This is correct. I think the Tribute is a 6V max and the Intact/Slagle is 8V. With the gyrator, if your mu is above 3, unless you use a step down at the input, it's probably not a great idea. To be fair though, I believe that the 6v-8v max is at 20hz. That sort of thing doesn't happen very often.
In Europe TME sells good (for example made by Indel) and cheap transformers, like this.
TS40/018 INDEL - Transformer: mains | TME - Electronic components
I usually use local made transformers:
Webshop
Reichelt also good source, but I'm not sure that theses have split bobbin.
https://www.reichelt.com/de/en/Tran...OUP=C56&GROUPID=7676&START=0&OFFSET=16&SHOW=1
TS40/018 INDEL - Transformer: mains | TME - Electronic components
I usually use local made transformers:
Webshop
Reichelt also good source, but I'm not sure that theses have split bobbin.
https://www.reichelt.com/de/en/Tran...OUP=C56&GROUPID=7676&START=0&OFFSET=16&SHOW=1
Any EI of dual bobbin (=splt-bobbin) construction, recently manufactured is good. Built & tested to some appropriate standard (e.g. EN60742 in Europe). The filament transformers run 100% load, full-time, so elderly junk-box finds are not ideal, and are less likely to be dual-bobbin.
Hammond are fine for the purpose, and I give the Hammond part № in the manuals, because they can be obtained anywhere in the World.
But I use (and recommend) locally-produced industrial power trafos, which are usually the best value. Many industrial cities have a factory making these. The easy to buy example here in the UK is JMS, which has an online form where you can enter a primary and secondary voltage, and have two or three secondaries without hassle.
150VA Premier – JMS Online
shop front:
Products – JMS Online
OK, so they are ugly, but you're going to mount them under the chassis, I suspect. Don't forget a litle ventilation.
Yes, you can use LCL, and it works, but the performance is not at all equivalent.
- The ripple and noise will not approach the µV-level rms measurements given by a properly-implemented filament regulator;
- the impedance cannot approach the KΩ-level impedance looking out of the filament (which is the key to best sound in DHTs), and most importantly the reactive impedance of a choke is not flat with frequency.
- The dc voltage output is at the mercy of the mains voltage, and its tolerance (±10% in the UK and other countries). sustained operation at +5% is enough to shorten the life of expensive DHTs, and +10% must never be contemplated.
- Cost and size may not be a problem for some constructors, but paying more and getting a lower-performing solution does not look like a great idea to me.
I used these - PCB mount transformer 1 x 230 V 1 x 15 Vac 50 VA 3.34 A 85/423 Weiss Elektrotechnik from Conrad Electronic UK Similar to the pair in Rui Lourenco's 26, which I heard and kind of copied. Poorly.
I'm sure that your filament regulator measures better, but the LCL-filter is completely silent on my almost 100 dB sensitive Altec VOTTs. No audible noise whatsoever.
The last choke does not approach the KOhm level below a couple of hundred Hz, but it still isolates the filament sufficiently to ensure good sound.
I set the filament voltage individually for each tube with a rheostat. Where I live (Uppsala, Sweden) the mains voltage seems very stable. I have monitored the filament voltage for days on end and there is no change at all, exept from 5 to 7 p.m. on weekdays when the voltage drops by 0,1V. Not a big problem in my view.
Whether my solution is "lower-performing" is open to debate. I obviously don't think so. I prefer the sound of purely passive filters, LT and HT, with a tube voltage stabiliser here and there. Commercial offerings with SS regulators everywhere don't float my boat.
Well OK, if it works for you in Uppsala, that's good. But that does not mean you say say to everyone that they do not need a filament regulator. For example, the mains voltage here varies by + or - 5% quite a lot, and low-frequency undulations are very frequent, so unregulated heating is (objectively) degraded performance - both for short-term and long-term heating stability and lifetime of the DHT.
The question about the impedance (looking out from the filament) is only partly to do with isolation. Any DHT has a differential music-signal across the filament (the heating voltage skews bias, and gm across the length of the filament). High impedance means that the differential signal remains in place; but if the impedance reduces, the signal is increasingly 'crushed' by the shunt-effect of the lower impedance. Frequency-dependent (reactive) impedance is undesirable, and if you compare chokes versus a solution with high (and flat) impedance, you will be able to hear the difference.
