|'); } // End -->|
This is just a generic all-purpose late-night-after-some great-listening post to say that after 3 months of eliminating everything but DHTs from my system, I've never been happier. Why anybody could stop using these miracles of sound I just don't know - decades and decades went by when they were just tossed into drawers and cupboards. I was lucky to have friends who turned me on to the DHT (Nic, Piotr..), and even luckier that they turned me on to the small tubes. After amassing a load of good but now useless IDHTs (I used to think these were the end of the road), I've simply come to the conclusion that it's DHTs all the way through the system. That's it.
I tried going back to my previous system of IDHTs yesterday and I could hardly listen to it - warm, vague.......... A system I'd enjoyed for years.
but I don't know what you mean by 'small tubes'.
Personally, I like big fat bright triodes. Anything with cylindrical plates and thoriated tungsten filiments is my clear preferance.
I totally agree, why would anyone want anything else?
I would. After protyping a 01A preamp and hearing it, there's no turning back. I've also this urge to change the drivers of my SE amps (I have 2A3, 10Y, 45, 50, and 300B) to DHT but that will be a major overhaul. I will just make an all-DHT amp later :-)
DHT is cleaner, better balanced, warmer in the right way, no lean clean ugly machine as so much dreck out there seems to be.
I certainly agree with Andy on the DHT merits and you on the loudspeaker problem. Try this out, go here http://www.positive-feedback.com/Issue21/standingwaves.htm
Read the entire thing and then we can start a thread over on speakers and irritate everyone... or not.
Hi Andy, sounds like you are extremely pleased, just as you should be.
I absolutley agree and would even go a step further. Among the DHTs the ones with thoriated tungsten filaments are the most transparent ones.
When you say you went all the way DHT. Does this include the phonostage ? CD or DAC output stage ?
I'd say up to and including linestage it's still a bearable effort to use all DHTs. When it comes to phono preamp it really gets difficult. I do have an all DHT phonostage right now but it's very sensitive to microphonics and all kinds of disturbances.
You're not allone with your preference!
What valves do you use in your DHT phono and are you able to tell it's topology?
The phono preamp has 4 stages. Each uses a circuit which I call "DirectPath". Here is an example of such a stage (A Line preamp in this case):
The preamp can be configured (by switches) for 4 different tubes in each socket: UX201A, UX112, 26 and 801A. 801A is my favorite.
For UX201A and UX112 the filament current is completely obtained from the B+ via the large resistor. For 26 or 801A additional current is injected at the more positive side of the filament via a choke.
Hope this describes it well. I need to draw up a circuit that shows this more clearly.
It uses all Lundahl transformers. EQ is done by 600 LCR network, placed between 2nd and 3rd stage.
MC step up is a 1681. 1st stage to 2nd and 3rd to 4th stage coupling is done by a 1660. the LCR EQ and output is driven by a 1680 9:1 step down. 1:5.6 step up via 7905 betwenn EQ and 3rd stage grid.
The first stage tubes need quite extensive damping for microphonics and also shielding with Mu metal tubes. Still it's very sensitive. Depending on placement it can pick up the signal from the speaker and produce horrible feedback.
It is not my phonostage for everyday use ;-)
So is your daily user the diff EC8020 all iron no C or R design?
The phonostage I currently use the most is EC8020 based, but not the differential one. It uses two stages of this circuit:
A parafeed circuit without coupling cap. The DC flowing through the primary is determined by the DCR of the winding. The transformer is a Lundahl 1677 used in reverse. I chose this one because it has a quite high DCR. B+ is around 90V. With this the DC current is within spec for the core.
600 Ohm LCR EQ between the stages.
You have some really interesting ccts in your collection. One thing that I've wondered for some time. There seems a trade off using the conventional series fed transformer vs conventional parafeed in that at some point a cap is involved. However the "theory" goes that the parafeed method can use a smaller cap (hence "better") and a more advanced core. OK you got rid of the cap but did you ever try say a decent cap (you might not think there is one :-) ) with a small amorphous core tx like Lundahl's 1544, 1676 etc?
What seems to you to be the attraction in the single ended EC8020 cct vs the differential one? Again, so many say how much better balanced is vs single ened. Your two ccts get about as close as you can to compare the two. What are your thoughts?
I arrived at the differntial circuit by a different way than people usually do. In my quest to minimize the components in the signal loop I finally had gain stages with 3 elements in the output loop. These were the SE circuit in ultrapath, or parafeed in ultrapath. These basically had the tube, transformer primary and 1 cap in the loop.
