Biasing MMIC amplifiers (GALI-5)

[[email protected]]

Can someone recommend a simple current regulator circuit that I can
bias devices like the Mini-Circuits GALI-5 with? These devices like to
run at a constant current (60 mA), and a series resistor is usually
used to approximate a constant current source. However, power
consumption is a concern in this design, so I don't want to dissipate
power in the series resistor. I was considering an LDO regulator until
I realized I need constant current, not constant voltage. Is there an
integrated current regulator that I could use that would address my
power consumption concern? Who makes it?

Thanks,

 

[Phil Allison]

Can someone recommend a simple current regulator circuit that I can
bias devices like the Mini-Circuits GALI-5 with? These devices like to
run at a constant current (60 mA), and a series resistor is usually
used to approximate a constant current source. However, power
consumption is a concern in this design, so I don't want to dissipate
power in the series resistor. I was considering an LDO regulator until
I realized I need constant current, not constant voltage. Is there an
integrated current regulator that I could use that would address my
power consumption concern?

** No.

An electronic current source dissipates heat just like a resistor - the
advantage is the current level is more stable with varying supply voltage
than a simple resistor and the effective impedance much higher.

Your best bet IS to use the recommended 40 ohm resistor with a 7 volt
regulated supply & use a switching regulator to drop the voltage from your
main supply.

http://www.minicircuits.com/pdfs/GALI-5+.pdf



.......... Phil

 

[Steve]

Can someone recommend a simple current regulator circuit that I can
bias devices like the Mini-Circuits GALI-5 with? These devices like to
run at a constant current (60 mA), and a series resistor is usually
used to approximate a constant current source. However, power
consumption is a concern in this design, so I don't want to dissipate
power in the series resistor. I was considering an LDO regulator until
I realized I need constant current, not constant voltage. Is there an
integrated current regulator that I could use that would address my
power consumption concern? Who makes it?

Thanks,

The bias inductor is often unable to provide a very high impedance, so that
the bias network shunts RF power (gain loss) and degrades output VSWR. In
those cases, the resistor also serves to raise the RF load impedance of the
bias network. Any current source you design will need to present a high
impedance beyond the frequency band of interest, unless your series inductor
can do the entire job below its self resonant point.

There are constant current designs using certain adjustable linear
regulators. The sense/set resistor is the "upper resistor" and the "lower
resistor" is your load. You'd need a regulator that has the fixed voltage
drop across the upper resistor, not the lower one. Assuming the typical
1.25V drop, at 60 mA, the resistor is only about 20 ohms. If you check your
design with a network analyzer, you might find that it maintains a high
impedance at your RF frequencies (I doubt it). Otherwise, the series
inductor has to provide all of the isolation.

Good luck
Steve

 

[[email protected]]

The bias inductor is often unable to provide a very high impedance, so that
the bias network shunts RF power (gain loss) and degrades output VSWR. In
those cases, the resistor also serves to raise the RF load impedance of the
bias network. Any current source you design will need to present a high
impedance beyond the frequency band of interest, unless your series inductor
can do the entire job below its self resonant point.

There are constant current designs using certain adjustable linear
regulators. The sense/set resistor is the "upper resistor" and the "lower
resistor" is your load. You'd need a regulator that has the fixed voltage
drop across the upper resistor, not the lower one. Assuming the typical
1.25V drop, at 60 mA, the resistor is only about 20 ohms. If you check your
design with a network analyzer, you might find that it maintains a high
impedance at your RF frequencies (I doubt it). Otherwise, the series
inductor has to provide all of the isolation.

Good luck
Steve

I think I'll try a current mirror with either PNP bipolar transistors
or P-channel FETs. I've done this before when I wasn't as concerned
about excess power disspipation, and I could use a fairly high bias
voltage. The problem in the current case is that excess bias voltage
leads to a proportional amount of excess power consumption that I want
to minimize. So I will have to determine how low I can go with the
supply voltage while maintaining a reasonably constant current.

 

[Joerg]

Can someone recommend a simple current regulator circuit that I can
bias devices like the Mini-Circuits GALI-5 with? These devices like to
run at a constant current (60 mA), and a series resistor is usually
used to approximate a constant current source. However, power
consumption is a concern in this design, so I don't want to dissipate
power in the series resistor. I was considering an LDO regulator until
I realized I need constant current, not constant voltage. Is there an
integrated current regulator that I could use that would address my
power consumption concern? Who makes it?

Check the LM317 datasheet. Such a regulator and one lone resistor can
make a nice current source. Personally I'd stay away from LDOs because
they are expensive and most have documented as well as undocumented
pathologies. Some of them of the kablouie category, pieces flying around
and stuff.

As Steve wrote you need to verify what load resistor the MMIC wants for
RF purposes. You can use an inductor plus load resistor but watch out
for resonances. Provide good bypassing at the other end of that.

The ultimate Uncle Scrooge method in power economy would be a switcher
that gets the current as feedback instead of a voltage. Then you could
keep the collector voltage at the far end of the inductor precisely
where it needs to be for your 60mA.

 

[John Larkin]

Can someone recommend a simple current regulator circuit that I can
bias devices like the Mini-Circuits GALI-5 with? These devices like to
run at a constant current (60 mA), and a series resistor is usually
used to approximate a constant current source. However, power
consumption is a concern in this design, so I don't want to dissipate
power in the series resistor. I was considering an LDO regulator until
I realized I need constant current, not constant voltage. Is there an
integrated current regulator that I could use that would address my
power consumption concern? Who makes it?

Thanks,

The extreme power-saving case would be to pull up the output through
an inductor, from a stiff supply, and sense/servo the operating
current by applying a bit of DC to the input. An opamp and a few
passives can do that.

Otherwise, any classic current source will work.


+------------+----------+----V+
| |
| |
| Re
LED a |
k |
| |
| e
+---Rb----b PNP
| c
| |
Rled |
| L
| |
| |
gnd to mmic


I think Zetex and maybe TI make integrated current mirrors that might
do this with less parts. Just be careful they don't oscillate, given
the inductor load.

Incidentally, most of these parts have a below-50 ohm input impedance
at recommended operating current, and go up at lower currents.

John

 

[John Larkin]

I think I'll try a current mirror with either PNP bipolar transistors
or P-channel FETs. I've done this before when I wasn't as concerned
about excess power disspipation, and I could use a fairly high bias
voltage. The problem in the current case is that excess bias voltage
leads to a proportional amount of excess power consumption that I want
to minimize. So I will have to determine how low I can go with the
supply voltage while maintaining a reasonably constant current.

A fet would probably have more output capacitance than a bipolar. One
good combination is a small PNP, a ferrite bead, then an inductor. The
whole mess will be a fairly good broadband high Z.

Be careful with Minicircuits. They outsource these parts and
occasionally change vendors, and the new parts only vaguely behave
like the old ones. And they tend to tell us to drop dead when we have
problems. I'd go with W-J or Sirenza or somebody dependable.

John