Guidance for a chip amp noob.

This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Joined 2011
The power ratings are not critical, since even the 20k feedback resistor only has 20mW dissipation worst case.
If you use an output Zobel network, the 4.7R could dissipate more. Typically use a 3W or more resistor there.
If the amp oscillates, the Zobel resistor will be likely to burn.
Last edited:
Just calculate V^2/R, if its more than 100mW consider a larger resistor.

For instance at 40V any value below 15k will be > 100mW, but at 10V you'd have to go below 1k to worry about it.

Below 100mW you can use 1/4W resistors without getting too hot, and these are the standard for through-hole parts.
With surface mount resistors you have to be more careful as common ratings are 100mW, 50mW and so forth.

And many resistors only see small voltages across them anyway, such as 2Vrms signal.

The high voltages in tube circuits mean that even comparively large resistances dissipate a lot of power, much of solid state is at far lower voltages - its current you have to be careful with, current is usually what blows things up with solid state.

Using a current limited bench power supply is very useful when commisioning a new circuit, as you can both limit and monitor the current to check all is as expected.
What Mark said!

I would add: I generally derate power resistors by 3-5x as they are designed to run way hotter than I'd like. 250 ºC is a common case temperature at the max rated power dissipation (assuming 25 ºC ambient). Some modern materials go up to 300 ºC. I don't like hot components in the chassis, so I generally specify a 2 W part for 500 mW dissipation.

There are exceptions. For example in the Thiele (L||R) network and Zobel network (R+C) on the output of an amp, significant power is only dissipated in the resistors if the amplifier misbehaves. So I tend to push those a lot closer to the max dissipation rating as they won't go even near that in normal operation.

For small-signal resistors - especially in an amp designed for music where the average power is much lower than the peak power - Mark's "100 mW Rule" is a good starting point. For the National chip amps it's usually only the feedback resistor from the amplifier's output to its inverting input that you have to worry about.


I am happy the LM-1876 is getting some recognition again. So far, I built two of these amplifiers from Ebay boards and found them enchanting. The 4.7 ohm resistors are spec'd on the data sheet supplied with the boards as 2 watt.
Two light green boards (photo, if it comes across) have done duty in my shop stereo (one amp is for the subwoofer) for 3-4 years. The LM-1876 is a good sounding amplifier, within its limits, even on Magneplanars.
The upright unbuilt board in the photo is a newer Ebay board but it has the same circuit and parts list.
I am currently fabricating a real chassis (formerly a plywood scrap) for my shop stereo which made it convenient for taking the photo after blowing off the heavy layer of sawdust.
I recommend the Ebay boards and also recommend (not in the photo) Cornell Dubilier 935 Polypropylene input capacitors, preferably 100 volt. You can find them in the Mouser catalog if you search long enough.
I use a 15V/15V transformer which should give you the needed 20v/20v on the board. I made my own heat sink mounting bracket since I have metal working capabilities. However, home centers have aluminum strap that you could bend and bolt together if want to make the chassis and heat sink a module like I did.


  • IMG_5288.jpg
    430.9 KB · Views: 133
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.