AudioAficionado.org

AudioAficionado.org (https://www.audioaficionado.org/index.php)
-   Mark Levinson (https://www.audioaficionado.org/forumdisplay.php?f=181)
-   -   The No32 vs No52 thread (https://www.audioaficionado.org/showthread.php?t=44740)

AudioIdiot 12-29-2018 03:56 PM

The No32 vs No52 thread
 
Dear Forum,
I would like to start my first thread here at AA that compares the Mark Levinson No32 and No52.

In the past, I designed high-end audio equipment for several brands like Sonic Frontiers, Classé audio, Transrotor, (Dutch Siltech owned) Sphinx and some others, but I have always been impressed by what ML did. When the No32 was introduced, I was extremely impressed with the design philosophy and the 32 was immediately a reference for me. I heard the No32 somewhere in 2004 where it sounded better than the design I made.

Now years later I am a new Aficionado. After a DIY speaker project that turned out quite well my stereo set is playing a lot. The sound has never been so good; streaming Tidal, via Roon with MQA files my audio life is good.
The upgrade virus got hold of me again, and the pre-amp is on my list of changes, so why not my all-time favorite a (second hand) No32?
So a few years ago, I noticed the follow-up unit of the 32, the No52 and I was not at all impressed. If I had judged (on paper) what would be the best unit I would have chosen the No32.

So what is there less to like (on paper, and as an engineer) on the structure of the 52?

• The chassis of the 52, although nice is not as impressive as used in the 32. The 32 is based on a single cast aluminum chassis, CNC machined to form this stable structure. The 52 is a more standard housing. I am not religious about a chassis made out of a single piece, but if you are trying to achieve the last bit of resolution all small things help.
• The first stage power-supply of the 52 is switching (nothing against switching if well designed, but often it’s outperformed by linear). The advantage is that it is auto ranging from AC 100V to 240V so no export models and convenience.
• The second power-supply stage of the 52 is a class-D design, nothing wrong either but building a noise free supply, the 32 had a better design.
• The 2nd power supply stage of the 52 runs on a 200Hz oscillator where the 32 uses a 400Hz oscillator. Both are in the audible frequency but 400Hz is used in many more (military) applications than 200Hz. 400Hz is twice as efficient as 200 and still low enough in frequency not to radiate. I would not have abandoned 400Hz.

Moving down to the audio part:

• The audio stage is virtually the same seen from the schematic but some differences are obvious.
• I have not seen the 52 audio chassis local supply stage, but the 32 has a great push/pull local regulated power-supply used to power the whole preamp stages.
• The relay structure looks identical with relays to remove the unused sources but in 2014, I would have ditched the tape output. Who in the world records analog sources nowadays? This would mean a lot less components and contacts in the signal path.
• The input stage of the52 looks identical with a 4 opamp instrumentation amplifier topology, but it uses small SMD parts instead through hole components. I am not sure what is best, SMD or through hole, but the latter allows for smaller designs and shorter tracks. A normal chip SMD resistor is not the best though I would use mini-MELF metal film resistors for all standard functions and high quality thick metal film, or tantalum-nitride in de volume control and gain stages. I am not sure what the 52 uses, but no Mini-MELF. The 32 uses high quality (through-hole) Dale resistors and film SMD resistors in the volume control. So why are resistors important? Well you use a lot in a pre-amp and they are all handling the precious signal.
• Volume control is done identical but in a smaller footprint in the 52. That should be better in the 52. Smaller is better here.
• Then the output stage is not even close to the great design of the 32. The discrete output stage with high class A bias and zero feedback was a revolutionary design. The 52 just has a simple op-amp based output stage. In my opinion an opamp is the device for an input stage and when you need gain. If you want to drive an output, a (long) cable and power-amp you need a discrete output stage with at least 20mA bias. I know no audio opamp with 20mA output stage bias. So the 52 in my opinion will only perform well with high input impedance power-amps.
• The 52 has the advantage of having two independent output stages (that are connected in parallel, so no individual gain). This in theory would be beneficial in a bi-amp situation like I have used in most cases for the last 25 years (Except for one two year when I owned Avalon Eidolon’s who couldn’t be bi-amped)
• But the 32 unbalanced output is just a mediocre opamp, not using the great class A discrete output stage. I think that ML with this decision thought that an unbalanced connection was not serious. If you own an 32 and use the unbalanced outputs because you have a nice (tube) unbalanced power amp try using the XLR connector with only pin 1 (gnd) and 2 (hot).
• On top of that, these 3 stages in the 52 are on boards connected via a fairly simple connector to the main board. Yes they are small but the signals still travel through at least 6 connectors extra.

