Re: 2N2222A vs 2N3904 sourcing
Tom,
Understood but the heavy mosfet device needs higher operating current to get over the high input
and output C.??For the driver its a maybe, but you need to heatsink it as its going to run at 100-250ma
for IMD and gain.? IRF510 at 50ma is still asleep.
A BS170 or 2n7000 (pair) will do a watt push pull, and that is what we need.??
It would be like using a 2n3055 (15 amp TO3) for an audio preamp!??
Allison
|
Terry,
while looking at it again, I realised that you'd also need to change lines 93 and 94.
These are the initial frequencies of VFO A and B. I suggest to set these to a frequency somewhere in the 80m band.
For example to 3600 kHz:
93 ?? unsigned long vfoA = 3600000UL;????????????????? // frequency of VFO A
94 ?? unsigned long vfoB = 3600000UL;????????????????? // frequency of VFO B
Also you need to change line 1734:
1734 ??? u.scan_start_freq = constrain(knob / 2 + 3500 + shift, u.LOWEST_FREQ / 1000, u.HIGHEST_FREQ / 1000);
This will allow you to set the upper and lower scan limits in the 80m band via the Function Button (SETTINGS menu).
After making these changes, remember to reload the default settings (keep Function Button pressed during power on).
73 Allard PE1NWL
|
Re: My list of Firmware, Executables and Nextion Screens for the ubitx
Marty - I seem to be having a problem with the ubitx_35.tft file for v1.1 seems it is producing an image for a smaller screen size. I think things have got mixed up. but I tossed my earlier version.
|
On Sat, Sep 15, 2018 at 10:34 AM, Warren Allgyer wrote:
I don't have a lot of experience in this but my gut tells me pushing interstage gains to the 1 dB compression point is too much. When I did that with the MMIC boards the IMD jumped dramatically at even 0.5 dB compression and was borderline unacceptable at 1.0 dB. 1 dB compression, in my experience, generally makes my goal of -30 dBc third/fifth order IMD unattainable. I would not plan for any measurable compression for interstage blocks..... but that is me.
Absolutely correct.? Once you hit gain compression even a little you going to get undesired products as your in nonlinear land.
That is true for any amplifying device.
The mar8 at 30db is a possible in the q90 slot.? It hits the -1db gain at 12dbm so it has more than enough dynamic range for q90
or as 45mhz IF.? ? Watch their noise figures some are very unspectacular.
The ROT I use is 3db (power) headroom minimum.? Most MIMICs 10dbm out is about it unless you want to pay! Many barely do that.?
For example the ERA5 (at 75ma per) is -1db compression by 18dbm so back off 3db you seeing maybe 15dBm or about 31mW,
not enough for the predriver slot.
Where do I get that 3db... two equal tones (power).? More is better, that is the minimum.
A 2n2222A biased at 20ma or better a 2n3866 biased for 40ma and 11db of gain will hit about 22-23dbm (over 150mW!)
without stress and do 3-30MHz.? That would drive a heavier stage easily.? IF that one also did 10db you are in the 1W
region comfortably.? We do not need four transistors to do that as one 5109 or maybe even 2n3866 can do that.
For efficiency two devices at 10-12ma idle current push pull will easily hit the 2W mark (plenty of headroom).
Allison
|
The MCL ERA-3SM draws 35 ma at 3.3 volts. It will be deployed with an onboard 3.3 volt regulator.
The amplifier at Q20-22 that the MMIC will replace draws 34.7 ma.
The layout is planned to be implemented on?50 Ohms Co-Planar Wave-Guide Over Ground Transmission Line boards that are rated to 1 GHz. Since they operated at different frequencies than do the high power PAs, the only feedback we need to be concerned about for preventing oscillation is the output to input coupling across this board and that is negligible. This could be an issue but tests using the MMICs I own show it not to be the case. Only tests of the final mod will tell for sure.
Not sure I see the issue with this MMIC.
WA8TOD
|
Hi Terry
The scan mode's lower and upper frequency limits are hard coded at lines 71 and 72.
You could modify these values as well for 80m band.
(Note that these values are kHz)
So? for example you could use
71??? SCAN_START 3500
72??? SCAN_STOP 3600
73 Allard PE1NWL
|
The BFR106 shows Vceo as just 16v. That is kind of low for our work. At 1/2 watt into a 50 ohm load, The voltage van swing +/- 7 volts, which is about 14 volts. We are almost at the breakdown point. How has your experience been??
