Re Jerry ..Donno... need to start making some transconductance measurements too anyway ?? I was looking at the gate voltage on the mosfets ..4 V at 100 ma bias ...More swing that +/- 4 volts and distortion rears its ugly head (more like 3 volts, but who's counting?
Jim,
Also the transconductance of the mosfets is around .2 S? fairly high.? You need to bias for about 100-300ma per device
and the Drain load must be low enough to stay on the load line for the device.? Keep in mind from about 2 to 6V gate voltage
you see at DC near 0 current to over 4A so it not like you need more than a volt maybe 2 P-P especially push pull.? That
requires a load at the output that does not allow the drain to swing to zero or 12V at full power out (linear load line).? To
get that?we need a low impedance load and translation to 50 ohms.? Typical load say 25 ohms and 20V PP swing
[why not 24V?? Never go to zero or max V saturation or voltage clipping.] barely?nets 10-11W. We move lower to
about 12.5ohms and we get 22w but we are still near saturation but at 10W we are not and the result is much
cleaner.? So we need to develop power from the mosfet with minimal gate swing and maximum safe drain
current.? Both happen with about 2A drain current and conditions that keep the gate at less than 4V peak
(at the device.).? ?For push pull that means each mosfet can go to near zero current point (best not to for IMD)
but not near the power supply max (also IMD).? Its also why parallel internal architecture hexfet, LDmos, Vmos,
Tmos] or hybrids of multiple discrete are used.? IRF510 is hexfet, essentially many parallel little MOSfets in parallel.
MOSFET is a voltage controlled current generator.? That's the general model where the current is Vgs* GM.
So when I see people getting high power I know one of two things either they have transformer loading the
drains to under 15ohms or they clipping heavily.? The IRF510 (is not alone with this) however runs better
at 20-28V where much more power is possible without exceeding peak current.? FYI the current stock
transformer barely has the drains Push-pull at 50 ohms which is way too high.? It should be a 2:3?
[maybe 4:6 turns] where 2:3 ratio nets about 22ohm drain load and it likely to have a chance at
10W IMD?being acceptable at 12V.??
Back of the napkin for current stock transformer is 50ohms and we push to the edge at 12V
(22V P-P) and we get... 9.68W peak (6.7rms)!? ? Its ohms law, power law, and that simple.
More than that it will be very unclean.
Things you have to do to generate /clean/ power.
Allison
|
Surprised by generous and kind act
#bitx40
Today I was greatly and pleasantly surprised.? I'd mentioned that I was thinking about converting my BITX20 to 40 meters, and I'd related some stories of frustration with the local ham community (generosity and willingness to help others seems to be a rarity in this area).
When I got our mail, there was a package in it - with a BITX40!? The housing it was in is very attractive as far as my wife and I are concerned - and perfect for our car.? The neat part was that you could see the circuitry in it - and the case also would lend itself to portable operation!
I won't embarrass the sender, but I will say it came from another state.? For many years, I've insisted that there was no such thing as Karma... but this suggests otherwise!
To the sender (not named) - a big Thank you! from both of us, and (laugh) no worry about "paying forward" - we do things all the time like that, whenever we can (we may not have much money, but we do have skills and knowledge we share).? Generosity and helping others is not a big price for gifts like this - in our opinion!
|
Allison
can the "wimpy ground paths" part be helped by overlaying fairly heavy wires in? parallel with existing wimpy grounds?
Arv _._
On Sun, Sep 9, 2018 at 9:06 AM, ajparent1/KB1GMX <kb1gmx@...> wrote: Jim,
Yes the big thing is developing a voltage across the capacitive gate, takes power to do that.
Also the impedance of the source (driver) has to be low at low HF as at 3mhz the gate looks
like about 220ohms reactive with whatever resistive values are applied across that.
At 30mhz its 22 ohms plus that same set if Rs and Ls.? A crappy driver can push 3mhz
but at 30 the layout and current for less than 22 ohms becomes significant.
FYI I have 4 of the the QSX amps and several I've built proving flat to 30mhz or beyond
with IRF510 is very doable.? However the board layout has to look like a VHF layout not
long thin traces with wimpy ground paths.? Also the drivers have to be able to push a
bit of power optimally about a watt.? While delivering the same gain at 3 and 30mhz.
All the at is to help you understand why it doesn't.? IF you fix drivers,, transformers,
and biasing, your still stuck with board layout.? That not to say it can't be improved greatly.
Work I did said 10-12W at 80m and 5W at 10M is very doable but to take advanage of
that you have to clean? up spurs and low pass at a minimum.
The problem was I was peeling a very pungent onion.
Allison
|
Jim,
Yes the big thing is developing a voltage across the capacitive gate, takes power to do that.
Also the impedance of the source (driver) has to be low at low HF as at 3mhz the gate looks
like about 220ohms reactive with whatever resistive values are applied across that.
At 30mhz its 22 ohms plus that same set if Rs and Ls.? A crappy driver can push 3mhz
but at 30 the layout and current for less than 22 ohms becomes significant.
