diyAudio (
-   Chip Amps (
-   -   LM1875 miniature desktop amplifier build (

Dr_EM 13th June 2010 08:07 PM

LM1875 miniature desktop amplifier build
I thought I'd start a build thread for this, though it's already well on its way. Pictures warning!

I wanted a desktop amplifier for powering computer speakers. These will be:

and the requirement is for Hi-Fi sound but minimal power. Actually, only about 1Wpc is required, so the LM1875 is overkill. I already had a +/-17VDC supply though and headroom is fine! I initially tried a discrete design, which I killed trying to set the bias :o . I had these chips around so thought I'd put them to use as a sort of plan-B!

I also had a small Hammond heatsink box. It's very compact at 100x90x50mm! I understand it won't provide sufficient cooling for high power operation, but the speakers bottom out with just 1W each so overheating is unlikely in practise.

Without a suitable small PCB, I set about a stripboard design:

here's what I came up with:

This is 1 half of the board populated; a complete mono amplifier. The other side will contain the other channel. The speaker connection is via the 5 way header, 2 pins per channel. The other headers are signal grounds and signal inputs. The board is mounted with a bar clamp and the chip is soldered on the underside along with the small bypass caps:

I finished this and the other half of the board a few days ago. They both tested fine in basic tests.

Today I have been working on the power distribution and rear panel connections. This amplifier is to incorporate a 'digital' power switch. This is mostly so I can use a light touch tactile switch to power it up, since the unit is so light a full current switch might push it back when used! This meant I needed a relay. I also wanted additional local smoothing capacitors, so included 1000uf per rail inside the unit. Here's what I came up with:

A central point for ground returns is included. The board is mounted to the rear panel by 5 nylon screws and 3mm nylon stand-offs:

The board mounted and all connections made. The relay is upside-down and affixed with a sticky pad! Time to link the amplifier to the power board and screw things down:

And that's where I'm up to now. It's a tight squeeze! Those long connecting wires didn't help, but I got the back on after some re-arranging of them. I've tested it briefly and it seems promising. I can power up the amplifiers with the relay and there is <1mv offset on each; connecting a speaker yields nothing but the faintest hiss, which is surprising! I suppose the lack of any AC inside the box helps.

What needs to be made is a small board at the front containing the logic gate latching circuit, the volume pot and a simple headphone amplifier. Watch this space :cool:

john blackburn 14th June 2010 01:27 PM

It is coming together very nicely indeed, I like the chip clamping idea.

The 1875 is a really nice sounding chip if you can live with the low power, I think you will be very happy with the result.


Dr_EM 14th June 2010 09:17 PM

Thanks! The power should be ample since the speakers are 87dB sens each and will be <1m away, desktop use.

Today, I built the latching switch circuit and affixed the volume control to the board. The headphone buffer is being built, but on hold due to the fact I found my 1/4" socket couldn't fit in the case so I'll need to obtain a 1/8" one!

This board then slots into the guides inside the case. It is held by the potentiometer nut also. You are seeing it upside down in this photo:

At this point I simply had to close it down. I love how these boxes accept standard M3 screws, so I could use nice allen cap head ones. I think I'd prefer a metallic knob, but I only had the black one around. Hope this isn't too many pics!

I used 2mm 'banana' sockets on the back, which surprisingly handle 5A! You can see the chip clamping bolts on the underside, they are tightened as much as I feel comfortable with, I am sure it's entirely possible to crack the chip with a clamp like this and allen bolts. The unit fits easily into my fairly average hand!

You can also see it in testing. All is promising so far. Both channels work and have very minimal hiss and no discrenable hum/buzz at all. The volume seems well calibrated, I used a lower amplifier gain than suggested to get more useful range out of the dial, I know pots can become unbalanced at the extreme end. Quality seems high, though currently only running into my cheap-as-chips test speaker!

I will check it over with my scope thoroughly tomorrow and run some sine waves into a dummy load to check for anomalies, probably only push it to 2W though. What I need now are the speakers, but that's another project :D :up:

Dr_EM 15th June 2010 02:01 PM

Ok, I've been testing this today. It all seems ok :yes:

Firstly, the output with the volume dial at 0. Scope on 5mv/div, so about 1mv RMS self noise. Also, the scope was set to DC, so this shows how minimal the offset it, it measures about -0.5mv. Quite respectable :)

Turning up the dial with the amplifier linked to my PC soundcard output quickly reveals a LOT more noise! A lot of digital hash on the output, as most internal soundcards seem to produce. I like to use my old MD deck as an external DAC when I can :)

Here's the output into an 8ohm resistive dummy load, 4W RMS power. Using PC as source (WinISD sig gen), 1KHz, the digital hash does show a bit on the waveform, I confirmed this was not related to the amplifier by scoping the output of the PC directly, revealing the same noise.

Clipping at 13.5V peaks (5v/div) into 8R load. About 11W RMS per channel, one channel driven:

The unit became quite warm after running an extended period 4W sinewave, but also cooled off quite quickly, so dynamic power should be comfortable.

Here's the external PSU. 75VA, 2x12V AC transformer, rectified and smoothed with 4400uF per rail (extra 1000uF local inside amp). Connection is via XLR umbilical:


AndrewT 15th June 2010 02:29 PM


Originally Posted by Dr.EM (
the output with the volume dial at 0. Scope on 5mv/div, so about 1mv RMS self noise

this looks quite high.
scope looks like 3 to 4mVpp and confirms your estimate of ~1mVac.
That should be 10 to 20dB less. Aim for <0.1mVac and if you can't achieve that settle for <0.3mVac with the bandwidth set to ~100kHz.

Dr_EM 15th June 2010 03:21 PM

I'm not sure I could do much to lower it seeing as it's all integrated into the chip? There's no hum/buzz component to it, it's just hiss. I've used mostly standard resistance values, only a smaller feedback resistor for lower gain, which should lower self noise if anything.

I can only hear it using my 87dB speakers with my ear about 3" or less from the cone and aside from that, turning the dial even fractionally upwards with the PC connected completely swamps the self noise with the soundcard noise, which becomes reasonably audible above about 3/4 setting.

I could check my regulated Symasym amplifier for noise, but I doubt even that is as low as 0.1mv!

All times are GMT. The time now is 09:49 AM.

vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2017 DragonByte Technologies Ltd.
Copyright 1999-2017 diyAudio

Content Relevant URLs by vBSEO 3.3.2