At this stage it's just an idea and there's a also a little back story here: my 6S4S amp is built with money I've made from working extra for my brothers company on weekends, money that is designated for hobby projects. The last time we spoke he mentioned more work in the near future and 300B has been on my list of potential tubes that could match my LL1688s.
Throughout the years I've had a myriad of ideas on this subject, including PL519 with local feedback and some rare zero bias transmitter triodes (RCA 808) that I have. A no-nonsense circuit with one of the most well-documented power triodes in history feels like a safer bet than most of the more creative ideas I've come up with.
Regarding power, my 6S4S project changed course from SE to PP after testing a few different amps with my speakers and realizing that they seemed happier with a little more power. I still fancy the sound from SE triodes though, so slightly bigger SET would be an interesting project. 845s running at "low" voltages would also be an option.
Throughout the years I've had a myriad of ideas on this subject, including PL519 with local feedback and some rare zero bias transmitter triodes (RCA 808) that I have. A no-nonsense circuit with one of the most well-documented power triodes in history feels like a safer bet than most of the more creative ideas I've come up with.
Regarding power, my 6S4S project changed course from SE to PP after testing a few different amps with my speakers and realizing that they seemed happier with a little more power. I still fancy the sound from SE triodes though, so slightly bigger SET would be an interesting project. 845s running at "low" voltages would also be an option.
Perhaps a little bit off topic but I have this idea that keeps coming back to me: How about an amp with two, three or even four different sockets (per channel, of course) for different outpout tubes? Say for example one 4 pin socket wired for 300B (with separate DC filament supplies), one octal wired for 6L6/EL34KT66-88-90 and maybe even a 5 pin socket for 807, as I have dozens of various 807s.
Perhaps even two octal sockets, one wired for straight triode operation and one for UL with CFB?
A proper driver stage (PowerDrive) should be able to drive the additional stray capacitances like nothing, and allow a wide range of negative bias voltages.
Perhaps even two octal sockets, one wired for straight triode operation and one for UL with CFB?
A proper driver stage (PowerDrive) should be able to drive the additional stray capacitances like nothing, and allow a wide range of negative bias voltages.
I have installed two sockets on a couple of amps. One was on my 30 to EL84 or 6V6 amp, and the other is my current build, which is a 45 or 46 amp with an SV83 pentode driver. Attached is a photo of this amp. With regular directly heated pentodes run as triodes, it is fairly trivial to install different sockets in parallel, with their own respective circuit because the filaments are usually all 6.3 Volts. The tube that is plugged in will work. With DHTs it is also doable, but somewhat more complicated because the filament requirements vary. I have a switch on my 45/46 amp that changes a pair of resistors on the Coleman filament supply to go from 45 to 46. They are both 2.5 Volt, so it is relatively easy. Changing voltage complicates things more.
Attachments
Nice project there, tizman!
A regulated, adjustable HV power supply would probably be a good feature in an amp like this. A makeshift solution could be to use two or more small caps between the rectifier and the choke and adjust the voltage by going from "almost choke input" to full C input. Should work fairly well in a class A amp.
The whole concept is quite interesting, I guess most of us have more than a few odd tubes that we would like to use but time and money gets in the way for building amps with them. I for one have a few used 6L6s of different models and dozens of 807s that don't belong to the matched quads I need for my 807 PP amps.
A regulated, adjustable HV power supply would probably be a good feature in an amp like this. A makeshift solution could be to use two or more small caps between the rectifier and the choke and adjust the voltage by going from "almost choke input" to full C input. Should work fairly well in a class A amp.
The whole concept is quite interesting, I guess most of us have more than a few odd tubes that we would like to use but time and money gets in the way for building amps with them. I for one have a few used 6L6s of different models and dozens of 807s that don't belong to the matched quads I need for my 807 PP amps.
As long as we use a driver stage that can override the capacitance added by the extra wiring and sockets, I don't see any problems with it. I thought long and well about building something like this with my 3,5k James transformers, focusing on slightly smaller tubes (6L6, 807, 6S4S and maybe 6P21S) but those transformers ended up in another project that I built on an impulse while waiting for parts for my 6S4S PP.
The LL1688s would probably be a better choice anyway, allowing bigger tubes and a more convenient 5,5k loadline (my speakers impedance is a bit on the low side).
807 in triode mode for non-critical background music, 300B operating at a high-ish Z plate load for serious listening sessions and KT88 in UL with a bit of CFB for Motörhead, could be fun
The LL1688s would probably be a better choice anyway, allowing bigger tubes and a more convenient 5,5k loadline (my speakers impedance is a bit on the low side).
