Topics

µBITX & QRP labs 50W PA

Arvo W0VRA
 

Anybody considering using a QRP Labs 50W PA with their µBITX?

http://qrp-labs.com/50wpa

There's no band switching yet.

Jerry Gaffke
 

That's a class C amp.
As it says on the web page you pointed to:
>    This amplifier is not intended to be a Linear (for SSB use) - it is designed for single-band CW operation only. 

Jerry


On Thu, Dec 12, 2019 at 08:49 AM, Arvo W0VRA wrote:
Anybody considering using a QRP Labs 50W PA with their µBITX?

http://qrp-labs.com/50wpa

There's no band switching yet.

Jerry Gaffke
 

Ok, I sort of take that back.

The webpage also says:

######################################

2. Can it be used as a Linear for SSB and other modes? 

NOT SSB as it is - it is designed as a Class-C amplifier, for CW operation (or FSK modes). Single Sideband (SSB) requires a Linear. If you want to experiment using this amplifier as a Linear, you can try setting the bias trimmer potentiometer to achieve a higher "idle" current during transmit, of around 250mA. This will dissipate some watts of heat even when there is no signal, and is normal for Linears. The efficiency will also be lower. Accordingly, heat dissipation may be higher too (depending on the average power output and other factors). So proceed with caution if you want to modify the amplifier for LINEAR operation - you will be in experimental territory. 

.......

6. Could it be modified for multi-band operation? 

 

Certainly. Just bypass the Low Pass Filter section of the design, and use an external switched Low Pass Filter for the bands of interest. There are no other band-specific components except for the Low Pass Filter components and it will work across all HF. As usual... everything gets harder at higher frequencies so expect the power output to start to fall off (see also the measurements section of the manual). 

##################################################



So multiband should be possible with just an appropriate LPF as a lump in your antenna feed line.
I would guess you will have to adjust the drive into the amp whenever you switch bands, the upper bands will require more drive.

As he says, linear operation for SSB might be possible but is not designed in.
You're on your own.

Jerry

MadRadioModder
 

Single band

 

 

From: BITX20@groups.io [mailto:BITX20@groups.io] On Behalf Of Arvo W0VRA via Groups.Io
Sent: Thursday, December 12, 2019 10:50 AM
To: BITX20@groups.io
Subject: [BITX20] µBITX & QRP labs 50W PA

 

Anybody considering using a QRP Labs 50W PA with their µBITX?

http://qrp-labs.com/50wpa

There's no band switching yet.


--

…_. _._

iz oos
 

Nice to know. I don't think it can be used for SSB as it is a class C amp unless some mods are done. Looking forward to seeing Allison's reply


Il 12/dic/2019 17:49, "Arvo W0VRA via Groups.Io" <arvopl=yahoo.com@groups.io> ha scritto:
Anybody considering using a QRP Labs 50W PA with their µBITX?

http://qrp-labs.com/50wpa

There's no band switching yet.

MadRadioModder
 

Ah ok….

 

Well if it’s “Class C” it’s not gonna work for SSB.  The only way to change that is to bring it to at least “Class AB2”… by reducing the bias and allowing the device more resting current.  That generates a lot more heat (because the devices are conducting more of the time) so the cooling needs to be increased.   You can make it happen but it’s gonna take some work.

 

 

From: BITX20@groups.io [mailto:BITX20@groups.io] On Behalf Of iz oos
Sent: Thursday, December 12, 2019 11:24 AM
To: BITX20@groups.io
Subject: Re: [BITX20] µBITX & QRP labs 50W PA

 

Nice to know. I don't think it can be used for SSB as it is a class C amp unless some mods are done. Looking forward to seeing Allison's reply

 

Il 12/dic/2019 17:49, "Arvo W0VRA via Groups.Io" <arvopl=yahoo.com@groups.io> ha scritto:

Anybody considering using a QRP Labs 50W PA with their µBITX?

http://qrp-labs.com/50wpa

There's no band switching yet.