When I talk about "lower performing", I mean objectively, from measurement. Noise, stability, and dynamic impedance, and flatness of dynamic-impedance, are all straightforward performance measurements. They are not like THD, which can be deceptive and very unhelpful.
Naturally, anyone can choose a "low performing" solution, and they are entitled to prefer it. But, until one has performed a careful comparison, claiming that the worst-measuring solution is better is likely to receive some challenges!
Enjoy those DHTs.
Balanced Differential #26 Pre-Amp
What a great thread!!
Unfortunately, it's a bit long and hard to find stuff. I suspect the questions I'm about to ask have been answered before so I apologize in advance (and also for the long post).
I am designing a preamp and need someone to look over my shoulder to see if I'm calculating things correctly.
Design Concept
Each channel of the preamp uses two #26 tubes in a balanced differential pair. The two kathodes of each pair are tied together and bias will be provided by a CCS in the tail.
Each tube will likely operate around 6mA and the anticipated plate resistance per plate is about 7400 Ohms. A drawing of the basic concept is attached.
I don't need much gain from the preamp; 2 dB will be plenty. This will allow use of a step-down transformer to lower the output impedance. The preamp will connect to a crossover with an input impedance of 25k Ohms. I'd like to get the output impedance of the preamp down to 2.5k Ohms, or less. I can fudge on this last requirement if necessary, as the cables will be very short.
Here are my calculations and questions:
1) Load from the Tubes. How do I determine the effective plate resistance of the differential pair as seen by the output transformer? Is it just double the Rp of a single tube? That would be (2 x 7400) = 14,800 Ohms.
2) Reflected Impedance. Assume (for the moment) that the step-down turns-ratio is 4.5 to 1. The reflected impedance from the 25k Ohm crossover will be the square of the turns ratio multiplied by the load: (4.5 x 4.5) x 25,000 = 506,250 Ohms.
3) Inductance Load. The load presented to the tubes from the transformer is dominated by the inductance of the transformer's primary coil. This varies with frequency.
Load = 2 x Pi x F x L.
F = frequency in Hertz
L = Inductance in Henry's
Assuming a transformer with 240 Henry's in the primary coil, the load presented to the tubes at 20 Hertz would be:
6.28 x 20 x 240 = 30,144 Ohms
This is about 2.0 times the combined plate resistance (assuming item #1, above, is correct). This seems low from my experience. A transformer primary with 295 Henrys would produce a load of 37,052 and about 2.5 times the tube's combined Rp. This seems like a more appropriate load for the #26.
4) Low-End Bandwidth. The -3dB frequency = Z/ (2 x Pi x L)
Where Z = "Reflected Impedance" in parallel with the combined Rp of the tubes.
Assuming I got item #1 correct, this should be:
Z = ((506,250 x 14800) / (506250 + 14800)) = 14,380 Ohms.
And, the -3dB freq (assuming 295 L) = (14,380 / (6.28 x 295)) = 7.8 Hertz.
5) Gain. If the tubes Mu = 8.3 and the transformer step-down ratio is 4.5 to 1, the gain for each volt of input will be:
8.3 / 4.5 = 1.84 volts out.
6) Output Impedance. How do I calculate this?
If anyone spots any errors I have made, please let me know. If others have built a similar preamp circuit I'd love to hear about your successes and failures. Advice on an appropriate output transformer will also be welcome.
Thanks!
What a great thread!!
Unfortunately, it's a bit long and hard to find stuff. I suspect the questions I'm about to ask have been answered before so I apologize in advance (and also for the long post).
I am designing a preamp and need someone to look over my shoulder to see if I'm calculating things correctly.
Design Concept
Each channel of the preamp uses two #26 tubes in a balanced differential pair. The two kathodes of each pair are tied together and bias will be provided by a CCS in the tail.
Each tube will likely operate around 6mA and the anticipated plate resistance per plate is about 7400 Ohms. A drawing of the basic concept is attached.
I don't need much gain from the preamp; 2 dB will be plenty. This will allow use of a step-down transformer to lower the output impedance. The preamp will connect to a crossover with an input impedance of 25k Ohms. I'd like to get the output impedance of the preamp down to 2.5k Ohms, or less. I can fudge on this last requirement if necessary, as the cables will be very short.
Here are my calculations and questions:
1) Load from the Tubes. How do I determine the effective plate resistance of the differential pair as seen by the output transformer? Is it just double the Rp of a single tube? That would be (2 x 7400) = 14,800 Ohms.
2) Reflected Impedance. Assume (for the moment) that the step-down turns-ratio is 4.5 to 1. The reflected impedance from the 25k Ohm crossover will be the square of the turns ratio multiplied by the load: (4.5 x 4.5) x 25,000 = 506,250 Ohms.
3) Inductance Load. The load presented to the tubes from the transformer is dominated by the inductance of the transformer's primary coil. This varies with frequency.
Load = 2 x Pi x F x L.
F = frequency in Hertz
L = Inductance in Henry's
Assuming a transformer with 240 Henry's in the primary coil, the load presented to the tubes at 20 Hertz would be:
6.28 x 20 x 240 = 30,144 Ohms
This is about 2.0 times the combined plate resistance (assuming item #1, above, is correct). This seems low from my experience. A transformer primary with 295 Henrys would produce a load of 37,052 and about 2.5 times the tube's combined Rp. This seems like a more appropriate load for the #26.
4) Low-End Bandwidth. The -3dB frequency = Z/ (2 x Pi x L)
Where Z = "Reflected Impedance" in parallel with the combined Rp of the tubes.
Assuming I got item #1 correct, this should be:
Z = ((506,250 x 14800) / (506250 + 14800)) = 14,380 Ohms.
And, the -3dB freq (assuming 295 L) = (14,380 / (6.28 x 295)) = 7.8 Hertz.
5) Gain. If the tubes Mu = 8.3 and the transformer step-down ratio is 4.5 to 1, the gain for each volt of input will be:
8.3 / 4.5 = 1.84 volts out.
6) Output Impedance. How do I calculate this?
If anyone spots any errors I have made, please let me know. If others have built a similar preamp circuit I'd love to hear about your successes and failures. Advice on an appropriate output transformer will also be welcome.
Thanks!
Attachments
Not much action on this thread recently so I thought I'd post a build I just did.
I built myself a 10Y line stage with filament bias at 1 amp so I could swap in a pair of 26s without much work. It didn't take long - just change the cathode resistors to 10 ohms and dial in 6mA on the gyrator. I attach the schematic.
I can say straightaway that the 10Y is a better sound. More clarity, more detail, more dynamic, tighter bass. I was surprised that the usual "warm and euphonic" sound of the 26 wasn't in evidence with the gyrator. It was clear and neutral in sound. But that extra dimension of magic with the 01A and 10Y wasn't there, alas. It stayed on earth. A good clear DHT sound nevertheless and would see off most other line stages. If you hadn't heard what a 10Y or 01A sounds like you'd probably look no further. I don't know if in the case of the 26 the gyrator is the best solution, but it makes demands on a plate choke with its higher Rp.
Where does the 2P29L figure in this? Again not a 10Y or 01A and I've only tried it out with a 126C as plate choke. But I like that combination, more than 26 with the gyrator.
I think the future is with the 10Y, once I figure out the best operating conditions. Thankfully I have a few of them. I still need to hear the 2P29L with the gyrator and also the 01A with Ale Moglia's new source follower board, since I'm driving a PSE 4P1L output stage.
My journey into DHTs started around 8 years ago with the 26 and I've only used DHTs since. I have a lot of affection for the tube, but this wasn't its best incarnation even though the operating points are pretty textbook.
I built myself a 10Y line stage with filament bias at 1 amp so I could swap in a pair of 26s without much work. It didn't take long - just change the cathode resistors to 10 ohms and dial in 6mA on the gyrator. I attach the schematic.
I can say straightaway that the 10Y is a better sound. More clarity, more detail, more dynamic, tighter bass. I was surprised that the usual "warm and euphonic" sound of the 26 wasn't in evidence with the gyrator. It was clear and neutral in sound. But that extra dimension of magic with the 01A and 10Y wasn't there, alas. It stayed on earth. A good clear DHT sound nevertheless and would see off most other line stages. If you hadn't heard what a 10Y or 01A sounds like you'd probably look no further. I don't know if in the case of the 26 the gyrator is the best solution, but it makes demands on a plate choke with its higher Rp.
Where does the 2P29L figure in this? Again not a 10Y or 01A and I've only tried it out with a 126C as plate choke. But I like that combination, more than 26 with the gyrator.
I think the future is with the 10Y, once I figure out the best operating conditions. Thankfully I have a few of them. I still need to hear the 2P29L with the gyrator and also the 01A with Ale Moglia's new source follower board, since I'm driving a PSE 4P1L output stage.
My journey into DHTs started around 8 years ago with the 26 and I've only used DHTs since. I have a lot of affection for the tube, but this wasn't its best incarnation even though the operating points are pretty textbook.
Attachments
Interesting observations. Perhaps you are right and that the mu-out of the gyrators doesn't enhance the "intrinsic" sound of the 26 tube, but somehow changes it to something different. I haven't been entirely happy with my build either (when do we ever?). For me it has become too detailed, clean and neutral. I have experienced something similar when I tried out some amorphous core input transformers from Lundahl years ago.
Yes - the 26 with the gyrator is a disappointment. Not sure why. The amorphous transformer analogy is also interesting. I have a few OPTs - an amorphous core one, a couple of standard Lundahls LL1620 and LL1664 and two O'Netics. Recently I've been listening to the O'Netics. Not the detail of the amorphous core, but I've been liking what I'm hearing, especially the smoother treble.
But the 10Y is something else - I can hear words clearly that I never heard before in choral music, modern gospel and the like. It's a step up in clarity. But again, I'm not sure if it lacks a little body and warmth. Need to sort out the right operating point and run it hotter.
You guys would hate my system, hybrid mu-follower in the strain gauge phono pre-amplifier and more amorphous core transformers than you can shake a stick at. I like the increased resolution and the measurably reduced distortion that the better transformers and mu-followers bring.
Possibly the improvements in the pre-amp performance are revealing problems elsewhere now that the substantial second harmonic sweetening is no longer occurring.
I have had moments like the one you describe. Almost 20 yrs ago I designed a 26 based line stage some of you might be familiar with. It was sweet and mind blowing at the time compared to anything I'd heard. Initially it was pretty forgiving of deficits in my sources, eventually it was the bottleneck in my system even with the switch to 12s and last year was replaced with a new one using EML 20A/20AM into 20K amorphous core opts made for me by Monolith Magnetics.
I built a system around the characteristics of that line stage, and while the new line stage was an immediately noticeable improvement in many ways I have spent the last year tweaking to get the balance I was looking. (Lots of changes to the GM70 power amps) The line stage above 100Hz is <=0.03% thd at any frequency, above 100Hz, while the response remains flat distortion increases quite a bit below 50Hz - possibly running out of primary inductance and an interaction with my tube based voltage regulators.
I ended up with a lot more resolution, better imaging, and a clean, low distortion presentation that never sounds harsh through my horn based speaker system.
You may find a path to better overall sound working with the new design, or not depending on your preferences. I know I developed a marked taste for the sweetening provided by a smattering of thd particularly second from a less than perfectly linear amplifying device. While I plan to stick with SE I hope to eliminate as much of the former as possible.
Possibly the improvements in the pre-amp performance are revealing problems elsewhere now that the substantial second harmonic sweetening is no longer occurring.
I have had moments like the one you describe. Almost 20 yrs ago I designed a 26 based line stage some of you might be familiar with. It was sweet and mind blowing at the time compared to anything I'd heard. Initially it was pretty forgiving of deficits in my sources, eventually it was the bottleneck in my system even with the switch to 12s and last year was replaced with a new one using EML 20A/20AM into 20K amorphous core opts made for me by Monolith Magnetics.
I built a system around the characteristics of that line stage, and while the new line stage was an immediately noticeable improvement in many ways I have spent the last year tweaking to get the balance I was looking. (Lots of changes to the GM70 power amps) The line stage above 100Hz is <=0.03% thd at any frequency, above 100Hz, while the response remains flat distortion increases quite a bit below 50Hz - possibly running out of primary inductance and an interaction with my tube based voltage regulators.
I ended up with a lot more resolution, better imaging, and a clean, low distortion presentation that never sounds harsh through my horn based speaker system.
You may find a path to better overall sound working with the new design, or not depending on your preferences. I know I developed a marked taste for the sweetening provided by a smattering of thd particularly second from a less than perfectly linear amplifying device. While I plan to stick with SE I hope to eliminate as much of the former as possible.