OK, I have to confess I've been having a bit of a problem with my heptodes lately... They all sound good... And it makes testing and comparing them very difficult. Add the three -I'm still trying to figure out a practical way to use the fourth one- strapping settings and you've got a bit of a headache ahead of you... I feel like I'm 3 feet deep in "heptode cases" to test!
Anyway, for once, instead of doing long burn-in periods -both electrical and brain burn-in- for each tubes, I tried to kind of "blitz compare" a few of my newest ones (the ones for which I didn't have a solid brain impression and conclusion for like the 5915); normally this wouldn't work since the tubes aren't exactly burnt-in, but it kind of does since many of those latest tubes have either had the level of detail and "musical glory" I seek right from the get go, or basically have not; burn-in just mixing things up a little (bass tightening, etc...).
So, here are a few conclusions after this (and, again, take my -and others- comments about strapping settings with a grain of salt, since the gear you have seems to have a pretty heavy influence on which setting you'll eventually settle for; at least that's the only pattern I can discern for now regarding strapping heptodes) and also a few thoughts after browsing Google way too much at night:
- The IBM 1680 (1950, made by RCA for IBM, basically a 6BE6 with grid 3 like a 6BY6, so for all intents and purposes a computer rated RCA 6BY6 prototype) I got have a very special, lovely and addictive -foot tapping- sound that is unlike anything I've heard from a heptode yet. Think heptode detail with pentode musicality; yeah, that sounds nice, doesn't it? You can get these on ebay in the US for about $3 a tube if you negotiate; honestly, if you live in the US, get a pair! Anyway, I've only used these for 3 hours, and the bass was still all flabby, but they made for an extremely pleasant listen! They sound more relaxed and realistic unstrapped than 1-7 strapped -on my gear and to my ears- btw; they immediately showed quite a treble bias, like a slip of every frequency towards treble, when strapped.
- My two kind of Haltmens (CV453 labeled and 6BE6 labeled) sound very similar, and I believe the slight differences I'm hearing are your basic placebo effect and expectation bias; that and one pair being more burnt-in than the other. Both pairs sound quite glorious, with the top-tier level of detail we've come to expect from heptodes, musical realism and focus.
- Or so I thought, but testing some unburnt-in Telefunken EK90, I have to admit that the Haltmens have some serious competition. Basically, the Telefunken seem to be at least on the same level in terms of detail, focus and realism, but have harder hitting bass and even better treble; trade-off seems to be that they are just a bit more forward and maybe treble biased -unstrapped and unburnt-in- but not as forward as my Tele EH900S, which were too forward imho. I'll actually give these a proper burn-in now, but I could swear they render more detail than the Siemens... Anyone else have both to compare?
- Thoughts: from google tube literature, we'd assessed that grid 3 on 6BE6/EK90 has remote-cutoff curves (meaning that, despite being the actual "control grid" in the tube -unlike the oscillator grid 1 we've been using as a signal grid- we shouldn't use it in our triode-strapped preamp gain stage application as a signal grid), which is why I won't even bother 1-7 strapping it to grid 1 in those tubes anymore, since it doesn't make sense and doesn't sound the best. Basically, grid 1 and grid 3 on those tubes have different characteristics and are meant for different applications; we've ended up using a non-control sharp-cutoff grid for another application than it was intended for; we didn't have a choice and it works fine, so great, let's keep doing that.
- 6BY6 tubes, on the other hand, are switching tubes, with two actual control grids, both of which have sharp-cutoff characteristics (and similar gain/mu, to the best I understand and have read). Both grids can, logically, be used as signal grids -grid 3 would be the one that is actually meant to have a negative voltage bias, like grid 1 in pentodes or triodes though- and we've been using grid 1 for that because it's easier. Working out a way to neutralize grid 1 are using grid 3 as the signal grid, as I've suggested before, would make perfect sense, but may not be worth the extra trouble, since we can't prove that one of the two grids would work better, as they are quite similar. So using grid 1 is fine here too. Actually, any setting seems fine for 6BY6 tubes, both in theory and in practice.
- Enter 6CS6/EH90 tubes. These have traits from both the above tubes: they're meant for some kind of TV application and not as switching or logic tubes, but both grids are control grids with sharp-cutoff curves. But, unlike a 6BY6, the two grids have very different gain/mu, with grid 3 having about three times the mu of grid 1 (high mu isn't necessary for good function, but is a good thing as far as I'm concerned; my only issue with all our heptodes so far is that they have a very low mu, and therefore low volume). Basically, grid 3 has better characteristics as a "virtual triode" signal grid for a hi-fi preamp than grid 1 does. So, for 6CS6/EH90, neutralizing grid 1 (strap to cathode) and using grid 3 as the signal grid (tying it somehow to socket hole 1) would make more sense than just using grid 1. No, it's not necessary, but it just seems like a better way of using the tubes, that I'm very seriously trying to figure out an easy way to test, short of an adapter. Strapping grids 1 and 3 together could work fine (anyone try it actually?), but I have no idea how tying two grids with different gain characteristics will behave; I don't think it can't sound bad, I just don't know.
Anyway, sorry for long babble, but I still think we need to make sense of how and why our different types of heptodes behave the way they do, if we actually want to eventually understand what we're doing here (up to now, we've just slapped some tubes in our amps, just expecting them to work; we're reaching the limits of that system with heptodes imo; and as some have pointed out, we just can't test dozens of tubes with two or three different settings every time...).