--

…_. _._

Jerry Gaffke
 

Like all of the stuff Hans sells, the manual is exceptional.
Lots of notes in there about design tradeoffs made.
And a bunch about what it would take to make it linear.

Should have a way to evaluate IMD if you give linear operation a try.

Price is right.
Has a very nice solution for QSK TR switching.

Jerry

Hans Summers
 


Hi all

Changing amplifier class is firstly a matter of altering the idle current. It requires an increase in bias to increase the idle current (NOT a decrease in bias). That is the first thing to try. 

But it would help to have capability to measure the results to see how good the linearity is and what the limitations are. Two tone audio test signal and RF spectrum analyzer. 

The circuit is essentially similar to my 10W Linear kit http://qrp-labs.com/linear 

Yes biasing the transistors for 125mA each (for example) would dissipate power straight away even with no input. That would be 2.5W per transistor at 20V supply. Whether it needs MORE cooling or not is not so straightforward to answer though. CW is a full power, intermittent carrier. SSB operates at a lower average level but without so many breaks (except when you take a breath etc). Any speech processing you have will increase the average power. Which is more, CW or SSB... I think it needs some experiment. Probably this amp can be biased to AB but I haven't had time to test all this as yet. 

73 Hans G0UPL 

Ashhar Farhan
 

Hans,
Even 25 ma idle current should be enough, given that it is a push pull amplifier and a large part of the power cycle keeps one or the other transistor switched on.
- f

On Fri, Dec 13, 2019 at 10:36 AM Hans Summers <hans.summers@...> wrote:

Hi all

Changing amplifier class is firstly a matter of altering the idle current. It requires an increase in bias to increase the idle current (NOT a decrease in bias). That is the first thing to try. 

But it would help to have capability to measure the results to see how good the linearity is and what the limitations are. Two tone audio test signal and RF spectrum analyzer. 

The circuit is essentially similar to my 10W Linear kit http://qrp-labs.com/linear 

Yes biasing the transistors for 125mA each (for example) would dissipate power straight away even with no input. That would be 2.5W per transistor at 20V supply. Whether it needs MORE cooling or not is not so straightforward to answer though. CW is a full power, intermittent carrier. SSB operates at a lower average level but without so many breaks (except when you take a breath etc). Any speech processing you have will increase the average power. Which is more, CW or SSB... I think it needs some experiment. Probably this amp can be biased to AB but I haven't had time to test all this as yet. 

73 Hans G0UPL 

ajparent1/kb1gmx
 

Farhan,

No its not.  You need to be more than 100ma to get sorta on the linear part of the IRF510 bias curves.

At minimum bias you are on the very curved part of the gate threshold curve and the IMD will be poor with
crossover distortion.  Also we are not switching as its linear from transistor to transistor its complementary.
 
---------------------------------
No direct email, it goes to bit bucket due to spam

Ashhar Farhan
 


Allison,
I am basing my report in wa2eby's recommendations. My own measurement lead me to believe that between 25 ma to 150 ma, there is not much of a difference in linearity. The linearity increases only around 500ma. By then, we are roasting potatoes.

- f

On Tue 17 Dec, 2019, 3:22 AM ajparent1/kb1gmx, <kb1gmx@...> wrote:
Farhan,

No its not.  You need to be more than 100ma to get sorta on the linear part of the IRF510 bias curves.

At minimum bias you are on the very curved part of the gate threshold curve and the IMD will be poor with
crossover distortion.  Also we are not switching as its linear from transistor to transistor its complementary.
 
