LM3875 Kit Finished - Could use some more bass

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hey all. I just finished building my first chipamp, the LM3875 kit from audiosector. It sounds pretty good, but it's a little bright and harsh in places where it shouldn't be. Vocals, occasionally, and cymbal crashes. I'm hoping that that will decrease with some burn in.

The bass seems lacking compared to the commercial amp I was using, and the bass I get in some cases seems distorted and uncontrolled. I was thinking that some additional capacitance might help with that, since the kit comes with only 1500uF per rail per channel. Would 3300uF capacitors be an appropriate replacement? And if so, what should I get? I don't see anything from Panasonic FC that goes that high. Thanks!

I should also point out that the chips don't really get warm at all. I have them on big heatsinks with good thermal paste. The transformer is an Avel-Lindberg 250VA 18V+18V toroidal.
 
The audiosector kits have a very minimal PSU. I would consider buying a (couple of) PSU kit from chipamp.com. That PSU is a snubberized design with large smoothing caps and I think most people agree it's a pretty decent PSU offering good performance without being too complicated or expensive.
I'd wait for some confirmation from someone with more experienced but I would be surprised if you didn't see improvement with that PSU
 
AndrewT said:
Hi,
point us to the schematic you have used/built.

I believe that PSU quality is critical to amp performance.
CarlosFM's implementation comes closest to this goal and uses +-27.2mF per channel not +-1.5mF per channel.
The schematic is the stock 'classic' implementation from Audiosector. Here's the .pdf - schematic is on page 3. http://www.audiosector.com/nigc_kit-users_guide.pdf

Yeah, I was looking at the Chipamp.com PSU, because 1500uF seems like a pretty small amount of capacitance for a power amp. The audiosector kit puts a 1500uF capacitor on each rail of each channel, so 3000uF per channel. The PS board just has the rectifier diodes and some 4.7uF caps. It is nice that the 1500uF caps are close to the amp boards, though. If I switch out this PS for the chipamp one, can I just leave the 1500uF caps on the amp board?

I'll try using the carbon feedback resistor instead - that may help too.
 
Yes you can...
I have been using 1000uF per rail at the amp and 16,000-20,000uF on the PSU board to good effect. As a matter of fact, it is usually suggested to have 1000uF or so right at the chip when using a larger bank of caps at the PSU


7/10

edit to add: Johan, I just noticed that you are a Seattle local. It would be interesting to compare implementations sometime...
 
seventenths said:
Yes you can...
I have been using 1000uF per rail at the amp and 16,000-20,000uF on the PSU board to good effect. As a matter of fact, it is usually suggested to have 1000uF or so right at the chip when using a larger bank of caps at the PSU
Ok, great. That's what I figured, but I wanted to be sure. Lots of smaller caps in parallel provides a lower ESR anyway, I think, so having those smaller caps close by should help.
 

Attachments

  • b3.jpg
    b3.jpg
    62 KB · Views: 1,405
Peter Daniel said:
You can use more capacitance with my PS boards, if you want. You can also implement snubbers and Zobels, if that's your cup of tea:

http://www.diyaudio.com/forums/showthread.php?postid=584429#post584429
Ok. I'm looking at the PCB here:
http://www.audiosector.com/images/lm3875_se_pcb.gif
And it's kind of hard to see what should go where. R3 is a resistor for the LED, right? C3 and C4 should be 0.1uF polypropylene, and what about Cz on the amp boards? What about R1/R2 on the PS board? And what terminals do the big 35mm caps connect to?
 
I'm not sure extra power supply caps is going to make your chip amp give better bass 'output' (which they seem to lack?), it might improve tautness and bass 'quality' though.

I have an LM3876T amp which is the same as 3875, but with a mute circuit (which I have bypassed). I agree with you that they have weak bass. Reading between the lines in many posts on this forum, this seems to be a common complaint.

Although lacking a bit of bass, the beauty of chip amps is their excellent overall tonality, good soundstaging and sweetness ...attributes that are rare among solid state amps when compared to valve amps (which nearly always sound nicer).
Therefore, a lot of people on this site use efficient speakers such as Lowther & Fostex as they bring out the best in the chip amps, which is in the region of midrange and treble.

Regards,

Steve M.
 
Ex-Moderator
Joined 2005
Everyone complains about a lack of bass, but has anyone actually measured the frequency response of their amps? If so, what results do you have that show the chip is the cause of the "lack of bass"?

It's simple to measure using a computer for a sine wave generator and a true RMS multimeter with decent frequency response. An oscilloscope makes things even easier.
 
Well, I can hear the frequency response, and it's not bad. The frequency response isn't the problem, it's a lack of control. Hard hitting bass passages, especially in classical music, sound lifeless and uncontrolled.

*Edit* Maybe lifeless isn't the right word. It's hard to describe...it lacks presence and authority. And every now and then it just sounds distorted.
 
BWRX said:
Everyone complains about a lack of bass, but has anyone actually measured the frequency response of their amps? I
Touche!

I personally haven't measured, but I have a Gainclone sitting next to a
"Linkwitz" chipamp, and the latter subjectively has much better bass response IMHO--tighter and less strained. Again, to each their own, but for my taste, the Gaincard circuit doesn't reproduce bass on my speakers as I would like.

PS I hate carbon comp resistors too! :D
 
Lack of bass could mean lack of driver control.

This aspect can be adjusted by the rails capacitance. Try 220uF onboard the amp. Bear in mind that the tweeters lose control first, and can resonate or squeak. So, in case the woofer is in control but the tweeter isn\'t, then the rails capacitance is still too large.
For reference, 2200uF is going to sound a lot like a 1970\'s transistor amplifier, with floppy (boomy) bass and a beautiful warm sound.
Halfway in-between, the tweeters will squeak, and just right is about the next step smaller. The \"right\" spot probably also brings the amplifier heat output to about half as much.

You can also adjust by the capacitance of the input filter cap, wheras too large is a problem for driver control, and too small isn\'t quite enough low bass.

Its also possible to use a 100v (or better) bipolar cap from + to - rails at the amplifier board. This really acts as a filter, and I\'m not sure what the recommended capacitance size is or even if the practice is okay or not. For sure it will make some sort of change to driver control.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.