I must admit I'm not a big fan of component comparisons. Once I identify a good component I tend to stick with it. I still use the same type 8uF cap which I got years ago. I did attend cap comparison tests but got bored of them. I find much bigger differences in circuit topologies.
So I never tried to optimize the cap size in a parafeed circuit. Can't comment on that.
When thinking about how to get rid of that last cap, I came up with two ways to substitute it. Replace it with a resistor or with another tube. The solution with a resistor burns a lot of power. The circuit with the second tube happens to be a differential scheme.
I always note biggest improvements when geting rid of a cap. The differential circuit brings an additional benefit by the large immunity to effects form the power supply. I can add some kiloohms of resistance in series to the power supply connection of the diff phono without noticing any degradation.
I like my current single ended EC8020 phono the best. It's as simple as it can get, just the tube and transformer in the ouput loop.
I'm with you on the cap comparison thing. In fact, for most purposes, I don't compare components at all as much like you (maybe), I've found a set of ingredients that seem to work well enough and there are bigger things to muck about with. I tend to be moving in a direction where more coils seem advantageous :-)
I note what you say about the diff cct being more immune to psu artifacts. I've been thinking that many people use diff ccts for that very reason however if the psu is shall we say "super", then the benefit of differential starts to diminish over single ended due to complexity. All other things being equal of course. Is this your experience?
BTW, I'd be interestd in the type of cap you mention. Reason being, up till now, I've used high uF power supplies over lower C ones as I found, for my taste, Black Gates NH and WKZ to be more suitable than the metallised film type I was using in my phono unregulated supplies.
our preferences seem quite similar :-)
I agree on the differential circuit and PS quality. The purpose of the diff circuit should not be to get away with a poor PS.
As mentioned I did try to hear the influence of the PS by just adding a 2 kOhm series resistance (with adaption of the voltage to compensate IR drop. The PS was a pretty good one with very low ripple.
I didn't hear any significant degradation.
The cap i use is a 8uF Matsushita MP. I guess that is metalized polyprop in oil. Bot I'm not sure.
I have a preference for low C values.
Best regards ... Thomas
When you say you went all the way DHT. Does this include the phonostage ? CD or DAC output stage ?>
Yes, DAC goes directly into the grids of my line stage, which is balanced. Balanced is a real help with DHTs - less microphonics.
I'm tempted to build a DHT phono stage - haven't got round to it yet. It would be balanced, I think, and transformer in. Any ideas for what DHTs to use? The 5767 has been used for example.
When you say DAC goes directly into the DHT linestage, does the DAC still have it's own outputstage, or do you go directly from the DA-Converter ICs to the DHT ? If not, there is still room to imrpove even further.
Would you share the schematic of the balanced linestage ? I attempted a differential circuit with DHTs, which had some noise problems. This came through the differential nature, which means the filaments have a high impedance connection to ground.
As mentioned above I like thoriated tungsten filaments best, so that's tubes like the UX201A, 10, 801A, 211, 845, etc
When you say DAC goes directly into the DHT linestage, does the DAC still have it's own outputstage, or do you go directly from the DA-Converter ICs to the DHT ?> >
Directly from DA converters to grids of line stage, or first stage of integrated amp, which is in both cases the first of my 3 amplification stages. The volume control is after the first stage.
Would you share the schematic of the balanced linestage ?> >
It's quite banal, and balanced all through - differential pair with anode resistors appropriate to DHT (27k to 47k, whatever) and CCS under it. Published designs for a CCS include Allen Wrights, on his site, and Morgan Jones on page 134 of his book. You can cap couple this to the next stage or preferably use a transformer in push pull - I have some Lundahl LL1660, LL1635, LL1671 for example. Then the volume after the first stage and another driver stage as above, cap coupled or ideally transformer in Push Pull. Then the output stage.
I attempted a differential circuit with DHTs, which had some noise problems. This came through the differential nature, which means the filaments have a high impedance connection to ground.> >
??? Not sure how balanced can be worse for noise - my amp's entirely silent.
I had some noise coupling between filament supply and B+ supply of the circuit. Since they did not share the same grounds, I got some noise. How is your filament supply connected to the B+ ground ? I assume since you use a CCS in the cathode (filament) to ground connection it's floating ?
I had some noise coupling between filament supply and B+ supply of the circuit. Since they did not share the same grounds, I got some noise. How is your filament supply connected to the B+ ground ? I assume since you use a CCS in the cathode (filament) to ground connection it's floating ?>
Yes, that's right. The filament supply is floating in all cases. I run each diff pair off one filament supply, in series or parallel - doesn't matter with a current source since if you take one valve out in the first case it goes open circuit and in the second the current stays the same. As you say, the CCS sets the current in the diff pair and that's earthed to the common audio earth, with -15vDC supply.
