Looks pretty nice Jack, very professional.
Joel
N6ALT
|
Re: 2N2222A vs 2N3904 sourcing
Allison,
Thank you very much. Very appreciated.
Tom, wb6b
|
Re: 2N2222A vs 2N3904 sourcing
Jim,
For that inductor the amps I have that perform well use it. All three use different
arrangements and all produce 10W and do it well.? One its a Ft50-43 with 8turns,
the other its 4 turns on FB43-202 and the third uses a stacked pair of
FB50-43 and 6turns bifilar #26.? ?
In all cases the output translation is 2:3 turns or more for a low drain load and the
feed inductor does not need to have an especially high inductance.? They are?
in the range of 4x the load impedance at the lowest frequency (3mhz).
One of the things to avoid is self resonance as well.
Allison
|
Henning,
Thanks,? One hopes that will help those that haven't a clear idea of whats needed.
That will go in the wiki.
Allison
|
Warren,
That proves first get the signal to the mixer clean and at a reasonable level.? IT also finds the needed level to the mixer?
and that is under -10dbm for the available LO drive (that can increase maybe 2-3db).
The power amp I've been writing about since May has issues, many, and you have cataloged them.
Those being poor gain flatness, 3 vs 30 mhz.? it currently sags about 11db.
Poor IMD, many reasons including stages going into gain compression and that point varies with frequency due to poor devices.
Also the 3904 has a gain vs current that is a decreasing value with increasing current so the device becomes increasingly non linear.
The IRF510 is operating with a Drain to Drain load of 50 ohms. the calculated non saturating power out is 2.88w at 12V.
This is consistent with your 3W (at 13V?).? ?Either way the Drain load is way too high.? The transformer T11 is both
undersized for ferrite volume and the turns ratio is 1:1 and we need more like 1:2 or greater.? That means a drain load of
must be lower, at least half more like 1/4.?
And so it goes for the entire power chain.
|
Well ya there's that too ..But I drink coffee so...
Jim
On Sunday, September 16, 2018, 7:49:47 AM PDT, Raj vu2zap <rajendrakumargg@...> wrote:
I like the designs on the napkin with beer stains!
Raj
At 16/09/2018, you wrote:
Indeed ..Real near what I
concluded on the old "back 'o the envelope trick"?
..
Jim
On Sunday, September 16, 2018, 1:42:26 AM PDT, Henning Weddig via
Groups.Io <hweddig@...> wrote:
Allison,
at the beginning of my professional career 38 years ago I learned that to
sketch level diagrammes will help a lot to understand the requirements of
individual stages concerning gain, outputpower and even give some insight
in IMD/ P1dB. A visiual display gives more information as jsut "bare
numbers".
So I just sketeched such a block diagramme with the power levels in
between, see the attachment.
From this point of view only the first stage can be critical in respect
of gain-- although a MMIC - (or its equivalent using discrete components
can do ?) can easily get this gain at "LF".?
Henning Weddig
DK5LV
Am 16.09.2018 um 02:13 schrieb ajparent1/KB1GMX:
Iz oos,
I was thinking whether a MMIC up to 10dbm and an RD06HHF as driver like
in the 10w qrp-pa-2008 might make sense
Whats that going to fix?? The transformers need to be reworked, the
prior stages in the amp are are poor.? And a MMIC
at 10dbm is near compression unless you get something ERA5 but then at 75
ma you only get 20db of gain.
Think this out.? To get from the mixer to the antenna you need
60db of power gain total!
Now if you used RD16hhf you get maybe 16db of gain more like 13 with
feedback so the gain is flatter.
For 1/10th the cost the IRF510 can do that.? Either way
13DB.
So you need 47 db of gain call it 48 as a round number.
Your going to get that from three stages.
NOTE:
The last stage must supply more than 500mw from 3-30mhz.? [likely
gain at that power under 13db]
The stage before that will have to supply maybe 70 (max) milliwatts.
[likely best gain at that power maybe 15db]
The stage before that at most 3-4 milliwatts. [gains of 20db possible at
low power]
Getting high gain at low power is easier.? Hard at higher
power.? Its bad to run at max gain as then
the results are variable from unit to unit.
FYI most all fets will produce a lot of gain down at 3mhz, that's not the
problem, its getting the same
{or close} gain from 3 to 30Mhz.
I've repeated this more than a few times in the past two days. All I hear
is what about this fix.
Go look in the wiki, its spelled out.? Don't like it do the research
and think system level.
And please do not bring up a monoband radio that may or may not perform
well as its not
relevant or even meaningful.? Its a monoband it does not not have to
work at any other
frequency and every time I look its clear it will not work in a 3-30
case.
