• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

#26 pre amp

For clarity: Illustration attached of my application for this.
OK, that is unusual, and not the simple circuit I thought you were using.

In your circuit, the power supply to the lower valve has its DC and LF impedance set by the cathode of the TOP valve. This is like a shunt regulator, but using the audio device as the shunt element. This is what sets the DC supply voltage, though it will vary for different valves.

The audio cathode is working to suppress LF noise coming from through the CCS, and this would be reflected in the output current. A small amount, but not zero -- depends on the CCS design, and the gm of the top valve. Cascoded depletion FETs would let the smallest amount of noise through.

I think it would work better with the top cathode disconnected from the CCS, and a shunt regulator installed.

A string of 5V6 or 5V1 zeners would also work quite well, and would demonstrate whether you like the general idea.
 
OK, that is unusual, and not the simple circuit I thought you were using.

In your circuit, the power supply to the lower valve has its DC and LF impedance set by the cathode of the TOP valve. This is like a shunt regulator, but using the audio device as the shunt element. This is what sets the DC supply voltage, though it will vary for different valves.

Drawing shows two DC coupled stages with stacked B supplies. Perhaps not obvious from the way it was drawn, but I couldn't think of a better way to draw it.

First stage: fixed Vg and Ip, and Vp is set by tube sample.

Second stage: Vg equals Vp (of first stage), bias voltage is Vp (of first stage) minus volt drop across choke (of first stage) and is therefore consistent regardless of input tube sample.

The second stage inclusion is to illustrate why I prefer the current reg to voltage reg in this instance, but gives no consideration to 'audio performance' (how else to say) ..

In isolation, the first stage is very likely the simple circuit that you had anticipated.

Thanks again.
 
Last edited:
Jogged by the latest posts here I converted an 01A stage that wasn't working very well to 26. I haven't used this tube in ages, but it has a special place in my heart since it was the tube that converted me to DHTs. My jaw literally dropped when I first heard it and I muttered "this is the tube I've been waiting for all my life!"

That was back in 2011, but subsequently I bought some 10Ys and that became and still is my tube of choice. Behind that tube are the 46 (close), 47, 112A and 26. A bit further back the 2P29L. The 46 has the best piano tone and is a leaner, more percussive sound. The 26 is a little plummy on piano but wonderful on vocals. The 10Y is the master of inner detail and subtlety, including the crystal-like treble. The 26 has good treble, but not 10Y class. But nothing really matches the 10Y for all round performance. All these have been used in filament bias with Rod Coleman regs. Big chokes in the filament supply and paper in oil and DC Link caps in the PSU. I use a 1:4 step up since this stage directly feeds my 2a3 SE outputs. The Hammond 1140-LN-C is excellent - one of their broadcast range and remarkably good value.

So here's the circuit I made. I like it and it will stay in my system for a while until I miss the 10Y too much. I sometimes put in the 46 stage, which is very realistic on jazz.

View attachment 1109494
Hi Andy,
I've been trying to find your definitive 46 preamp schematic, but I've been unsuccessful to date. I've got a bunch of 46 valves and I like jazz, being from pre-Motown Detroit, and having been collecitng jazz LPs since 1958. What does your 46 Pre schematic look like?

Best, Robert
 
Hi Robert

This is what I used for the 46. I'm now using the 10Y but the 46 is good too. I mounted the cathode resistors above the top plate - they get hot. I used a combination on a tag strip. If using Rod Coleman regs V9, the supply should be 17V. Rod will always help with designs. Lyrima is his website. The 470K on the output was because it was driving a 2a3 stage. For that the coupling cap only needs to be 0.1uF FT-2. For larger values as a preamp I'd use a Russian KBG-MN 1uF which you can get from the old satellite Soviet Union countries like Bulgaria, Ukraine etc.

46 filbias4.jpg
 
Running a reduced voltage across the filament (starved) is good in thoriated tungsten types (01A, 10Y ...) and so long as the sound, and the gain are still the same as with the nominal filament current, and voltage all will be fine. With these tubes, the lifetime is often improved, as well as the lower microphony.

With oxide coated filaments (most other receiving tubes) Running the filament low is likely to cause more trouble than it's worth.

Oxide coated filaments have a nominal heating power which has been carefully determined to maintain a dynamic flow of active ingredients that support emission. It's an equilibrium of supply: just enough to support emission. (Overheating simply burns up the active ingredients too fast, and shortens the lifetime of the tube).

Underpowered heating - "starved" is a good name for it, as low filament power starves the supply of sub-surface barium for the active surface, disrupting the dynamic equilibrium; emission soon falls. After a time, the diminished emission will affect the available current peaks and gm; this is certainly not good for the sound.
Running 2A3s at 2.2V is more than 10% down on voltage and probably 20% down on power (as the current also drops). I suspect that this will cause a fairly rapid loss of performance, and dull sound, like that of a worn out tube (which is what it has become).
I would not do it, and recommend adjusting the supply for as near to 2.5V as possible.
 
  • Like
Reactions: 1 users
Member
Joined 2004
Paid Member
Running a reduced voltage across the filament (starved) is good in thoriated tungsten types (01A, 10Y ...) and so long as the sound, and the gain are still the same as with the nominal filament current, and voltage all will be fine. With these tubes, the lifetime is often improved, as well as the lower microphony.
I'm not sure, that TT filament lifetime more depends on filament voltage, than power on/off cycles.
The later occurs most of filament breakage, due to the Miller-Larson effect at 600 to 700C (both at heating -power on- or cooling -power off-).

See:
Valve Amplifiers, By: Morgan Jones
Power Vacuum Tubes: Handbook, Author: Jerry C. Whitaker

This type of filament usually has 1900-2400K temperature.
Emission at 2300K is -about- 25 times larger, than at 1900K.

There are no objection, if you use this tube as VAS stage, but as power stage it's problematic.
 
Last edited:
Yes, power ON cycles are a lifetime limit too. It depends on how often you power ON: if you do this ten times every day, then maybe tube will fail from this first.

But look at this diagram from EIMAC's App note AB-18, for power tubes.
It show that there is a BIG extension in lifetime if the filament is controlled to -5% of nominal voltage, for most of its life. It suggests a 2.5x increase in lifetime! This can't be ignored, for typical audio use!

You can see how destructive overheating is too, even at only Vf = +5% overvoltage: Lose 40% of the emissive life. Beware of mains tolerances, for AC or Raw rectified DC, applied to these expensive tubes.

Thoriated--Life-vs-Uf--EIMAC-aus-AB-18.png
 
  • Like
Reactions: 1 user
I am not about to read all 266 pages, so hopefully my understanding this has not been discussed before is correct?

Just before the website went offline, downloaded the schematic. However, didn't copy the text so using it as inspiration especially as the first two valves have problems. Some even seem to say the 26 can't be used as an output tube.
csb_riaa.gif


Here is my very rough draft. Have not been able to run it, the 6JC6 script from the forum has an error on line seven (inductance card missing).
Wilhelmina_Phono_Amp_SL7_Input.JPG


12SL7s paralleled for the Technics SL-10 uses a MC (design is for either, though figure be an improvement). Don't think this circuit includes RIAA equalization, as I understand the aforementioned TT already has it.

Have an idea or two regarding valve order to try once get the model to simulate.
 
Here you go.

* type26cun LTSpice model
.subckt type26cun P G K
CGK G K 3.5p
CPK P K 2.2p
CGP G P 8.1p
* koren 4
Bp P K I=(0.0169871083m)uramp(V(P,K)*ln(1.0+exp((12.39904937)+(12.39904937)(8.189852358)V(G,K)/V(P,K)))/(12.39904937))*(1.40553086)
  • koren 8
  • Bp P K I=(0.01505035691m)uramp(V(P,K)*ln(1.0+(0.04617378061)+exp((13.71786498)+(13.71786498)((8.563661196)+(9.663362929m)V(G,K))V(G,K)/sqrt((-1.856325497e-005)**2+(V(P,K)-(-5.468070728))**2)))/(13.71786498))(1.432783777)
.ends type26cun
Well, tried this model and getting very low gain, 2,9e^-05. I have read and search, not found an update.

Removing the type 26 gets a reasonable amount of gain, so thinking maybe it is the circuit. Help appreciated, certainly not expected. :D
 
Hi,

I conclude that if I wish to build a line preamp with type 46 valve, I do not want to build the circuit of message number 5304. That circuit is a power amp driver stage. With three different filament voltages indicated on the same page...? ? ? With a 1:4 input step up transformer. With an un-needed input grid stopper. And, what resistance should that screen stopper be? I think I would do much better starting from zero with plate characteristics for whatever valve 26, 46, 801A, 2P29L I might select. I think Andy's schematics need to be carefully relegated to their context, and carefully regarded or discarded.

So, I think I can go ahead with an open mind. Happily, I think I can do it myself.

Best, Robert
 
Thanks Demonkleaner. That's correct. I don't use a preamp. All my designs and builds are for the driver stage of a 2a3 SE amp which I've used for several years. So the output in my case is a 0.1uF FT-2 teflon cap and 470K resistor. But these circuits can all be used as preamps by increasing the output cap to e.g. 1uF and using a 1M or 2M resistor to ground after it.

Be aware that all my DHT stages are in filament bias, up to and including 10Y and 46 which get quite hot and need the cathode resistors mounted above the chassis. This gives me the best sound I've ever built with. I briefly used SIC diodes but they had a hard edge I couldn't eradicate. I only use DHTs in filament bias, nothing else. Rod Coleman V9 regs these days, previously V4-8. I use a 1:4 SUT in front to get the gain I require, in my case a Hammond 1140 LN-C or LN-D. Very happy with those.