TomB

Nebby - just curious - are you talking about Doug's newest scheme with the transistor bias or the WE connection with the higher CCS current?

Yep - or 6J6-A, 6J6-W, or 6J6-WA. There's also 5964 and 5844, but so far with the WE connection and higher bias, I prefer the 6J6 versions. Looking forward to trying the new setup Doug just posted this weekend ... Nice work on the amp, Nate! Marc, that pic of yours looks great, too!

Actually, Doug may be able to alleviate some of that increased distortion by changing to a higher voltage PT. It's possible that those AKG's are still pushing it more than it likes. As for the bass hump, did you opt for the lower value on the coupling caps? I'm using 3.8uf (instead of 4.7uf) and have heard that artificially produces a bass hump with a parafeed.

Well, I haven't fiddled with the LED changes yet, but Doug says doing this does the same thing: 1. Lower the CCS resistor to ~56R. Those are the R4L and R4R resistors, as distinguished from the solitary "R4" resistor that's with the LM317 in the heater power supply section. Anyway, R4L and R4R are 75R. To get them to ~56R, just solder a 221R resistor in parallel with each one. That value was very convenient for me, because we commonly use it in some of Colin's diamond buffer designs. 2. After that, switch the cathode bias to the Western Electric connection (WE*) by switching all the jumpers - all 4 of them to the opposite pin. * WE Connection - the jumpers connect the primary to the cathode with the bias done via resistor

Looks really awesome, Nate! It's great that you'll be testing for the heat around the PT wit those slots, too. I agree that it would be nice to get rid of the heat sink. I tried several dozen smallish holes in the same location as the heat sink, but no bottom or side venting - it didn't work and I put the heat sink back on. We were trying to keep from putting any holes in the bottom of the case and I got fed up with trying to drill through the sides with all the slots. (If you don't hit a slot dead-on, the hole is miserably screwed up.) If you haven't done it yet, I'd definitely go with Doug's LED mod or use the higher CCS current with the resistor bias - I think Doug is calling that the WE connection? It's a different amp when you do that. In any event, the plans are definitely to include a custom-machined case with the eventual kits. That should alleviate the issues for those of us who are tool-challenged.

Personally, I'd rather see some of you try this new bias option before you start voting. About the 6J6 tubes - they are as plentiful as just about any tube, but dead cheap, to boot. They're just not marketed very much: high heater current and apparently some idiosyncrasies that Doug is quickly becoming expert in solving. Most vendors I've dealt with talk about having thousands of these things. There's also quite a list of tube equivalents, although admiitedly, most are variations on the same tube rather than different tube types. I listened more to it today just using KSC75's at work and will not back off of my previous post. I've got a bit of a mid-bass hump, but that's because I used the smaller 3.9uf Clarity Cap I think. Actually, with the mid-bass hump, it sounds very much like these esoteric CD players that everyone seems to crave at Head-Fi meets (Meridian?). I am very happy with this amp again and it's much better at powering other mainstream phones (other than Grado's) than before. Maybe a combination of Options - two amps?