I expect that the No52 is lower cost to build, lower cost of material and lower cost in labor. So what does this all say? Nothing if the 52 sounds better than the 32 ….. but does it?

Cheers,
Peter

gadawg 12-29-2018 04:16 PM

Peter,

Great write up and I certainly agree on many points but one thing that I'm a little confused about is that you mention the discrete Class A output of the 32? What output transistors are in the 32 and where are they? I notice what appears to be the same Op-Amp board in the 32 that was in the 326 is why I'm confused. Also, ML I believe does bias the Op-Amp section of the 52 to Class A operation. I don't see any heatsinks for the discrete transistors in the 32 output section. Not arguing for sure as I'm not an engineer but when I look at the inards of the 32 I can't find the discrete transistors like I can in the new 523 with its associated heatsinks. Let me know where to look as I'ld love to check it out. Also ... the 52 sounds amazing for sure as did the 32 just different. I would imagine that depending on other system components, synergy could determine which preamp any particular user might prefer. In my system the 52 is the clear choice but I can imagine someone else liking the 32 better in a different system. I agree that the 52 seems to be much higher priced for no real apparent reason.

All the above said I'm guessing we will see a new Ref product range in the next year or two and I'm guessing it might combine the 2 box preamp approach with the discrete approach similar to the new 523 but with fewer compromises made. I'm hoping the 53 goes away and they bring back a monster pure class A ref amp. That would be fun for sure!

Really enjoying some conversation about Levinson again! Always interested in learning more about our hobby!

George

Update:

another reason I thought the 32 was using the op-amp output section is the review in stereophile from the day ... "Fully balanced instrumentation amplifiers are used in the gain stages, and each input can be separately adjusted for 0, 6, 12, or 18dB of gain." This would mirror the 326 and the 52.

AudioIdiot 12-29-2018 05:38 PM

George,

The 32 (and partly) 52 amp topology is as follows:

A 4 OPAMP instrumentation amplifier is placed directly after the input selector.
It is used by both the XLR and RCA inputs, where the RCA ground is a bit floating (not hard grounded, to get an unbalanced to balanced conversion.
This input stage is where the gain is adjusted. (o, 6 12 and 18dB)

This stage feeds the floating volume control board. A 16 bit R2R ladder network with a al the SMD resistors and fet switches into an AD797 I to V conversion. (this partly SMD board in the 32 is reused in the 320 and 326s)

The output of the balanced volume control directly feeds two identical partly discrete output buffers.
It also feeds in parallel an opamp that feeds the RCA output.

The output stage of this buffer uses a matched pnp and npn transistors (LS3250A and LS3550A) in a metal TO71 can. (you can see them close to the output connectors on a picture where the input relay board is removed)
These transistors are the class A output transistors but to keep them cool they applied a (genius) cascode stage based on a current-feedback video opamp ADEL2020.
The beauty of this configuration is that the miller capacity (base to collector) is almost removed and the power dissipation is moved to the ADEL2020 (only 0,3W).
The DC offset of this stage is removed by an integrator dual opamp in the middle.
So these two buffers sit between the output XLR/RCA connectors and the floating volume board over the complete width of the mono board (left and right) so about 1/3 of the total board space.

It is a bit technical sorry, but this is how it works in the 32.

Peter

AudioIdiot 12-29-2018 07:25 PM

An extra note:
This discrete output stage was also used in the no38/380 series of pre-amps.
They used SSM dual transistors that became obsolete end of the 1990's.
But these units used a ladder network DAC as the volume control. So an IC, unlike the discrete unit in the 32/320/326s and in a new form in the 52.
I think it is also used in the 523/526 but not on a floating PCB, but direct on the main board.

