Does anyone have a view on whether the two decoupling caps (in post1) should be connected together first and then taken to the Audio Ground?
AndrewT. I hadn't thought about taking the grounds from the caps separately, as I always thought the considered way was to put a stout piece of wire as the "power" ground.
http://i73.photobucket.com/albums/i239/saxonsex/GC1.jpg
http://i73.photobucket.com/albums/i239/saxonsex/GC1.jpg
AndrewT said:
My view is that they SHOULD be connected together, near the chipamp, and should also connect to ground very close to each other, there. Otherwise (or even additionally), another small cap should connect directly between the V+ and V- power pins.
Hi Gootee,
I have already sent him a private message with instructions on where to find the less dangerous disconnecting network and how to contact the moderators to remove the dangerous diagram and substitute the safer one.
It's just a pity he does not realise how many newbies may have read his first post in the intervening 17days of inactivity.
I have already sent him a private message with instructions on where to find the less dangerous disconnecting network and how to contact the moderators to remove the dangerous diagram and substitute the safer one.
It's just a pity he does not realise how many newbies may have read his first post in the intervening 17days of inactivity.
AndrewT said:
The image is hosted on photobucket. All that needs to be done is to change the link url in the post to the url of the new image. Post the url of the new image and we can change it.
The updated diagram is :
http://i188.photobucket.com/albums/z72/widhibrata/audio/GaincloneGrounding2.jpg
How do I update the 1st post? ...
http://i188.photobucket.com/albums/z72/widhibrata/audio/GaincloneGrounding2.jpg
How do I update the 1st post? ...
Mike514ml said:
Hi Mike, I don't understand about wire gauges so I posted that picture (I have medium-size thumb 😀). Basically the largest wire is from good speaker cable about 4mm in diameter, the smallest is jumper cable about 0.5mm diameter. The middle size is about 1.5mm diameter.
Mine is built using single supply board for 4 channels. I don't know how using 4 PSU will help more as my amp is very quiet ie. no hump/thump...
gainphile said:
You can only edit your posts for up to 30 minutes (I think) after posting. After that you need to ask a moderator to do it.
I've updated the url in your first post so the correct diagram should now show up.
Thanks for updating the diagram 😎
Now, about "private messages" ... where is this located? I can't find private messages as AndrewT mentioned above?
Now, about "private messages" ... where is this located? I can't find private messages as AndrewT mentioned above?
or go to the bottom of a posting and choose to click on "Email" or "report"
Email goes to the poster and report goes to the Moderating team.
Email goes to the poster and report goes to the Moderating team.
gainphile said:
Gainphile,
Thanks! The speaker return wire is bigger than I thought. I can work with this info.
Of course, it's kind of ironic that your amp sounds better if you don't use gigantic wire for the speakers! 😀 At least internally.
I think I'll go ahead and use a single supply board for my BrianGT amp, since I'm using a single (center tapped) transformer. It'll simplify the ac wiring. Of course, I'll stuff the largest wire I can into the ground connections!
Thanks again,
Mike514ml
In my own amp I have as follows:
Each channel's rail decoupling capacitor grounds have a common ground trace. This include the ground pin of the LM3886. This is returned via one fairly thick wire (same thickness as the power supply wires) to the central ground.
Signal grounds are seperated and connected at one place - the input jacks. This in turn is connected to central ground via a 10 ohm resistor.
Speaker grounds are returned direct to central ground. The zobel networks are on the output posts, so the speaker ground is also the return for the zobel.
The central ground is on the capacitor side of the supply - not the rectifier.
Central ground is connected to chassis earth via a 10 ohm 3W resistor bypassed with a 100nF capacitor. I could add diodes but I suspect this combination would allow more than the 30mA trip point of the household RCD to flow, so it is safe.
No hum heard at all 🙂
Each channel's rail decoupling capacitor grounds have a common ground trace. This include the ground pin of the LM3886. This is returned via one fairly thick wire (same thickness as the power supply wires) to the central ground.
Signal grounds are seperated and connected at one place - the input jacks. This in turn is connected to central ground via a 10 ohm resistor.
Speaker grounds are returned direct to central ground. The zobel networks are on the output posts, so the speaker ground is also the return for the zobel.
The central ground is on the capacitor side of the supply - not the rectifier.
Central ground is connected to chassis earth via a 10 ohm 3W resistor bypassed with a 100nF capacitor. I could add diodes but I suspect this combination would allow more than the 30mA trip point of the household RCD to flow, so it is safe.
No hum heard at all 🙂
Hi jaycee,jaycee said:
I'm here in a little doubt since, for example, Mick Feuerbacher on his site http://dogbreath.de/Chipamps/ThreeResAmp/ThreeResAmp.html (comment with figure 6) connects the ground pin (pin 7) of the LM3886 to signal ground.
What should be the better approach?
No hum/noise with ear close to the woofer/tweeter with, let's say, a volume pot up to 12 clock?
Thanks,
Baka
Baka said:
I never liked point to point construction. I also dont agree with a number of things in his design, for example, he has an outboard power supply consisting of the transformer and some 30,000uF of capacitance per rail - but then he has a small power supply connector with tiny pins.
His star grounds are at the chip. He doesn't really connect pin 4 to signal ground - he just doesnt have any isolation between the two grounds.
You can try it, but I would think it's more prone to hum myself.
Baka said:
No hum - some hiss, but no hum 🙂
Baka said:
On my system there's no hum with the pot at maximum, the amp is dead silent. I use a LM3886 at +/-30V, 3x4700uF/rail, 330uF/rail on the boards. Speakers have 92dBm efficiency. I will make a diagram to explain better.
jaycee said:
Thanks, jaycee.
I have already implemented the point to point construction like it was explained on this thread, but on the gainphile's schematic ground pin on the chip and the way it should be connected is missing. Maybe LM3875 doesn't have the ground pin.
Other than that I discovered that a volume pot body itself (Alps blue) generates a strong hum and noise disregarding the properly connected its ground pins! When I touched the pot audible hum and noise could be heard through the speakers. I have tried then to connect a metal part of body to a central ground and eventually solved the problem.
Should I, maybe, connect the body via disconnecting network (R||C||D1||D2) to earth rather than power and signal central ground since the pot is additionally connected to extension rod and aluminum knob?
With my ~85dB efficient speakers with a volume pot up to 12 clock I can detect some hiss but close to the tweeter.
- Status
- Not open for further replies.