# A switching boost converter biasing a tube preamp - saturation issue

I am designing a 9V-battery powered tube preamp (for electric guitar), based on a Russian rod tube (1j24b). The latter is a pentode, and works with about 45V on its screen grid and at least 30V on its anode. I have used a LT1172 switcher to implement a switching boost converter, in order to convert the 9V from the battery to about 60V, 1mA for biasing the tube.

This is the boost converter:

simulate this circuit – Schematic created using CircuitLab

The voltage from the boost converter goes to the tube, which is biased this way:

simulate this circuit

(sorry for the bad symbol for the tube - I did not find anything else).

Here, a: anode f+: filament, positive f-: filament, negative g1: control grid g2: screen grid g3: suppressor grid SCN: internal screen

Actual voltage readings from a digital multimeter are shown. The anodic current is about 1mA. No current is absorbed by grids and screen; the 1.2V filamentary power supply is obtained from the 9V battery through a voltage divider; the filament current is about 13mA.

The issue is that the output signal is saturated in the positive semiwave. A screenshot from an oscilloscope reading, with 500mV p-p input, 500 Hz, and 2V/div:

If the resistor on the anode is reduced, the gain reduces, but saturation always occurs at the same output signal amplitude (higher input signal). This suggests that the problem is not with the control grid, but with the anodic circuit.

Also, since saturation is in the positive semiwave, it happens when the input signal is in the negative semiwave and the anodic current is lower than its bias.

At the onset, saturation happens also in the negative semiwave, then slowly that part of the signal smooths out and the positive semiwave saturates further, up to the reported waveform.

Lastly, I note that the p-p output amplitude is roughly 8 V, which is the actual reading of the battery voltage when I powered up the circuit and took the screenshot. This seems to suggest that the amplitude of a varying signal at the output (thus, powered by the boost converter) cannot exceed the supply voltage to the boost converter. It seems to me that the problem is with the boost converter, but I do not understand it, nor have found any mention of such an issue in the design manual - so I do not know how to overcome it.

Any idea?

• Is the scope dc coupled to the anode? Commented Jan 28, 2017 at 11:49
• @Andyaka: No, it is coupled to the point where I put the label "V_out". Commented Jan 28, 2017 at 12:02
• This little tube was used in Sputnik’s emitter circuit ;)
– greg
Commented Jan 28, 2017 at 16:06
• @greg Yeah, I knew that it was military intended. I didn't know about the Sputnik. What was that emitter circuit? Commented Jan 29, 2017 at 0:04

The screenshot is so cleanly cut-off that we really can rule out biasing problems with your tube – things would look a lot "smoother"/rounded off before saturation if you just hit some magical voltage barrier in your amp.

In other words, you're observing a measurement artifact. Probably your scope clips at some point. Try to observe the voltage across a voltage divider instead.

• I didn't think of this - yes, it may be. It is not a professional scope. I will try and let you know. Commented Jan 28, 2017 at 12:04
• Ok, it was the scope. With the voltage divider it's perfect. Commented Jan 28, 2017 at 12:22
• @Enrico how come it didn't clip negatively i.e. symmetrically if it was AC coupled? Commented Jan 28, 2017 at 12:27
• @Andyaka First I quote myself: At the onset, saturation happens also in the negative semiwave, then slowly that part of the signal smooths out and the positive semiwave saturates further, up to the reported waveform. I mean: when I increase the input signal level up to saturation, the output signal clips both positively and negatively; but it doesn't settle there: it "moves up" smoothing in the negative part and clipping further in the positive one. I cannot imagine whatever it is that causes this behaviour; I only note it is a capacitor-like behaviour. Commented Jan 29, 2017 at 0:00

Since negative (input) peaks are turning it off, you need to turn it on a little harder.

Add a little positive grid bias - 11 Meg from grid to heater V+ (1.2V) would give +0.1V as a starting point, then season to taste.

Grid bias is generally negative, but when you're using a tube at such low voltages and (relatively) high anode currents, a small positive bias is not unusual. 1mA may not sound high, but in comparison with 13mA heater current it really is.

Also note that the mean potential on the cathode is probably +0.6V (assuming this is a directly heated tube) so the grid is currently biased negative.

• Already tried - no improvement, just lower gain, but with saturation at the same output level. Commented Jan 28, 2017 at 12:05

Your circuit has little to no bias. This tube can take just a little positive grid bias before clipping as you've shown. You need to supply the heater with a higher voltage and insert a resistor between f- and ground to maintain a max of 1.2v. This will bias for more headroom. If you supply f+ with 1.5v and insert 22r between f- and ground, your heater will see 1.2v and you'll be able to apply more than a volt input to g1 with no clipping. As it stands - you're running the tube too hot so when you bias as explained above you'll need to adjust/increase your plate resistor to around 60k. It's a really nice sounding tube btw, and super efficient. Add a buffer to the output and it makes a nice headphone amp.