---------------------------------
No direct email, it goes to bit bucket due to spam

John AE5X <ae5x@...>
 

At $300 the HR-50 supports the uBITX on all bands and modes. And it does not require a PTT In signal since it has a carrier-operated relay:
https://www.hobbypcb.com/index.php/products/hf-radio/hardrock-50-hf-power-amp

John AE5X

padmanabhan cattamanchi
 

Most of the the VU2s use a laptop smps with a bit of filtering to give nearly 20 watts out. I don't think there is a need of another amp since the devices are the same

de vu2pep 

On Tuesday, December 17, 2019, 05:28:23 PM GMT+5:30, John AE5X <ae5x@...> wrote:


At $300 the HR-50 supports the uBITX on all bands and modes. And it does not require a PTT In signal since it has a carrier-operated relay:
https://www.hobbypcb.com/index.php/products/hf-radio/hardrock-50-hf-power-amp

John AE5X

 

Hi,

I think linearity will depend on a lot of things, most importantly supplied voltage and bias current (and likely input signal level as well).   I would think it would also depend on the design of the amplifier and not only on what output device it uses.

In WA2EBY's article, he doesn't show or talk about any linearity measurements that he might have made, only that "No indications of instability, CW key clicks or distortion on SSB have been reported." which is not really a reliable test.

So I would either say as Jerry does and have a way to measure IMD distortion, or go with what the designer recommends.  For SSB, unless you are really long-winded, is it probably is OK to have higher idle currents; 250mA at 20V is only an idle of 5W so not so bad if the heatsink is good for intermittant 50W.  Regardless, it would be nice to see some test results showing IMD versus idle current so it could be set it as low as it can. 

73,


Mark.




ajparent1/kb1gmx
 

Since I run an EBY amp since 2006 you must have mis-read.  In the files section
I've even posted his articles back around 2010 and copied to this iteration years
ago. Still there.

The ubitx is not run that low, spec is 100ma per. 

The irf510 like all fets has a considerable hook at low gate to source voltages.
If the gate voltage is platted against the drian current the linear window is above
100ma and below saturation or about 4A.  Below that in the low end under 100ma 
drain current he gate voltage is not linear at all from start of conduction to the 
100ma range and  that makes for very non linear near cross over.  Class AB1
is more like 80-150ma but it is very toplology dependent.

His (WA2EBY) minimum was 50 ma per device for CW and testing put it higher
near 100ma per device for SSB to get a decent IMD at 28V.  I find mine
comfortable at 125 ma per device and the two tone testing is good enough at
about -28dbc.  However that design has the features needed for linear operation
including independent device bias,  current sourcing transformer for the drains
and large ferrite for the output transformer.

At 100ma per you only talking a standing power of 2.8W at 28V per device,
if that gets warm the heatsink is inadequate. 

So saying a specific current, especially low under 100ma, is inadequate
for linear use without testing questionable.  Especially for an amp designed
for class C use.  I do not say it cant be linear but at low bias its highly unlikely.

I have one on the way so testing will happen.

Allison

--
---------------------------------
No direct email, it goes to bit bucket due spam

Gordon Gibby
 

Allison, great information!   That sheds a great deal of light on the subject, thanks!


On Tue, Dec 17, 2019 at 11:26 AM ajparent1/kb1gmx <kb1gmx@...> wrote:
Since I run an EBY amp since 2006 you must have mis-read.  In the files section
I've even posted his articles back around 2010 and copied to this iteration years
ago. Still there.

The ubitx is not run that low, spec is 100ma per. 

The irf510 like all fets has a considerable hook at low gate to source voltages.
If the gate voltage is platted against the drian current the linear window is above
100ma and below saturation or about 4A.  Below that in the low end under 100ma 
drain current he gate voltage is not linear at all from start of conduction to the 
100ma range and  that makes for very non linear near cross over.  Class AB1
is more like 80-150ma but it is very toplology dependent.

His (WA2EBY) minimum was 50 ma per device for CW and testing put it higher
near 100ma per device for SSB to get a decent IMD at 28V.  I find mine
comfortable at 125 ma per device and the two tone testing is good enough at
about -28dbc.  However that design has the features needed for linear operation
including independent device bias,  current sourcing transformer for the drains
and large ferrite for the output transformer.