OK - I'm totally clueless on that. Meanwhile ... I tried the new CCS awhile back with a 221R resistor in parallel. First thing I noticed was really punchy bass. Impressive, but the honeymoon was soon over - highs rolled off and the bass was obviously mid-bass exagerration. I preferred the original configuration, to tell the truth. So ... I was ready for a change. Reading back to Doug's recent post about changing the bias point, he stated that if you had installed the resistor for the new CCS current suggestion, all you needed to do to try the new bias point was to switch the jumpers. Sounded easy enough - so, I did that last night. Lo and behold, more detail and better highs than with the original configuration, IMHO. With Concrete Blonde's Mexican Moon (a favorite test of mine for highs), the tamborine bells were strong and detailed. That's the thing that this change has seemingly accomplished - it's brought back the smooth and clear highs that I heard with the original CCS, but now there's an obvious increase in detail and transient response that seems even better than before. To pursue this further, I took the amp and several tubes to work today with the premise that I'd report back after extended listening determined that this wasn't any fluke. I can't tell you with specific measurements - the M-Audio Transit I have has had no working drivers since before Vista. Unfortunately, my plan didn't turn out very well - I never even had the chance to flip the power switch. So, I listened some more at home tonight. My setup right now is a simple GrubDAC for the source running FLAC files. I'm running two Sylvania 6J6A's with halo top getters. They're not matched - I just grabbed a couple with identical construction and plugged them in. I'm using Foobar 2000 and run several music selections from different genre that I typically use to test things with my ears (terrible test device, but it's all I have at the moment). These include passages that highlight transients, midrange, bass response, and highs. For instance to me, the cymbals on Porcupine Tree's Deadwing, In Absentia, and Stupid Dream are among the best in modern music. So, I listen for whether those cymbals are clear - is there a clear swish without breakup, is there a metallic component, and can you hear the drumsticks tap as they make contact with the metal? The answers are all yes. Similar for Diana Krall - can you hear the midrance in her voice overwhelming all the other sounds? Does Rihanna's Rude Boy (Rated R) come through with impressive bass slam? There are other selections I can mention, but I'm probably beyond Head-Fi lingo by now. ;-) I've tried the Senheiser HD600 with Cardas cables, an HD580 with 650 cable, the HD 25 II, Sony V6's, AKG K701's and the Grado SR225, HF-1, and HF-2 - even the KSC-75. With every headphone save the Grado's, the upper switch position sounded the most forward and most dynamic. Even KSC-75's sound best with the switch up. That tells me that anything higher than 32 ohms should use the high switch setting. There's plenty of bass punch, too - except for the K701's. Admittedly, the amp may not be a powerhouse. With the K701's, the bass sort of futzes out and never manifests itself. Detail, midrange and highs are all there, but the bass just loses it before it starts. Not so with lighter loads - the HD 25's sound detailed and bass-punchy; the KSC-75's are ready to jump off of your head. The real payoff is with Grado's, however. I've said it before so will say it again (perhaps with a bit more authority now that I've been to CanJam) - there's no other amp I've heard - except for Doug's higher level offerings - that sounds as good with Grado's than this amp, especially at this new setting. Bass is strong and punchy, midrange is strong, detailed, and lush, while the highs are as smooth as silk with no breakup and no harshness. That's something I've never experienced with any other amp using the Grado SR225's, for instance. I've never had a headphone as hot and harsh as those things. They're unlistenable (IMHO) with a lot of equipment, and usually end up giving me a huge headache with any extended listening. With the Torpedo and this new bias setting, the detail and highs are exquisite. Bass is sufficiently punchy, while the midrange is lush - simply wonderful with Grado's. So, please - some of you with RMAA and other tools - give it another shot with this new bias setting that Doug recommends and let us know what you think. I know some of you are going to try K1000's, Audeze's and perhaps some other esoteric Orthos. Like I said, it may not come through like a powerhouse with some of those phones, but right now, it sounds fantastic with many other mainstream phones.

Hmm ... Doug's really got something with this tube bias thing. More later ... I don't want to jump to conclusions, but - well, more later ...

Well yeah, see - anybody that measures them with that much precision, I figured I'd better increase the odds to get you a matched pair. The non-6J6 tubes are pot-luck, though, but maybe you'll have a good time trying out the differences, anyway. I really like the 5964's - at least with the original CCS current. They're not as punchy as the 6J6's, but the highs are very nice. To be honest, I've not had much time on the 5844's, so I don't know how they do.

I think I would agree with that, but most of my listening is with Grados further down on the dial. Maybe we can talk Doug into making a switched CCS setting for the production version - or perhaps ganged with the impedance switch?

Just an FYI - but after more listening, I've decided I like the original CCS current better and will change it back. Chris - your tubes should be on the way. Marc - can you PM with your address and I'll send you some more tubes, too?

"Flat out" assumes a lot - that gain is within the controllable testing parameters, for instance. With tubes, you could have gains well into the 20's, 30's, and higher that may have no realistic operating points at the maximum volume position.

AFAIK, this was a practice started by AMB. I'm not sure there's any basis in a standard other than that. AFAIK, Tangent doesn't test his amps that way and I don't think Pete Millett does, either. I wasn't going to go into detail, but Doug hit on a key point. When I tested Millett Hybrids/MAXes/MiniMAXes, I essentially got the same thing when attempting to set the volume to 100% - too much distortion for RMAA to function. As Doug said, tube amps (and hybrids) are not necessarily designed to run at 100% volume. There are certain tubes and bias settings in a Millett Hybrid that will result in the amp's output being "racked" into opposite ends of the signal wave, if that makes any sense (something to do with the voltage swing on the tubes over-taking the voltage swing available for the buffer, I think). In any event, I eventually realized that the amp wasn't able to operate in that regime and it was pointless in attempting to test it at that.

Just glancing at the results, it looks like dynamic range and noise floor improves, whereas distortion does not. I think one of those tubes definitely looks bad in the final set of results/plots, though. That may be skewing the overall numbers. As far as the channel imbalance - even when well-matched, I think a 1 dB channel variance is pretty good for tubes. BTW, I got mixed results with the 5964's at the new CCS setting - they may sound a bit worse than before? Hopefully, I'll get a chance to do more listening in the next few days. I would try RMAA, too, but I only have an M-Audio Transit and they never got the drivers fixed beyond Windows XP. When the Transit worked, I had instances where RMAA refused to give results because distortion was too high on other tube amps/hybrids. JMHO, but I don't think RMAA was created with tubes in mind.

Thank goodness. I didn't think I'd send somebody a tube missing the getter flash. It must've used it up fast! This goes back to what I was referring to with Pars. If the tube sat on the shelf for 40-50 yrs, gas sometimes infiltrates into the tube. I guess if the getter flash is thin enough, it gets burned away leaving clear glass. This is different than when you break one - the getter flash turns white in those cases, but doesn't go away. OK - anybody else need more tubes while I'm at it?

Yep - 5964 is the one that's "over-driven" but sounds quite nice. I haven't tried it again since upping the current in the CCS, though, but I will give it a go tonight. The other one that's in relatively plentiful supply and is still cheap is the 5844. Many of the others listed on TDSL are just slight variations on the 6J6, but those two have noticeably different plate construction. Here's the basic list that shows up as "close or identical performance" on TDSL: 1216, 5844, 5964, 6030, 6045, 6099, 6101, 6535, 6927, 6CC31, 6J6A, 6MNN3*, CK6101, CV5046, CV8160, CV8231, ECC91, M8081, T2M05

My apologies on the tubes. I'll send you a couple more. What happens is that some tubes may break-in with the few minutes/hours of use in an amp and the output could be quite different than when tested. That doesn't happen very often, but it does happen. Some say it's bits of gas that infiltrate over the 30-40 years of sitting on a shelf. A few hours of running in an amp allows the getter to burn out the gas, changing the characteristic of the tube. Then again, the tester may have just blown it. I can confirm what Doug says - the amp has always been optimized toward Grados and low impedance cans, but running the CCS at the different current made a huge difference with mid-to-high impedance phones. Mine has 3.8uf coupling caps and I've read where that should produce a bass hump of sorts. To tell the truth, I never noticed until making this change with the CCS (I soldered 221R resistors onto the 75R of R4L and R4R). Now, my little HD25-1's sound like they have a subwoofer. HD600/580's sound remarkably better now and the impedance switch has a noticeable effect - much more "forwardness" with the Senns than at the lower switch setting, if that makes any sense.

