ECP Torpedo III [indexed in first post]

With ZDS as a preamp to the T3, I'm just guessing it'll give the DSHA2 a run for its soundstage depth and height money based on what I'm hearing right now.



Pardon the dust.

 

A power transformer is not really going to do much in the way of sound quality. To replace the existing one, you'd have to get one made with the same pins which is probably near to impossible. Certainly not worth the cost. And the Lundahl PTs are huge.

The inductance of the OPTs is high enough that you should be able to run the 80 Ohm impedance off the low winding and not really have any bass rolloff until 20 or 25 Hz. It is worth a try. As discussed, you can also try parallel resistors (and I would not worry abot getting the total resistance all the way down to 32 - shoot for 50 perhaps). Alternatively, you can use a series resistor and run it off the high tap. Again, it is nominally a 300 Ohm tap, but lower is fine here. It might change the FR, but you can EQ that out :)

Yes, either method or resistors will soak up power.

Thanks for the extremely kind words. I am not quite sure what to say. DSHA2 is for sale on a made to order basis for $2200.

 

Nice, what's the nominal bandwidth of these Cinemags in the absence of any mismatch? Rolloff starting at 25 Hz isn't worth doing anything about to me, ahhhh but what the hell, I'll still try it and test with my ears. If it's really 25 Hz, I aint gonna hear a thing but what else do I have to do. Heh.

 

Some of you are not going to like this post - especially Morten - but I feel obligated to give you the information and my opinion.

After discussing the matter of the Amtrans resistors with Doug, he told me that the smaller ones actually have a sufficient power rating (3/4W) for the T3 circuit. Since they were obviously smaller, I discussed this with Brencho and we agreed to give them another try. I ordered a batch of 10 from an ebay seller in Vietnam, in anticipation of wanting to install them in my own personal T3 after trying them out in Brencho's.

Here's a pic that shows you the difference in size:


I carefully bent the leads this time and installed each one with about 1/16" clearance above the PCB. To be honest, these were small enough that the installation was a night and day difference over the big ones. Even with a liberally curved bend at the end of the cans, they are small enough that the width almost fits the position exactly. There were no pops or crackles this time.

Pictured below is Brencho's T3:


Hopefully, you can see that the fit is superior and I would have no qualms about installing these in other T3's - except, the sound is not as good.

The comparison between Brencho's amp and my personal T3 (only difference now is the Amtrans resistors) is rather obvious:

• Bass is fuzzied up and not as resolute,
• The metal shimmer at the trailing end of a cymbal hit is gone, and
• The slight trailing hiss in a singer's "S" (without sibilance) is missing.

In general, the resistors put a "soft fog" over the micro-detail and dynamics that the Output CCS's bring.

I tried not to have any pre-conceived notions about this and gave it an honest try - a more than honest try. However, one pre-conception I'll admit to having is that every documentation I can find says metal films are superior to carbon films (not just carbon comps) in low noise and temperature stability. Here are just a few:

http://components.about.com/od/Components/a/Types-Of-Resistors.htm
http://www.resistorguide.com/carbon-film-resistor/
http://www.analog.com/library/analogDialogue/archives/31-1/Ask_Engineer.html

I'm sorry, but I cannot recommend these resistors in combination with the Output CCS's. Maybe they're great without the CCS's, but I don't know. Anyway, that's my opinion.

 

Yeah, I remember being impressed by the DSHA2 prototype you had at CanLanta, but when I got this one in my home system in a quiet environment it blew me away. Part of me is tempted to sell the TIII and just go ahead and buy the DSHA2, but I'm still loving the TIII so much that I'll probably just wait until early next year to buy the DSHA2. I feel like if I can get just the right tube in the TIII it could be really close to the performance of the DSHA2, but with its own character. They're both stunningly good values. If the DSHA2 had some ugly glow colored balls on it and a laser engraved logo etched into the black aluminum case, you could charge $10,000 for it. Insane value on both the TIII and DSHA2.