Sorry, this isn't an answer to your question. I ran across your post just as I was thinking about doing a DHT amp myself, and so have to ask you how you've done it.
Still not definite on which output I'll try (but likely either 300B or PX-25) I've been wondering about DHT drivers and also the consideration of filament supply that DHT's warrant. Would you mind sharing any tips on what you've learned with regard to good DHT drivers (the small tubes you mentioned ?) and also anything you've found particulary good in doing the fils?
To choose a driver , first you need to know the requirements of the output stage and if you're going to use an active linestage . Then you can work backwards to find suitable driver candidates . With DHT's you don't have a lot of choice , typical DHT driver valves have mu=3-6 Ra 2-5K , mu=8-12 / Ra 6-8k . There are also valves which have mu=30 but much higher Ra and a few unusual thoriated types such as 15E and 3C24 which have mu=20 Ra 15k . Try to avoid multi stage amps using battery valves , these have sub 200mA filaments and can be lead to microphony . Also on the same note ensure that there is not a hideous excess of gain as microphony on a two stage driver DHT amp can be a real problem . For your driver you will probably need to use DC , I would recommend LCL filaments , which is the bulkiest but IMO best sounding . Alternatively the Ronan reg style setups are ok (I think Andy uses these) , with valves with sub 250mA filaments a DN2540 depletion mosfet makes a good current source for filament heating . Other recommendations : use rubber isolation bushes to suspend the valves under the chassis , also make the chassis heavy and damp/brace extensively as a top-plate that rings also makes the valves ring . Once you go with DHT's , nail the quirks and enjoy the music , there's no turning back ;)
Try to avoid multi stage amps using battery valves , these have sub 200mA filaments and can be lead to microphony .> >
My answer to this is use balanced throughout, with constant current sinks under each stage.
Also on the same note ensure that there is not a hideous excess of gain as microphony on a two stage driver DHT amp can be a real problem.> >
Three stages should do it. Fatbottle's right - just use the amplification you need and no more. I'm OK with 3 stages, 3a5=mu 15, 1H4=mu10 and 2a3 outputs. Try and maximise the gain in each stage, e.g. active loads (haven't tried this with a CCCs under, but Morgan Jones got it to work, it's in his book), transformer coupling.
For your driver you will probably need to use DC , I would recommend LCL filaments , which is the bulkiest but IMO best sounding . Alternatively the Ronan reg style setups are ok (I think Andy uses these)> >
Yes - filament supply is essential. I don't think there are any really good published designs on the Net - the Ronan reg style setups work well. I found the circuit somewhere - forget where. You need a current source - sounds audibly better than a voltage source.
use rubber isolation bushes to suspend the valves under the chassis> >
I haven't done this. I did put rubber rings (plummers ones) round the 3A5.
also make the chassis heavy and damp/brace extensively as a top-plate that rings also makes the valves ring> >
Yes, yes, yes. My chassis plates are 4mm alu. That's pretty massive. Forget thin alu - must be rigid and not ring at all. Otherwise you have a bell.
Once you go with DHT's , nail the quirks and enjoy the music , there's no turning back ;)> >
A word on small DHT tubes. My findings:
1G6G - mellow, but rather bland. Quite usable. Available cheap
3B7 - didn't like this one - harsh in my setup. Available cheap
3A5 - surprisingly good with real sparkle - top is bright, however. Tame the top and you have a very good sound. Available very cheap
30, 1H4 - lighter but detailed sound, good treble. Plentiful
VT-67 - 30 on steroids. Uncommon and pricey.
31 - lighter, detailed, useful lower anode impedence. Available
26 fat warm sound but detailed, some complain of dull extreme treble. Very plentiful
1G4 - mellow, warm verging on a bit tubby. Available
5676 - balanced, rather bland but very usable miniature. Available
12A, 01A - good allround favourites - 01A is a bit lighter sounding. Available
X199 - unusual, haven't tried it. Pricey
49 strapped as triode. Available, good value
46 strapped as triode. Available
71A. Available, pricey.
Haven't tried the above enough to comment.
