Linear lab amp DC-30MHz?

[Richard Kayman]

What are the most cost effective options for buying or building a
linear amp for general lab work that will do DC up into the early tens
of Mhz?

I would prefer to use a single unit, rather than several to cover the
entire frequency range. It would be powered by a commercial signal
generator. Maximum output required would be 20W into a non-reactive
load.

Richard

 

[Tim Williams]

What are the most cost effective options for buying or building a
linear amp for general lab work that will do DC up into the early tens
of Mhz?

Find your scope (you ..already.. have a Tektronix scope, so no cost at all)
and find the vertical amplifier output. Voila.

Maximum output required would be 20W into a non-reactive load.

Oh...nevermind then.

;o)

Tim

 

[Rene Tschaggelar]

What are the most cost effective options for buying or building a
linear amp for general lab work that will do DC up into the early tens
of Mhz?

I would prefer to use a single unit, rather than several to cover the
entire frequency range. It would be powered by a commercial signal
generator. Maximum output required would be 20W into a non-reactive
load.

into 1 Ohm ? 50 Ohm ? 1kOhm ? adaptive ?

Rene

 

[Reg Edwards]

Buy a second-hand radio transmitter.

 

[Winfield Hill]

into 1 Ohm? 50 Ohm? 1kOhm? adaptive?

Richard, you know of many types of wideband DC amplifiers going
beyond 30MHz, ranging from those in scopes to wideband opamps,
but none can meet your 20W requirement. Furthermore, most any
say 2 to 30MHz amplifier you find that can deliver 20 watts or
more will do so only under restricted conditions, e.g. into a
50-ohm resistive load.

I have made a 50W 1-30MHz amplifier, using military surplus
RF stages. It heavily employs large hideously-expensive RF
transistors, operating class-A with special thermal-bias-set
feedback. It uses balanced transmission-line transformers
in and out, wound on large RF toroids to get down to 1MHz.

It's designed for 50 ohm loads, but it's fairly tolerant of
non-resistive loads, and has been used continuously for many
years to make swept magnetic fields to cool atoms in light-
stopping experiments. The amplifier won't deliver "50 watts"
into an inductive load or a capacitive short at 30MHz, but it
will deliver several amps of current, creating a wideband
1-30MHz tunable magnetic field.

I imagine my RF amplifier could be mated somehow to a DC-3MHz
power amplifier (I have two candidate designs that have gone
through maturing iterations over the years, as we've made about
30 of them), but even thinking about such a thing is highly
unappealing to me. It would be one ugly, fragile messy beast.

 

[Andy]

Andy replies:
1) Go to the Communications Concepts website. They have these
modules, and you can buy the parts there, but you have to
assemble them yourself.

2) Check copies of The Radio Amateur's Handbook. Windband
amps in the 20 watt range driving a 50 ohm load have been
around a long time. Expect to pay $50-$100 in parts, but,
like Comminications Concepts, you have to build it yourself.

3) Check our ebay for ham band linear amplifiers.

4) Check your local amateur radio club..... You will probably
need a ham license before anyone will sell/trade you one.

Good Luck,
Andy w4oah

 

[Ken Smith]

into 1 Ohm ? 50 Ohm ? 1kOhm ? adaptive ?

Or how about massively overdesigning it so it could do about 2KW into 50
Ohms.

 

[Joerg]

Hello Win,

I imagine my RF amplifier could be mated somehow to a DC-3MHz
power amplifier (I have two candidate designs that have gone
through maturing iterations over the years, as we've made about
30 of them), but even thinking about such a thing is highly
unappealing to me. It would be one ugly, fragile messy beast.

That is why I stocked up heavily on LH0063 buffers before they became
extinct, so I don't have to go through that messy beast scenario. Nice
heavy-duty TO-3 style 8-pin package that can be bolted onto a heat sink.
The only downside is that their supply abs max limit is 40V. Riding the
supplies of one on the outputs of two others is something I have never
dared to do with these since they oscillate easily.

The 1982 databook officially calls them 'damn fast buffers'. Which they
are, rolling off around 100MHz. I guess they couldn't write 'damn' into
a data sheet these days anymore.

Once I had to drive a transducer of a few ohms really hard. So I
paralleled a bunch of them and they actually blew up the transducer.

Regards, Joerg

 

[Joerg]

Hello Richard,

What are the most cost effective options for buying or building a
linear amp for general lab work that will do DC up into the early tens
of Mhz?

I would prefer to use a single unit, rather than several to cover the
entire frequency range. It would be powered by a commercial signal
generator. Maximum output required would be 20W into a non-reactive
load.

If you don't want to build it yourself check out companies like these:

http://www.wavecontrol.com/index2004.php?p=pantalla&pant=RF Amplifiers: DC - 1GHz&idm=385

They claim that the 50WD1000 does 50W from DC to a GHz. It is probably
really expensive, the size of a book cabinet and you may need a crane to
get it into the building.

Regards, Joerg

 

[John Fields]

What are the most cost effective options for buying or building a
linear amp for general lab work that will do DC up into the early tens
of Mhz?

I would prefer to use a single unit, rather than several to cover the
entire frequency range. It would be powered by a commercial signal
generator. Maximum output required would be 20W into a non-reactive
load.

 

[Fred Bloggs]

OP is from Australia, ignore it.

 

[John Devereux]

Richard, you know of many types of wideband DC amplifiers going
beyond 30MHz, ranging from those in scopes to wideband opamps,
but none can meet your 20W requirement. Furthermore, most any
say 2 to 30MHz amplifier you find that can deliver 20 watts or
more will do so only under restricted conditions, e.g. into a
50-ohm resistive load.

I must admit I find it surprising that this is so difficult. In these
days of GHz opamps and CPUs, who would have thought it would be such a
challenge?

Makes me want to fire up LTSpice to see what is so hard...

 

[Rene Tschaggelar]

I must admit I find it surprising that this is so difficult. In these
days of GHz opamps and CPUs, who would have thought it would be such a
challenge?

Makes me want to fire up LTSpice to see what is so hard...

You don't neet a simulation tool.
20Watt means the amplifier takes at least 60W
and the difference goes up in heat.

And 20W into 1 Ohm is 4.5V @4.5A, but into 50 Ohms it
is 30V @0.6A, and into 1kOhm it is 140V @140mA.
This means you have to provide an amplifier with 140V
and 4.5A just to cover 1 Ohm to 1 kOhm.

Rene

 

[Ken Smith]

Winfield Hill <> writes: [...]
I must admit I find it surprising that this is so difficult. In these
days of GHz opamps and CPUs, who would have thought it would be such a
challenge?

Makes me want to fire up LTSpice to see what is so hard...

You don't neet a simulation tool.
20Watt means the amplifier takes at least 60W
and the difference goes up in heat.

And 20W into 1 Ohm is 4.5V @4.5A, but into 50 Ohms it
is 30V @0.6A, and into 1kOhm it is 140V @140mA.
This means you have to provide an amplifier with 140V
and 4.5A just to cover 1 Ohm to 1 kOhm.

Did someone quit making great big heat sinks, fans, heat pipes, water
cooling etc? 150V 6A supplies can be had. I've got one on my work bench.

I think the biggest problem is getting fast, high voltage parts. At only
a few hundred volts, I'd look at www.ixysrf.com to see if they have whats
needed. If the upper frequency was only 1 or 2 MHz, I'd suggest TO-220
power MOSFETS in parallel.

 

[Joerg]

Hello Ken,

I think the biggest problem is getting fast, high voltage parts. At only
a few hundred volts, I'd look at www.ixysrf.com to see if they have whats
needed. If the upper frequency was only 1 or 2 MHz, I'd suggest TO-220
power MOSFETS in parallel.

When I was in college we built RF amps up to 30MHz with large FETs but
back then these were expensive. The total power output was pretty much
limited by the trade off between the FET budget and the beer budget.
Beer had a higher priority ;-)

Regards, Joerg

 

[Fred Bartoli]

Hello Ken,


When I was in college we built RF amps up to 30MHz with large FETs but
back then these were expensive. The total power output was pretty much
limited by the trade off between the FET budget and the beer budget.
Beer had a higher priority ;-)

With tubes you wouldn't have had such loose-loose choice. It's rather:

tube/beer or not tube/beer.

 

[Joerg]

Hello Fred,

With tubes you wouldn't have had such loose-loose choice. It's rather:

tube/beer or not tube/beer.

Actually, with tubes it was worse because of the needed infrastructure.
Microwaves weren't old enough back then to find sufficient quantities in
the scrap yards, to get the HV caps out of them. I remember having to
plunk down about $20 for one 6uF 5kV capacitor. Ouch. That really hits
hard if you are in college.

In contrast, the scrap yards in that university town had tons of old
mainframes flying around. You could scavenge 10,000uF+ caps to your
hearts desire out of these, plus big transformers and so on. Typically
you paid by weight and not by what you scrapped out. I still have a
100,000uF cap from a RAM backup module. Paid next to nothing for it. And
some diode arrays with huge lag bolts on them.

Regards, Joerg

 

[Ancient_Hacker]

Do you need a wide-band amp or will a tunable amp do?

If tunable with a 5% bandwidth will do:


A 600V power supply, a single 8417 tube and a pi-network will more than
do the trick.

 

[Rich Grise]

What are the most cost effective options for buying or building a
linear amp for general lab work that will do DC up into the early tens
of Mhz?

I would prefer to use a single unit, rather than several to cover the
entire frequency range. It would be powered by a commercial signal
generator. Maximum output required would be 20W into a non-reactive
load.

What kind of phase distortion can you tolerate? If you don't mind
waiting a few microseconds or so for your gazintas to get to your
gazoutas, you could use a distributed amplifier. They can be made
amazingly wideband, but they're real power hogs.

Good Luck!
Rich

 

[Rich Grise]

Do you need a wide-band amp or will a tunable amp do?

If tunable with a 5% bandwidth will do:


A 600V power supply, a single 8417 tube and a pi-network will more than
do the trick.

Gonna take a HELL of a big pi-network to go down to "DC". ;-)

Cheers!
Rich