As a note, as much as I've raved about the DSHA2 in comparison to the TIII, they sound 98% identical. It's still the same overall character. It's just a handful of extremely subtle ways that the DSHA2 pulls ahead. I think in the audiophile world we tend to exaggerate that last little bit of performance. To put it in clearer terms, I think the TIII maxed out is a bigger improvement over the Mjolnir than the DSHA2 is over the TIII. I honestly think at this point the biggest thing holding the TIII back is finding just the right tube. Like I said earlier, with the DSHA2, it's almost like Doug was able to bring out all the benefits we see with tubes, but with the dynamicism and extension we associate with a world class solid state system.

Had a friend over last night who subscribes to the NWAVguy theory of "if it's got enough mW and the output impedance is low enough, all amps are the same" theory. Let him listen to my setup. Both the TIII and DSHA2 changed his mind.

 

That's interesting, @TomB . I certainly don't have those observations with the resistors after proper run in. And maybe it's also hard to believe, but these resistors do have a ~10-12 hour break in period. Given your first bullet point, I don't think you've run signal through these for very long? I had almost removed mine the first hour after installation for the same reason (at least that first bullet point for sure), but instead I let the music keep rolling over night and things had settled down in the A.M in to the sound signature I've previously described.

BTW, @brencho just heard my amp Thursday night (with AMRG and output CCS boards), and I don't think he heard any of those bullet points you list either. But he's certainly free to comment, of course.

PS - We've at the very least removed any doubt that resistors can make a difference, now, right? Hehe.

Edit: By the way, you'll find quite a lot on resistors within DIYForums, their usage in specific applications, etc. While it is true that carbon films will be higher in noise, that should be relatively small compared to the tubes themselves in this application. Heat dissipation, i suspect, is what AMRG is trying to manage with the metal tube and epoxy. And, I don't know, maybe they also affect noise in some second order way. But I do hope those 3/4W versions suffice because I do know that my 2W AMRGs got hot (not warm) to the touch when I first installed them. Haven't checked recently, but now with output CCS boards, the entire amp is really toasty.

 

Thanks for the honest feedback tom. I was considering the amtrans but will probably hold off now. Like others have theorized it may be that the Output CCS have made the amtrans unnecessary. Or maybe it's just some peculiarity of how the Output CCS and the amtrans interact that isn't as good as either of them in their own.

I think as long as everybody gives their honest perceptions and is respectful of the opinions of others, any and all feedback is appreciated. One of the things that has made the TIII become the world class amp it is today is the fact that we have the world's best amp designer (Doug), somebody very experienced with all the iterations of the build (Tom) and then a community that's able to try out and test more stuff than any one person or company could.

I think we also probably have the TIII to the point now where all the obvious objectively better improvements have been made and most changes now are just flavorings (or extreme reworkings). But maybe there's some next clear cut huge improvement waiting out there.

I think after I (begrudgingly) give the DSHA2 back on Tuesday I'm going to really immerse myself in some tube rolling now that I have a clear sense of what I like and don't like about the three tubes Tom sent with my TIII (that being said, I'll be 100% satisfied if the stock 12AY7 is my endgame tube with the TIII as well. It's the basic idea of what I want). Also the DSHA2 is nice as a reference of the ECP house sound because it gives me a clear sense of what I want out of my idealized tube.

 

@bazelio is right, we didn't hear a lack of resolution or softness in his T3 comparing it side by side to drfindley's T3 without amrg (and also without ccs). @TomB it might be worth checking to see if a half day of burn in does something, as baz points out. But I'm not a builder or designer so I don't claim to know what that would do in practice!

 

I'll give it a chance and run it till tomorrow, just to be sure. However, I've already done a couple of more hours listening to it. I even set up and powered my own T3 right next to it so I only have to change the plugs real quickly to compare. That comparison is still obvious.

