"Yuanjing" Gainclone 3886 - eBay amazing value ?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Another good looking ChipAmp board..

Hi all..

I'm a newbie and just wondering if anybody have information how to
contacts the owner of this web: www.hifizone.net

I'm really interested on their 3 good looking PCBs:
1. ALEPH3886 : http://www.hifizone.net/shop_view.asp?id=964
2. X-PP : http://www.hifizone.net/forum_view.asp?forum_id=14&view_id=1619
3. X-PT : http://www.hifizone.net/forum_view.asp?forum_id=14&view_id=1620

I mailed them many times but it seems they ignore me.. :( hiksz..

If anybody can't help me getting those stuff, I'd really appreciate.



Blessings,

- Cruzs -
 
Yuanjing LM3886 + NE5532 gainclone PCB

The kit may be dodgy, but the bare PCB is available and easy to assemble. I got one, and it's OK for the price. However, there are some issues with the design.

1) No star grounding. It uses two ground planes instead for the NE5532 and LM3886 sections, and the planes are connected at a single point. However, this is not good enough, since the main PSU filter capacitors are located and grounded on the NE5532 side, and the 7812/7912 regulators for the opamp supply are located and grounded on the LM3886 side. Significant ground-bounce issues are a certainty, and there's no easy fix for this.

2) No lifted ground for the input near the NE5532 - this is what contributes to the hum. However, this easily fixable with minor mods:

i) Remove the 10uF, 25V DC blocking cap for the feedback network for the input gain stage. Replace it with a 10 ohm, 1W metal-oxide resistor, mounted vertically as shown in the attached pictures. The upper lead of this resistor will be the lifted signal ground. The shield ground from the input connector is soldered to this lead (instead of being connected to the input terminal block), while the input goes to the terminal block as before.

ii) The input resistance from the non-inv terminal of the NE5532 gain stage is grounded at the lifted ground, not the PCB ground. This is done as shown in the attached picture - the resistor is (physically) diagonally bridged over another intervening resistor.

3) Lower frequency cutoff is too high. This is fixed by changing the values of the input resistance and feedback resistance to 47k instead of 22k. This changes the gain from 3.2 to 5.7, and improves the bass response for a given input capacitor.

4) The NE5532 opamp is not an optimal choice - it can easily be upgraded to a better opamp using 8-pin DIP sockets to permit rolling. I used an LT1208 lying around, but better choices are available - probably LM4562, LT1213 or OPA2134 will sound better.

5) Upgrade the DC blocking capacitors. I've used both Siemens MKL 2.2uF/630V and Pilkor MKP 1uF/275 VAC X2 with good results.

6) The bypass caps for the LM3886 are not close enough to the input pins. I soldered 100 nF MLCCs between +Vs and GND, -Vs and GND close to the pins on the bottom of the PCB.

With these mods, it sounds acceptably good for the price, but nowhere close to a decent audiophile-class gainclone like a Pavel Dudek or a Mauro Penasa Rev C.
 

Attachments

  • gc1_10x7.jpg
    gc1_10x7.jpg
    122.6 KB · Views: 3,802
Last edited:
One other mod (optional): solder a 47p to 470p Styroflex or Silver Mica cap between the non-inv and inv inputs of the LM3886 on the solder side. This helps suppress some RF interference, as well as helps the LM3886 recover better from some kinds of clipping, as per the NatSemi Application Note. I used 68 pF, but the use of this cap is entirely up to the taste of the builder. You can omit it if the LM3886 is never going to be driven to clipping.
 
Pre-assembled LM3886 success story

I have purchased from Ebay and used several China sourced LM3886 chip amps. They have all worked perfectly. I didn't check for star grounding or anything like that - it was just plug and play.

They look like this (below) and can be sourced from several vendors. Just search LM3886 in the Electronics>Home Audio>Amplifiers category. I would avoid the "all in one" boards that have on board PS.

LM3886.jpg


-Charlie
 
schematic is from Datasheet

That little unbuffered one is probably ok, within the constraints of its topology. Could you post its schematic, if you have one? I'd like to see if a lifted signal ground is implemented.

If you are referring to my post, the schematic is pretty much straight from the LM3886 datasheet. There is no muting switch, it is permanently connected to be "on" (no muting). There is a 1 ohm series resistor on the output, which I plan to replace with a link or the parallel resistor/inductor.

-Charlie
 
I bought one of these Yuanjing boards, and i'm quite pleased with the results. I think I will try linuxguru's mods though. I would like someone to explain the pre-amp. Is it a buffer and pre? Should the pot go where the jumpers are or are they just for bypassing the pre. I changed the op-amps for 4562's and it is much better. In fact it sounds very good, although bass is lacking. I considered getting a Pass B1 buffer and not using the op-amp, would this be an improvement? Any help would be appreciated. Cheers Barry
 
... I would like someone to explain the pre-amp. Is it a buffer and pre? Should the pot go where the jumpers are or are they just for bypassing the pre. I changed the op-amps for 4562's and it is much better. In fact it sounds very good, although bass is lacking. I considered getting a Pass B1 buffer and not using the op-amp, would this be an improvement? Any help would be appreciated. Cheers Barry

The pre-amp is a 2-stage amp, consisting of a gain of 3.2 stage, followed by the jumper block, followed by a unity-gain buffer. I used the jumper block for a dual-ganged pot, but you can just jumper it (actually, you need to jumper it or use a pot to get any sound at all). The jumper just connects the output of the gain stage to the input of the unity-gain buffer.

