[bartas]

Hi everyone,
I think every CORNET builder would be interested in different opinions about upgrade points in cornet 2. I`am not even talking of differences between specific parts manufactures (there are many posts about caps or resistors or etc.) but more about "What CORNET likes" The start point could be basic signal path components and lets see where it takes us:

I would firstly concider parts marked in red. Or `am I wrong?

What do you think? 

[tubesforever]

First, I replaced the pcb rca chassis pieces with Vampire direct gold plated oxygen free copper pieces.  These were 20 dollar a pair. 

Second, I replaced the red colored resistors with Vishay S102 nudes.  These cost me 11.95 each from Percy Audio.  The Texas Component pieces are identical from what I understand.  You can save a few dollars there.  I would use the same resistors for the 47k loading as well. 

Third, if I were building all over again, I would probably try the PRP 1 watt resistors in every other resistor point other than the H+ stepdown.  For that I use a Mills MRA 12 volt resistor.   I understand that sonically these are very close to the sonics of the nudes and are 65 cents each from Partsconnexion.

I found that using a teflon cap at C203 and a Mundorf Silver in Oil  in parallel with a second Teflon cap at C208 gave me outstanding silence and great dynamics.  I am trying some Jupiters right now at C208 and these have incredible vocals and crisper highs.  Caps are sonic filters, so select the ones that help you system match to your cables, amps, speakers and room.

Hope this helps!

[ronpod]

bartas,
It seems you are asking about the function of each component in the circuit. Then from a functional understanding, can each component be specified according to a reasonable assignment of importance of that function. Is this is what you mean?

You have highlighted the direct signal path in red. Several other members of the DIY community have asked about what components make up the signal path, so you have done the community a service. I would agree that these should be considered the most important components because all of your music signal must pass through each of these components. This is why Tubes suggests not sparing any expense to get the least noisy resistors and best sounding coupling caps in this signal path. Certainly, the tubes are very important also because they provide the gain for amplifying the signal and are directly in the signal path.

The first shunt resistance (R200) is 47k ohms, and is important because it creates input signal matching. In other words, what ever you connect to the input of the Cornet will see 47k ohms so you would do well to make sure that there is a match. The resistor is only slightly less important than the signal path resistors, so the quality of the Cornet could improve if you carefully select a good resistor for this position.

Each tube in the Cornet represents a gain stage (except the 12AU7 which is a buffer stage). The tube acts like a variable resistor in a resistor ladder that works in a way that creates the gain. So, although the resistors above and below the tubes may seem to be even slightly less important than the others mentioned above, the signal must be clean (free of spurious noise). The caps in the tube power supply line may help provide quick discharge impulse if needed for good transient response.

The shunts to ground all along the signal path chain create the RIAA shaping of signal necessary for the use of a needle cartridge. So these also have a reasonable amount of importance. Jim Hagerman's design is rather spare, mean there are no components in this circuit that are not needed. (That's why Jim is successful.)

A reasonably good match of values between the two channels helps ensure more equal gain.

Hope this helps,
ronpod

[hagtech]

parts marked in red

Classic left-brain thinking.  However, a signal path is not like driving roads.  Forget everything you thought in coming up with the red parts.  It will lead you in the wrong direction.  In reality, every single component in that drawing is part of the signal path.  Sorry.  That's probably not what you wanted to hear.  It is difficult to become selective as to which components are more important than others.  Indeed, some are.  But it is all grey scale.  There are no right or wrong choices.  Only preferences.  Fortunately, the circuit is simple enough to respond well to experimentation.  Almost every component change will affect the sonics.  I would say R220 is the least important.  Then maybe R215.  Most important either C208 or V200.

jh

[bartas]

I totally agree with Jim. It is a CIRCUIT, and it is a really good one Every part is important in it`s place. But I still think that some of this "importance" can be graded due to the impact on the sonics.
I started a new chart as Ronpod and Jim stated some places of importance:

This is getting really interesting! I think we could get a overall picture of how CORNET does what it does. I can redraw the chart at some discusion points. For me "newbee" it is very important to get to know how different sections of amplifier do the sound magic.

[hagtech]

Nah, you're still headed on the wrong path.  Much of the signal path is affected the shunt elements, far moreso than most of those series components. 

jh

[bartas]

Help me out here Jim.
I`ve reloaded the drawing with some of the shunts marked. Could you be more specific about "red/orange" signal path parts and shunts. I`ve marked basic RIAA components, as i understand from you posts these are not that easy definable, but anyway. Please show us the wright way.

[MusicMtnMonkey]

Should we be asking which component is best for which position instead of which position most affects sound quality?  For example, should we be discussing which capacitor is best for a power supply as opposed to coupling use??

[mingles]

I don't claim to know much about circuit design, but I recently learned the following about resistors:

- Resistor noise is proportional to current flow. The higher the current, the greater the noise.

- Noise contribution is greatest at low-level stages.

- Wirewound resistors are the quietest followed by metal film, metal oxide, carbon film, and lastly, carbon composition.

In lieu of these considerations, it would make sense to look at R202 and R209 b/c they feed the 12AX7s with power. Mills MRA-5 might be a good option here.

I recommend reading Randall Aiken's tech note:

Resistor Types - Does It Matter?
http://www.aikenamps.com/ResistorNoise.htm

[PatOMalley]

mingles,

It looks like the best way to beat noise in carbon film or carbon composition is to get the highest wattage you can find? We are told that adding 10ohm resistors at R223 along with the 1.8 will deliver the right amount of power [if needed]. I bought two 5 watt Kiwami's for this purpose and matched the 5 watt only to keep it in spec. Looks like I might have stolen a march on noise as well. They will dwarf that cool looking Mills.

What about 2 watt Tantalum's?

The only drag is that 1 watt PRPs do not fit nicely in the positions with one end up. The 1/2 watts fit like they were made for it.

[analog97]

So, if somebody made a bigger Cornet2 pcb and replaced all the resistors with Mills or other exotic/expensive low-noise type, how much of a sonic difference would there be? 

[tubesforever]

You can bend over the leads of the bigger resistors and they will still fit on the board.  Just use some pliars to hold the leads so you don't damage the resistor body as you make the necessary bends. 

I am using the 5 watt Mills wire wounds for the B+ chores on the Clarinet.  These are 5 dollars each.  I doubt I would ever use them again.  They do not sound night and day better than a 2 watt Dale resistor. 

I plan to try some PRP resistors. These suposedly sound nearly as nice as the Vishay nudes and cost 0.65 cents each.  That is a good deal for a high quality metal film resistor. 

BTW Jim is using the Dale 2 watt Vishays on his Trumpet board.  Stereophile gave that piece a Class A sound rating!   I consider this proof enough that passive parts make a difference.

When it comes to distortion we actually  have to deal with two types of distortion.  One is electrical distortion.  The second is from resonance.  Only the Vishay nudes work to eliminante both types of distortion to allow the clearest signal.   How good are these.  Ask any sound guy on a submarine.  They can hear the digestive tracks of Whales hundreds of miles away.  That is impressive!

I like using the Vishay nude S102 resistors in the direct signal pathway simply because the voltages are low enough to stay within their limits. 

I cannot use them everywhere simply because their 0.60 volt ratings do not allow them to work well in the B+ and grid positions.  You need a solid 1 watt rating for these resistors.  If you stress a Vishay nude it creates distortion.  I have tried them as B+ and Grid stoppers and they sound broken in that environment.

You can select other resistors that timbre match the nudes.  Dales would be a good choice and I hear the PRP resisitors are also very similar sounding. 

Lately I have been experimenting with the Kiwami 2 watt resistors.  They are found in the ultra expensive top line Japanese electronics.  I don't find them a perfect timbre match to the Vishay nudes. 

I plan to try some PRP resistors soon.  I am hoping they will better match the Vishays.