807 in triode mode for non-critical background music, 300B operating at a high-ish Z plate load for serious listening sessions and KT88 in UL with a bit of CFB for Motörhead, could be fun
Thanks! Long ago, when I first started building tube amps, I went to the home of an older gent who had been building for many years, and his amps were all split into sections. He had power supplies and signal circuits on separate chassis. This isn’t that unusual, but he also had a few OPTs on their own chassis and separate from the signal circuit. On a shelving unit he had three or four power supplies, a few complete signal circuits on their own chassis and then a few OPTs on their own chassis, and a few signal circuits without OPTs on their own chassis. Confused by this at the time, I asked him about it, and he explained that when he was using expensive, high quality OPTs, he didn’t want to commit them to just one circuit in one amp. I have been thinking about splitting off OPTs from the signal circuit for some time, but haven’t ever done it. In the case of the amp above, the OPTs are nice Electra-prints that could be used for a number of other tubes and circuits. I’m now just splitting power supplies and signal circuits. Baby steps.
Plenty of good ideas! I've also contemplated the possibility of building an adjustable, regulated power supply that would cover the needs of pretty much any normal tube amp. I'm sure it would save money in the long run, as quality iron is expensive and surprisingly hard to find. My latest three amps actually uses the same external power supply, although one of them is single ended and has a hum problem when powered from the PSU that was initially built for a PP amp.
I read about "recycling" output transformers in a thread somewhere here just the other day, someone suggested they could be be put in boxes with bases from scrapped octal tubes as connectors. Sure an elegant way to make it easy to move them between different amps. I have a dozens of old 10 pin connectors rated for 500VAC 16A or something that could serve the same purpose.
I read about "recycling" output transformers in a thread somewhere here just the other day, someone suggested they could be be put in boxes with bases from scrapped octal tubes as connectors. Sure an elegant way to make it easy to move them between different amps. I have a dozens of old 10 pin connectors rated for 500VAC 16A or something that could serve the same purpose.
I've also had this idea of building a "generic SE driver stage" and a "generic PP driver stage" as separate modules. Would come handy when experimenting with oddball output tubes, the kind of experiments that tend to become semi-permanent. It would require some serious connectors though, possibly some high quality XLR ones or something.
I concur. I played briefly with the holy grail a pair of WE 86 amps, 300B PP with interstage transformer. They sounded nice but not THAT nice for their ungodly price. To me 2A3's bass is fuller and has a more robust tone. The 86s didn't sound great until matched with a pair of WE TA-7381 preamps.
That could have been me!! I have separate PSUs, output stages and input stages. I have 3 or 4 of each. So I have my "best ever" amp, which is 10Y into 2a3 with a 1:4 step-up in front. And because the 10Ys are rare and expensive, and 2a3s aren't cheap, I have a "daily" amp which is around 95% as good. It has all the tonality I need on acoustic instruments and just lacks the extreme clarity and micro-detail of my best amp. It has an odd combination of type 47 input and EL12n outputs (mu = 18 in triode), which you can buy new in Europe still. By sheer luck this combination sounds marvellous and you get the DHT clarity and tone in the input stage rather than the outputs. Just 2 stages, so very clean sound.
Andy: I don't have any 2A3 on hand. Have you tried any new production 2A3s that you like?
I use Hammond 1140-LN-C. No doubt the 1140-LN-D is OK too. Studio range and very transparent. Better than an extra tube stage. The SUT enables me to build a 2 stage all-DHT amplifier which was always my goal. You just need a DHT with a gain of around 8, so 10Y, 47 in preference and in filament bias. Also possible are 2P29L, 112A and 4P1L but they're not as good.
I just have EH 2a3s. I also use 6S4S which are Svetlana and very nice.
andyjevans: I have read about your efforts with step-up input transformers to get enough gain in your system. Transformers are picky when it comes to source impedance, perhaps it would be beneficial to use discrete Jfet buffers in front of the transformers to minimize the influences from using different sources?
Such buffers can be built with very few components and are very transparent. The Pass Labs forum should have plenty of information on this subject, one example would be the FirstWatt M2 where a complementary Jfet pair is used to drive an autoformer that does all the voltage gain in the amplifier.
Such buffers can be built with very few components and are very transparent. The Pass Labs forum should have plenty of information on this subject, one example would be the FirstWatt M2 where a complementary Jfet pair is used to drive an autoformer that does all the voltage gain in the amplifier.
Indeed - a 1:4 SUT requires a low impedance source. My DAC is OK for that, but I have very short cables made of twisted pairs.
I do use the 1140-LN-C in reverse and single ended.
Here's a draft of my 47 > EL12n amp. My daily amp. Excellent tonality, not quite a 10Y > 2a3 for micro-detail but close and much cheaper.
I do use the 1140-LN-C in reverse and single ended.
Here's a draft of my 47 > EL12n amp. My daily amp. Excellent tonality, not quite a 10Y > 2a3 for micro-detail but close and much cheaper.