FYI I have 4 of the the QSX amps and several I've built proving flat to 30mhz or beyond
with IRF510 is very doable.? However the board layout has to look like a VHF layout not
long thin traces with wimpy ground paths.? Also the drivers have to be able to push a
bit of power optimally about a watt.? While delivering the same gain at 3 and 30mhz.
All the at is to help you understand why it doesn't.? IF you fix drivers,, transformers,
and biasing, your still stuck with board layout.? That not to say it can't be improved greatly.
Work I did said 10-12W at 80m and 5W at 10M is very doable but to take advanage of
that you have to clean? up spurs and low pass at a minimum.
The problem was I was peeling a very pungent onion.
Allison
|
There have been many posts regarding end fed?antennas. I hope these posts were helpful.? My favorite portable antenna are?various loops, either full-wave?loops,?quads, or Miniature Magnetic Loop Antennas. Why do I like loops:
- Because they are a closed circuit resulting in very low noise, far less than a dipole, verticle, or end-fed wire (NOTE: You can't work what you can't hear)
- Because the can be polarized vertically for low-angle radiation, horizontally, or even a hybrid of these
- Because they are easy to build and match to a 50-ohm coax
But for portable operation, nothing beats a Minature Magnetic Loop antenna. This is because it is light-weight, directional, and can literally sit on the ground and still work perfectly well! I wrote an article about Miniature Magnetic Loop Antennas for the Elecraft.com team way back in 1999 when they were not yet so popular (i.e., I was a builder of the early K2 and beta tester for the 100 watt?linear). Check it out at?. You can build these antennas with materials from any home improvement store. I hope they inspire some antenna building from among the uBitx?building family. David Posthuma, WD8PUO
|
I appreciate all the support I have received and thanks. I will work on some of these but the whip antenna will not work I'll the magnetic loop antenna.? Thanks for sharing.
Richard Chan?
On Sun, Sep 9, 2018, 9:24 AM Warren Allgyer, < allgyer@...> wrote: Daniel
You are receiving varied and solid advice in answer to your antenna question and I hope it helps.
Aside from that, as a new ham, I urge you to read through this board and be sure you understand some of the issues with the uBitx as it is delivered before you put it on the air. If you choose to go ahead and use the unmodified radio on the air, and many have and continue to do so, at least be sure you understand the ramifications of that decision.
Please ask questions on anything you do not understand and you will get immediate and enthusiastic response. Welcome to our hobby.
WA8TOD
|
Re: Lcd nextion generating qrm
with 5 ohms and with 100 brightness level , is there no qrm, Junior !!!\ ?Congrats. sarma ?vu3zmv
On Sun, Sep 9, 2018 at 7:44 PM J¨²nior PY2ADA via Groups.Io <aaj280174= [email protected]> wrote: The nextion 2.8 "display is generating qrm, when I adjust the brightness control to less than 100 I have qrm at the reception, if it is 100 without problems. I installed a 5 ohm resistor in series with the power of the display, it was nice the brightness of the lcd.
?
73 de J¨²nior
|
Lcd nextion generating qrm
The nextion 2.8 "display is generating qrm, when I adjust the brightness control to less than 100 I have qrm at the reception, if it is 100 without problems. I installed a 5 ohm resistor in series with the power of the display, it was nice the brightness of the lcd.
?
73 de J¨²nior
|
Yep?? all true ... minor circuit/device constraints
Jim
On Sunday, September 9, 2018, 6:30:33 AM PDT, ajparent1/KB1GMX <kb1gmx@...> wrote:
jim,
input C for irf510 is in the 150-180pf range, drain to gate about 20pf.
Classic case for Triode miller effect.? Feed impendance must be low of you get the?
output voltage divided across the input 20pf and 180pf as negative feedback.
You seen this will all mosfets only the degree of it varies.
FYI the drain to gate C varies with DC drain voltage for all mosfets.
The circuit for the driver has issues:
Output impedance is not constant or low enough.
The stage voltage saturates (high emitter resistor)?
THe stage current limits due to high Collector L(transformer).
The 3904s loose gain with increasing frequency.
Allison
|
And a correction to clear up any confusion: I wrote: " Clearly we need these transformers to be impedance matched in the circuit." It should have read: "Clearly we need these FILTERS to be impedance matched in the circuit". Wish I could blame autocorrect but it was lack of proofreading.
WA8TOD
|
John
It is 7:2 or a 3.5:1 ratio. The impedance ration is the square of the turns ratio.
I honestly don't know the answer to your question about paralleling filters. I guess you could try it..... but some bad karma and icky things might result. I would have to test it to be sure. If you are going to use a transformer anyway I would just adjust the turns ratio and use one.
WA8TOD
|
I don't think paralleling the filters will impact the nominal input impedance of each filter. If you have a SA and have something to generate a 45Mhz signal you could sweep the filters in parallel and see what it does. tim ab0wr On Sat, 08 Sep 2018 12:41:16 -0700 "John KC9OJV" <greusel@...> wrote: I rolled the dice on a pair of filters and received Toyocom 45E1A9Fs
which show up here
as
two pole filters at 1200 ohm termination impedance. If I parallel
those will I bring that impedance down to 600 and should that be a
goal? Should I try both in parallel? I do have a modest spectrum
analyzer to take a look at the results with.
