开云体育

The LTSpice schematic file to simulate the Bitx40 final is also up there, not just the pdf:?
? ? ?/g/BITX20/files/KE7ER/bitx40pa.asc

I'll have to think about this worst case 180v pk-pk of yours, I may learn something here.
Mostly caught up on sleep during the class that covered transmisssion lines oh so many years ago,
my interest by then was in things digital.

Here's the LTSpice simulation I did of the Bitx40 final one year ago:?
? ? ?/g/BITX20/files/KE7ER/bitx40pa.pdf

I cooked up a simulation that covered the rest of the board in LTSpice, not posted there.
The mixers seemed to work as expected, but I couldn't get anything to go through the crystal filter.
Best guess is that there are some LTSpice settings that need to be adjusted to get crystals to behave,
didn't follow through on that to figure it out.

Jerry

?

Hi again Jerry;

?

It looks like you caught me, or I failed to catch myself on this, but I wasn't clear on my comment: “At 20W output, the switch could theoretically see 180V P/P when presented with a perfect open circuit infinite VSWR”. In my defense however; I did follow that with “I'm not entirely certain that I'm simulating the operating environment realistically,...”

?

So here's my line of thinking on this... In order to experience 180V P/P at the source of a 20W transmitter, the fundamental and reflected waves must add algebraically, be simultaneously present at the transmitter's output port, both at the exact same amplitude, and be precisely aligned in phase. Assuming a zero impedance source driving a 50 ohm transmission line as it's load, and by applying Ohm's law (at time zero... when the load “looks” like it truly is 50 ohms), 20 Watts output at the source yields a source voltage of 90 Vpp. This forward voltage wave must now travel precisely 360 degrees in time before presenting itself at the source in order to achieve the feat of actually performing the said “in phase” algebraic addition. It must also do this in a lossless manner in order that the magnitude of 180 Vpp can be realized.

?

One way to induce this condition is to terminate the transmission line in a short circuit at 90 degrees, forcing the forward voltage to zero, which in turn reflects the wave back to the source, and arriving there as a reflected voltage maxima. Assuming a lossless transmission line and zero ohm short, this would be 90 Vpp, and also in phase with the source.

?

Any lossless 90 degree phase shift network terminated in a short is an alternate means of inducing an infinite VSWR condition. Alternating shorts and opens at 90 degree intervals perform this task as well... assuming this can be done in a lossless manner... but I think you get the idea.

?

What I struggle with however; is how to define a realistic mismatch model at the transmitter output. I've not yet been made aware of an infinite (open circuit) mismatch condition that would induce the theoretical doubling of the source voltage at the source. However; in the case where a transformer is employed, and an inductively coupled winding is not terminated, the flyback voltage induced in the unterminated winding can be many times higher than the induced voltage due to Q multiplication. At the same time, I believe this too is limited by the loading impedance of the primary, and any reflected energy in this case is sensed at the source as a reflected current rather than a voltage. Perhaps the decoupling inductor in the collector/drain does this, but I've not attempted it... Something fun to try when I feel like letting the smoke out of an otherwise good transistor. LOL!

?

I routinely shunt the collector/drain load of my amps with 470 ohms to reduce the Q and minimize any tendency to build up to a dangerous flyback potential. 1K as you suggest I'm sure is probably just as effective.

?

Acknowledging the fact that I'm repeating myself, I'm doing so intentionally... it being an important caveat to convey:

?

“I'm not entirely certain that I'm simulating the operating environment realistically,...”

?

Caveat aside, some of what I'm observing in my attempts at simulating this without the ability to sanity check any of it on my bench, could be gotcha's on what I'm doing here. I may not be simulating precisely what I actually have working on the bench and tested, I may be making incorrect assumptions, or any of a number of other inconsistencies between what I'm playing with here in simulation, and what I have working and measured is suspect at the moment, and may not get sorted out until I return home. I sometimes get bogus results between LTSpice runs also that I haven't yet been able to resolve as well. There's also the risk that my simulations are valid, and they're simply telling me something I don't wish to hear. <sigh!>

?

?