At 100ma per you only talking a standing power of 2.8W at 28V per device,
if that gets warm the heatsink is inadequate. 

So saying a specific current, especially low under 100ma, is inadequate
for linear use without testing questionable.  Especially for an amp designed
for class C use.  I do not say it cant be linear but at low bias its highly unlikely.

I have one on the way so testing will happen.

Allison

--
---------------------------------
No direct email, it goes to bit bucket due spam

ajparent1/kb1gmx
 

Mark,

For SSB high idle current are not a factor (unless unrealistically high) as
the average current is likely going to be higher and if heat is an issue
than larger heatsinks or fans are the solution.

Keep in mind speculation is not measurements.  For 50W out 
your input DC at 20V has to be in the range of 70 to 90W
assuming 70 to 50% efficiency.  The difference between output
power and input power is resulting in heat.  Power amplifiers 
above the milliwatt range are all concerned with heat and
managing it.

I have used the same heatsinks for the QRPLabs 10W SSB amp 
at higher voltages and power there is a lot of heatsink there.  Running
bias in the 100Ma per device range is fairly trivial at 4W.  Even at 250ma
we are at under 10W.    Considering at 50W out the likely current  is
3.6A at 20V and converting 30% of that to heat for classC and much
Less (typical is 55%, average) for class AB1.

For SSB running to 50W peak the average is likely closer to 20W
and the heat load is maybe near 20W average as well assuming
50% efficient and that would be low.  So heat should not be an issue
with supplied heatsinks and reasonable airflow (no sealed chassis)
and not over driven.

It should be noted that for SSB the peak power is around 50W the
average will be lower and should be to avoid splatter from clipping.
That means the SSB drive power should be lower than CW.

Allison
---------------------------------
No direct email, it goes to bit bucket due spam

Alex Netherton
 


On Thu, Dec 12, 2019 at 11:49 AM Arvo W0VRA via Groups.Io <arvopl=yahoo.com@groups.io> wrote:
Anybody considering using a QRP Labs 50W PA with their µBITX?

http://qrp-labs.com/50wpa

There's no band switching yet.

wnpauls
 

That would work, but it will need band pass filters and the
associated band switching.  The same amp is available (unassembled)
for about 38 to 40 US dollars shipped. 

Paul K0ZYV 


From: BITX20@groups.io <BITX20@groups.io> on behalf of Alex Netherton via Groups.Io <anetherton@...>
Sent: Tuesday, December 17, 2019 9:06 PM
To: BITX20@groups.io <BITX20@groups.io>
Subject: Re: [BITX20] µBITX & QRP labs 50W PA
 

On Thu, Dec 12, 2019 at 11:49 AM Arvo W0VRA via Groups.Io <arvopl=yahoo.com@groups.io> wrote:
Anybody considering using a QRP Labs 50W PA with their µBITX?

http://qrp-labs.com/50wpa

There's no band switching yet.

Alex Netherton
 

Is there a cost effective way to add band switching? The scope trace showing on the page seems pretty good.


On Wed, Dec 18, 2019 at 12:36 AM wnpauls via Groups.Io <pschumacher=winona.edu@groups.io> wrote:
That would work, but it will need band pass filters and the
associated band switching.  The same amp is available (unassembled)
for about 38 to 40 US dollars shipped. 

Paul K0ZYV 

From: BITX20@groups.io <BITX20@groups.io> on behalf of Alex Netherton via Groups.Io <anetherton=gmail.com@groups.io>
Sent: Tuesday, December 17, 2019 9:06 PM
To: BITX20@groups.io <BITX20@groups.io>
Subject: Re: [BITX20] µBITX & QRP labs 50W PA
 

On Thu, Dec 12, 2019 at 11:49 AM Arvo W0VRA via Groups.Io <arvopl=yahoo.com@groups.io> wrote:
Anybody considering using a QRP Labs 50W PA with their µBITX?

http://qrp-labs.com/50wpa

There's no band switching yet.