My first LM3875 gainclone

maurycy

Member
2005-12-24 4:34 am
NJ
Actually few months ago I have completed my first gainclone based on BrianGTs LM3875 stereo kit. Besides test light bulb turning on after assembly, I have not encountered any major problems. Removing all parts out of the case and putting them back on fix the bulb problem. Kit was pretty easy to assemble but I wish the manual was updated to the latest revision with better description of the pot and CT trafo instalation (for noobs like me). I had to buy the kit, case, trafo and all the small accessories so the total cost was way more than I have anticipated. Still worth it though.

My current setup is Denon DRA-295 stereo receiver, Yamaha DVD/CD player model DV-S5750 and Mission m30i speakers with Infinnity BU-1 subwoofer. Nothing fancy but it suits me well. When I have replaced the Denon receiver with gainclone I was running only the Mission speakers without the subwoofer. After the test run, I was pretty surprised at the bass response of the speakers without the subwoofer. I must say that I like the sound. The highs and lows are incredible (comparing to the Denon receiver). In my opinion, the only thing the gainclone lacks are the mids. Gainclone sounds more dynamic but the lack of midrange is noticeable. Still blows away Denon.

So that's what I wanted to share. Pictures are attached. You have to excuse the mess of wires as it was my first project. Space in enclosure is very tight because I think I went overboard with heatsinks. After 4-5 hours of playing the heatsinks are not even warm (did not try touching chips).

I just wanted to thank everyone who replied to my posts asking for help. Special thanks to AndrewT for his patience. Without you guys I wouldn't be able to finish the gainclone (remember, I am total noob when it comes to electronics).
 

Attachments

  • 1.jpg
    1.jpg
    59.9 KB · Views: 1,164

johnmax

Member
2008-01-22 8:38 pm
No photos from me, but I just finished my first pair of gainclones. One is an integrated amp driven by a 160VA 2x25V transformer, regulated down to 28V DC using LM338T regulators, using a passive pre-amp (rotary selector and alps blue 10k pot). The other is a dual mono power amp, with twin 120VA 2x25V transformers, regulated to 25V (for my 6R speakers). Both are non-inverting.

The dual mono uses Panasonic FC caps - 1000uF on each rail before the regulators, then Pana FC 1000uFs on each chip, with an ELNA Starget 4.7uF as the input dc blocking cap. All resistors are maplin metal film types.

The integrated is the same, but uses EVOX-Rifa PEG124 series 1000uF caps on the chips, which were harder to implement as they are axial, but I think sound better than the Pana FCs.

Both amps use the optional Ci from the NatSemi datasheet - one runs at 0.1mV of offset, the other has 0.0mV according to my (admittedly cheap) multimeter.

Anyway - basically wanted to thank everyone here as I have found the solutions to all my build issues in the forum here - its been an invaluable resource, and thanks to you all both amplifiers sound amazing - much better than the NAD C320 that I was using before (driving B&W 601 S2s and a set of Quad Lite satellites - which I highly recommend!)

So, thanks, and now I am buying in Black Gates and Allen Bradley resistors for a higher spec 3rd time lucky one :)
 
johnmax said:
No photos from me, but I just finished my first pair of gainclones. One is an integrated amp driven by a 160VA 2x25V transformer, regulated down to 28V DC using LM338T regulators, using a passive pre-amp (rotary selector and alps blue 10k pot). The other is a dual mono power amp, with twin 120VA 2x25V transformers, regulated to 25V (for my 6R speakers). Both are non-inverting.

The dual mono uses Panasonic FC caps - 1000uF on each rail before the regulators, then Pana FC 1000uFs on each chip, with an ELNA Starget 4.7uF as the input dc blocking cap. All resistors are maplin metal film types.

The integrated is the same, but uses EVOX-Rifa PEG124 series 1000uF caps on the chips, which were harder to implement as they are axial, but I think sound better than the Pana FCs.

Both amps use the optional Ci from the NatSemi datasheet - one runs at 0.1mV of offset, the other has 0.0mV according to my (admittedly cheap) multimeter.

Anyway - basically wanted to thank everyone here as I have found the solutions to all my build issues in the forum here - its been an invaluable resource, and thanks to you all both amplifiers sound amazing - much better than the NAD C320 that I was using before (driving B&W 601 S2s and a set of Quad Lite satellites - which I highly recommend!)

So, thanks, and now I am buying in Black Gates and Allen Bradley resistors for a higher spec 3rd time lucky one :)


Hi John! What cap did you use for Ci? I mean, what's the brand, value, and voltage rating? That Ci has so much effect on sound, I'd love to know of a good cap to use there.
Thank you!
 

johnmax

Member
2008-01-22 8:38 pm
The C1 caps on both amps are ELNA Starget 4.7uF 50V electrolytic caps (£0.33 each + the tax in the UK from RS min of 5 i think) - I use them in my headphone amps as well, really think they sound great for the money, my next amp is going to use the NX Black Gates though, I might try getting an extra set so I can compare with the Stargets in my existing amp.