Not sure what's going on here, but I agree with luvdunhill. Mine worked fine with the bezels, but would not without - at all (or with the backplates on the inside surface of the bezel). For instance, without the bezels, the pot sticks out far enough that you'd have to drill for the outside boss on the RK27. If you do that, there's nothing for the mounting nut to tighten down on except the pot itself (the endplate becomes superfluous). The only downside with the clearance that the bezels give you is that the locating pin on the RK27 sticks out enough that it rubs the inside diameter on my Partspipe knob, making it feel as if it's binding. As for the headphone jack and the impedance switch, there are plenty of spacers to make things fit with the Neutrik and once you drill a hole big enough for the switch lever boss, the clearance is elemental. So, the only issue is with the pot. The extra clearance that the bezel gives is enough to allow you to clamp the endplate between the pot's boss and the shaft nut. On the backplate - with the bezel, I actually had to add #4 washers as spacers behind the backplate for the IEC mounting screws. Other than that, it fit perfectly. (IOW, the RCA jack assembly should not be flush with the IEC - more like 1/32" behind the mounting surface of the IEC. All the mounting screws are 4-40's, but you have to tap the plastic mounting on the RCA jack assembly - no nut on the back of that 4-40.

Yep. Just an FYI, but Hammond won't let you order the chokes directly from them - even if you're a dealer. Their transformers and chokes are governed through strict distribution arrangements - been there, tried that.

It looks like PartsConnexion has them in stock. I entered up to 100 and never got an "out-of-stock" notice. Whether their website is sophisticated enough to account for their inventory, I don't know - but it seems a good bet they have several. Newark says they have 31. Mouser's lead time is only two weeks - it might be faster than that at Allied and DigiKey ~10 days.

Yep. It was in the text-parts list in one of the first e-mails I sent. It was not on my Mouser BOM, though, because I stock them for MiniMAX kits and had quite a few on hand.

For those that didn't see the latest iteration at CanJam, maybe this pic will help clear things up: No comment on switches ...

Yes, good point. However, that's what I was referring to about clearance. The sink is not nearly flush with the lid. AFAIK, you'd to have mount the sink to the case lid instead of to the PCB, with leads from the LM to the pads, etc. Besides wanting to keep everything PC-mounted, we want the lid to be as easy to remove as possible. that's one of the reasons Doug is strongly suggesting that we may use the Heeger cases, because the entire top half of the case is easily removable, as opposed to the slide-in/slide-out arrangement of the Hammond. What happens is that so little of the tubes stick out of the top, you can't grip them very easily to remove them. With 1" diameter holes, you can just barely get a wiggle on them to finally work them out, but it's definitely not as easy as it should be.

Nebby - are you saying that the switch won't work? The cutout shape means little unless the pins don't line up. Red doesn't have much to do with anything, either, I would think. If it doesn't work in the present design, though, let me know and I'll pick another one out. That won't help for those who've already ordered, but maybe we minimize the error for a few. About the heat sink on top of the case. I told a few people at CanJam, but I actually tried drilling 36 holes in the case lid above the PT on mine, but it was too hot. The heat sink works well. Assuming all series-heat transfer, the PT radiates to the inside surface of the lid, conducts to the outside surface, and then to the heat sink. If you can lower the surface temperature of the lid around the general area of the PT, then heat transfer is greatly enhanced and the entire area runs cooler. I really think the curent design reflected in the drawings have the heat issue solved. We may just add some slots on the sides of the tube for convenience when we go to production (assuming the slots don't cost too much). Same for the cooling around the heater supply sink - those will be slots/chevrons, too. It's just difficult to do much more than holes by hand (or drill press) unless we had access to Nate's mill.

IMHO, the case is rejecting all the heat it can already. Besides, there's plenty of clearance, even with a 1-1/2" sink. It's the corner and top slot/edge lip that interferes with the clearance of the heat sink. We also wanted to keep everything mounted to the PCB. Once we have the cases machined for kits, it won't be an issue. No. It's designed for the Hammond slots. I can't speak for Doug, but the side holes are SOP, even for slot designed PCB's. There's always the off-chance that someone will want to build their own case and then they'd be needed. With the Hammond, only the center two mouting holes are necessary with standoffs to the bottom. This supports the tubes when you plug/unplug them and reinforces the center of the PCB from the torque on the ends with the xfmrs.