I just found out about this very interesting thread. I am planning to build a #26 buffer like the Phaeton (VT52.com) and from my experience with choke loaded DHT buffers (300B/2A3) I consider hum as the biggest problem. I am using batteries for the filament but there is still some minor hum from B+. My question: did anybody already try John Broskies' approach with a signal transformer input in order to create a voltage divider that cancels out the hum?
"The plate and cathode resistor values must be selected carefully. The formula is an easy one:
Rk = (Ra - rp)/(mu + 1)
and
Ra = rp + (mu + 1)Rk
where Rk is the cathode resistor; Ra, the plate resistor; rp, the triode's plate resistance; and mu, the triode's amplification factor. If the plate and cathode resistor values follow the above ratios, then the triode and its cathode resistor effectively creates a two-resistor voltage divider with the plate resistor, whose division ratio is 50%. With this ratio, and only with this ratio, the anti-phase power-supply rail noise cancel at the output. "
http://www.tubecad.com/2015/08/blog0328.htm
"The plate and cathode resistor values must be selected carefully. The formula is an easy one:
Rk = (Ra - rp)/(mu + 1)
and
Ra = rp + (mu + 1)Rk
where Rk is the cathode resistor; Ra, the plate resistor; rp, the triode's plate resistance; and mu, the triode's amplification factor. If the plate and cathode resistor values follow the above ratios, then the triode and its cathode resistor effectively creates a two-resistor voltage divider with the plate resistor, whose division ratio is 50%. With this ratio, and only with this ratio, the anti-phase power-supply rail noise cancel at the output. "
http://www.tubecad.com/2015/08/blog0328.htm
I think you're exaggerating the problem of hum and noise. Have a look at the Bartola Valves Gen 2 01A preamp. I've built two already and they're silent. DC filament bias. Ale's Gyrator is a real success - works for the 26 as well, but I'd give the edge in sound to the 01A. I've built countless 26, 4P1L, 10Y etc. preamps and the Gen 2 is in my system now.
I am talking about using a single DHT as a buffer with gain like the Pheaton LC for example. I agree with what you say about preamps but this is something else. In my experience a simple circuit with a choke loaded DHT produces the most natural sound but it comes with a lot of hum. The reason for this is also explained in the tubecad blog.
Now here's a thing. I built my 26, by copying the layout of Rui Lourenco's pre-amp, and with a lot of advice. Everything measured OK, 156V HT, 875mA on the filaments, and it sounded clear. But I had a problem with sibilance, no matter what I tried with my speakers, various power amps and cables. When I took the 26 up to the workshop to check voltages with different 26s, I put a TVC in its place. A softer sound, sure, but no sibilance!
I have just bought a 'scope (which I don't really know how to use very well, but can measure waves), and using a cheap Chinese sig gen (sine, triangle and square), I put various levels of waves through the 26, particularly square. The output of the sig gen looks clean enough, but the output from the 26 is horrendous!
10kHz from sig gen.
Same after passing through the 26. 2V output, volume pot wide open.
From 100Hz to 20kHz is similar. I tried loading the output with 33k ohms (same as my power amp); no difference. Admittedly the leads I used are a mixture of BNC to crocs, or RCA, but surely that wouldn't do this?
How do I start hunting down whatever I have done wrong? Using a scope with a sig gen is new to me, so any advice would be welcomed.
This is the circuit, except the limiting resistors in the raw filament supply to the Rod Coleman regs are before the choke, not after, and there is a 2.4k not a 1k resistor in the HT to get 156V HT. I've tried 2 different brands of 26, and both are the same.
I have just bought a 'scope (which I don't really know how to use very well, but can measure waves), and using a cheap Chinese sig gen (sine, triangle and square), I put various levels of waves through the 26, particularly square. The output of the sig gen looks clean enough, but the output from the 26 is horrendous!
10kHz from sig gen.
Same after passing through the 26. 2V output, volume pot wide open.
From 100Hz to 20kHz is similar. I tried loading the output with 33k ohms (same as my power amp); no difference. Admittedly the leads I used are a mixture of BNC to crocs, or RCA, but surely that wouldn't do this?
How do I start hunting down whatever I have done wrong? Using a scope with a sig gen is new to me, so any advice would be welcomed.
This is the circuit, except the limiting resistors in the raw filament supply to the Rod Coleman regs are before the choke, not after, and there is a 2.4k not a 1k resistor in the HT to get 156V HT. I've tried 2 different brands of 26, and both are the same.
Could you give me some idea of how a "buffer with gain" differs from a line stage with low impedance output?
I guess in the end it's more or less a matter of how to name something rather than a difference in its function. What's normally called a preamp (active) has a volume control and in most cases a dedicated input and output stage combined in one chassis. A buffer - in most cases unity gain- simply buffers the signal from the input to the output (impedance) and would be ideally located very close to the source of the signal, like a DAC for example. I can use a DHT without input stage and get gain while I buffer the signal from input to output (impedance) and the volume control would be digital in the case of the DAC. So I could call it a DAC-preamp, or a DAC with a buffer with gain (=amplifier) with a digital volume control, or a line stage for a DAC. It all works. My original point, however, was rather that I was talking very specifically about hum in a very simple circuit like the Pheaton LC which has a choke loaded anode and not about hum in any DHT line stage or preamp in general with a gyrator for example. And in this context (choke loaded anode/hum) I wanted to know if anybody had ever tried the hum cancelling method as explained in the tubecad blog. I know that it's possible to get almost no hum with a gyrator. But in my experience it sounds different than a choke loaded anode which I prefer.
Now here's a thing. I built my 26, by copying the layout of Rui Lourenco's pre-amp, and with a lot of advice. Everything measured OK, 156V HT, 875mA on the filaments, and it sounded clear. But I had a problem with sibilance, no matter what I tried with my speakers, various power amps and cables. When I took the 26 up to the workshop to check voltages with different 26s, I put a TVC in its place. A softer sound, sure, but no sibilance!
I have just bought a 'scope (which I don't really know how to use very well, but can measure waves), and using a cheap Chinese sig gen (sine, triangle and square), I put various levels of waves through the 26, particularly square. The output of the sig gen looks clean enough, but the output from the 26 is horrendous!
10kHz from sig gen.
An externally hosted image should be here but it was not working when we last tested it.
Same after passing through the 26. 2V output, volume pot wide open.
An externally hosted image should be here but it was not working when we last tested it.
From 100Hz to 20kHz is similar. I tried loading the output with 33k ohms (same as my power amp); no difference. Admittedly the leads I used are a mixture of BNC to crocs, or RCA, but surely that wouldn't do this?
How do I start hunting down whatever I have done wrong? Using a scope with a sig gen is new to me, so any advice would be welcomed.
This is the circuit, except the limiting resistors in the raw filament supply to the Rod Coleman regs are before the choke, not after, and there is a 2.4k not a 1k resistor in the HT to get 156V HT. I've tried 2 different brands of 26, and both are the same.
Strange looking waves you got there Richard. Sorry for not having a clue of what's going on. Keep us posted on your progress. It should not be possible for the square waves to look similar from 100Hz to 20 kHz....
I may have stated that badly. I meant they were all "dirty". Not experienced with this, unfortunately. I expect someone with experience could tell quite a lot from these traces, but I can't!
This is 100Hz.
1kHz.
20kHz.
Could you check frequency bandwidth with the sinus generator?
The 100 Hz square wave shows that there is quite some low frequency loss.
The 20 kHz square wave shows resonance; check 20 kHz and higher (up to 100 Khz to see where these resonances are, and how severe).
By the way: don't expect great results anyway with a high Rp tube (Rp of 26 is some 8 k) in combination with a step down transformer.
The 100 Hz square wave shows that there is quite some low frequency loss.
The 20 kHz square wave shows resonance; check 20 kHz and higher (up to 100 Khz to see where these resonances are, and how severe).
By the way: don't expect great results anyway with a high Rp tube (Rp of 26 is some 8 k) in combination with a step down transformer.
Looks like transformer resonance peaks.
Looking at the LL2745 specification
http://www.lundahl.se/wp-content/uploads/datasheets/2745.pdf
It expects to be driven by a 3500Ω anode, the 26 is somewhat more than that.
You may be able to tame the (frequency) peaks with a network on the secondary, e.g. a R-C series (zobel), or by an even heavier load.
I would send your scopeshot and circuit to Lundahl, and ask. This should be the way to get the most accurate advice.
Looking at the LL2745 specification
http://www.lundahl.se/wp-content/uploads/datasheets/2745.pdf
It expects to be driven by a 3500Ω anode, the 26 is somewhat more than that.
You may be able to tame the (frequency) peaks with a network on the secondary, e.g. a R-C series (zobel), or by an even heavier load.
I would send your scopeshot and circuit to Lundahl, and ask. This should be the way to get the most accurate advice.
Thanks for all the replies. I will try all those suggestions; I only looked at this last night and haven't gone any further without some advice.
I bought the Lundahls from Thomas Mayer; they should be the 8mA version, I believe. I can only assume they are correct. They are connected for minimum output impedance, Alt R, as I don't need the gain.
vt4c; would that be in series, between the plate and the transformer?
I bought the Lundahls from Thomas Mayer; they should be the 8mA version, I believe. I can only assume they are correct. They are connected for minimum output impedance, Alt R, as I don't need the gain.
vt4c; would that be in series, between the plate and the transformer?
![WP_20161205_001[1].jpg](/community/data/attachments/541/541101-9e19ed467d9ffecee657d4685b36e2db.jpg)
![WP_20161205_002[1].jpg](/community/data/attachments/541/541119-3a893db8477e2bc9147fe93f7afa19c3.jpg)