• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

My Experience with "Music Angel" EL34 PP

Last week I took delivery of one of the chinese "Music Angel" EL34 Push Pull Amps.

I fired it up on the weekend and initial impressions were that it was Ok without being anything special. Stereo imaging was'nt particularly good and it was a bit "covered" indicating too much 2nd harmonic distortion.

So dig out the schematic to look at modification possibilities (see attached):
Hang on - IIRC the larger anode resistor for a cathode coupled phase splitter should be on the cathode coupled side, NOT the "straight through" side as shown in the schematic.

Off with its lid too have a look:
The 83K are in fact 4 x 332K in parallel, the 73K6 is in fact 75K consisting of 4 x 301K in parallel and the 55K "Tail" resistor is 75K (4 x 301K) as well.

Checking the first channel I find that the 83K is in fact on the cathode coupled side, that is the amp is correct and the schematic is wrong.

Checking the other channel and HEY the 83K and 75K are opposite to the first channel. A manufacturing fault.

I corrected this problem and Wow! a totally different amp. The stereo image is now quite acceptable and the "covered" sound has cleaned up significantly.

So last thing, a bit of output tube rolling
- tried Winged C Svetlana EL34s, Much better

- on a whim, I tried some "Fat Bottle" chinese EL34s I bought 5 years ago and had left sitting on the shelf as I had decided they were probably not muich good - WRONG these tubes were simply stunning in this amp, significantly better than the Winged C Svets.

So with 4 or 5 hours effort I have a lovely, cheap EL34 amp.

A few other comments:
- I run it in triode mode all the time
- The input tube is not a 12AU7 but a 6N1 instead. The extra gain affored by the 6N1 is compensated for by simply leaving out the 220uF/6V3 cathode bypass cap.
- full power output is 20W per channel
- frequency response at 10W out is 7Hz to 60kHz
- Absolutely no indication of any transformer saturation at 10Hz 10Watts out telling me that the output transformer is better than average quality
- Zout is 2.3 Ohms
- Gain (to 4Ohm output) as shown is 33.54
- Gain (to 4 Ohm output) with feedback disconnected is 42.85
That means there is ONLY about 2dB of global feedback

Cheers,
Ian
 

Attachments

  • 500_schem.jpg
    500_schem.jpg
    68.5 KB · Views: 7,154
Yes - this is the amp being offered out of HK by people like the eaudioguy.

I did some more tube rolling on mine - the 6N8P were replaced by EH Gold 6SN7GT and the 6N1 was replaced by a "Winged C" Svetlana 6N1P (both modern production tubes).

Both made significant improvements.
I've ordered some JJ EL34s to try in it. (The JJ EL84 are simply stunning, better than any NOS EL84 or 6BQ5 I have tried - I'm hoping the JJ EL34s will be in the same class).

Even with the shipping costs and a complete re-tube this amp is good value for money. Purchase of parts to build something the same would cost MORE.

The components are all good quality (Dale resistors, Wima Polypropylene coupling caps etc.)

Is it the Dynaco of the 21st century? - well, you judge for yourself. For what it has cost me, I'm extremely happy - and if I get sick of it then its a great candidate for a rebuild of the phase splitter/driver.

Cheers,
Ian
 

rdf

Member
2004-06-21 8:04 am
big smoke
gingertube said:
I've ordered some JJ EL34s to try in it. (The JJ EL84 are simply stunning, better than any NOS EL84 or 6BQ5 I have tried - I'm hoping the JJ EL34s will be in the same class).

In the few circuits I've compared them the JJ EL84's are substantially better than Sovteks and completely trounce the EH in measured performance. Compared to the latter less than a quarter the distortion wasn't unusual.
Did you find JJ's EL34 to be in the same class?
 
Another tweak to this amp - see schematic posted above.

Last weekend I deleted the 75K and 83K anode loads of the phase splitter. Replaced both with 50K (2 x 100K Dale 1W in parallel).
At the same time I replaced the 75K "tail resistor" with a cascode current source set for 3.7mA. The cascode was a red LED referenced BC547B/MJE340 with 270R as the current set resistor in the emitter of the BC547B. This improved the sound significantly.

As a last check I looked at the HV rail and the anode voltages on this stage. The extra current in this stage plus the extra 1.5mA I was putting through the current source LED reference had dragged the rail below the 360V at which the zener regulator was operating. Dropped the HV series resistor until the zeners were once again operating and rail was nominally 360V (3 x 120V zeners in series). This improved the sound significantly again - so maybe there is something to Allen W's shunt regulator theories!!!

Cheers,
Ian
 

SY

diyAudio Moderator Emeritus
2002-10-24 10:19 pm
Chicagoland
www.SYclotron.com
rdf said:


In the few circuits I've compared them the JJ EL84's are substantially better than Sovteks and completely trounce the EH in measured performance. Compared to the latter less than a quarter the distortion wasn't unusual.
Did you find JJ's EL34 to be in the same class?