At the moment; my confidence is high based on linear simulation results and bench measurements that show good agreement; but my inability to mirror those results in transient simulations is a bit disturbing. As I give this some thought however; some of what I'm observing may be due to limitations in the linear device models, and the fact that I'm driving the switch at a peak to peak voltage many times greater than the 13.8V switch bias supply. I was on the cusp of testing this latter point just before leaving for vacation; although a preliminary quick first look didn't reveal anything that caught my attention at that time. e.g. My simulations suggest the series connected 2N7000 dissipates 7W for a 20W output transmitter. I would have detected that much power with my calibrated finger thermometer. :-) I also got sidetracked when I discovered Allard's code wasn't yet supporting Full QSK and ran out of time to do more robust testing at the BITX output power level.

?

?

I obtained excellent agreement between measured and simulated linear performance before leaving for vacation. As I discovered in the audio stages, a transient analysis can reveal other artifacts that can become issues if not properly addressed, such as circuit stability, noise, and even ground loops if properly and adequately modeled. Something similar may be at play here with my RF switch as well. For example; the enhancement of switch isolation through output matching isn't predicted in simulation. This could be a good thing or a bad thing, and its an important detail to understand before assuming switch survival is the end goal of the design. Since the switch is also a direct feedback path to the input of the transmitter chain, it's isolation may not be sufficient to prevent the transmitter chain from breaking into self oscillation. This is my biggest concern, and the first item I hope to put to rest when I have Full QSK code to work with. If I identify this as an issue, it could well send me back to the drawing board with respect to the RF switching part of this feature. This may well be the case, and perhaps my transient analysis results are telling me what I don't want to know and I am just stubbornly keeping myself in denial. :-(

?

LTSpice allows one to make differential measurements, and thus verify whether or not device limits are being such as Vgs, and Vds, are being exceeded. It also has the utility to assess the Pd of components that you feel may be at risk of excess thermal stress. It's an incredibly powerful tool, and the only cost is the effort expended to learn how to use it. I think I've simulated everything but the TX chain in the BITX, and I may end up having to do that as well if what I'm seeing in my RF switch analysis proves to be valid. It's as much a troubleshooting tool as it is as a design tool.

?

72/73

?

Gary, N3GO

?

On Mon, Jan 29, 2018 at 07:09 pm, Jerry Gaffke wrote:

?

Erik,

On the hfsignals wireup page? ?? ? it says:??
??DANGER : If you power up the ?BITX without the pull-up resistor the ?BITX can randomly go into CW transmit

Do you have the 4.7k pullup to 5v on the CW key pin?
If not, it could be going into transmit mode.

Erik,

I suspect a relay. k3 mostly..

Check the continuity between M1 and M2. It should be short on RX and open on TX.

Check if you have 12V on the TDA2822 pin 2

Raj

Jerry makes an excellent point which should be kept in mind.

RV1 on the BITX40 is for adjusting the voltage level on the gates
of the IRF510. This should never go beyond about 3v or so.
Rv2 and RV3 are used on the uBITX for the same thing.

It is a bit unfortunate in the nomenclature as it is easily confused.

The corresponding device on the BITX40 to the RV1 on the uBITX
is R136. These only should be fiddled with. Once the gate voltage
on the IRF510 is set, forget about adjusting it.

john
AD5YE

--
Best 73
KD8CEC / Ph.D ian lee
kd8cec@...
(my blog)

Twiddling RV1 is for uBitx owners, it's the pot that determines how hard to drive the power amp on transmit.
On the Bitx40, RV1 sets the bias voltage to the IRF510 gate, definitely not something to mindlessly twiddle
or you will smoke the IRF510,? It's R136 on the Bitx40 that sets the drive level??

On the uBitx, it is RV2 and RV3 that set the bias voltage to the IRF510 gates.
Should be adjusted for around 100ma per IRF510.
A wee bit too far and suddenly you have several amps through the IRF510, not good.

All these pots are backwards, fully clockwise is minimum.

RV1 is, in effect, the bridge between the power amplifier
and the rest of the radio. It sets the drive level to the output
amplifier.

If it is too low, then there is not enough "oomph" in your power
output. If it is set too high, then there is a lot of distortion from
"overdrive".

The signal input to RV1 should be about 1-2 mv. Measure with an RF
probe or detector. The drive level should be about 1 mv, though that
may vary a great deal depending on impedances and output power level.

You will know when the level is right by the reports you get in a QSO
about audio quality. Upping the power output will mean upping the drive
level to some degree...but this is not always the case; sometimes there
is enough drive to accomodate any reasonable power level. But this is
the exception rather than the rule...

