LM3875 outputing DC

h3ad

Member
2007-09-25 4:14 am
hey all, recently got hold of two LM3875's and wired one up as per the diagram of the basic circuit in the datasheet how ever with out an input with -+17v power (for testing will get bigger trans) the amp outputs around 15-16v DC! what have i done wrong ?
 
haydn, have you drawn the schematic yet?
Sometimes I just begin building a circuit by just having a more or less similar circuit from a book or article on my mind.
That is a bad habit and intodruces all kind of errors.
Take pen and paper and draw the circuit yourself. This helps focusing and maybe you will recognise the error.
You can also post your drawings here.
Regards
 

h3ad

Member
2007-09-25 4:14 am
no i havnt drawn a scematic, i am bulding it simple point to point using the "typical application" in the datasheet, and even with a load it still shows high DC voltage
i have tried 2 chips, and both do the same thing... and both power rails are working fine.... im stuck please help :(

thanx heaps haydn
 
AndrewT said:
If a chip amp loses a power rail, does it output DC?

Its possible that a capacitor is backwards. Check to see if one is warmer than the other. Check with a different power supply, even if its a cheapo.

Check the NFB (gain loop) connections. Be sure to use at least 1k, not the common 680 ohm mistake. Be sure to use a capacitor from 22uf to 47uf. If raising the gain is your thing, change the "other" resistor, which does that job with less DC output.

Also, do use a nice clean gel flux along with an always-clean soldering iron. You can keep the iron brilliant with a ball of #0000 grade of steel wool (a furniture refinishing supply). Just a quick swipe makes the iron bright. The NFB parts are critical.

DC, up to 1.6v can be caused by a potentiometer; however, your 16vdc is too severe to be caused by the input circuit.
I think its either lopsided power or a funky connection (or wrong value or defective/missing component) in the NFB.

Some kits do not have the output zobel. That can allow the amp to send out DC under certain conditions.

The chips do vary in manufacture. Some put out more DC than others. That's why the cap in the NFB and the output zobel aren't optional. Otherwise its only luck if two chips match up for stereo.

Still. Swap that power supply and check the NFB. Those two areas could do your 16v dc.
 

h3ad

Member
2007-09-25 4:14 am
i habe checked the NFB and tried new resistors, and we have tried 2 different power supplys at different voltages and capacitors... and still no luck, it just gives out what ever the positive rail is :S, BUT when it hits about 5vdc after been turned off it goes into -5v then drops... im stuck


thanks haydn
 

paulmac

Member
2004-10-23 3:36 am
G'day h3ad

You mentioned "without an input" in your first post.

Does this mean you have the 10k pot in circuit?

What happens when you actually connect a source up to it or ground the input?

What I'm getting at is that without an input, ie if you just left the 1k Rb floating, your amp may oscillate, or as you've found, sit hard up against a rail.

My 2 cents worth.
 
h3ad said:
i habe checked the NFB and tried new resistors, and we have tried 2 different power supplys at different voltages and capacitors... and still no luck, it just gives out what ever the positive rail is :S, BUT when it hits about 5vdc after been turned off it goes into -5v then drops... im stuck


thanks haydn

Does it do this if you were to borrow the exact input circuit from the LM1875 PDF and use it on your LM3875?

Do you have an output zobel? You can also borrow that from the LM1875 PDF, because its a bit more effective.

Does your NFB contain a 1k resistor, a 22uf cap, and a 22k resistor? (optional is use of 47uf cap and 56k resistor)

Have you reassembled one of them, using flux to put it back together?