Dear members, I am a german High-End enthusiast since several years, but to date I had not enough money to realize my own high-End dream. In the last time I bought really good loudspeakers (JamLab Electra 5000Euro). I tested them with a really good CD-Player (Sony SA50) and a integrated full line valve amplifier (Octave V40 3500 Euro) in the reference class as a Mark Levinson. This combination sounded extremely good and I thought to buy a better SACD (Marantz SA8400) and the same valve amplifier. Then I read of the LM3886 amplifier. Because of my know how in electronics I decided to test a self made amplifier before I spent 3500 Euro. Because of my 4Ohm systems I built a parallel bridged inverting LM3886 with SMD-outfit, DC-correction, snubberized power supply with ultrafast diodes, paralleled Silmic and 20000uF capacitors. The volume control was a alps 10K pot. All was ties as small as possible together. For first tests I used short wires and the same connectos as in the reference systm. The first sound comparison to my reference combination was amazing. It sounded like a valve amplifier with extretemely good warm and crystal clear sound. But when I tested the bass response with extremely good recorded tracks (down to 20Hz, extreme impulses) it was in comparison to the Octave amp as crispy, but the volume was flat and not deep. Do I need a buffer (OPA 627) behind the alps pot or is this a main problem of the LM3886?
Hi Manuruthen
I don't think that "flat bass" is a main problem of the LM3886. If anything it is the opposite.
Without a schematic it is not easy to help you, but I will it a shot.
My first “Item on the list” is the ”DC-correction” that you talk about. This can be done in several ways and some of them act as high-pass filters (DC-servo and capacitors in the lower feedback arm.). Is it possible that you have miscalculated or simply used the wrong value in the circuit you use? It is not that difficult to get it wrong by factor x10, that might move it into the low-frequency range in question?
I don’t think that a buffer will solve your problem (it might solve others). If you don’t have a buffer where you really should have it, it most offend results in high-frequency roll-off and not low-freq. roll-off.
Is your circuit paralleled LM3886 or both paralleled and bridged LM3886??
Have fun
Thomas
I don't think that "flat bass" is a main problem of the LM3886. If anything it is the opposite.
Without a schematic it is not easy to help you, but I will it a shot.
My first “Item on the list” is the ”DC-correction” that you talk about. This can be done in several ways and some of them act as high-pass filters (DC-servo and capacitors in the lower feedback arm.). Is it possible that you have miscalculated or simply used the wrong value in the circuit you use? It is not that difficult to get it wrong by factor x10, that might move it into the low-frequency range in question?
I don’t think that a buffer will solve your problem (it might solve others). If you don’t have a buffer where you really should have it, it most offend results in high-frequency roll-off and not low-freq. roll-off.
Is your circuit paralleled LM3886 or both paralleled and bridged LM3886??
Have fun
Thomas
manuruthen said:
First of all, if you are using an inverted GC it might be better to isolate it with a buffer. But I don't think that is related to the bass response.
What certainly is, is the cap value you are using at both GC inputs. Both value and quality of this cap might have a great influence on the bass you are getting.
Have you mated all resistors, including the output ones?
How much voltage are you using on your supplies? Are you sure your transformer is up to the task?
How does the amp behave at lower volumes?
4 ohm speakers can be problematic, as they might have dips exactly on the lower frequencies you are talking about.
Have you looked at your amp on an oscilloscope, and tried it with test resistors at the output?
Carlos
I tend to agree with Thomas. Post the schematic with the actual values (including the PS), so we don't just guess what it could be. Check the values of the caps... I think if you're happy with the upper range you can get the lower range the way you like it, no-problem. In my case I have 6 ohm/90dBm speakers and a single LM3875 per channel. The base is excellent. No complains at all.
Greg
Greg
Thank you for all your replies,
I have no plans that I can post you, because I built the layout with power point (will post a picture later) with SMDs as possible, because they offer me the best possibillities for extremely short wire lenghts and bypassing. The design is the same as in the manuskipt Overture PA100 with DC-Correction at the noneinverting input. No coupling capacitors anywehere in the signal path.