- f
toggle quoted messageFor Q90 BFR106 is fine for the predriver and driver you need something else. As is Q90 sits there with 12V supply to the box and about 2V on the emitter and about .5V drop to r84. so its barely seeing 10V.? Also 2n5179 or the mmbt5179 would work. As to using it for power beyond a few milliwatts no,? Wrong part.? BUT the system is wired to get 18+db of gain from q90 and the 3904 can't do that past maybe 10mhz. For Q10-12 and Q20-22 its also fine as its not running much power and the amplitudes are under a volt.? If there were concern r10 and 28 can be 270 ohms and 9.1v zener to ground to regulate the voltage. All stage here are low level milliwatt maximum. Allison
|
Iz oos,
The big problem with MMICS is power...? The devices we are talking about eat 50-60ma each
and prefer to be run from a fixed voltage.? For a say 8V source that on the order of 400mW
each for receive or transmit.? Not friendly for battery powered systems.
The other price is layout has to be good or they will oscillate.
Allison
|
Thanks Warren, that seems a very good and promising news. MMICS really simplify circuits and will probably improve the overload specs of the receiver too. If ground loops might be possibly a problem simple 1:1 isolating transformers might be used. I look forward to seeing this tricycle better than some Ford in terms of IMD!
Il 15/set/2018 17:19, "Warren Allgyer" < allgyer@...> ha scritto: As an aside, I tested this particular MMIC board for its susceptibility to oscillation. I connected a 60 dB attenuator between the input and output while I fed a signal to the input. I gradually reduced the attenuation one dB at a time until oscillation was visible on a 1.5 GHz spectrum. The amplifier broke into oscillation very reliably when the attenuation was stepped below -14 dB. Higher than that and it was completely stable.
In the course of these two tests the board was hanging unshielded about 3 inches from the PA heat sinks and the output was a little over 4 watts. In both cases I saw no indication of oscillation.
WA8TOD
|
DIFX Transceiver.? Does it compliment or compete with BITX.? The design sounds a lot
like a BITX.? Might be interesting to see how the DIFX performs for spurs, harmonics, and IMD relative
to a BITX.
Arv? K7HKL _._
|
Wouldn't it be simpler to just add an AC feedback RC network ? A series capacitor (100nF) and a series resistor (680 ohm to 1K?) between the "base" end of R21 to the "collector" end of R21 on Q20 as discussed in several messages? earlier (Wes' design) ? Then do the same to Q40. Make R21 4.7K.
Make up any gain loss with the 2N2222A PA driver pair replacement with 10 ohm emitter resistors.
73 Kees K5BCQ
|
Allison, The BFR106 shows Vceo as just 16v. That is kind of low for our work. At 1/2 watt into a 50 ohm load, The voltage van swing +/- 7 volts, which is about 14 volts. We are almost at the breakdown point. How has your experience been?? - f
On Sat, 15 Sep 2018, 23:04 Warren Allgyer, < allgyer@...> wrote: I don't have a lot of experience in this but my gut tells me pushing interstage gains to the 1 dB compression point is too much. When I did that with the MMIC boards the IMD jumped dramatically at even 0.5 dB compression and was borderline unacceptable at 1.0 dB. 1 dB compression, in my experience, generally makes my goal of -30 dBc third/fifth order IMD unattainable. I would not plan for any measurable compression for interstage blocks..... but that is me.
WA8TOD
|
Looked at that and the output stage is the Softrock-Ensamble amp a proven
design for 3-30.
Also Hans QRPlabs used the BS170 (*2 P-P) as driver for IRF510s to get near flat 3-30mhz.
I used a 2n5109 single ended driving IRF510s also remarkably flat.
The real trick is the driver as if that not flat the amp is then degraded.
Thats a variation of Miller calculation for triodes from way back and applied
to modern devices.? The rough outline of the calculation is input C (c gate to source)
and the feedback C? (c gate to drain) and the other effector is gain.? more gain
more the effect of feedback C and of course the frequency.? ?Its why pushing
the IRF510 or any MOSFET for more gain at high impedance (input and output)
results in poor gain with increasing frequency or even instability.
The use of a bifiler choke at the output at 1W is likely of little value.
It does allow for other transformer possibilities that might not provide
a convenient center tap such as transmission line transformers.
In short you want the output as a current source driving a low impedance load
and the input as a low impedance voltage driver.? ?Then the MOSFET gain
and usable frequency range is limited by other factors (lead length and
operating parameters).? As always the trade is gain vs bandwidth.
Allison
|
Re: 2N2222A vs 2N3904 sourcing
Hi Allison,
I was not suggesting that the stage be run at 7 watts. Only one of the issues is the marginal drive ability of the current TX chain to properly drive the finals of the uBITX. The fact that the BITX40 can drive 7 watts at the equivalent of the pre-driver for the uBITX, would suggest that using a mosfet in a similar configuration to the BITX40 may provide sufficient gain so the rest of the chain can be run at more relaxed levels. The IRF150s are cheap and already in use, so may be as good of a choice as any other mosfet in that spot.?
Tom, wb6b
|
I don't have a lot of experience in this but my gut tells me pushing interstage gains to the 1 dB compression point is too much. When I did that with the MMIC boards the IMD jumped dramatically at even 0.5 dB compression and was borderline unacceptable at 1.0 dB. 1 dB compression, in my experience, generally makes my goal of -30 dBc third/fifth order IMD unattainable. I would not plan for any measurable compression for interstage blocks..... but that is me.