Allison
|
Re: 2N2222A vs 2N3904 sourcing
I was specifically referring to BIFILAR inductor ....Just like the big guys do it
Jim
On Sunday, September 16, 2018, 6:48:33 AM PDT, ajparent1/KB1GMX <kb1gmx@...> wrote:
Hi Jim,
Proof, it has ot make sense and consistant with my memory as i'm writer.? If you want to write a page you can.
As to the L8/L9? depends if its separate inductors as is or a bifilar inductor where the impedance can be much
lower as its balanced and DC only at the center plus its magnetically self canceling.
Allison
|
Thank you first and foremost? I apologize to the group. My signal has increased chu 7.85 nice strong my wire feeds off my cable attached to bnc connector got 200 feet 50 ohms type. I'm in apartment owner really did like amateur radio operators, any wire or antenna $300?next eviction in lease agreement. surrounded by apartment's . they are tolerant of patio furniture I read somewhere about a antenna umbrella. my patio size of a piece of sheetrock uncut I don't know the size 8 foot by 4 foot. any kind suggestions are welcome.?
On Sun, Sep 16, 2018, 10:57 AM James Lynes < jmlynesjr@...> wrote: Resonant antenna is not required on receive. I've heard over 40 states and 20 countries with 65ft wire up 10ft in my rafters with my ZZRX-40 DC receiver.?
Check the BNC connector center pin. It's been noted that some have a loose fit.
Check for the Raduino clocks by using another receiver.?
James KE4MIQ
|
Resonant antenna is not required on receive. I've heard over 40 states and 20 countries with 65ft wire up 10ft in my rafters with my ZZRX-40 DC receiver.?
Check the BNC connector center pin. It's been noted that some have a loose fit.
Check for the Raduino clocks by using another receiver.?
James KE4MIQ
|
I like the designs on the napkin with beer stains!
Raj
At 16/09/2018, you wrote:
toggle quoted message
Show quoted text
Indeed ..Real near what I
concluded on the old "back 'o the envelope trick"?
..
Jim
On Sunday, September 16, 2018, 1:42:26 AM PDT, Henning Weddig via
Groups.Io <hweddig@...> wrote:
Allison,
at the beginning of my professional career 38 years ago I learned that to
sketch level diagrammes will help a lot to understand the requirements of
individual stages concerning gain, outputpower and even give some insight
in IMD/ P1dB. A visiual display gives more information as jsut "bare
numbers".
So I just sketeched such a block diagramme with the power levels in
between, see the attachment.
From this point of view only the first stage can be critical in respect
of gain-- although a MMIC - (or its equivalent using discrete components
can do ?) can easily get this gain at "LF".?
Henning Weddig
DK5LV
Am 16.09.2018 um 02:13 schrieb ajparent1/KB1GMX:
Iz oos,
I was thinking whether a MMIC up to 10dbm and an RD06HHF as driver like
in the 10w qrp-pa-2008 might make sense
Whats that going to fix?? The transformers need to be reworked, the
prior stages in the amp are are poor.? And a MMIC
at 10dbm is near compression unless you get something ERA5 but then at 75
ma you only get 20db of gain.
Think this out.? To get from the mixer to the antenna you need
60db of power gain total!
Now if you used RD16hhf you get maybe 16db of gain more like 13 with
feedback so the gain is flatter.
For 1/10th the cost the IRF510 can do that.? Either way
13DB.
So you need 47 db of gain call it 48 as a round number.
Your going to get that from three stages.
NOTE:
The last stage must supply more than 500mw from 3-30mhz.? [likely
gain at that power under 13db]
The stage before that will have to supply maybe 70 (max) milliwatts.
[likely best gain at that power maybe 15db]
The stage before that at most 3-4 milliwatts. [gains of 20db possible at
low power]
Getting high gain at low power is easier.? Hard at higher
power.? Its bad to run at max gain as then
the results are variable from unit to unit.
FYI most all fets will produce a lot of gain down at 3mhz, that's not the
problem, its getting the same
{or close} gain from 3 to 30Mhz.
I've repeated this more than a few times in the past two days. All I hear
is what about this fix.
Go look in the wiki, its spelled out.? Don't like it do the research
and think system level.
And please do not bring up a monoband radio that may or may not perform
well as its not
relevant or even meaningful.? Its a monoband it does not not have to
work at any other
frequency and every time I look its clear it will not work in a 3-30
case.
Allison
|
?Warren --- very helpful measurement!
Cheers,
gordon
toggle quoted message
Show quoted text
This is going to be way over the top for "nerd-ishism". Probably not the cup of tea for many........
