I was digging around in my basement and I found an old gainclone project that I never comleted. There is 6 PCB's from briangt that I built but never put in a box. I have some 6 ohm speakers and I have though about picking up some 4 ohms.
Can they handle it, they are the panisonic black gates
Can they handle it, they are the panisonic black gates
Gainclone sounds good with 6-8 ohm speakers for me, but 4 ohms not bad. You can use two 4ohm speakers in series. Which chip you are using see datasheet.
For what it's worth, the only trouble I've had with with Lm3886 based amps at 4 ohms is a tendency toward instability. I've built enough of them and tried enough layouts to feel pretty darn certain the issue isn't my own. Pay attention the the 0.7uH inductor in paralel with the 10 ohm resistor shown in the "test" circuits in the specs. First of all, many circuits I've seen leave it off, and in truth it's not as critical at 8 ohms, so they get away with it. I have miswound a few, believing I had achieved the correct inductance. Second, especially if you intend to have two circuits for a stereo amp, or use the two in a bridge configuration, build a preamp into the box, and especially if your main intent is for 4 ohm loads, consider raising this inductance level. The .7uH recommended will add another 4 ohms at 1MHz, where this chip seems to like to oscillate if given the chance. As for me, I like to up that to around 2uH, creating that same added impedance at around 330kHz, which is still only about 0.2 ohms added at 20Khz. If you don't buy an inductor, you may have to do something ugly like winding a coil around a nail to get the higher value. That part is just my opinion, and probably a bit of paranoid overkill. The chips tend to protect themselves well if they do oscillate. But I am speaking from experience, not a theoretical problem. And as used these chips in a lot of performance amps, sometimes with speakers and cabling I won't get to test till I get to the performance, I'd rather minimize my chance of any instability at the worst possible time. The tiny high EQ boost I thought I'd need to compensate was never even audible. But even if it were, it would have been a small sacrifice for the added peace of mind.
check the datasheet for maximum recommended PSU voltage for the speaker impedance you may want to use.
I run the 3886 at near maximum driving 8ohm speakers, but I do not recommend pushing a 4ohm version to near maximum. I suggest you aim for 30W to 40W into 4r0 or 60W into 8r0.
I run the 3886 at near maximum driving 8ohm speakers, but I do not recommend pushing a 4ohm version to near maximum. I suggest you aim for 30W to 40W into 4r0 or 60W into 8r0.
LM3886 run on 4 ohm load without problems but not at full power. I have some experience with the LM3886 GainClone (I've made 4 amplifiers) and they work very well on 8 ohms load ( I've got 50W on 8 ohms). On 4 ohms they've got too hot. Just be carefull!
Last edited:
I will have to check my amps again, it may be the 3875 that was sold on the group buy years ago.
These instructions look really similar to what I used before
http://www.audiosector.com/nigc_kit-users_guide.pdf
I would like to use my parasound's chasis and torroid for this setup, heck even the heatsink
http://www.diyaudio.com/forums/solid-state/144947-parasound-1200-ii-left-channel-static.html
I figured I would make three gains per channel.
These instructions look really similar to what I used before
http://www.audiosector.com/nigc_kit-users_guide.pdf
I would like to use my parasound's chasis and torroid for this setup, heck even the heatsink
http://www.diyaudio.com/forums/solid-state/144947-parasound-1200-ii-left-channel-static.html
I figured I would make three gains per channel.
Well, despite all I said, now having made a dual channel LM3886 based amp that is driving 4 ohms well at rated capacity, I disagree that it shouldn't be attempted. Again, excessive heat can often be caused by oscillation, and going a little beyond the spec's inductor requirement will definitively solve that problem. As for normal heating due to load, one thing I have been doing with all my amps (not just chip based) is arranging my construction so I can mount some fans directly on the heatsync. With pentium series space heaters being what they are in PCs, there are plenty of good choices. As a bonus the added cooling often helps you gain space by using a smaller heatsync. Anyway, with air cooling and the extra inductors, I've bench tested my LM3886 amps into 4 ohm loads up to and even a little beyond the 68 watts RMS they are rated for, and continued the test for hours. I'm also sure when I use the unit as a stage monitor, I push way past rated power on the peaks. I've use those chips in at least a dozen projects in stand alone and bridged configurations and I swear by those LM3886s now ( even though in the beginning I tended more swear AT them! 🙂
Last edited:
Im running my LM3875 chips into 4 ohm loads, but you HAVE to reduce the power supply. You can still get 50 watts into 4 ohm loads but you have to make sure they are heatsinked very very well, and try to not use any kind of thermal shims.
I have my chips directly screwed to pentium 4 heatsinks that are fan forced.
If you try to use LM3875 on 4 ohm loads without reducing the power supply rail to around +-28volts. If you keep them at the normal +-35 volts, they will hit their internal protection before you get full power, and it will sound like crap.
believe it or not, the only real difference between the LM3886 and the LM3875 is their SPIKE protection current limit. the '75 has a 2 amp protection limit, the '86 is around 4.5 amps, which gives them the ability to drive 4ohm loads at their normal voltages. But as peter pan said, you have to make sure they are heatsinked well. The SPIKE protection derates due to temperature, so if your '86s are getting really warm, they will hit protection at lower amperages and cause distortion. You will also need to make sure all your power and output leads are fairly thick and short and to the point, otherwise oscillation will get you.
I have my chips directly screwed to pentium 4 heatsinks that are fan forced.
If you try to use LM3875 on 4 ohm loads without reducing the power supply rail to around +-28volts. If you keep them at the normal +-35 volts, they will hit their internal protection before you get full power, and it will sound like crap.
believe it or not, the only real difference between the LM3886 and the LM3875 is their SPIKE protection current limit. the '75 has a 2 amp protection limit, the '86 is around 4.5 amps, which gives them the ability to drive 4ohm loads at their normal voltages. But as peter pan said, you have to make sure they are heatsinked well. The SPIKE protection derates due to temperature, so if your '86s are getting really warm, they will hit protection at lower amperages and cause distortion. You will also need to make sure all your power and output leads are fairly thick and short and to the point, otherwise oscillation will get you.
Last edited:
- Status
- Not open for further replies.