LM3875 - newbie gainclone project

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
P.S. that cap on the input is one of my favourite Nuuk GC gripes... although it will certainly work, it only needs to be 4.7uf, which makes it enormously much more affordable to change it to a non-polar film cap.

That 56k resistor to ground is normaly soldered directly from + to - on the input RCA socket, makeing the negative tab the audio groundstar...

Regardless of input cap, you should not be seeing those output DC at rail levels, usual suspect is bad solder connection on one leg of a cap in the power supply...

At least its not a short between a power rail and the input, or you'd have fried a chip nice and toasty... makes quite a nasty sizzleing sound.... if it doesn't blow like a little grenade...

Other notes: the 100nf caps are completely optional... listen without, then with, and form your own opinion and habbits...

P.S. Welcome to the GC club! :D :D ;)

On the little ground network, (the 220nf and 100R resistor on the bottom, you HAVE to paralell this with 2 BIG say 8amp diodes each connected in oposite polarity to the other...

I have build tonnes of those babies, if you leave out the diodes the 100R resistor will burn (possibly catch fire) if there are any shorts in your amp... although if you are close enough to the power switch to flip it, when the smoke starts boiling off it will certainly save your amp... worked for me... with the diodes, the mains trip, without any burned resistors...
 
Hi,
post19 referred to non-inverting chipamp.
The schematic in post20 is inverting.
Your comment that gain goes through the roof is correct. It is limited by the Rs seen by the chipamp and could be very low ~100ohms depending on what preceeds the chipamp.

However, the capacitor presents a high impedance at DC and so the gain reduces to very low at DC. Did you omit the cap as well?

Nordic,
4u7F and 10k (=47mS RC) gives adequate bass. Have you tried 100mS RC at the input?
 
No, the 56k sets up the input impendance the source sees and provides a path to ground for DC on the input signal, the higher the easier to drive... (less current needed - ohms law)

The 10k and 4u7 cap forms a highpass filter which passes frequencies above a preset limit, the specific combinatiion is 3db down at 3Hz, (means half as loud as other higher frequencies) and drops further towards 0.
 
Fisrt go doublecheck if it is + or - voltage potential...
As always double check solder joints... use a multimeter continuity test... (the beep)

It may just be a not so good chip... (i.e. the reason for the spread between 7 and 11A max output) they are just not all equal...

That said 300mv is a bit excesive, and probably audiable as hum... but it may be a sign of oscillation too..

Does the chip get very hot, with no signal...? anything else getting hot?

Lets hear from you before we check further...
 
I moved the GND source from caps of left channel to central ground point(I admit, I was to lazy to do that before...but somewhere I read that gcs are very,very sensitive, so I done that)

Than I checked the fuse, it was burnt because of unknown reason(2A, fast blow)

Now it both channels seems to work, nothing is hot (before it wasnt too) BUT Im afraid of instability...I dont want to burn my $$$ speakers

EDIT:I noticed buzzing, but it is gone if I connect GND with earth.Can I do that?
 
Nordic said:
P.S. that cap on the input is one of my favourite Nuuk GC gripes... although it will certainly work, it only needs to be 4.7uf, which makes it enormously much more affordable to change it to a non-polar film cap.

In NUUK's site, Nick makes it clear that he used 47uf because that is what he had in his parts bin at the time. He recommends 2.2uf to 4.7uf film polyproplene as ideal.

I used a Vishay X2 4.7uf polyproplene in mine, but recently removed them and got a better sound. I am using a Kookabura as a preamp.

Weird thing happened my DC offset reduced on removing the caps previously I was getting 22-34mv (depending on which of 4 chips I measured) but now my offset measures 5mv on each chip. Not sure why???

If your dc offset is such that you dont need the caps, definately try the amp without. I would not recomend the caps that I tried on mine, if I had to use a cap, I would try something like a sonicap or clarity cap.

One other thing I noticed. The background noise on my amp dropped when I removed the caps. On one channel, some noise had started to be audible. Now its gone.
 
Ex-Moderator
Joined 2005
justblair said:
Weird thing happened my DC offset reduced on removing the caps previously I was getting 22-34mv (depending on which of 4 chips I measured) but now my offset measures 5mv on each chip. Not sure why???

With the coupling cap blocking DC at the input of the amp, any DC offset you see at the output is due to the output offset voltage of the chip amp. When you remove the cap you are now DC coupling the input signal - any DC voltage on the input will be amplified at the output just like the audio signal. If your source has a 1mV DC offset and your amp has a gain of -20V/V then you should see -20mV DC at the output. Combine that with the ~25mV offset of the chip and you have your 5mV offset. It will be different for every setup. If your source has a small offset voltage I'd recommend trying the amp without the input cap. That's one less component in the signal path to color the sound.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.