• 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.

A newbie question about 12AX7 tube

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hi folks,

I have here a question, because I'm newbie with tubes need some advice. The question is how to improve the circuit on the 12AX7 schematic. The idea is to increase the plate voltage from 48V to 140V using an external Cockcroft-Walton voltage multiplier, 140V to 250V is the recommended plate voltage on 12AX7 for operation.
Mic2200.PNG


The big question is if the voltage divider on the grid with resistors of 47K that gives approx +24 grid voltage will affect the operation of the tube as cathode follower (a strange design that?). Should this circuit part be changed? Also I will replace the 1uf decoupling caps with better ones, using 250v polyester caps. The cathode resistance at the output of 100K is maybe to high, I want to lower it to 18K using a 22K resistor in parallel, this change increase the output current lowering the reflected impedance to op amp IC18B.

Any suggestions to improve the tube circuitry are welcomed.
 
Administrator
Joined 2004
Paid Member
This is a cathode follower, boosting the B+ slightly will probably help linearity, other changes you propose may not make the improvement you seek. The +24V on the grid sets the operating current of the follower, under normal conditions this would result in about 250uA of cathode current, because of the low plate voltage it could be and likely is substantially less.

This incidentally seems to be a "token" tube intended to add some "tube colorations" by running a tube under sub optimum conditions. Fixing that would be relatively close to removing it altogether, it's being used as a tone shaping element.

The 4580 is a fairly mediocre op-amp, but that is another can of worms.

What is the end application? If guitar this thread should move to the I&A forum.
 
The application is for a microphone stereo preamplifier. Is a rack unit from Behringer the MIC2200 for high quality recordings. What is strange here is not only the lower 48 plate voltage same as the phantom power maybe to reduce costs, also the 100K cathode resistor. I think that lowering the value to 18K will reflect a better matching impedance and more tube coloration, but all that is experimental. The big question is the voltage of +24V at the grid which seems to be high.
 
I will increase the plate voltage to +140 volts lowering at the same time the cathode resistor to 18K. Better polyester decoupling caps will be not significant in sound quality but is maybe a must do change on a preamp tube stage. The maximal plate current on 12AX7 is 6ma at 250V plate, with 140V plate, 24V grid and 16K ohms reflected cathode resistance (100K // 150K // 22K) the tube max current is 1,5ma... should work fine with that values on 12AX7.
 
Last edited:
I'll wager my lunch money that is a piece of a Behringer guitar amp schematic.

What do you plan to accomplish here? The cathode follower tube stage has no gain, so increasing the voltage won't increase any gain. As to the grid, with 48v, the grid is set at half that, 24v. And that will have the net effect of setting the cathode voltage at about 24v as well, the point is to set the idle voltage at the cathode to half the supply voltage. That leaves the most headroom, 24v either way.

Considering the lack of gain, and the fact the stage is driven by an op amp, which cannot put out beyond 15v at its output, your grid and thus cathode will have a maximum 30v p-p of signal. (Ignoring details like the op amp not quite reaching rails etc.) That 30v p-p won't change, so if you triple the tube supply, the cathode follower will still only pass through that same 30v p-p. So what is there to gain? ...pun intended.

Your pair of 47k resistors at 48v draw maybe half a milliamp. At 140v, a bit more, and they dissipate about 200mw, 100 apiece, so that is probably OK if they are 1/4w axials. If they are surface mounts, you may want to change the values.

Nit pick: those are coupling caps, not decoupling caps.


Are you trying to improve specific performance shortcomings you hear in performance? Or is this an exercise on paper?
 
My question is if 48V plate on 12AX7 is to low considering that on most preamp applications the plate voltage is between 150 and 250 volts. We are taking about preamplifiers stages, with very low output signals (below 10Vp-p max) despite that on most design the plate is at +150V or more. I know that the performance will be the same with 48V or 140V plate. The good point of a cathode follower is the linearity-stability after tube aging and the much easier circuit design avoiding oscillations and other problems of the plate follower configuration. Also for tube high drive coloration a cathode follower is not appropriate at all. But on most preamp designs a plate voltage of 100 to 250 volts is the standard configuration and the reasons about my idea to change that part. Increasing the +voltage on the grid input divider from +48V to +140V to have +70V on grid can lead to damages on the op amp and will put the 12AX7 at maximal current specs lowering its durability, is much better to let that part at +48V (+24 volts grid).
 
Last edited:
You also have to choose the plate current for the Vpk voltage you want.
For Vpk # 150V, a good value for Ip is 1 mA and the Vgk voltage for that point is -1,15V

The static voltage on the cathode will be Rk*0,001, so 18 V with Rk = 18k, as you suggest.

The B+ voltage shall be 150+18V=168V

The grid voltage shall be 1.15V under cathode voltage : 18-1.15 ~17 V. So you have to define a voltage divider from the B+ voltage (168V) providing 17V on the grid.
 
My question is if 48V plate on 12AX7 is too low considering that on most preamp applications the plate voltage is between 150 and 250 volts.
It is low compared to all-tube circuits, but in this case the tube is being used to generate 'pleasant' distortion. If you increase the voltage you will reduce the distortion, which defeats the object of having a tube in the first place.

Increasing the +voltage on the grid input divider from +48V to +140V to have +70V on grid can lead to damages on the op amp
Yes a higher voltage may damage the opamps, espceially at switch-on when the capacitors charge up.
 
.... The cathode resistance at the output of 100K is maybe to high, I want to lower it to 18K using a 22K resistor in parallel, this change increase the output current lowering the reflected impedance to op amp IC18B.

The reflected impedance to op amp is not the 100 k resistor only but the 100 k in // with the output tube internal resistor, which is Ri/µ, so for the 12AX7 in the current situation around 100k/100 = 1k.