John
KC9OJV
|
Should the pad after the balanced modulator be changed to add more loss? That should be easy to do. tim ab0wr On Sat, 08 Sep 2018 12:39:46 -0700 "ajparent1/KB1GMX" <kb1gmx@...> wrote: Warren,
1) 120 mVrms is far above the audio level that caused unacceptable
IMD before the filter mod. IMD must be checked and the audio levels
adjusted to make it acceptable.
The problem is the balanced mod easily takes that but the gain after
it means the 2nd mixer gets hit with a very high level.? With that
the output of that mixer is high and that leads to the next mixer
getting overloaded.? There is too much gain after the balanced mod.
This is why the balanced mod needs such a low input.
Allison
|
Warren,
Is that a 3.5:1 winding ratio or 3:5? Also to my previous question- does a 1200 ohm impedance filter become 600 ohm if you parallel two of them? I realize it narrows the BW but I've never seen information on the effect on termination resistance.
John
KC9OJV
|
jim,
input C for irf510 is in the 150-180pf range, drain to gate about 20pf.
Classic case for Triode miller effect.? Feed impendance must be low of you get the?
output voltage divided across the input 20pf and 180pf as negative feedback.
You seen this will all mosfets only the degree of it varies.
FYI the drain to gate C varies with DC drain voltage for all mosfets.
The circuit for the driver has issues:
Output impedance is not constant or low enough.
The stage voltage saturates (high emitter resistor)?
THe stage current limits due to high Collector L(transformer).
The 3904s loose gain with increasing frequency.
Allison
|
Daniel
You are receiving varied and solid advice in answer to your antenna question and I hope it helps.
Aside from that, as a new ham, I urge you to read through this board and be sure you understand some of the issues with the uBitx as it is delivered before you put it on the air. If you choose to go ahead and use the unmodified radio on the air, and many have and continue to do so, at least be sure you understand the ramifications of that decision.
Please ask questions on anything you do not understand and you will get immediate and enthusiastic response. Welcome to our hobby.
WA8TOD
|
[email protected]Also the flower pot thing is a poor design and a only qualifies as a? better than rubber duck antenna? for 2M portable.? As to 6M vertical dipole is easier and only 10ft(2.8M) tall. However for HF you need what you can do allowing for space. FYI A cb atntenna is way to short for 20 (barely .26 wave at about 16.3ft). While anything is better than nothing generally short antennas are inefficient and need a good ground to work against. Allison
|
More data for those interested. The first is a comparison of the onboard 45 MHz filter as supplied, with the eBay filter I have listed previously. This image was taken with the filters simply connected as is between the 50 ohm output of the spectrum analyzer tracking generator and its 50 ohm input. The blue trace is the filter supplied and the purple is the eBay filter. The second image uses the same two filters but with a 2:7 turn binocular transformer on the input and a 7:2 on the output. The transformers are identical with the 7 turn windings connected to the filters and the 2 turns to the input and output. 7:2 is a 3.5 turns ratio which gives a 12:1 impedance transformation, matching the 50 ohm in and out to the nominal 600 ohm filter. Clearly we need these transformers to be impedance matched in the circuit. It is also interesting to me how the center frequency shifts between the matched and mis-matched conditions.  
|
You have not left the trailing part of the message you are replying so I am not sure what
you want me share ?
Raj
toggle quoted message
Show quoted text
At 09-09-18, you wrote: Actually - that makes a lot of sense Raj b/c the two ends of the filter seem to be capacitively coupled internally according to the spec sheet. It would be great to see the ultimate results of that. Could you share those results?
|
Kees,
I tried with and without R27. With R27 the 45MHz signal @ mixer drops a lot and only 6db without.
But the power out drops very little with or without. R27 reduces the spurs a little.
I had suggested removing the C22 and placing it there. I learnt a lesson the hard way.. the xtal filter
pin shorted to ground and the transistor failed !!! So the cap came back. I chose to place a 120pF
cap at C22 as that is enough to pass 45MHz without much attenuation. This RC roll off at 3 MHz
will be at least 20db.. I propose to replace the other coupling caps and emitter bypass with a
calculated lower value to attenuate as much below 30MHz as possible.
I think it will be helpful to have the 45M IF stage respond less to HF so that its pickup of the output
signal and amping it is further reduced. This the thought that brought out filter as quick solution!
The harmonics below 20Mhz is worrying me.
Raj
Ps Please leave in the parts that you are replying to! I read mails in Eudora and not on the web.
toggle quoted message
Show quoted text
At 08-09-18, you wrote: Raj,
Doesn't leaving R27 in place give you better impedance matching ? Did you try cutting the trace in between and just putting the filter between C22 and R27 ?
73 Kees K5BCQ
|
David Posthuma
J¨²nior PY2ADA
Raj vu2zap