I'd recommend pairing a light sounding tube like the 30 or 3A5 with a warm sounding tube like the 26. I haven't gone into this in detail, but I can see it's the way to go. I presently have 3a5 into 30 and it's a bit too light and bright - you need some warmth lower down to round out the sound. There again, with the 26 you need to brighten it up a touch in the treble. Consider this with your choice of output tube - with 300b you might want brighter drivers, with 2a3 you'd be looking for at least one stage with more body and warmth. And ultimately this depends on your personal preference - I like a lighter more transparent sound, others like a full warm sound with the detail still there but less in your face.
Hey thanks guys.
Still kicking tires. For the last couple of years I've been listening to a 6BX7 driving 6336 Parafeed. I love the sound and might be happy staying with it but the amps are huge (four channels for biamp) and hanging out all over the place and need a more contained rebuild if I want to make them keepers that I feel OK letting my friends near. Before investing the time and energy I'd like to give a linear DHT with about the same power output a try. My system setup is pretty well matched gainwise and to maintain that, it works out that with a PX25 as output, the driver mu would still be best at something around 10. For that, it would probably be easiest to use the 6BX7 again but you know how it is, new horizons beckon and all that, so a DHT driver IS tempting. . .
This is where I start to wonder though because with a biamped system the supply considerations are doubled.
For a single stereo amp my leanings would be to something like your LCL fils Fatbottle, but eight DHT fils done that way would make amps I'd need a fork lift to get on and off the bench. I know there's got to be a good way of doing it, it's just taking a while to arrive at compromises I like the look of.
I've been searching around to see if anyone has tried using just a single choke as the L in an LR filter - the fil providing the R value - but haven't found anything more than a single hint and no clear indication of the degree of success obtainable. I'll give it a try on the breadboard.
Andy, thanks for your mention of the 12A. I ran across that tube a few years ago but never tried it. It's because of you that I got around to building a circuit with the 2C22 - which I really liked - so I'll probably give the 12A a go in the new amp - with or without DC fils.
Anyhow, no hard data here, I really just signed on to say thanks for your replies.
If either of you happen to run across any interesting schemos or pages on DC fils would you please post a link?
Thanks, I'll post again with results once something gets built.
Lots of good information there. I especially like your comparisons
of the various tube types. I have a few of most of these types and
am looking forward to experimenting with them.
I don't have an all DHT system yet but I am working on it. My next
design is going to be all DHT p-p amp to go with the all DHT pre
I have already built.
DHT's gotta love em.
Glad to help. I should have mentioned 6J6G in the double triodes - balanced neutral sound, quite nice, bit bland, but very usable. Handy octal base, mu 14. Would probably go rather nicely with 1H4, another octal, which could be used as a driver, mu 9.5.
Any experience with the 3C6 Loctal? I have a few of these somewere.
Probably microphonic due to the low filament current but looks interesting.
I too liked your posts. Studied it and looked up a few tubes. Happy too that NJ7P tube database is back after a few days of hinky operation.
You mention 6J6G. I know the 6J6 to be an IDHT twin triode.
Do you mean 1J6G. I have been looking high and low for information on Class A operation of 1J6G. I finally bought some and tested on my RAT tube tester. However, I got a mu of 2- - 21. Now maybe I tested wrong -- you indicate a mu of 14, assuming you are talking 1J6. Do you have a data sheet on it?
However, I got a mu of 2- - 21
Sorry. Left hand broken - bad for typing.
I meant 20 to 21. I have a Steve Bench design RAT tube tester I built three years ago. However, I did not build his supplemental power supply for DHTs, so the ony DHT I have tested was 1J6. I calibrated first with known IDHT tubes, but my mu readings were stable and consistent.
Anyway, my starting point was a search of the (now) many good web sources for datasheets to find some Class A info on 1j6, but to no avail.
This page links to a number of data sheets, but I didn't see any with class A.
Indeed - all the online data looks like class B. I don't know where I got mu of 14 from. Kurt Strain was using them - maybe something he wrote? Sounds about right, though. It's a smooth sound, but not holographic like the best DHTs.
"I don't know where I got mu of 14 from."
I don't know for sure anymore, but 14 sounds about right to me, too. These are nice sounding, but a bit warm and fuzzy for a DHT, after looking into other DHT's out there. Very easy filament to heat, though, not a lot of current. They were designed for battery operated class B output stages drawing low power.
because he used 1J6s as Class A drivers. I studied a lot of his posts, but never saw a statement of mu for Class A purposes. I never asked him, concluding instead to procure and test. Have not tried them in a circuit yet.
I built a PP KT-88 amp using 1J6G as input and driver. Mu is indeed *about* 14, but it' hard to be accurate as there's no class-A data about. I just extrapolated the class-B data.
Never Heard Of Them.
This post is made possible by the generous support of people like you and our sponsors: