• 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

Andy, I assume the Vitreous part was
perhaps a noninductive (wound in a Ayrton-Perry style)

I wonder how much an issue the winding type is in real life. The specifications of either standard or non-inductive are rarely given. I've measured a few by making a voltage divider with a wirewound and metal film in series. Then I connect the signal generator across the pair and the scope across the WW. Up to 1MHz, I can't see any difference measuring across the WW (regular white ceramic types) or MF. Above that I get some generator artifacts, independent of resistor type. I can't see them affecting frequency response. Maybe oscillation issues in some positions?

Sheldon
 
I wonder how much an issue the winding type is in real life. The specifications of either standard or non-inductive are rarely given. I've measured a few by making a voltage divider with a wirewound and metal film in series. Then I connect the signal generator across the pair and the scope across the WW. Up to 1MHz, I can't see any difference measuring across the WW (regular white ceramic types) or MF. Above that I get some generator artifacts, independent of resistor type. I can't see them affecting frequency response. Maybe oscillation issues in some positions?

Sheldon

The frequency response over audio is not likely to be a problem. But inductance is usually unwanted, especially at low-level audio. I'll have to degrade into hand-waving to say why; measuring the effects would be hard to do.

First, even a few uH will act like a search coil and increase the coupling effect for any nearby fields. Capacitor-input dc heater supplies might send peak currents of 5A into the first electrolytic in the supply. How long is the circumference of the loop of wiring between the trafo to rectifier to 1st Capacitor? The area of this loop is an electromagnetic moment (leverage), which expresses its ability to couple to other conductors. Power-tube ac heat, radio stations, B+ supply cap charging (E-field and B field) will all couple easier into big wirewounds.

The other effect, as you noted, is oscillation. ECC88/6DJ8 can amplify near to GHz! other valves have wide bandwidths too. If you've chosen a stopper (noninductive carbon composition, for sure!) correctly to limit bandwidth to 100kHz or less, and mounted it right on the pins, your circuit might be immune to VHF instability. But if instability happens, often only during a small part of the signal waveform, it could feed to the next stage and be detected like an AM receiver.

Whether the winding style of your wirewound makes a difference depends on design and layout of the rest of the circuit, I suspect. But I always look out for the old Ayrton-Perry wound resistors when I'm 'decommissioning' some old electronic junk!
 
How would you know an Ayrton-Perry winding if you met it on a dark night?

Andy

For new parts, only the data sheet knows - see pic!

for old style, there's no endcaps, just wire entry, with a stub poking out next to the entry-point. I'll get a pic when I'm out with the camera. Sad to say, nobody made these, since the 1980s, as far as I know.

See how they're wound:

Ayrton-Perry winding - Wikipedia, the free encyclopedia
 

Attachments

  • APwinding.jpeg
    APwinding.jpeg
    74.3 KB · Views: 661
I have one or two resistors that look like what you are describing. The wire comes straight out of the end, plus a short second piece of wire which is kind of doubled over to the case. Does that sound like it?

I found a few like this in my parts drawers - 10 ohms at 30W. I can touch it very quickly when it's conducting. Do you think 30W in free air is safe? I could parallel/series them for 60W but that means using 8 in total. Perfectly possible if 30W isn't enough, but maybe a bit over the top?

This resistor sounds great, incidentally. Best yet.

Andy

Andy
 
Hi Andy,

Did you ever get around to trying battery grid bias (with the cathode grounded)? I'm running my 26 pre in that configuration and it also sounded better than cathode bias.

I might give filament bias a shot, but wanted to clarify the calculation of the resistor.

On the AA thread Thomas indicated “The resistor is sized such that the voltage drop caused by the filament current provides the desired bias voltage.”
Resistance of filament = 1.5V/1.05A = 1.4 ohms
Bias voltage = -10V/1.05A = 9.5 ohms
How does the current regulated filament voltage come into this?

Rich
 
Last edited:
I haven't tried battery bias - don't know if I fancy a capacitor in the signal path. May be an option with a transformer.

Re calculation of cathode resistor, I think the point Thomas was making is that the current through the tube is tiny compared with the current through the filament. So in the case of the 26, we have 1 amp through the filament and 6 mA through the tube, so effectively you just use the filament current to calculate the resistor according to the bias voltage you want.

andy
 
I have one or two resistors that look like what you are describing. The wire comes straight out of the end, plus a short second piece of wire which is kind of doubled over to the case. Does that sound like it?

I found a few like this in my parts drawers - 10 ohms at 30W. I can touch it very quickly when it's conducting. Do you think 30W in free air is safe? I could parallel/series them for 60W but that means using 8 in total. Perfectly possible if 30W isn't enough, but maybe a bit over the top?

This resistor sounds great, incidentally. Best yet.

Andy

Andy

Andy, I'm not surprised that one sounds the best - it's the proper noninductive wound type, probably Zenith or Welwyn 1960s type. At rated power, they might hit 200 deg C, so watch nothing gets burned. May need high melting point solder.

happy listening

Rod
 
You certainly know your resistors, Rod! It is indeed Zenith. So do you think I can go ahead with 30w? I can barely touch it - have to pull my finger away fairly fast. But it is just about quickly touchable. Haven't got a thermometer to measure it - how do you in fact measure the temperature of resisters? Parallel/series would give me 60W which would certainly be better or even (?) necessary if I wanted to put it inside the chassis.

Andy
 
You certainly know your resistors, Rod! It is indeed Zenith. So do you think I can go ahead with 30w? I can barely touch it - have to pull my finger away fairly fast. But it is just about quickly touchable. Haven't got a thermometer to measure it - how do you in fact measure the temperature of resisters? Parallel/series would give me 60W which would certainly be better or even (?) necessary if I wanted to put it inside the chassis.

Andy

The IR thermometers from Maplin will measure resistor, and probably valve temperature.

If it says 30W on it (must be a very large one) then should be possible in free air. Two 5K/W heatsinks either side of it, gently clamped with four screws, ought to make a safe job, if you want it to last a long time.

30W from any source will be bad news under a chassis, need to vent it somehow, or have a heatsink sticking out the back of it!
 
Graphite! That guy is radical.

I'm just starting the fine tweaking process getting the voltages right for filament bias. I'm using a 150v glow tube into the 126C OPT, so that's a fixed part. So I want about 9.5v bias for about 6.5mA. Looks like with the 26 tubes I've been trying, this means a 11 ohm cathode resistor. All the tubes are slightly different. I'll report as I get more of a handle on this.

Andy
 
Hi Andy,

Did you ever get around to trying battery grid bias (with the cathode grounded)? I'm running my 26 pre in that configuration and it also sounded better than cathode bias.

I might give filament bias a shot, but wanted to clarify the calculation of the resistor.

On the AA thread Thomas indicated “The resistor is sized such that the voltage drop caused by the filament current provides the desired bias voltage.”
Resistance of filament = 1.5V/1.05A = 1.4 ohms
Bias voltage = -10V/1.05A = 9.5 ohms
How does the current regulated filament voltage come into this?

Rich

Same here Rich. My 26 pre also gird biased and I felt it is better sounding than the cathode baised which I used initially. Like to try the filmament bias but still trying to understnad how it works :)
 
Like to try the filmament bias but still trying to understnad how it works :)

It's easier to understand if you think of current loops, and ground (or common) reference points as independent issues.

There is a current loop for the filament. The current from the filament supply must return exactly to the filament supply.

There is a separate current loop for the current through the tube, which also must return exactly to its supply.

Let's say you are using cathode bias. In this case the current through the tube must go through the cathode resistor before it can return to the supply. And in cathode bias, the bottom of the cathode resistor is at ground (or common).

Let's take an example, of a 5V, 1amp filament, and 10mA plate current and a 1k cathode resistor.

Let's assume that the negative end of the filament is connected to the cathode resistor. With 10mA and a 1k cathode resistor, the potential at the top of the cathode resistor will be 10V. Now, the negative side of the filament will be at 10V.

If we want to use filament bias, we take the filament return voltage from the bottom of the cathode resistor, instead of from the top of the cathode resistor. Now, the filament is in series with the cathode resistor, so both the plate current and filament current must go through the cathode resistor. The total current through the cathode resistor is now 1.01R (1A+10mA). If we want the same 10V at the cathode resistor, we must calculate the new resistor size (V/I=R or R=10/1.01=9.9R).

So to get the same bias, we have reduced the cathode resistor by a factor of 100. The advantage is that we can eliminate the bypass resistor, since this low value resistor will cause very little cathode degeneration. The disadvantage is that any noise due to the filament supply ripple over the cathode resistor, will be amplified by the u of the tube.

Sheldon
 
Last edited: