where is the boucherot cell on the output? I would recommend it since it is called for by the manufacturer on the datasheet.
Why the high impedance on the input (50K pot). Large impedance on the input means the potential for electrical noise pickup. Use shielded cable on the input leads inside the chassis if impedance has to be this high.
The 50k pot in series with 10K resistor means gain can be below 10 and manufacturer recommends gain of 10 or better for stability.
Where is the boucherot cell on the output? I would recommend it since it is called for by the manufacturer on the datasheet.
Wrong! The gain is set at 20 in his amplifier (26dB) and has nothing to do with the potentiometer. Amplifier input impedance varies beteen 8.33k (10k || 50k) and 50k, so it's fine to have a buffer.
The ground is always the starting point. Even though you used star grounding, the ground loop is created through the earthing connection.
Think about it. Gain=Rf/Rin and the resistance of the pot is clearly part of Rin for an AC signal. With Neg feedback, part of the signal is sent back to the inv input to counter gain. The POT is clearly in this FB circuit and must be considered as part of gain calc.
Putting a pot in the negative feedback loop seems to be a bad idea as the amplifier gain could be set less than 10 and the chip become unstable.
Flip the inputs around so that the non inverting input is used as the signal in and negative feedback loop is not disturbed.
Putting a pot in the negative feedback loop seems to be a bad idea as the amplifier gain could be set less than 10 and the chip become unstable.
Flip the inputs around so that the non inverting input is used as the signal in and negative feedback loop is not disturbed.
Last edited:
the inverting topology includes the volume adjusting pot inside the feedback circuit.
Use non-inverted and put back in all the missing components.
Now, you don't need that buffer.
The three grounds on RCA, P1 and R3 are signal grounds. These must be connected together.
Now a choice:
either connect the signal ground directly with a single wire to the main audio ground
or
Connect the signal ground to the main audio ground with a 10r resistor AND in parallel a pair of inverse parallel power diodes, i.e. 3components in parallel.
BTW,
if you insist on the inverting topology, then R3 should be approximately equal to R2. Now check whether the noise performance has deteriorated.
Use non-inverted and put back in all the missing components.
Now, you don't need that buffer.
The three grounds on RCA, P1 and R3 are signal grounds. These must be connected together.
Now a choice:
either connect the signal ground directly with a single wire to the main audio ground
or
Connect the signal ground to the main audio ground with a 10r resistor AND in parallel a pair of inverse parallel power diodes, i.e. 3components in parallel.
BTW,
if you insist on the inverting topology, then R3 should be approximately equal to R2. Now check whether the noise performance has deteriorated.
Last edited:
Hi John,
The solid state buffer sometimes helps with underpowered sources (such as my iPhone). I prefer the sound of my portables with it, so I thought I could build it inside my enclosure.
I did not add a pot to my design. I removed it from the topology. I also added a Zobel (0.22 uF / 1 Ohm) at the output.
Now, the weird thing.
I took the amp home and tested it. Unpowered iPod: fine. POWERED iPod: FINE! (???). I took the iPod cable and plugged it into my laptop: FINE!. I then used a CDP.. fine... every music source I tried sounded fine.
My problem happens only at my office, to devices plugged in my computer. Which is bizarre, because I have a small amp there and it works OK.
So I checked my grounds. The chassis was not connected to the ground plug, so I added the cable.
Now the unit has a slight hum with all sources. I haven't tried it yet at my office. But at home it works perfectly well. I can't figure out why it does not work at my office, butI guess I'll have to find out somehow.
Regarding the hum - I suppose I need to follow your instructions Andrew. Hopefully this is the solution I'm looking for!
Thanks for all your help! It's really appreciated.
The solid state buffer sometimes helps with underpowered sources (such as my iPhone). I prefer the sound of my portables with it, so I thought I could build it inside my enclosure.
I did not add a pot to my design. I removed it from the topology. I also added a Zobel (0.22 uF / 1 Ohm) at the output.
Now, the weird thing.
I took the amp home and tested it. Unpowered iPod: fine. POWERED iPod: FINE! (???). I took the iPod cable and plugged it into my laptop: FINE!. I then used a CDP.. fine... every music source I tried sounded fine.
My problem happens only at my office, to devices plugged in my computer. Which is bizarre, because I have a small amp there and it works OK.
So I checked my grounds. The chassis was not connected to the ground plug, so I added the cable.
Now the unit has a slight hum with all sources. I haven't tried it yet at my office. But at home it works perfectly well. I can't figure out why it does not work at my office, butI guess I'll have to find out somehow.
Regarding the hum - I suppose I need to follow your instructions Andrew. Hopefully this is the solution I'm looking for!
Thanks for all your help! It's really appreciated.
Last edited:
Hi Andrew, thanks for your comments. This is the first Gainclone I built, and I'm looking to learn as much as possible from them. I knew I possibly started with a sub-optimal design, but I decided I needed to start somewhere, too!
I have a question. When you mention RCA, P1 and R3 grounds wired together, by P1 do you mean the speaker output?
The way I have it wired right now is: I have a common ground point for everything (RCA, speaker output, power supply ground, signal ground) on each circuit. Should I then separate RCA, speaker output and signal ground from power supply ground and then use a heavy gauge wire to connect this ground to the power supply one?
p1 is taken from your post1 schematic. It is the ground pin of the pot.
You have said there is no pot but you have also said you added a buffer/pre-amp to the diagram.