Like Jim says, every darn part on that board is going to effect sound.  Every single piece is critical and important.  Use great parts and you can play side by side with 5,6,7 and higher dollar phono stages and leave them lacking.  I would not trade my Cornet 2 for anything I have heard on the market place.  The Cornet 2 rocks!

Jim has done a great job with his design.  It is up to us to match the passive parts to system match the sound to our amplifiers and speakers.  This is a personal thing and as such -- subjectivity rules.

Cheers!


 

[SoundBound]

Okay, it is a 'minimal' design and every component does affect performance.  Agreed.  We know on which capacitors we should pull out all the stops and spend a few more dollars on an upgrade to make the best sonic investment (ie interstage cap and output cap).  So, if we applied the same logic, and given that all resistors are important, which resistors would be the best 'investment' of pulling out the stops and spending a bit or a lot more to get the best return on dollars spent (while trying to not sacrifice to much quality on the rest)?  Obviously because you can buy resistors that range in price from less than $1 to $7-8 each.  Using the most expensive ones for all resistors does not make economical sense unless budget is unlimited.  Any ideas?

[ronpod]

Nah, you're still headed on the wrong path.  Much of the signal path is affected the shunt elements, far moreso than most of those series components. 

Jim,
You got me thinking. I'm so glad that this hobby allows these wonderful opportunities to learn (and I have plenty of that to do). I had totally forgotten that tubes are voltage devices. I would guess that very little current is passed through the tube grid so the series resistors R201, R208, and R214 (R219 as well) pass very little current. They are all 220 ohms so their voltage drops as well as their currents are likely to be small. The resistances passing the most current are mostly in line with the tubes (cathode and plate) and some of the other shunt paths. This is a really fun circuit to experiment with! It is a good thing that there are very few components in this circuit so those of us that like to "upgrade" don't have to consider the credit institutions. 

At one point I envisioned that the entire circuit block feeding a particular tube grid was a voltage conditioner. (And that the tube was a voltage amplifier.) But I may be getting too left brain again (no coulombs or synaptic stimuli crossing the Corpus Coulosseum). Jim, thanks for the fun of exploring...

ronpod

[mingles]

The other thing to consider with resistors (along with current flowing through them) is their value. Large values like 150k have more thermal noise than smaller values like 220 ohms. This is why I believe R202 and R209 are ripe candidates to be replaced with wirewound Mills MRA-5. R216 might also be a candidate.

Jim Hagerman, any thoughts on this?

[Brinkman]

The other thing to consider with resistors (along with current flowing through them) is their value. Large values like 150k have more thermal noise than smaller values like 220 ohms. This is why I believe R202 and R209 are ripe candidates to be replaced with wirewound Mills MRA-5. R216 might also be a candidate.

It would be nice to know if the inductance of the wirewounds would be something to be concerned about. I did a cursory internet search but could not find any measurement parameters for "non-inductive" (re: Mills). From what I can tell, a small amount of inductance is not likely to affect the audio band, so we're probably fine there.

I've used the PRP resistors in a Bellari VP-129 upgrade I performed (matched pair loading resistors, grid stoppers, output and other locations) and noticed an definite improvement. In full disclosure, I also upgraded a pair of miniature (ceramic? mica?) coupling caps with matched Sonicaps, and I gotta say the tolerances on all the stock components I upgraded was surprisingly sloppy. We're talking .063uF where the schematic specs .22uF Don't know if that alone can account for the increase in detail and Left-Right balance, but that was the improvement.

I would suggest carbon comp (lowest inductance) for the grid stoppers (R201x, R206x, R214x), as the current through them is so little that thermal and contact noise is not really an issue. However, I've read claims that Kiwames (KOA?) carbon comps do not exhibit frequency drift over time but I'm not quite convinced of this.

Now, I understand a manufacturer could push a resistor to it's rated wattage limits (and beyond) to simulate long-term use and then measure it's value, but without simulating the long-term humidity and heat an average component's subjected to, I'm not sure how definitive these claims can be. Perhaps the manufacturer pre-cooks them and assumes that's sufficient. Maybe it would be. Maybe drift wouldn't affect the Cornet2 in those positions anyhow.

Also, the five 47uF electrolytic caps (C103, C201x, C207x) could be upgraded with polypropylenes. They're a bit oversized, but if one was DIY'ing a chassis anyway it might be a worthwhile consideration. I've used the Jantzen crosscaps from Parts Connection.

Some good advice I've read here that I'd second is in regards to matching components (resistors, caps). I think buying a surplus of common resistor values, even if they're precision PRP, and then going through and measuring each one for tolerance and then pairing them into their respective channels is likely to be the the most cost-effective tweak out there. $15.00 buys a couple Vishay nudes, but it also buys a ton of PRPs.

Just thinking out loud.

-Brinkman

[mingles]

I have a quick question that should be easy for someone to answer. What do you call the 47 uF electrolytics C201 and C207? Is there a name for the role they play?

Brinkman, do you think a poly cap like the Jantzen Crosscap offers better performance than an electrolytic in this position? They aren't cheap at $13.71 each:

http://www.partsexpress.com/pe/pshowdetl.cfm?&PartNumber=027-948&DID=7

Jim's parts list calls for a $1.87 Nichicon. FWIW, I used Panasonic's (P13563-ND).

thx in advance!

[MusicMtnMonkey]

So when I purchased my TX2575 resistors, they were offered at .1% tolerance or better.  Is the .1% tolerance equivalent what you were suggesting: to measure and then pair the resistors based on tolerance?  Would this in theory be an advantage to ordering from Texax Components if they do very precise measuring and matching for you??

A really good question relating to this thread may be does tolerance matter more or less for the RIAA, Shunt, or Signal path resistors??  What positions would be best served by precisely matched resistors and are there other positions in the circuit where matching resistors doesn't make much difference.

advice I've read here that I'd second is in regards to matching components (resistors, caps). I think buying a surplus of common resistor values, even if they're precision PRP, and then going through and measuring each one for tolerance and then pairing them into their respective channels is likely to be the the most cost-effective tweak out there. $15.00 buys a couple Vishay nudes, but it also buys a ton of PRPs.

Just thinking out loud.

-Brinkman

[Brinkman]

Brinkman, do you think a poly cap like the Jantzen Crosscap offers better performance than an electrolytic in this position? They aren't cheap at $13.71 each.

Hmm. They went up a couple bucks in the last year.
If you're on a budget, it may be better to just upgrade the 1uF caps right next to those 47uF electrolytics. Even cheaper, just bypass said 1uF caps with .1uF or .01uF boutique caps. There's a thread here about this topic, but in regards to the Clarinet.

I would personally use the Jantzens, but perhaps I've been hanging around the anti-electrolytic crowd too much. When an electrolytic can be avoided, I try to avoid it.

[hagtech]

wirewound

Wirewounds can be good, but I think the bulk foils are the best.  That's probably what the S102 are.  The metal films are good, bulk films better (PRP), and then wirewounds and bulk foils on top.  The wirewounds come in various flavors, different metals for different ohmic ranges.  Usually they are hard to find about 10k or 25k.  Some can be junk.  The metallurgy and ceramic insulation / heat dissipator can make a difference.

Is the .1% tolerance

Nothing in this circuit requires a tight tolerance.  Remember, the equivalent plate resistance from a tube can vary 20% easily.  What good is a 0.1% resistor in series with that?  The important thing is the sonics, not the tolerance.  It is the construction method and material types that make the difference.  The tight tolerance is an extra benefit that comes for free.

Is there a name for the role they play?

Yes, they are power supply decoupling and filtering.

inductance of the wirewounds

Nothing to worry about in this circuit.  The inductance is minimal (like 1uH) and they are usually made with counter-rotating spirals (or the Caddock hairpin turn patent) such that magnetic fields are cancelled and the self-inductance minimized.

jh