Extremely Low Current Buffer options

[Bill Mountain]

I want to replace the preamp in my active bass with a simple discrete buffer.

The chip used in the preamp is the LM4250 which can be programmed to operate around 10uA.

I'm not positive but I think the low current consumption could be why the preamp is always distorting.  I pulled the LM4250 and the current programming resistor and tried a few different opamps which all yielded much higher headroom but also heavily reduced battery life due to much higher current consumption.

The preamp was bad news in other ways (jack was wire backwards making it hum, there was way too much noise, pots were scratchy and ineffectual, etc.).

I simply want to replace the preamp with a buffer or low gain booster (this preamp doesn't boost much anyway and the EQ isn't active) but I don't know how to search for low current discrete preamp or buffer designs.  I know that you can manipulate the biasing resistors to get lower current but I'm not sure what transistors would be best.  I'm also not sure if lower current designs will sound as good without the proper transistor.

I had considered the Tillman preamp but with my batch of J201's to get similar current consumption and to bias it properly I have to use a much larger drain resistor which increases gain too much leading to unwanted distortion.

I'm not against a good BJT design either.

Just looking for ideas.

Thanks!

[bool]

A tl061 or a tle2061 should be good candidates for an onboard pre. Have you tried those? The tle chip beig a "robo" version of the 061.

[Bill Mountain]

A tl061 or a tle2061 should be good candidates for an onboard pre. Have you tried those? The tle chip beig a "robo" version of the 061.

Those are nice chips but I don't currently have any.  I also fear that being low current chips they will behave like the 4250 that I have.  I am most likely wrong in my assumptions.

[Johan]

R.G had a design about a year ago for a low current draw buffer

[R.G.]

Bill, see http://geofex.com/FX_images/Onboard_Preamp.pdf

It's designed to solve a similar but not quite identical problem, so it may not be perfect, but it's close. In the design I looked at a number of possible low-current opamps, but none offered currents as low as the discrete design with similar good performance.

As you've found, the crux of the problem with onboard electronics is the need for power and how long it lasts. I chose lithium coin cells for the design because they are both smaller than conventional 9V batteries and have a higher energy density - they last longer. In the case of a bass, the signals are bigger than in a guitar, so there may be issues with simply not having enough power supply voltage at 6V. However, for a similar *volume* of battery space used in your bass, you might be able to use four lithium coins and get 12V for this (or any!) circuit and increase your headroom so you don't hit clipping so easily. I can advise you about what to change to optimize the circuit for 9V or 12V if you want to experiment.

And it will be an experiment.

[Bill Mountain]

That looks cool but seems complicated.  Is there I reason I would need to use four transistors?

A single transistor could get even less current draw right?

[R.G.]

That looks cool but seems complicated.  Is there I reason I would need to use four transistors?

A single transistor could get even less current draw right?

No, a single transistor would not necessarily draw a lot less current. Three of the transistors use the SAME current, mostly. The top two transistors are set up in a feedback pair where all the gain is generated. The PNP uses most of the current, and the NPN a bit more than 1/hfe of the PNP, just the base current. The NPN immediately below the PNP is set up as a current source. It acts like a much -MUCH!- bigger resistor for better gain, without eating either the voltage or the current that a real resistor would, and not generating as much thermal noise in most cases.

The only real extra current is the current into the collector of the lower-left NPN. This provides the base current to the current-source transistor and the lower-left NPN shunts this away from the lower right NPN base to regulate. This current is about (6V-1.4V)/470K, or about 10uA. This forces the current in the 5.1K resistor to be about 100uA, so the whole circuit pulls about 110uA.

[Bill Mountain]

That looks cool but seems complicated.  Is there I reason I would need to use four transistors?

A single transistor could get even less current draw right?

No, a single transistor would not necessarily draw a lot less current. Three of the transistors use the SAME current, mostly. The top two transistors are set up in a feedback pair where all the gain is generated. The PNP uses most of the current, and the NPN a bit more than 1/hfe of the PNP, just the base current. The NPN immediately below the PNP is set up as a current source. It acts like a much -MUCH!- bigger resistor for better gain, without eating either the voltage or the current that a real resistor would, and not generating as much thermal noise in most cases.