This complicates the issue.
Take out the buffer pre-amp.
Connect the two remaining signal grounds together. Then follow the choices.
Once you have the amp working correctly you can investigate replacing the buffer/pre-amp. But you really should put the buffer on the output of sources that are inadequate, not on the input of a good power amplifier.
You have said there is no pot but you have also said you added a buffer/pre-amp to the diagram.
This complicates the issue.
Take out the buffer pre-amp.
Connect the two remaining signal grounds together. Then follow the choices.
Once you have the amp working correctly you can investigate replacing the buffer/pre-amp. But you really should put the buffer on the output of sources that are inadequate, not on the input of a good power amplifier.
fjhuerta, Once you solve the problem, you'll get the bug and want to build more amps. I've done 8 and many more never got off the breadboard.
Anyhow, It sounds like it could be a ground loop issue causing a small amount of AC hum getting in. If the other amp is quiet, it could be an issue with your PSU. At the office, do you connect using the same electrical outlet at the other sources? Do you get hum with battery sources at the office?
Also, with portable audio sources, you can lower the input impedance. I use 1K for my ipod/walkman amps. This reduces the potential for stray electrical noise pickup (fluorescent lamps can be nasty noise makers.
I'd also put .1uf film caps on the PSU pins to ground close to the chip.
Please download the data sheet and use the schematic provided. It makes more sense than that "gainclown" mess. I think people here sell good pre-etched boards for the chip as well.
Anyhow, It sounds like it could be a ground loop issue causing a small amount of AC hum getting in. If the other amp is quiet, it could be an issue with your PSU. At the office, do you connect using the same electrical outlet at the other sources? Do you get hum with battery sources at the office?
Also, with portable audio sources, you can lower the input impedance. I use 1K for my ipod/walkman amps. This reduces the potential for stray electrical noise pickup (fluorescent lamps can be nasty noise makers.
I'd also put .1uf film caps on the PSU pins to ground close to the chip.
Please download the data sheet and use the schematic provided. It makes more sense than that "gainclown" mess. I think people here sell good pre-etched boards for the chip as well.
fjhuerta, Once you solve the problem, you'll get the bug and want to build more amps. I've done 8 and many more never got off the breadboard.
...
Please download the data sheet and use the schematic provided. It makes more sense than that "gainclown" mess. I think people here sell good pre-etched boards for the chip as well.
Thanks for your comments, they are certainly helpful. I didn't know this design was so bad... at first I thought it was the simplest circuit I could find, and therefore, the best (as I thought this was the deal with Gainclones). I did wonder why the original designer took out the Zobel, since the datasheet insisted it was necessary. As for the buffer, it was merely an excercise. This wasn't meant to be my final Gainclone - only a small step towards learning how to build a proper amp. So I guess I tried things the hard way - point to point wiring, adding the buffer for complexity, etc.
I already have the chipamp.com LM1875 and electronics123.com boards, in order to start experimenting, anyway
OK, this is pretty scary.
I read the star ground thread in its entirety. I figured a couple of things that were wrong with my amp. I've fixed some of them, but I noticed that something was definitely wrong - my neutral was in continuity with Safety Earth!
After much thought and some measurements I figured out what was wrong. I always took measurements with the top cover of the amp off and everything was fine. One of the screws in my enclosure punctured the MAINS power cable (!!!!!!!!!!) and put a short between Safety Earth and Neutral!!!!
Really, really, really scary. I'm actually very lucky nothing happened, but I'll be extra careful next time. I corrected the error, assembled the amp, took tons of measurements, and everything is fine now.
I read the star ground thread in its entirety. I figured a couple of things that were wrong with my amp. I've fixed some of them, but I noticed that something was definitely wrong - my neutral was in continuity with Safety Earth!
After much thought and some measurements I figured out what was wrong. I always took measurements with the top cover of the amp off and everything was fine. One of the screws in my enclosure punctured the MAINS power cable (!!!!!!!!!!) and put a short between Safety Earth and Neutral!!!!
Really, really, really scary. I'm actually very lucky nothing happened, but I'll be extra careful next time. I corrected the error, assembled the amp, took tons of measurements, and everything is fine now.
Yes Andrew. I read the thread and found all of your posts, and THAT was what helped me with my issue! As I said, one of the screws cut through the insulation of one of the mains (scary, scary, scary). I replaced the wire, re-routed it and insulated the mains terminal with shrink tape and then coated everything with hot glue. The amp now works perfectly. I'll add the SLB network you described today (bridge, cap, resistor).
Ultimately, I found it conceptually easy to prove a design is safe now - but quite honestly, had I not found the grounding thread, I'd still be wondering if everything is correctly assembled. That thread should be *required* reading for everyone trying to build a chipamp.
Thanks for all your input. Honestly, it may have saved a couple of lives.
Ultimately, I found it conceptually easy to prove a design is safe now - but quite honestly, had I not found the grounding thread, I'd still be wondering if everything is correctly assembled. That thread should be *required* reading for everyone trying to build a chipamp.
Thanks for all your input. Honestly, it may have saved a couple of lives.
The disconnect network worked beautifully. The amp has the lowest noise floor I've yet measured, and now I know everything is safe. I learned a lot from this amp, and I definitely can't wait to build a "proper" (no buffer, high quality parts, nice enclosure) GainClone. Thanks you all for your help, it was really appreciated!
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.