To improve the bass, increase the input and feedback resistances to 47k (from the 22k shown). That will change the gain to 5.7, which doesn't hurt. Also consider larger-valued input caps (say 3.3 uF or 4.7 uF) if you can fit them.

To remove the 50/60 Hz hum, use the lifted-ground mod that I showed earlier - it will get rid of about 95% of the hum, and the remaining hum is not noticeable.
 
Hi, I done the mods, but just got a very loud hum. I turned off immediatly to try to save my speakers. Checked everything, but can't see whats wrong. :(

Check all your joints again, look for solder bridges/shorts:

i)The 10-ohm resistor is connected to the ground plane on one end, and to the feedback resistor network on the other - which is the "lifted" ground.
ii)The input bias resistor (47k) is connected between the non-inv input of the opamp and the lifted ground.
iii) The shield ground from the input signal cable is connected to the lifted ground.

This is repeated for the other channel also. Everything else is the same as before on the op-amp circuit.

Check carefully for shorts or bridges if you've soldered bypass caps on the underside of the PCB near the +Vs, -Vs and GND pins of the LM3886.
 
Hi, I done the mods, but just got a very loud hum. I turned off immediatly to try to save my speakers. Checked everything, but can't see whats wrong. :(
Tried the same mods and got the same result (loud hum), plus a knackered speaker. Have now completely bypassed the 5532 by soldering inputs directly to the second pair of wima's on the board. Now no hum at all!
 
The mystery deepens - the assembled Yuanjing board appears to have subtle differences from the bare board from Jim's Audio on EBay. In particular, the location of the bypass caps for the NE5532 differs, and the electrolytics next to the NE5532 are unpopulated. It's possible that the layout of the resistors in the vicinity of the NE5532 differs between the two boards. Ensure that you have a non-inverting gain of 3.2 stage (22k + 10k), followed by the jumper block, followed by a non-inverting unity-gain buffer before proceeding with the mods.

To repeat: the gain stage has a 10uF, 25V DC-blocking capacitor to ground in the negative feedback network, which I remove and replace with the 10-ohm ground-loop breaker to create the lifted ground. This will increase the DC offset at the output of the gain-stage, but this has no effect beyond the unity gain buffer, which drives the DC-blocking Wima at the LM3886 input. The shield ground from the input connector on the gainclone case is connected to the lifted ground, as is the input bias resistor from the non-inverting input of the gain stage.

On the Jim's Audio bare board which is shown in the picture I posted, the mods I detailed work perfectly. Check the traces in the pre-amp section of the assembled Yuanjing to ensure that it's the same before proceeding. Before plugging in the source, do a touch test on each of the signal inputs of the signal cable and see if you get some mild (not loud or heavy) hum pickup.

Bypassing the op-amp completely is probably the best option, in any event. You may have to give up the volume-control pot option on the gainclone box, but that's no big loss.
 
Well iv'e managed to unmod it and all seems ok. I was too woried about my Linn speakers to mess around anymore. Perhaps there are different variations on the board. This is my second chipamp, and I still think there is something lacking. I have been looking on ebay at a site run by Indrasep in Indonesia. He has a 50w dual mono amp that looks to have a darlington pair output similar to my John Shearne amp. Does any one know about these. It would use the same power supply, but I don't know if it would need a bigger heatsink. I am using a bar of 300cm x 40cm x15cm at the moment. What do you think Linuxguru?
 
But Borninthefifties, I tried one side with and one side without the pre and thought the extra gain and clarity to be superior all be it with a little hum. By bypassing it you could get the gigaworks pair for £15 which most people say is superior to this board. That's what I used on my first gainclone and the results were very good. What about using the pass B1 buffer copy on ebay for about £25? I might try one and let you know. Barry
 
I noticed a Sanken discrete monoblock board on Ebay, from Indonesia (probably from Indrasep). I simulated the schematic shown in the Ebay listing, and the simulation numbers weren't great either for distortion or stability. In general, a darlington output will have worse sonics than a discrete darlington emitter follower.

The best bets at the entry-level are still gainclones, either LM3886 or TDA7294. Another alternative is to go with a hybrid gainclone + discrete, like the LM4702 + discrete output pair boards often seen on EBay.
 
Thanks for the advice, What about the naim nap140 clones. Iv'e heard with a bit of tweeking these are quite good. I'm just looking for the best amp for little outlay.

I haven't tried the Naim clone PCBs from EBay, but I've seen them on EBay and read about them here. There's a lot more room for modding and tuning the Naim clones, but the results can be either much better or much worse than a basic gainclone. Consider also that there seem to be several variants of the NAP-140 clone PCBs, some of them with errors in silkscreen outlines. If you have the patience to work around all these issues and try different components/mods, the Naim clones can potentially give much better results than gainclones.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.