• 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.

current servo not voltage.

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

I have a hybrid who use a servo for offset nulling, it works trough the cathode follower input, it do work fine, but.

I did read something about a current servo who has less impact on sound eventually, I did this tryout by using a fet between the current source tube cathode and gate, see schematic, if i do simulate it it works fine, but in the real world not, I did see that I eventually has to change tho opamp inverting input, because on the negative side of the amp there is no inverting or how I do call that in english. ps I did try to invert the opamp input, no result.

maybe some has a idea, voltage servo,s do have impact, and so I want to try what current does.

thanks.
 

Attachments

  • servo-current test.jpg
    servo-current test.jpg
    54.7 KB · Views: 179
Last edited:
Hi Kees,

It looks to me like you either need some kind of level shifter between your op-amp and your PMOS, or you need to replace your PMOS source follower with something that can work with a control voltage low enough for the op-amp.

Your present circuit as an op-amp of which the output can't swing below -16 V. It drives a PMOS source follower whose source must end up a few volts above -120 V. That means its gate has to swing close to -120 V, while it can't go below -16 V.

If you are satisfied with a very limited control range, maybe you can put a resistive voltage divider between the op-amp and the source follower: a big resistor from op-amp to PMOS gate and a much smaller resistor from PMOS gate to -120 V.

Regards,
Marcel
 
Hi Kees,

It looks to me like you either need some kind of level shifter between your op-amp and your PMOS, or you need to replace your PMOS source follower with something that can work with a control voltage low enough for the op-amp.

Your present circuit as an op-amp of which the output can't swing below -16 V. It drives a PMOS source follower whose source must end up a few volts above -120 V. That means its gate has to swing close to -120 V, while it can't go below -16 V.

If you are satisfied with a very limited control range, maybe you can put a resistive voltage divider between the op-amp and the source follower: a big resistor from op-amp to PMOS gate and a much smaller resistor from PMOS gate to -120 V.

Regards,
Marcel

Hi Marcel

And if I try a drain follower there? or even a bipolair transistor, a level shifter feeding with the -120 volts can work best?

I don,t now if I am right with the impact, but I think current way of nulling offset is better.

Be the way, do you agree that windows 7 get,s not better after updates, I have crashes the last time, 3 times i did lost a schematic (blue screen) and reset of bios, but maybe you like me not so pc minded.

The schematic in this post is the version I did use for some years now. R31 is not used in present scheme.

thanks.
 

Attachments

  • working servo.jpg
    working servo.jpg
    76.7 KB · Views: 165
Last edited:
some food for thought as they say.

If I do that, then cathode resistor of lower tube 100 ohm has 47 volts that is much, but I think the simulation is not oke.

use a zener work, but there is 47 volts on cathode for 21 mA current, maybe a voltage driven current source is a good way to go,
maybe a hybrid current source who can be voltage driven.

amp is dc coupled and current need 20 mA for driving mosfets, you have a scheme who has much lower current, but I presume
that it is as a example.
 
Last edited:
I have test now with a P mosfet, it,s measurement give now 67 uV on schematic.

A hybdride current source will maybe the way to go, it is now a kind of the same
but to simpel, anyway it do work in multisim, but not in real.
 

Attachments

  • second servo test.jpg
    second servo test.jpg
    108.9 KB · Views: 111
Hi All again.

I have done it this way (see picture) a mosfet to eart and let it snoop some current from the tube ccs cathode, but maybe a danger, if the mosfet fails the resistors will burn, but this afcourse is only if mosfet turn fully on, that is why a resistor of 1 M is from gate to earth, what do you think, it has a offset with it at 8 uV on simulation, but like first in real time what will it do.

thanks
 

Attachments

  • danger servo-but works.jpg
    danger servo-but works.jpg
    152.8 KB · Views: 43
Hello Kees,

I see two inversions in your loop (op-amp and PMOS Q5), so doesn't this circuit latch up?

If you replace the integrator with a non-inverting one, then it should work, or at least then I don't see a reason why it shouldn't anymore. You can then protect the op-amp against excessive input voltages by using a resistive voltage divider between the amplifier output and the plus input of the op-amp.

By the way, you could add a resistor in the drain lead of Q5 to limit the current if it should somehow get into the fault condition you described.

Best regards,
Marcel
 
Last edited:
Hello Kees,

I see two inversions in your loop (op-amp and PMOS Q5), so doesn't this circuit latch up?

If you replace the integrator with a non-inverting one, then it should work, or at least then I don't see a reason why it shouldn't anymore. You can then protect the op-amp against excessive input voltages by using a resistive voltage divider between the amplifier output and the plus input of the op-amp.

By the way, you could add a resistor in the drain lead of Q5 to limit the current if it should somehow get into the fault condition you described.

Best regards,
Marcel

Hi Marcel

I have only simulate it there it did work but maybe if the offset get positive it don,t.

I have see it, I will make the opamp non inverting it is easy then simulate again, a n fet is not the option.

thanks
 
Hi Marcel

I have only simulate it there it did work but maybe if the offset get positive it don,t.

I have see it, I will make the opamp non inverting it is easy then simulate again, a n fet is not the option.

thanks

Most simulators don't care whether a DC bias point is stable or not when simulating a DC operating point. When you simulate a schematic of a digital latch, it will typically end up in the metastable state (logic value 1/2), even though a real-life latch always leaves the metastable state within a few nanoseconds. You normally don't have that kind of problem in a transient simulation with an input signal.
 
Most simulators don't care whether a DC bias point is stable or not when simulating a DC operating point. When you simulate a schematic of a digital latch, it will typically end up in the metastable state (logic value 1/2), even though a real-life latch always leaves the metastable state within a few nanoseconds. You normally don't have that kind of problem in a transient simulation with an input signal.

Wel it is clear your explanation, I have to try real time, if the mosfets arrive I start. just try to make the opamp again non inverting, then the offset do keep walking upwards, if i do make it inverting again I get a stable offset in the uV region (8) and keeps there.

You are dutch, so, maybe you now a good electronic component supplyer here in the netherlands.

thanks.
 
Last edited:
Wel it is clear your explanation, I have to try real time, if the mosfets arrive I start.

You are dutch, so, maybe you now a good electronic component supplyer here in the netherlands.

thanks.

There used to be a very good one in the Hague, Stuut en Bruin, but they stopped some time ago. Radio Twenthe in the Hague is very good in historic components (valves, valve holders, germanium transistors et cetera). Farnell supplies a lot of modern stuff, but it could be that they only sell to companies (I'm not sure about that). And of course there is Eurosound in Haarlem and DIL in Rotterdam.
 
Hi Marcel

I now Dil not eurosound, I have order some time ago in china, but it is a gamble sometimes, and get parts is more difficult the last years, I think because a lot of people being stopped with electronics.

A small update about the servo, you are right, it has to be not inverting, I have put a other part in simulation, who did not work, component database of multisim is not so good, the part I had in schematic was already non inverting, but ic house was wrong, now I have put a nieuwer part in it, the behavior of the simulation is as expected now,
so it will proberly work.

thanks, en nog een prettige week.
 
Wel I have stop the attemp to use current servo, I stay with my own first way of doing it, only I have use a dual opamp with low pass extra, I did found it on the internet and has a very low point of working (1Hz) this do work also in amp, and it seems audio get much less stressed.

Maybe I can better chamge the zeners for two antiparalell diodes?

thanks.
 

Attachments

  • low passed servo.jpg
    low passed servo.jpg
    132.9 KB · Views: 35
  • offset-amplifier.jpg
    offset-amplifier.jpg
    67.8 KB · Views: 37
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.