So output impedance is 1k//100k # 1k.

Using Rk=18 k instead of 100k will only decrease the output impedance to 1k//18k# 947 ohm.
 
Furthermore, after switch-on, the plate current is very low till the cathode temperature reaches its operating point.

So after switch-on the voltage at the cathode is nearly 0 during the first 30 secondes , but the voltage on the grid is already at 21V . This will reduce the lifetime of the tube.

With B+ = 140 or 150V it will be worse.

So you need a device to postpone the B+ after the heaters.

Or you need to introduce a large time constant (2 or 3 minutes) on the voltage applied to the grid. This will also solve the risk to damage to opamp as said by Merlinb.
 
So after switch-on the voltage at the cathode is nearly 0 during the first 30 secondes , but the voltage on the grid is already at 21V . This will reduce the lifetime of the tube.
21V is harmless. You can have up to 100V between grid and cathode (cold) before damaging the tube.

So you need a device to postpone the B+ after the heaters.
A diode between grid and cathode would do the job -no delay needed.

Or you need to introduce a large time constant (2 or 3 minutes) on the voltage applied to the grid. This will also solve the risk to damage to opamp as said by Merlinb.
This would protect the opamp from the start-up conditions, but not from high-voltage signals, noise clicks etc. You would need clamping diodes at the input of the opamp for reliable protection.
 
21V is harmless. You can have up to 100V between grid and cathode (cold) before damaging the tube.

Sorry, I am affraid not to share your point of view.

The voltage between grid and cathode is POSITIVE at start-up, when not plate current, so the grid will take high current from the B+ through the voltage divider, till plate curent increases. Not very good for the lifetime of the tube...

Am I wrong ?
 
I found this idea:

FIG-4A-17.gif


Factory design simplified is circuit a, I do some circuit changes to adapt it to circuit b. Circuit b increase the input resistance taking all the advantages of a tube, also more plate voltage increase linearity. Unfortunately I was not able to include the Rk1 resistance on cathode (approx 10K ohm), but because is a cathode follower output voltage is less (approx 0.98 gain factor) avoiding any types of feedback oscillations. Rk2 remains at 100K ohm and Rk3 at 560K ohm. The B circuit is know as modified cathode follower, the less voltage on cathode bias the grid positive, increasing the grid input impedance to a factor of megohms.

I tried to find a way to include the Rk1 (10K) resistor but was impossible, anyway without that resistance should work fine.

In the case the circuit is unstable is very easy to rollback to factory design. Right now I´m finishing the installation of the plate voltage multiplier and see whats happens.
 
Last edited:
Disabled Account
Joined 2013
Hi, what do you use the preamp for? Found the bottom end has rolled off , if you need good bottom end response you need change the grid bias resistor or the increase the coupling cap. It's possible to adapt the included sch sim for other model.
 

Attachments

  • hybrid 12ax7 4580 1.png
    hybrid 12ax7 4580 1.png
    265.3 KB · Views: 146
  • hybrid 12ax7 4580 2.png
    hybrid 12ax7 4580 2.png
    274 KB · Views: 166
  • hybrid 12ax7 4580-b.asc
    4.9 KB · Views: 36
I have good news, seems to work pretty well the circuit improvement. The added tube warmth in noticeable thanks to the higher plate voltage, now its running at 140 volts plate versus 48 volts factory (use the same +48V mic phantom power supply). At the same time the grid input impedance is at megohms range, another improvement. The removed RK1 don't affect the tube stability.

I will put some audio sample on Youtube. Thanks for all the tips and the circuit simulation.

Regards.
 
Furthermore, after switch-on, the plate current is very low till the cathode temperature reaches its operating point.

So after switch-on the voltage at the cathode is nearly 0 during the first 30 secondes , but the voltage on the grid is already at 21V . This will reduce the lifetime of the tube.

With B+ = 140 or 150V it will be worse.

So you need a device to postpone the B+ after the heaters.

Or you need to introduce a large time constant (2 or 3 minutes) on the voltage applied to the grid. This will also solve the risk to damage to opamp as said by Merlinb.

Here again the original schematic and a question:

Mic2200.PNG


On my last circuit following the (b) figure the added coupling cap that I use (C1) is 47nf, a very low cap value that is possible due the high grid impedance in the range of Megohms on such configuration. R1 is set at 560 K ohms and RK2 remains same as factory (R160) at 100K ohms. Audio quality is very good and much better. No noticeable distortion with approx 20 sec of tube warmth up.

Will this new circuit suffice the requirements for a safety tube warmth up without any damages on the grid?

FIG-4A-17.gif


From technical point of view is much better as factory design, on which the grid is forced at +24V (48 phantom power / 2). Remember that I have removed RK1 because was impossible to add that part due the small SMD devices and the possibility of PCB damages on its cooper traces. The 1uf in/out coupling caps are the same electrolytic from factory and fortunately after 10 years remains in tolerance (1,1 uf measured). Also I don´t removed the previous grid +48 (R160 / R130) voltage divider because the SMD parts are to small for a safety unsoldering.

To build a plate B+ turn on delay of 90 seconds is not difficult at all, maybe is the best solution for a long life tube run... is this a must do ??? my big question...
 
Last edited:
The voltage between grid and cathode is POSITIVE at start-up, when not plate current, so the grid will take high current from the B+ through the voltage divider, till plate curent increases. Not very good for the lifetime of the tube...

Am I wrong ?

Unless I am missing something, when you have that 21v positive on the grid, the cathode is cold and not conducting, so what is the current path from the grid? The cathode only conducts once it is hot, and certainly there would be no current from grid to plate. So any grid current would be very brief pulse after the cathode comes up towards operating temp.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.