Amp/Preamp Asylum

Ming Da MC-2A3 preamp schematic

68.183.226.193


[ Follow Ups ] [ Post Followup ] Thread: [ Display  Email  Next ] [ Amp/Preamp Asylum ]

This Post Has Been Edited by the Author




Just want to share with fellow audiophiles a schematic of the Ming Da MC-2A3 preamp and my experience as a curious owner.

I have always been a believer of preamps being responsible for setting the tonal quality and power amps just delivering power with transparency and stability.

I have been a fan of 6SN7-based tubed preamps such as the MFA Luminescence, Melody 1688 Signature, and the Canary CA906. Violin music just sounds magical with those 3. preamps. I also briefly owned a Cary SLP98P which I was not too impressed with due to the blurry bass and somewhat AM radio-like mid-highs.

The description of the preamp @ the manufacturer's website sucks.
There has not been a review of this unit on the several major audio-rags or sites like 6moons or enjoythemusic. The unit looks somewhat of a monstrosity.

The Ming Da MC-2A3 struck me as an interesting design (tube rectification and regulation, dual stage chokes, and of course… 6SN7) and I had the opportunity to pick up one at a bargain price. The preamp arrived with the cage detached from shipping and created a ding on one of the front support “pillars”. No big deal, just cosmetics.

Stock tubes included Shuguang 2A3 rectifiers, and the rest of the tubes are EH.

Initial impressions

Out of the box and as expected, there is a sweetness to the violins that makes this preamp very desirable. However, careful listening reveals the same weaknesses as the SLP98P I owned previously. Flabby and lethargic bass, a bit veiled mid-highs, no top end extension. The somewhat narrow soundstage destabilizes when playing complex passages. Still the musicality is there that makes it an enjoyable preamp.

Off comes the bottom cover

OK let’s take a look at what’s inside the unit. Wait… where are the Sprague vitamin Q gumdrops and the Mundorf M-caps in the advertisements? All I see is run of the mill SCR-MKP caps. Oh well no problem, I was gonna swap them out with my collection of Mundorf Silver Oil, Dynamicaps, Jensen Copper-oil, or Hovland Musicaps anyway.

The B+ power supply sure looks impressive. L-C-L-C-[series-pass-regulator]-C-R-C+bypass. Wow… but wait… how’s the filament power supply hooked up?
There seems to be 2?! Nice bridge-C-R-C to the 6SN7s but wait… it supplies it with an unregulated 13V to be share serially among 2 6SN7s! That means my left channel 6SN7 has filament voltage of 6.5 on a good day at 0-6.5V and the other tube has it raised to 6.5-13V! That’s not a good design  The 6922 side is even worse, 20 volts shared between 3 6922s, that means one channel would have its filament at absolute potential starting as high as 13.5V.

Looking at the circuitry, the 6SN7s are in SRPP config, no surprises there. The middle 6922 is the cathode follower, sounds logical. The 2 6922s on the side are a bit idiosyncratic. ½ of each “side” 6922 serve as the constant current source to each half of the the middle 6922. the other ½ of each of the “side” 6922s is the top tube of the bootstrapped cathode follower for the middle 6922. Seems convoluted enough but sure convinced me that a lot of thought has been put into the circuit (well… read on).

Shotgun Approach

I decided to start with a shotgun approach by taking out the SCR MKP output caps ( was 2.2uF paralleled with some PIO-like 0.01uF) and replace them with Mundorf 2.2uF Silver Oils. The 0.22uF caps between the cathode follower and the top 6922 was also replaced with Jensen 0.22uF Copper Oil. The sound opened up quite a bit and every sounds more well defined. However, it came at a price of a pronounced mid-high and the bass is equally as flabby. I believe the change has more to do with the output cap than the cap between the cathode follower and the upper follower tube’s grid.

Schematics

Thanks to a fellow audiophile who graciously shared with me a copy of the schematics.
I looked at the 2 High Voltage bypass caps closest to the 6SN7s and right at the regulator outputs and according to the schematic they are 100uF 400V rated. So in went to Blackgate 100uF caps in place of the blue 150uF caps marked Philips and BC (they looked identical and based on the markings on the cap I have reasonable suspicion that they are some Kung Pao Mushu counterfeit caps). As expected the bass got deeper and fuller but it still sounds slow and muddy. There is a slight incoherence in the mid-high that I associated with the 0.1uF bypass cap across strapped from B+ to ground. I came across several writeups that Blackgate lytics do not like to be bypassed so I took it out. Marginal improvement I called it.

Colored sound and lethargic bass

“So why didn’t you wait a few weeks to let everything run in?” You might ask.
While I believe in running in I refuse to believe in the 400+ hours of break in time. If something does not sound decisively better than the old component a after few hours of run-in. It’s likely your mind playing games on you cos your ears are getting used to being pierced. Anyway, the preamp still sounds lethargic and presents a somewhat colored midrange.

Why so finicky about tubes?

The same audiophile who provided me with the schematics was very kind and generous to loan me some of his prized collection of tubes. So the MC-2A3 is now equipped with RCA 2A3s, an uber expensive Philips Miniwatt EL34 as regulator series pass (yes!), my trusty RCA 6SN7GT smoke as comparator, and mega-expensive pinched waist 6922s as followers and constant current source (yes!).

For the 6922 in the input/gain stage I tube-rolled among aged Sovtek (don’t laugh, they ARE competent), Ken Rad, RCA-smoke, Sylvania stubby chrome top, Raytheons, RCA redbase, and CBS 6SN7GT and CBS 5692.
Each set of 6SN7s gave it a different sound as expected, and I won’t go into the details. But what I found was that with certain 6SN7s the preamp sounds WEIRD. Also with some known good, known quiet 6SN7s the preamp started to hum as if grounding is a problem. The midrange gets additional harmonic coloration which I know from using the tubes in other applications is definitely NOT a characteristic of that tube. So onto the bench it goes and we checked for gremlins.

Oscillation!

We looked at the output of the 6SN7 to the 6922 middle tube and found there is SERIOUS oscillation when RCA smoke-glass 6SN7 gain stage mates with my (couldn’t resist so paid the big bucks) pinch-waist 6922 follower. A quick look at the schematics reveal that Ming Da uses a 220Ohm grid stop resistor between the 2 stages, that’s about 1/10th of what we’d expect in a circuit like this. Out goes the stock 220Ohm and in comes a 2kOhm grid stop. No observable high-end roll off there but the strange noise is gone and the voltage measures normal. Now I can at least seriously evaluate the tube combos without this tube-capacitance-matching problem.

Is that the output caps?

So back to the pronounced mid-highs. After running the Mundorfs for a few weeks the preamp still shows this tonal imbalance so I swapped in 2 2uF Dynamicaps. Initially it sounded more buttoned down but the air and decay went away. Not my cup of tea.

Faulty Jensen, a blessing in disguise

At around the same time the preamp started to emit a “puff puff putt putt” sound and it was traced to a faulty Jensen cap in the boot-strap-follower upper tube. Out goes that cap and the preamp actually sounds cleaner. Swapped in an Auricap 0.33uF, and then an Obligato 0.1uF with no success. Everytime that cap (actually it’s 2, one per 6922 on each side) is in place the coloration comes back.

Looking at page 168 of “Valve Amplifiers, Third Edition by Morgan Jones” the upper tube of that bootstrapped cathode follower is for distortion cancellation but even Morgan Jones commented that it “invites further problems” and “requires a separate heater supply” which the MC-2A3 certainly does not have.

So I decided to leave that cap OUT.

Experiments with NFB

It’s common assumption that preamps with NFB gives a punchier sound. So I proceeded to investigate whether a small amount of NFB, implemented using a cap and a resistor looping either at the output cap or before the output cap might serve to tighten up bass.

I tried resistor values ranging from 25 to 80kOhms at 10k steps and various capacitor sizes without success. Too high a resistor value (low NFB) gives no effect, and once the value is small enough (50k) to tighten the bass it makes the preamp sound nervous and constricted. No go with NFB.

Scrutinize the schematic

Running out of ideas, I decided to spend some time looking at the schematic. The unit that I got as well as the units depicted in THIS picture
http://www.pacificvalve.us/images/MC_2A3_Inside.JPG
do not show the 5.1kOhm resistor per the schematic on the B+ rail separating the 6SN7 stage and the 6922 follower stage. The way my MC-2A3 was wired the resulting B+ to the 6SN7 be the same as the 6922s top triode.
According to “Beginner's Guide to Tube Audio Design by Bruce Rozenblit”, this kind of an arrangement could even result in coupling between the 2 anodes.
Perhaps I should put in a 5.1k resistor as the schematic suggests.
The schematic also agrees with European units shown here:
http://www.audiocostruzioni.com/r_s/pre/ming-da-mc-2a3/dentro.jpg
http://www.audiovideofoto.ro/index.php?option=com_remository&Itemid=63&func=fileinfo&id=3292

My unit has the green 300Ohm resistor just as illustrated here:
http://www.pacificvalve.us/images/MC_2A3_Inside.JPG
but the unit depicted here
http://www.audiocostruzioni.com/r_s/pre/ming-da-mc-2a3/dentro.jpg
does not have that resistor.
The 300Ohm resistor in my unit (and possibly all units coming into the US) separates the regulated B+ from the rail going to the tubes. This drops the 220V regulated output to about 202V or so. No problems here. BUT if you take a closer look, this means the 100uF filtering capacitor (now a blackgate) at the output of the regulator has to go through the 300Ohm before getting to the plates of 2 6922 and 2 6SN7. With this arrangement, the only bypass cap directly serving those tubes would be one single 100uF lytic ( Blackgate WKZ 100uF 500V).

The photo of the European version looks different. There is no 300Ohm resistor separating the output of the regulator and the 6922 anodes. A 5.1kOhm resistor separates the 6922 anode rail from the 6SN7 anode rail. This means (and it agrees with the schematic) that the 6922 sees 220V B+ and the 6SN7 sees 190V. This prevents anode coupling between the gain stage and the cathode follower. The 6SN7s also get full access to 100uF of bypass and the 6922s get their share of a full 100uF of bypass. This all makes a lot more sense. It’s time to follow the schematic and start making corrections.

With this change to make the circuit a big step closer to the schematic, the midrange coloration (an occasional overtone noticeable in complex passages) is completely gone.

Film caps

I’m sure everyone has come across postings bashing the lytic caps. Sure I’d like to use film caps for bypass but how do I fit a total of 200uF or more into that chassis? Also since Blackgates are known for their deep bass will I be going the wrong direction by attempting to replace the 2 blackgates with film caps? There is definitely enough bass from the MC-2A3 now but the problem is with the quality of bass I am getting. I am looking for tight bass with slam, and a clear after-decay, not the “farty” bass I am getting.

Fix the obvious and snub them Zeners!

While trying to figure out how to squeeze 4 33uF film caps into the chassis, I notice that there are 2 (well technically 3) places where zener diodes are used. Zeners are known for their diode noises but they are a necessary evil for voltage reference to the constant current source and the main B+ voltage regulator. Zener noises are usually snubbed by placing a cap across the zener. The zener used in the 6922 constant current source for each channel has 2 stage snubbing. But 70V 220uF lytics are used. I duct-taped it with a cascade of film caps so as to ensure no high frequency zener noise gets into the CCS. Also I noticed the zener snubber in my unit is wired across 1 diode only and not both as depicted in the schematic. I ADDED another cap just to make very sure there is no zener noise getting into the EL34 series pass. The result was not dramatic but noticeable. The noise floor was lowered and I can hear more details.

Film caps for bypass

After I figured out how to physically squeeze in 3 Solen 33uF caps, I used them in parallel for post-regulation B+ on the 6922 rail before hitting getting to the 6SN7s which are still bypassed by a Blackgate. The net capacitance there is not lowered so I did not expect any drop in bass response.

Even without any warm up I notice that there is an improvement that is not subtle. The lethargic sound is gone. Dynamic contrast is outright scary with absolutely no sacrifice in smoothness or coherence. This is a much bigger improvement than swapping in Mundorfs and Blackgates in place of the stock caps.

1 More RC filter stage may not improve sound

Time for another fancy idea. Why not put back the 300Ohm resistor and put back the Blackgate 100uF lytic to ADD another C-R filtering stage before the 6922 rail? Due to space constraints, I’d end up with only 66uF servicing the 6922 (2 Solens instead of 3).

The result? A slight but noticeable incoherence and roughess in the mid-highs. That’s exactly the opposite of what I’d expect from an additional C-R filter.

I want to keep 100uF bypassing the 6922s anyway so I ended up taking out the 300 Ohm resistor which is not in the schematic anyway.

Ripple-phobia

I was also surprised at how the preamp reacted to different EL34s as regulators. A good regulator should dissipate incoming power and provide smooth constant voltage to the B+. The EL34 is not “playing music” so why is the preamp sensitive to that tube. Also why did a Groove Tube E34LS high current EL34 which is supposed to have great heat dissipation sound bad as a regulator in my preamp?

I was suspicious with the quality of the choke-filtered high voltage going into the regulator and I came across this posting

http://user.tninet.se/~vhw129w/mt_audio_design/tweakpage_dac.htm

So I decided to squeeze a small film cap ( 0.47uF Auricap) right between the EL34’s plate and ground. In theory this cap should serve as a fast reacting capacitor filtering out any high frequency noise or small ripple going into the regulator.

Drum Roll….

Voila! Another MAJOR improvement. Now I get copious amounts of detail, air, decay, limitless dynamic headroom and soundstage stability. Never had I thought that DC supply pre-regulation is THAT important. (Are those “BC lytic caps” fake). The preamp us now uncolored and coherent.

I guess what I really should have done is replace the 2 blue 300uF 400V lytic caps with Nichicon. Duct tape approach works… at least for now.

This was indeed an interesting journey in audio. The fact that the MC-2A3 uses point-to-point wiring gave me much confidence in trying certain things. I absolutely need to give credit to the authors of these 2 books
Beginner's Guide to Tube Audio Design by Bruce Rozenblit
Valve Amplifiers, Third Edition by Morgan Jones





Edits: 06/13/08   06/13/08   06/13/08   06/13/08   08/10/08   08/17/08   08/17/08   08/17/08   08/17/08   08/17/08

Follow Ups: