LM3875 Amplifier Voltage Issue

[kblsnbts]

I am attempting to build a LM3875 amplifier using the schematic found on the National Semiconductor web site, http://www.national.com/mpf/LM/LM3875.html.

I'm using a 200V - 8A bridge, 50V 4700uf capacitors for +/- DCV, and the feedback values displayed in the schematic. I also have a 100k resistor on the input to ground. The volume pot displayed in the schematic was omitted, as I'm using a dedicated line stage for volume control. The capacitor on the feedback loop was also omitted.

My issue is this. When I'm playing a CD at normal volume the voltage measured at the + / - rail is within normal values… 25V. When I turn up the volume, though, the voltage sags to 15V or lower. With no volume, the voltage rises to the chip limits of 35V. Is this normal?

I resolved the issue by creating a shunt load - a 150 Ohm power resistor was connected from the +V rail to ground and a second connected to the -V rail to ground, both between the rectifier and power supply cap. This stabilized the voltage on the rails to +/- 20V at all volume levels.

I don't like the additional current shunted into the ground plane, though. Is there a way around this? BTW - I don't believe this is a transformer issue… I've tried two, both with more than enough amps to drive this chip.

Any help would be greatly appreciated. I’m new to this, so please forgive me if my technical lingo is not correct. Thanks.

[BWRX]

If you've tried two different, adequately rated transformers and still see the rails sagging that much under load then make sure the transformer's primary connections are nice and solid. It sounds like the transformer is being starved for current. Have you checked to see if the line voltage sags under load as well? How about the transformer secondary voltages? You need to track down where the voltage is sagging.

As always, be extra careful when probing around line voltages

[kblsnbts]

Thanks for the reply, Brian. If the transformer was starved for current, though, wouldn't it also sag when the shunt load was in place? With the 150 ohm power resistors in place (connecting the +/-V rails to ground), there is no sag when the volume is increased. I can turn up the volume to levels that hurt my ears and the music plays fine, and more importantly, the voltage stays steady at 20V.

I did take my setup over to a friends house, who has a lot of experience in audio design. We tested many things, including the primary and secondary of the transformers and he was stumped as well. As an afterthought he decided to try the shunt load, which stabilized things.

One thing I'll mention, not sure if it's important, is that my transformer is not center tapped. I don't actually have the secondary, or the transformer per se, connected to the ground plane. My transformer has two secondary wires. The wires are connected to the two AC pins on the bridge rectifier. The +/- pins on the bridge are then connected to their respective power supply caps. The other side of the caps are then star grounded with the input ground, output ground, and the feedback ground point. The 100k resistor on my input pin is also star grounded to the same point.

Thanks again.

[Leolabs]

Did u try another caps???

[BWRX]

Originally posted by kblsnbts
With the 150 ohm power resistors in place (connecting the +/-V rails to ground), there is no sag when the volume is increased.

One thing I'll mention, not sure if it's important, is that my transformer is not center tapped. I don't actually have the secondary, or the transformer per se, connected to the ground plane. My transformer has two secondary wires. The wires are connected to the two AC pins on the bridge rectifier. The +/- pins on the bridge are then connected to their respective power supply caps. The other side of the caps are then star grounded with the input ground, output ground, and the feedback ground point. The 100k resistor on my input pin is also star grounded to the same point.

That bit of info is very important! Without having a supply ground your supply rails are essentially floating with respect to the input signal voltage, which is why your rails fluctuate without the resistor loads on the rails. Those resistor loads on the secondaries are basically defining the supply ground voltage, because they are evenly loading down the rails. You should be using a transformer with a center tapped secondary.

[raudio1969]

A grounded center tap is needed. When you added the shunt resistors, an artificial ground was created. This is acceptable if the wasted power can be tolerated. If you don't have a center tapped transformer, you can use one with dual secondaries of the same voltage/current and connect two of the pins together to make a center tap, but watch the phase. You could also use two equally rated transformers, each with a single secondary and connect the same as the dual secondary.

Richard

[AndrewT]

Originally posted by kblsnbts
BTW - I don't believe this is a transformer issue…

no, the builder is to blame.

[Puffin]

I have used this PS with dual secondaries to very good effect. If you read the article on Power Supplies at TNT Audio you will see that the layout he feels is optimum for Gainclones is virtually identical.



http://vikash.info/audio/p2plm3886/

[peranders]

Just wondering, how many VA does your transformer have? How much uF do you have as smoothing caps.