WA8TOD
|
Re: 2N2222A vs 2N3904 sourcing
On Sat, Sep 15, 2018 at 08:54 AM, Tom, wb6b wrote:
At that final stage in the BITX40 it can put out 7 watts. Would there be any advantage in reworking the uBITX so that the stage with?Q92-Q93-Q96-Q97 was replaced with a IRF510 and the transformers and bias rewired to be like the BITX40 final? Then let that reworked stage drive the Push-Pull IRF510 PA stage of the uBITX.
FYI all of the images I have of the schematics for ubitx and Bitx40 are sufficient resolution that its my eyes not the printed image.
No.? IT like installing a 400CI engine for the wiper motor.
For the q90 a better device with higher FT.
For the predriver and driver a better device might be BS170 and maybe others.
I'm keeping track its not the 10th time repeated here....
you need about .5W from the driver the IRF510s (to 10W).
you need about .07W to drive the driver
and you need a milliwatt or three to drive that stage.
Not much of that requires big devices with heatsinks.
Allison
|
Dunno. I do not plan to use that amp. On the way are 10 pieces MCL ERA-3SM MMIC and 8 pieces of SV1AFN transmission line PC experimenter board (??). The board will fit the space currently occupied by the transmit bi-directional amps. There is enough space on the experimenter board to mount the MMICs, isolation caps, and, if needed, an attenuator. If it works when tested then I would plan to scrape the current amp components off the uBitx and seat the boards in their place.
We shall see. Lot's of things have to work out, including proving the 70 cent MMICs are genuine, for this to happen.
WA8TOD
|
Yeah, got that. Perhaps, an inductive load would have done better.?
On Sat, 15 Sep 2018, 22:53 ajparent1/KB1GMX, < kb1gmx@...> wrote: Farhan,
the gain tested correct with a sag in gain attributable to frequency but the -1db compression
was at a lower than expected level at 45mhz than at 12mhz.? About the same difference as
the actual measured gain or about in the test case breadboard 6db.
What I got was 19.6 (almost 20db) at 12mhz, at 45 it was down to 13.9db and the
-1DB point tracked that.? ?Also the S11 degraded? significantly for return loss of?
16db at 12mhz dropping to less than 10db at 45mhz.??
With 2n2369 those improved both gain, flatness, and return loss.
Went back to 3904s and Wes's circuti and set the gain to 15db and saw near flat?
gain (-2db at 45mhz), improved -1db compression and better than 14db return
loss at 45mhz.? It was still better at 12mhz only by a 2db.?
Simple answer your pushing for too much gain from a device that cannot do it.
also higher operating current had a negative effect on compression point as
it was not linear.
The 3904 is a great utility device.? It has two limitations 300mhz FT but at
low gain its fine.? The other is at more than 10mA its HFE falls, this means
increasing the current does not help its gain and may lower its compression point.
Allison
|
Folks,
by lurking thorugh the internet I found an interesting post from UT5FFR. He builds up a 1 W PA from 1 to 30 MHz using BC548, 2N2222 and BS170.
Interesting: he is using the formulae to calculate input impedances etc fo the BS170 using the well known (?) Philips app notes!
One personal remark: I always had thes eapp note in mind!
One thing i would modify: use a bifilar choke at the aoutput of the push pull stages to supply the DC voltage instead of the center tapped transformer.
Who is willing to thest this circuitry and in addtion calculate the post amp with the IRF510?
As Allison already stated and proven in the calculation for UT5FFR ist the necessity to drive the FET?s with a low impedance in order to compensate for the big gate to source and feedback cpaacitances!
Henning Weddig
DK5LV
|
Farhan,
the gain tested correct with a sag in gain attributable to frequency but the -1db compression
was at a lower than expected level at 45mhz than at 12mhz.? About the same difference as
the actual measured gain or about in the test case breadboard 6db.
What I got was 19.6 (almost 20db) at 12mhz, at 45 it was down to 13.9db and the
-1DB point tracked that.? ?Also the S11 degraded? significantly for return loss of?
16db at 12mhz dropping to less than 10db at 45mhz.??
With 2n2369 those improved both gain, flatness, and return loss.
Went back to 3904s and Wes's circuti and set the gain to 15db and saw near flat?
gain (-2db at 45mhz), improved -1db compression and better than 14db return
loss at 45mhz.? It was still better at 12mhz only by a 2db.?
Simple answer your pushing for too much gain from a device that cannot do it.
also higher operating current had a negative effect on compression point as
it was not linear.
The 3904 is a great utility device.? It has two limitations 300mhz FT but at
low gain its fine.? The other is at more than 10mA its HFE falls, this means
increasing the current does not help its gain and may lower its compression point.
Allison
|
Allard PE1NWL