The JJs are perhaps the best EL84s I've ever used. That they are relatively cheap and widely available is just a bonus. I've got a set of EL34s in my Curcio ST-70 and they perform flawlessly after a year of heavy use.
 
I agree that the JJs are the best EL84 going around - tried everything from NOS 6BQ5s, Sovteks, EHs etc. in my Morgan Jones Bervois Valley Amp and the JJs were a cut above the rest. It wasn't just better matched pairs either. On the BV amp I run ring of 2 transistor current sources in the cathodes of each EL84 (instead of 270R resistors) with 470uF/50V Blackgate bypasses. This guarantees idle current balance and the cathode voltages at idle give a good indication of how well the tubes are matched.

I pick up some Matched Quads of JJ EL34s to try in the Music Angel amp this weekend. Its back running Winged C Svetlanas at the moment. Will report next week on JJ vs Winged C Svets.

Cheers,
Ian
 
Jaap,
I've NOT used the Allen Wright Scheme (Yet). He turns the output stage into a diff amp by connecting the Output tube cathodes together and then having a current source to ground. This scheme enforces Class A operation, if you try to push it into Class B it will simply clip like a SS Amp.

What I did was instead of using separate cathode bias resistors (of 270R) was that I fitted separate "Ring of Two " transistor current sources set for 38mA. I don't have a schematic to post BUT draw this on a bit of paper from the description and you'll get the drift.
- Tube cathode to collector MJE340 and 6K8 resistor
-other end 6K8 to MJE340 base and BC547B Collector
-MJE340 emitter to BC547B base and 16R resistor
-other end of 16R 0V
- BC547B emitter to 0V
The 16R sets the current
Still need a bypass cap to allow the stage to run in class AB1.

For the Diff Stage pure Class A Output Scheme I would expect that a better current source (higher Zout) would be required. Also the stage would probably want lower anode volts and higher idle current to suit Class A operation. There will be about 1 watt dissipated in the current source so heatsinking may be an issue. Any of the Morgan Jones cascode current source circuits should work OK or use the Allens MOSFET current source shown on schematics at VSE Webpages. Oh and I would put a small (say 10R) resistor in each cathode to provide a bit of balance by degeneration rather than just connect the cathodes together directly. then attach the current source to the junction of the 2 x 10R.

Hope this makes sense and is helpfull.
Cheers,
Ian
 
I checked the "bervois valley" Amp on the weekend - no eveidence of plate blackening on the EL84s.

I also picked up 3 matched quads of JJ EL34 for some experiments. I tried them in the "Music Angel" Saturday, swapped back to Winged C Svetlana Sunday and finally back to the JJs Sunday afternoon - I listened to a bit of everything from Elke Brooks to Scissor Sisters. My view is that there is NOT a lot to pick bewteen the JJs and the Winged C Svetlanas. The JJs seem to have a little more detail BUT they also had a bit of an agressive/hard edge, the Svetlanas were just warm, smooth and gorgeous. I've left the JJs in the amp for now to see if they "break in" a bit after some use (the Svets I am using for comparison have about 100 hours use).

First impressions are that the Winged C Svetlana are marginally better than the JJs.

Cheers,
Ian
 
I checked the "bervois valley" Amp on the weekend - no eveidence of plate blackening on the EL84s.

I also picked up 3 matched quads of JJ EL34 for some experiments. I tried them in the "Music Angel" Saturday, swapped back to Winged C Svetlana Sunday and finally back to the JJs Sunday afternoon - I listened to a bit of everything from Elke Brooks to Scissor Sisters. My view is that there is NOT a lot to pick bewteen the JJs and the Winged C Svetlanas. The JJs seem to have a little more detail BUT they also had a bit of an agressive/hard edge, the Svetlanas were just warm, smooth and gorgeous. I've left the JJs in the amp for now to see if they "break in" a bit after some use (the Svets I am using for comparison have about 100 hours use).

First impressions are that the Winged C Svetlana are marginally better than the JJs.

I run my JJ >300V at plate. And After 3 weeks it showed signs of 'blacking'.:smash:

Sonicwise, quite neutral in P-P, but can be a bit edgy sometimes.
 

amperex

Member
2005-02-20 1:42 pm
MI
gingertube

In six EL84 based amps & trying about every EL84 tube ever manufactured (I am a tube dealer), the GE 6BQ5 grey oval plate outperformed the other tubes. This includes tubes such as the early Mullard 'D' getter, EI Elite, RFT, JJ, a few versions of RCA, Sylvania, etc. The tests were at various bias points as well trying to tweak each individual tube. Try the GE, you will be amazed.
 

rdf

Member
2004-06-21 8:04 am
big smoke
Amperex, have you tried the 6n14n-EB? Sixteen were waiting in the mail yesterday and frankly I'm stunned by what I hear them doing at 250 volt and 40 ma into a 5k SE James transformer. And UL at that. If I remember to grab it test gear will come home with me tonight to compare perfomance against new JJs and Sovteks.