lm1875 psu design

Status
Not open for further replies.
Ok well I had some time to work on the PSU for my amp today. I've poured through all of the psu guides and I think I am close. The only thing I'm not sure of is how it will work with the toroid I am trying to get (25-0-25 dual seconds 160va from a group buy). I'm not sure if the 1875 can take the rails that will result.

anyway here is the schematic for the PSU NOTE: I an using mur860s in it but all eagle had was the 820. Same size just didn't want to make a new part because of a name difference

An externally hosted image should be here but it was not working when we last tested it.


The bulk caps are nichcon ones and all of the others are wima poly film.

Here is the layout thus far
An externally hosted image should be here but it was not working when we last tested it.


Hows it looking?
 
Aaaaak … the LM1875 is only rated for 60V (+/- 30V) max. The dual 25 volt transformer will yield approximately +/- 34 volts (68 V total) of rectified DC. :hot:

I’d go for a transformer with dual 18 volt secondaries, which would yield approximately +/- 24 volt rails. At that voltage the chips can be expected to idle at 3.36 watts each typical (4.8 watts max).

On the layout I think I’d beef up the ground bus too. Also, the D5-D8 diodes appear to be backwards (I think, unless you plan on mounting them upside down, or will the diodes stand up?). Looks like they're backwards in the schematic too.
 
kaos said:
I’d go for a transformer with dual 18 volt secondaries,
..if your speakers are 8 Ohm. With 4 Ohm speakers use a 2*15 V transformer.

kaos said:
Also, the D5-D8 diodes appear to be backwards (I think, unless you plan on mounting them upside down, or will the diodes stand up?). Looks like they're backwards in the schematic too.
:smash:

Did you use the correct package for C1-C8? They look quite big for 22 nF.
 
I've upped the ground trace to .15 inch so that should be better. I may just used this for a lm3886 instead. In which case I will make a seprate board for the cap bank and up them to about 10k uf per channel.

On the diode r+c network what should I use for the resistor to pair with a 22nf cap?

The 22nf caps are 630v with a 10mm grid. Its just a large size cap I double checked the package size (4.3x13.3)
 
a 5mm package will have about half the inductance of a 10mm package if it's been produced as a low inductance type.

Don't worry about resistor values just now.
They will need adjusting once you get a scope on the supply rails.
 
ok I switched to wima PP film caps in the 5mm grid so now they are about 7.5x7.5 square. What resistors should I use for the filter? Carbons are massive at 6x15 mm. Is there something smaller that I can use and what wattage for them? Carlosfm made reference to carbon or film resistors in a rc filter but i'm not sure if those are carbon film or metal film.
 
Hi BionicSniper,

Here's a similar layout. Note that diodes D2 and D4 are flipped over.

regards
 

Attachments

  • diode bridge.gif
    diode bridge.gif
    25.5 KB · Views: 540
ok I think I have it now.

All of the diodes have been fliped acordingly and I fixed the reversal of the sch/pcb. Even better I may be able to get these boards fabed for free. I'm joining the solar car team here at Iowa State as they need more EE's so I should be able to get anything I want made in with their orders

The diodes are onSemi 860's

The RC caps are Wima Polypro Film 100V .1uF 5%
The RC resistors are 1/2 watt 1% 100ohm metal film

the bulk caps are nichicon 1500uf 63v audio (LKG1J152MESZBK)

and the decoupling caps are Wima polyester 1uf 2.5% 63v

An externally hosted image should be here but it was not working when we last tested it.

An externally hosted image should be here but it was not working when we last tested it.
 
ok I switched them to series. every connection off of the diode+rc needs to be from a diode lead correct?

Also I cant find a heat sink anywhere near small enough for the space that I have. I have never seen a heat sink used on any of the other bridges aswell
 
Status
Not open for further replies.