First: The loudspeakers are easy to drive by a weak amp, because they have a flat impedance curve over the whole frequency range which never dropps down under 4 ohm. They have no complex phase reversals, are highly efficient (90dB/W). With 1W you hear loud music, with more than 10W you produce earthquakes.
My power supply is directy near as possible on the amp board:25V/160W Torroid-Transformer, fast MUR360 diodes bypassed with 47nF, BC cap 22mF with 1 ohm 200nF snubber across its pins and directly near the paralleled amps a 470uf Silmic. Of course several bypassing 100nF ceramic-SMD-caps directly across the pins of the amps. Al resistors (gain 20:200K-10K) and the paralleling resistors of 0,1 ohm are matched to 0,05%. I have no capacitor at the input that will possible iintroduce a low frequency rolloff. DC-correction is made like in the Overture manual in the base of the noninverting input (2,2uF, 470k:high pass around 1Hz). I will test if it is the same when I switch this feature off. I tested the amps in frequency response: No problem from 3 to 100kHz in sine and square wave, rise time in the range of 2us for 30V, no problem with resistive loads down to 2 ohm and capacitive loads to 100nF.
I have no plans that I can post you, because I built the layout with power point (will post a picture later) with SMDs as possible, because they offer me the best possibillities for extremely short wire lenghts and bypassing. The design is the same as in the manuskipt Overture PA100 with DC-Correction at the noneinverting input. No coupling capacitors anywehere in the signal path.
First: The loudspeakers are easy to drive by a weak amp, because they have a flat impedance curve over the whole frequency range which never dropps down under 4 ohm. They have no complex phase reversals, are highly efficient (90dB/W). With 1W you hear loud music, with more than 10W you produce earthquakes.
My power supply is directy near as possible on the amp board:25V/160W Torroid-Transformer, fast MUR360 diodes bypassed with 47nF, BC cap 22mF with 1 ohm 200nF snubber across its pins and directly near the paralleled amps a 470uf Silmic. Of course several bypassing 100nF ceramic-SMD-caps directly across the pins of the amps. Al resistors (gain 20:200K-10K) and the paralleling resistors of 0,1 ohm are matched to 0,05%. I have no capacitor at the input that will possible iintroduce a low frequency rolloff. DC-correction is made like in the Overture manual in the base of the noninverting input (2,2uF, 470k:high pass around 1Hz). I will test if it is the same when I switch this feature off. I tested the amps in frequency response: No problem from 3 to 100kHz in sine and square wave, rise time in the range of 2us for 30V, no problem with resistive loads down to 2 ohm and capacitive loads to 100nF.
''The loudspeakers are easy to drive by a weak amp, because they have a flat impedance curve over the whole frequency range which never dropps down under 4 ohm. They have no complex phase reversals, are highly efficient (90dB/W). With 1W you hear loud music, with more than 10W you produce earthquakes. ''
I think that LM3886 is capable of driving your loudspeakers without paralleling/bridging( check the graph power vs. suply voltage).I also think that sound quality will not benefit from paralleling/bridging.
I would first try the basic non inverted LM3886 schematic(datasheet-page1) with 1000uF caps on IC pins without Ci cap and without any bypass caps.
I would also remove bypass caps on power suply MUR diodes.
After burn-in period and listening impressions I would try adding bypass caps step by step and listening if this improves sound.
Regards;
Igla
I think that LM3886 is capable of driving your loudspeakers without paralleling/bridging( check the graph power vs. suply voltage).I also think that sound quality will not benefit from paralleling/bridging.
I would first try the basic non inverted LM3886 schematic(datasheet-page1) with 1000uF caps on IC pins without Ci cap and without any bypass caps.
I would also remove bypass caps on power suply MUR diodes.
After burn-in period and listening impressions I would try adding bypass caps step by step and listening if this improves sound.
Regards;
Igla
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.