And to make things clear, I favour the 32 based on a theoretical analasys, not on a listening test. I do hope I will be in the position to perform that test one day.
Peter

rnrmf1971 12-29-2018 11:26 PM

With you guys discussing the 32, I've been reading about it and it seems like a pre I'd love to hear, someday, to compare to my 52, as well.

gadawg 12-30-2018 12:14 AM

Peter,

Thanks for the clarification ... no doubt the No32 is a great performer and to be clear its been a while since I've listened to the 32 for sure. Depending on the rest of the system I can imagine some liking one and others liking the other based on the different sonic signatures of the two units. I'll be interested to hear about any direct comparison you are able to do.

George

Cohibaman 12-30-2018 12:17 AM

Quote:

Originally Posted by AudioIdiot (Post 946738)
George,



The 32 (and partly) 52 amp topology is as follows:



A 4 OPAMP instrumentation amplifier is placed directly after the input selector.

It is used by both the XLR and RCA inputs, where the RCA ground is a bit floating (not hard grounded, to get an unbalanced to balanced conversion.

This input stage is where the gain is adjusted. (o, 6 12 and 18dB)



This stage feeds the floating volume control board. A 16 bit R2R ladder network with a al the SMD resistors and fet switches into an AD797 I to V conversion. (this partly SMD board in the 32 is reused in the 320 and 326s)



The output of the balanced volume control directly feeds two identical partly discrete output buffers.

It also feeds in parallel an opamp that feeds the RCA output.



The output stage of this buffer uses a matched pnp and npn transistors (LS3250A and LS3550A) in a metal TO71 can. (you can see them close to the output connectors on a picture where the input relay board is removed)

These transistors are the class A output transistors but to keep them cool they applied a (genius) cascode stage based on a current-feedback video opamp ADEL2020.

The beauty of this configuration is that the miller capacity (base to collector) is almost removed and the power dissipation is moved to the ADEL2020 (only 0,3W).

The DC offset of this stage is removed by an integrator dual opamp in the middle.

So these two buffers sit between the output XLR/RCA connectors and the floating volume board over the complete width of the mono board (left and right) so about 1/3 of the total board space.



It is a bit technical sorry, but this is how it works in the 32.



Peter



Love the technical side. Thanks!

sefischer1 12-30-2018 12:38 AM

I disagree about dumping the tape loops. While it's true that most people don't use a tape recorder these days, there are still many applications for a processor of some type, like EQ or noise reduction, etc. This is why I will discount many new preamps if they lack one or more "Processor" loops. I currently use a DEQX and would like to integrate a Sweet Vinyl SC-1, with my 32 preamp.

tdelahanty 12-30-2018 06:19 AM

I spent 33 yrs. in military electronics and have often found my thoughts and preconceived notions in conflict with the real world of audio electronics. I'm well aware of the costs of designing and manufacturing such devices. Companies are constantly pressured to come up with better product for a lower price to reach a broader market and stay in business. These preamps should be judged by the end result, sound quality.

I would love to own both these preamps. A used 335 amp was traded in at my dealer's shop. I owned this amp years ago and still think it the best I've owned, it destroyed some of the latest offerings from Pass. ML's newest amps sound wonderful but I don't think they can afford to build another design as lavish as the 335.

AudioIdiot 12-30-2018 10:11 AM

Quote:

Originally Posted by sefischer1 (Post 946788)
I disagree about dumping the tape loops. While it's true that most people don't use a tape recorder these days, there are still many applications for a processor of some type, like EQ or noise reduction, etc. This is why I will discount many new preamps if they lack one or more "Processor" loops. I currently use a DEQX and would like to integrate a Sweet Vinyl SC-1, with my 32 preamp.

Hi Sefischer!,
Indeed I didn't think about that. However I would probably go pre-processor-amp, but you indeed lose the flexibility to by-pass it.

I do however need a surround sound bypass function with zero dB gain to integrate the stereo in the front channels of a surround set (like I do).
All the new ML units can do that (32, 52, 320, 326)


All times are GMT -4. The time now is 03:55 AM.

Powered by vBulletin® Version 3.8.10
Copyright ©2000 - 2024, vBulletin Solutions, Inc.
©Copyright 2009-2023 AudioAficionado.org.Privately owned, All Rights Reserved.