The question I wanted to answer for myself was: "So.... if we can linearize the SSB generator signal chain so that it produces no appreciable intermodulation products, what would be the capabilities of the transmitter from the final mixer through the PA?" To
answer that question I needed a 45 MHz, two-tone signal with no IMD. If I have that and I inject it into the new filter followed by the mixer, then any IMD measured on the output is the result of that chain and not of the SSB generator and the bi-directional
amps. In essence, "what is the clean power capability of the PA section of the radio?"
The answer is "About 3 watts."
So I made a 45 MHz two-tone test signal by setting one output of my signal generator to 44.996000 MHz and a second to 44.997000.? Both signals are more or less centered on the 45 MHz filter passband and are the same level, 1 KHz apart. Those two signals are
put into a hybrid combiner and then through an attenuator to the injection point in the graphic, the input to the new filter. By varying the attenuator (with RV1 set for maximum) I could control the SSB output of the radio.?
IMD was measured at the 1 watt out level for reference and then the level increased until IMD measured -24 dB. The power level was different on each band but was a little over 3 watts for the lower three bands. For 10 meters (not shown) I was only able to get
2.1 watts at that IMD threshold.
For each band the purple trace shows intermodulation (refer to the table) when the injection level is set to produce one watt. The blue trace shows the power level in dBm that resulted from increasing the injection level until the IMD reached -24 dB.
So what does all of that mean? It means if we can completely clean up the SSB generator, and that looks reasonably possible, we have a 3 Watt SSB transceiver until someone tackles linearizing the transmit power chain. Lots of work to do!
WA8TOD
|
On Sun, Sep 16, 2018 at 07:10 AM, ajparent1/KB1GMX wrote: Allison
"It would be adequate...? the only problem I see is the mixer has an injection of about 5dbm so any level more than
about -2db has it fully saturated.? I'm saying in compression, I mean saturated."
Yup. That is kind of the point of the last post. This is not a 10 watt SSB radio. It is not a 5 watt SSB radio. It is a 1.5 - 2 watt SSB radio unless someone re-designs (and tests!) the RF chain. WA8TOD
|
Carl,
That was exactly the point. The RDxxx does not supply any majik and people keep applying that like it does.
Dropping them in the holes does nothing useful till you fix all the problems around them both before and after.
IF you do the things needed to fix the driver then that means input and out put transformers and then move
back to the predriver and repeat. and then back to q90 and repeat.
Henning put up a good block diagram of whats needed and the existing design can do that at 80m but at
20M barely, and at 10M not so much.
Allison
Allison
|
Warren,
Following is a table of injection levels required into the new 45 MHz filter in order to achieve 5 and 10 watts on each of the primary 5 bands. From this point forward would be the final transmitter mixer, pre-driver, driver, and PA. Power output was measured with no filters following the PA. It would seem the MMIC with a rated output of +13.5 dBm will be adequate to drive the transmitter to a full 10 watts on all bands. There is a real question as to whether or not that will be a good idea
It would be adequate...? the only problem I see is the mixer has an injection of about 5dbm so any level more than
about -2db has it fully saturated.? I'm saying in compression, I mean saturated.? Typical level there should not exceed
maybe -8 or more dbm and -10 or less for a reasonably clean output.??
Allison
|
Allison,?
I believe you missed the point of Iz oos's question.? The question was not of using the RD06HHF? for the finals, but for the driver.?
From my perspective, this was an entirely new question, not an oft repeated one.
73,
Carl,? K0MWC
|
Re: 2N2222A vs 2N3904 sourcing
Hi Jim,
Proof, it has ot make sense and consistant with my memory as i'm writer.? If you want to write a page you can.
As to the L8/L9? depends if its separate inductors as is or a bifilar inductor where the impedance can be much
lower as its balanced and DC only at the center plus its magnetically self canceling.
Allison
|
Re: 2N2222A vs 2N3904 sourcing
Tom,
I apologize.?
Too late too tired and moo many repeats.
|
This is going to be way over the top for "nerd-ishism". Probably not the cup of tea for many........ The question I wanted to answer for myself was: "So.... if we can linearize the SSB generator signal chain so that it produces no appreciable intermodulation products, what would be the capabilities of the transmitter from the final mixer through the PA?" To answer that question I needed a 45 MHz, two-tone signal with no IMD. If I have that and I inject it into the new filter followed by the mixer, then any IMD measured on the output is the result of that chain and not of the SSB generator and the bi-directional amps. In essence, "what is the clean power capability of the PA section of the radio?" The answer is "About 3 watts." So I made a 45 MHz two-tone test signal by setting one output of my signal generator to 44.996000 MHz and a second to 44.997000.? Both signals are more or less centered on the 45 MHz filter passband and are the same level, 1 KHz apart. Those two signals are put into a hybrid combiner and then through an attenuator to the injection point in the graphic, the input to the new filter. By varying the attenuator (with RV1 set for maximum) I could control the SSB output of the radio.? IMD was measured at the 1 watt out level for reference and then the level increased until IMD measured -24 dB. The power level was different on each band but was a little over 3 watts for the lower three bands. For 10 meters (not shown) I was only able to get 2.1 watts at that IMD threshold. For each band the purple trace shows intermodulation (refer to the table) when the injection level is set to produce one watt. The blue trace shows the power level in dBm that resulted from increasing the injection level until the IMD reached -24 dB. So what does all of that mean? It means if we can completely clean up the SSB generator, and that looks reasonably possible, we have a 3 Watt SSB transceiver until someone tackles linearizing the transmit power chain. Lots of work to do! WA8TOD 
|
Indeed ..Real near what I concluded on the old "back 'o the envelope trick"? ..
Jim
On Sunday, September 16, 2018, 1:42:26 AM PDT, Henning Weddig via Groups.Io <hweddig@...> wrote:
Allison,
at the beginning of my professional career 38 years ago I learned
that to sketch level diagrammes will help a lot to understand the
requirements of individual stages concerning gain, outputpower and
even give some insight in IMD/ P1dB. A visiual display gives more
information as jsut "bare numbers".
So I just sketeched such a block diagramme with the power levels
in between, see the attachment.
From this point of view only the first stage can be critical in
respect of gain-- although a MMIC - (or its equivalent using
discrete components can do ?) can easily get this gain at "LF".?
Henning Weddig
DK5LV
Am 16.09.2018 um 02:13 schrieb
ajparent1/KB1GMX:
Iz oos,
I was thinking whether a MMIC up to 10dbm and an RD06HHF as driver
like in the 10w qrp-pa-2008 might make sense
Whats that going to fix?? The transformers need to be reworked,
the prior stages in the amp are are poor.? And a MMIC
at 10dbm is near compression unless you get something ERA5 but
then at 75 ma you only get 20db of gain.
Think this out.? To get from the mixer to the antenna you
need 60db of power gain total!
Now if you used RD16hhf you get maybe 16db of gain more like 13
with feedback so the gain is flatter.
For 1/10th the cost the IRF510 can do that.? Either way 13DB.
So you need 47 db of gain call it 48 as a round number.
Your going to get that from three stages.
NOTE:
The last stage must supply more than 500mw from 3-30mhz.? [likely
gain at that power under 13db]
The stage before that will have to supply maybe 70 (max)
milliwatts. [likely best gain at that power maybe 15db]
The stage before that at most 3-4 milliwatts. [gains of 20db
possible at low power]
Getting high gain at low power is easier.? Hard at higher power.?
Its bad to run at max gain as then
the results are variable from unit to unit.
FYI most all fets will produce a lot of gain down at 3mhz, that's
not the problem, its getting the same
{or close} gain from 3 to 30Mhz.
I've repeated this more than a few times in the past two days. All
I hear is what about this fix.
Go look in the wiki, its spelled out.? Don't like it do the
research and think system level.
And please do not bring up a monoband radio that may or may not
perform well as its not
relevant or even meaningful.? Its a monoband it does not not have
to work at any other
frequency and every time I look its clear it will not work in a
3-30 case.
Allison
|
Re: 2N2222A vs 2N3904 sourcing
So just as a question and as (I assume) this is an ongoing project wiki ...How does one (a) offer solutions beyond those presented and what KIND of proof is required ... as a for instance L8 and L9 replacement/modification is NOT mentioned ...Curious given that I recently duplicated someone elses effort (from may or june ?) ...MEASURE the inductance of your exhisting L8,L9 ...Those are there to keep RF off the PA power lead ...Keeping RF off the power lead to the REST of the radio ...as designed, the inductance is supposed to be 60 uH ..as measured on mine 20 uH AFTER DC current alterted the u of the ferrite ...10 ma current altered them to 10 uH ...guess what 100 mA current (standing bias current on the PA MOSFETs ) does to the inductance.... Writing and maintaining wiki's is at best a thankless task ...How can it be made easier for you?
Jim
..
On Saturday, September 15, 2018, 8:02:02 PM PDT, ajparent1/KB1GMX <kb1gmx@...> wrote:
For Pete and repete:
Known problems and this was as of mid June.? IF your not aware of this log in to the form web version and read.
History is amazing and repeated, repeated, repeated, repeated, repeated, repeated, repeated, repeated, repeated, repeated, repeated.
Amplifier power vs frequency is poor.
IMD extremely poor for power over 2W.
Output Low pass filters, they don't. board layout.
Spurs, unwanted and unfiltered mix products that show above 20mhz.
Band pass filter after the mixers not existant.??
Carrier leakage that is not related to modulator or 12mhz filter.? Again board layout.
Nothing new here.
|
Joel Caulkins/N6ALT
Raj vu2zap