My dual mono is up to 20 hours playing now and I don't know if its just that I am getting used to it, but it does seem to be getting more musical. I have added active cooling to the chips now as well - running CPU fans from an old pc under each heatsink - blowing a ready stream of cool air over them - they are rated at 5VDC, but got them running from 3V at the moment (2xAA - haven't worked out the best way to run them from the transformer yet!) and I can't hear a thing - they keep the chip temperature down to roughly room temp, so hopefully no risk of meeting SPIKE :)

Next project is some active pre-amps/buffers for them.... (the amps, rather than the fans!)
 

johnmax

Member
2008-01-22 8:38 pm
Doh! - Ci not C1 - need to find my glasses!

I have a Vishay BC 100uF 35VDC cap on there - I tried it without and got 70mV offset, soldered it in and bingo - 0.0/0.1mV offset and a large smile. I haven't even experimented on that one - just used what I had to hand, and took the recommendation that it needed to be a bit substantial from somewhere (maybe a search here? or NatSemi datasheet? can't think). That is one that I am going to experiment with on my next amp.

I'll try and audition some various options and report back once I have built it.

John
 
Hi John!

On the fans, most computer fans will run happily from a resistor. Zalman fans are made to.

On Ci, that makes a heck of a lot more difference than most of the other components. So, I'd be glad to hear about some options. I also bought several bags of likely caps. So far, 22uF, 33uF, of 100v models have been working for me. Sometimes its nice to bypass with 0.01uF to 0.006uF for a bit more "air" breathy effect.

I'm still testing. Thanks for the clues! I'll try some Elna "made for audio" versions and some Vishay BC's. Maybe we can share some more info on Ci caps later on?
 

craigg4c

Member
2008-02-11 6:32 pm
Ci musings ...

Because of some very silly stuff on the web, inverted configurations and Ci-less designs, the whole Ci issue has been blown out of proportion. 30 dB of gain at DC is a disaster waiting to happen, and these forums are full of beginners who've believed the hype and built capacitorless amps which they then discover have humongous DC offsets.

According to the National design spreadsheet, with most reasonable values of Rf & Ri you can get a -3dB point below 10 Hz with less than 20uF -- e.g.

20k/1k/20uF gives 8Hz
47k/3.3k/10uF gives 5Hz

These are values where experimenting with film capacitors of the actual nominal value might be worthwhile.

The reason oversized electrolytics (40uF and up) are typically used is that according to such respected engineers as Douglas Self (http://www.dself.dsl.pipex.com), electrolytics may introduce subtle distortions which can be avoided by using much larger caps than theoretically necessary.

There are real advantages to having a -3dB point around 10Hz -- for one thing, the amp won't waste power amplifying subsonic garbage (record warp, tuner muting thumps, etc.), and for another it's easier on your speakers when the cat jumps on the turntable or your baby nephew pulls out the input cable...
 
Re: Ci musings ...

craigg4c said:
There are real advantages to having a -3dB point around 10Hz -- for one thing, the amp won't waste power amplifying subsonic garbage (record warp, tuner muting thumps, etc.), and for another it's easier on your speakers when the cat jumps on the turntable or your baby nephew pulls out the input cable...
fiddling with the frequency response of the Power amp is not the correct place to cure the ills you have described.
Far better to get that Power Amp operating wideband and solve each of the individual problems at or near the source.
 

craigg4c

Member
2008-02-11 6:32 pm
Re: Ci musings ...

AndrewT said:
Far better to get that Power Amp operating wideband and solve each of the individual problems at or near the source.

Perhaps. OTOH, chip amps generally -- due to their usually compact size and the pride of construction -- lend themselves to being plugged in for demonstration purposes to e.g. the beer buddy's iPod and the girlfriend's Walkman; more generally, if I'm publishing recommendations or circuits for the unit, I have absolutely no control over how it might eventually be used or abused and what it may wind up trying to amplify. So it seems to me the better part of valor to design with a view towards minimizing risk.

In addition, I've never seen any credible engineering arguments in favor of designing an audio amplifier for normal use that is capable of full power at e.g. 10 Hz -- the only good reason I've ever seen is the limitation on DC blocking electrolytics referred to above. "Wideband" is a vague term; the question is "how wide is appropriate?"

Another relevant fact is that several widely-lauded small amplifiers -- e.g. the NAD 3020, the Van Alstine/Dynaco Stereo 70, and the amp section of the Advent 300 -- purposely rolled off the subsonic range for precisely the reasons I mention. So your point may be well-taken, but it is less than self-evident.