The only real extra current is the current into the collector of the lower-left NPN. This provides the base current to the current-source transistor and the lower-left NPN shunts this away from the lower right NPN base to regulate. This current is about (6V-1.4V)/470K, or about 10uA. This forces the current in the 5.1K resistor to be about 100uA, so the whole circuit pulls about 110uA.

Thanks for the explanation.  That is definitely the type of current drain I'm looking for.  I'll mess around with it on the breadboard when I get a chance.

[R.G.]

I neglected to mention - the NPN/PNP is a "super emitter follower", with hfe times higher current gain, in a tight feedback loop for a gain that's very, very near unity. The current source load would improve any single-transistor emitter follower, but it really makes the complementary feedback pair shine. And the current is fixed. The dribble into the base/collector of the current source regulator changes with different voltages, of course, and so does the bias point. But those are completely controllable by changing resistor values. The current source makes the main current path fixed for any supply voltage.

[pee-j]

you might be able to use four lithium coins and get 12V for this (or any!) circuit and increase your headroom so you don't hit clipping so easily. I can advise you about what to change to optimize the circuit for 9V or 12V if you want to experiment.

And it will be an experiment.

it'd be great to have that experiment!
would you please outline how to go about it?

I don't have an oscilloscope, nor the EE knowledge, but I'm okay with pedal building,
my goal is to make an onboard buffer for a bass, without clipping

[Clint Eastwood]

Since you say you have some j201's, maybe this variation on R.G.'s circuit would be worth trying:


At least, using a J201 as current sink would save you some parts/space.

[Steben]

The simple jFET stage has the benefit of being phantom powered through a mono cable omitting the need for saving battery power.
It DOES need a part of the circuit always on first in line.

[Steben]

Since you say you have some j201's, maybe this variation on R.G.'s circuit would be worth trying:


At least, using a J201 as current sink would save you some parts/space.

Isn't the bias feasible with the classic 2 lower resistor values and a large one to the gate?

[antonis]

Isn't the bias feasible with the classic 2 lower resistor values and a large one to the gate?

One item more for some amount of noise less..

[Clint Eastwood]

Isn't the bias feasible with the classic 2 lower resistor values and a large one to the gate?

Yes, you are right. But I thought if it is battery operated only, maybe that wouldn't be necessary. But I'm not sure, for I did not actually build and test this circuit.

[m4268588]

If the input cap is large, there is not much difference.

[antonis]

would you please outline how to go about it?

To stay on R.G.'s design, for 12V supply just double 470k resistor value (820k to 1M)..
You'll keep the same total current consumption while obtaining the double output swing..

P.S.
A perfectionist should also alter 1M/1.5M ratio a bit (1M ->1.2M) for 2N5088 Emitter biased at exactly +12V half-way but, IMHO, the 500mV "off-set" (6.5V instead of 6.0V) should be a more realistic bias point for symmetrical output swing both due to current source BJT V_CEsat + V_E and finite output impedance..

[Steben]

Isn't the bias feasible with the classic 2 lower resistor values and a large one to the gate?

Yes, you are right. But I thought if it is battery operated only, maybe that wouldn't be necessary. But I'm not sure, for I did not actually build and test this circuit.

Noise from DC through the resistor is the same. Only the hum goes away with battery (not that it is not pleasant).

[Clint Eastwood]

Well, let's get rid then of the bias resistors with dc running through them! My favourite buffer is this:


If you choose a JFET with a cutoff voltage of about half the supply voltage, you end up with a bit less than half Vsupply at the source. For  a 9 volt supply, most J111's will do. I use a BF246c, wich is the same part with a different name. For 6 volts, go for the J112.
This buffer works really well for me. And I like that it doesn't come simpler than this.

[antonis]

My favourite buffer is this:

If you choose a JFET with a cutoff voltage of about half the supply voltage, you end up with a bit less than half Vsupply at the source. For  a 9 volt supply, most J111's will do. I use a BF246c, wich is the same part with a different name. For 6 volts, go for the J112.
This buffer works really well for me. And I like that it doesn't come simpler than this.

Could I ask the reason for a JFET buffer of about 120μA working current..??
(not to mention the input/output impedance ratio..)