I don't want to belabor this, but it sounds sort of like the T3 without the Output CCS, as if the Amtrans cancelled out the tweak. I can see how some would think it helps tonality, because it sort of rounds off all the edges of things. If you think that a tube amp should be fat and smooth, then this gets further there. The Amtrans may address what Jun was complaining about a while back, too, but that's not what I want in an amp.

Further, if you've been reading fjrabon's impressions about some of that close-miking subtlety that makes you sit up and take notice all of a sudden, those things are gone. When Frank says he's been having to fiddle with the volume more than usual, that's a great characteristic of the T3. It's quite similar to Tyll's description of the Elear, as in, "Turn it up too loud and a piano strike will punch out one of your eyeballs." It's not that extreme, of course, but that subtlety that makes you wonder if you've got the volume set right seems gone with the Amtrans.

As I noted before, bass is less distinct and certain peaks are definitely muted by these resistors: the shimmer trailing off from a cymbal hit, the almost crystalline pluck of a metal guitar string - those things are gone or subdued to the background. I'm not saying the pluck is completely gone, just that when a pluck rises to the level of a glass crunch - that part is gone - and I don't think it was distortion before, just the opposite.

There are times with the HD800 and the T3 that I have to look around my house to see if someone dropped a footlocker or if a truck went by and shook the house. Then I realize it was a bass hit that was not heard, but felt. I can get that from my turntable using the T3 and HD800, too. I've never had that happen with any other combination of amp and headphone. That is also gone from the Amtrans version: you hear it now as a muddier, higher bass note, but you don't "feel it."

It would actually be easier for me to just leave it and send it back to Brencho. He wanted the Amtrans, they're in there now, and the amp "operates" just fine. To most people, it probably still sounds like a world-class amp, just not as outstanding as before, IMHO. Anyway, this will be the 5th time I've disassembled Brencho's amp to do work on it, if I put back in the metal films. I'm not exactly looking forward to it, but it's not as good as it could be - or as it was.

Sorry - we'll just have to agree to disagree with the Amtrans. I guess maybe I did belabor it, some.


P.S. BTW, if some of you are interested, I've been listening a lot lately to what I think is an excellent recording for listening tests: Gretchen Peters' Blackbirds. It's got it all - lots of echo, presence, punchy electric guitars on "Blackbirds," a driving, very deep bass and drum line on "The House on Auburn Street," great dueling acoustic guitar string plucks on "Pretty Things," and even eyeball-punching piano on "Jubilee." Just my opinion, of course, but it's simply a sterling, very versatile album with excellent production quality.

 

Alright, well, let us know after more burn in, @TomB . There is no question in my mind that there is a big tonality shift after some period of time. I suggest stepping away and letting it burn for a while and then coming back. Either way, I'm certainly intrigued, and if 3/4W is sufficient in those spots then it makes life easier and opens up other possibilities like some of the nicer radial options. I'm almost ready to socket those resistors and really put things to the test because clearly this is a place where the overall sound signature can be affected.

 

I am staying agnostic on the resistor debate But as for the power rating: the regulated voltage is 225V or so. The tail of the CCS pulls ~2.2mA which is split between both sides of the tube. That means each 100K plate resistor has ~1.1mA through it and drops V = IR = 100000 * 0.0011 = 110V Power is W = VI = 110 * .0011 = 0.121W. With signal you get instantaneous peaks (and valleys) and transfer some of the heat from the tube to the resistor, but even then, if you have all 2.2mA going through one side, it is still less than 0.5W peaks (220V * 0.0022mA = 0.484W, and of course 0W valleys). We used the 2W resistors for the kit as they are cheap and provide a huge safety margin and voltage handling. But they are a bit of overkill.

 

I don't have the measurements Cinemag did during design, but suffice it to say that driven from the low impedance as they are, they are flat from well below hearing range to well above it. That was actually the case even when we did a worst case scenario and drove them from a 10K source.

 

Thanks, Doug. Sounds like we're talkin about roughly 0.342W RMS (< 0.5W peak). This is great to know.

It sounds like @TomB will send back @brencho 's amp with the AMRG resistors removed. We plan to get together in the next week or two and compare amps. I suppose it's possible that the lower wattage (i.e. smaller) resistor (3/4W vs 2W) has greater contact noise and a different sound signature than mine as a result of an increase in distortion. Just a WAG, actually. But the AMRG resistors came about prior to CCS output boards in an effort to improve tonal balance, to compensate for some leanness. And it seems reasonable that noise / "attractive" distortion is at play. But the CCS boards seem to be achieving similar results (by different means), so it'll be interesting to hear the two amps again and see if my AMRGs need to come back out. I have no emotional attachment to resistors, just music.

 

Some reading for resistors choice and tube amps for anyone interested. I do not know if this guy is "legit" in a sense that matters but I have read these posts many times these past few months.

http://www.aikenamps.com/index.php/technical-q-a
http://www.aikenamps.com/index.php/resistor-types-does-it-matter

Right now I am resistor-type agnositc as there are places where the choice matters more and places where choice matters less. There are those that believe mixing types is the best practice, those that believe carbon comp is the old school and only way to go, and those that believe metal foils everywhere is the best way.

Note, we do not know in which context he's giving this 2W 750V recommendation for plate load resistors and I think it's an odd rating to have to shop for.

 

Returning to the PT.. I know next to nothing about design reqs, methodology, theory, etc... it's a bit embarrassing since my dad's an EE (who owns an Engineering manuf. firm). He's a freakin' wiz and I'm certain he could (and would) whip up a custom PT. @dsavitsk.. if we wanted to (meaningfully) improve the PT in the TIII, do you have a design schematic (or benchmark #s.. or whatever else is necessary) I could provide my dad so he could manufacture a few for us to test, etc.? Sourcing any particular part for the PT, however difficult, shouldn't be a problem. If we're successful, I'll look to have as many as possible made for the TIII owners in this thread (available for free.. or as close to free, as possible).

 

Would transistor rolling make a difference? What order would matter most for rolling? Can we can get a list or table of transistors that work in the Torpedo III?

Default:

1. Q2, Q3, Q4, Q5 - 2SC4793, BJT NPN 230V 1A, TO-220 http://www.mouser.com/ds/2/408/toshiba_2SC4793-587292.pdf
2. Q6, Q7, Q8, Q9 - KSP42BU, BJT NPN Si Epitaxial, TO-92 http://www.mouser.com/ds/2/149/KSP42-889409.pdf

Here's an example of the resources available for the discrete diamond buffer project regarding transistor rolling. My CKKIII rolling/mods were based on these observations on the pioneers before me:

 

I'm not sure what you mean by "OOP," but only genuine Toshibas have been supplied with the T3 - kit or assembled - for Q2, Q3, Q4, Q5.

The problem with Steinchen's table of output transistors is that it was used for the revMH Millett Hybrid Diamond Buffer. Like your CKIII, the voltages in the revMH MIllett Hybrid run from 24VDC to maybe 30VDC in the extreme case. The voltage rating of many of those transistors is about 1/4 of what is needed in the T3. The Toshibas supplied with the T3 are rated for 230V, exceeding anything in that table.

That said, rolling different output transistors might be a worthwhile exercise, albeit a pretty big hassle. The table is still a good example of how much different transistors can vary in sound quality - when they are not wrapped in a feedback loop.

 

Corrected. I understand these are for way different voltages, it was meant for illustrative purposes for different transistors that actually are compatible with the Torpedo3. Transistor swapping was not that easy, but still doable. I believe it remains a viable option that should be explored as was suggested before me.

On another note, i rolled these 12az7 in today,

 

The power transformer is not a secret. http://edcorusa.com/Content/Manuals/XPWR083A_120-240.pdf

The TO92 parts probably matter more. You'll have to search the usual sources to find some that will work and investigate the datasheets to get a sense of whether they are appropriate.

 

Cool. The 12AZ7 continues to be my favorite tube in the T3.