Occasionally, there is not enough drive to handle a greater power level.
In that case, the RF level into RV1 (or whatever adjuster is used) has to
be increased. I should think that a simple RF amp (with a 2N5109 or
equivalent) would be enough to provide the extra gain. On the uBITX,
that would mean moving RV1 to the RF amp output and maybe an attenuator
at the input to match things.

The biggest contributor to increased power output is increasing the IRF510
drain voltage from 12v to 24 or 36v. Usually, simply increasing the drive by increasing
RV1 a little bit is enough to fix the created problem.

Ordinarily you can get by with a simple adjustment of RV1 for most power levels
on the BITX. Pay attention to the heat generated and mind the thermal sinks.
And watch for oscillation and high levels of SWR. That should do it.

As to saturating the T30-6, build a new LPF with T50-6 cores; they should easily
handle anything up to 100W or so. Watch out for RF leakage and feedback at
higher power levels. Shielding may be required.

john
AD5YE

It ups the power.

Increase RV1 until received audio starts getting distorted, then back it off a little.
Alternately, if you don't have a big enough heat sink, increase RV1 until the IRF510's blow, then back it off a little.

If you turn up RV1 all the way and need a bit more, you can try increasing the gain of the microphone amp by decreasing the value of R63.
Doesn't really matter much which end of the transmitter you raise the gain, as those diode mixers and bidi amps have?
some serious dynamic range.

I have no idea where, but at some point all those 2n3904's in the driver and pre-driver will not have enough oooomph
to drive the capacitance they see in the IRF510 gates hard enough, and the amp will distort.
Allison has cautioned that not driving those gates hard enough can blow the FET's due to
the drain to gate capacitance raising the gate voltage with each RF sine wave enough to exceed Vgs-max.
Farhan has said that if you want much more than 20 watts out of the uBitx, the IRF510 final should be
on some other board.? Otherwise, I assume it will interfere with the low level stages up front.

I've been thinking that it would be interesting to build the WA2EBY amp,
see how hard the uBitx driver at Q92,93,96,97 can drive it.
Could get an idea by removing PA+12 from the IRF510's and checking to see how much RF
develops across the 100 ohm load at R97+R98.
Perhaps use something like the IRF510 final from the Bitx40 as the WA2EBY driver stage,
if you can figure out how to get it to work at 30m.


> Ill have to stay under 20W anyway as the T30-6 in the LPF will start to saturate any higher?

I can only suggest that you look hard at what WA2EBY did with his amp.
He spent months getting that thing to work correctly, and he's a lot smarter than I.

> Why? well becuase its fun, and I like playing... and I have a packet of 20 spare 510s... and the priceless look on the XYLs face when one does go off will be hillarious.

Hope your XYL has a good sense of humor.
Reminds me of scaring the crap out of my little sister when I was 12
by discharging several hundred volts from a big bank of electrolytics salvaged from old TV sets.

Jerry, KE7ER

Think about it...what does a volume control actually do?
It controls the drive of the stages following it -- it has no effect on
prior stages, and really has little effect on following stages other
that regulating the drive level.

Almost all volume controls are audio taper, 100-250K units. They are
audio taper because that is the way our ears naturally respond to sound.
They are about 100K or so because that is what provides a reasonable range
to drive to the following stages.

It is not complicated in this case. Everyone who has ever used a radio
or a music player has used a volume control. It is a simple potentiometer.
If one used a higher value, then one has to turn it up a bit higher to get
the result one wants. And audio taper is used because it provides a
greater dynamic range than a linear one...but a linear pot will work.

Use what you have; if it is unsatisfactory, then use another until you get
the effect you want. It is one of the most common pots out there, available
nearly everywhere.

john
AD5YE

Just thinking, Ill have to stay under 20W anyway as the T30-6 in the LPF will start to saturate any higher?
Thoughts?

73, Nick, VK4PLN

HI Group.
I have been looking at upping the power, what effect does RV1 have on this?
I understand it is for adjusting the input drive level to the second pre-driver.
Can anyone elaborate a bit more on its adjustment?

73 Nick, VK4PLN

They're nice cases. It was a tight fit but I used this one for my BITX20 V3B


Be aware that the panels are steel so they're harder to work with. I had an alu one cut to replace the front one.

Paul ZS2OE

Ok, So my next mods for the uBITX is to install a voltage boost for the PA.

LARGE heatsink + ALU case mount the 510, even a thermostatically controlled fan...
The boost converter can go up to 30v. I understand that the electro caps C98 and C99 will need to be upgraded too?
Any other thoughts?

Why? well becuase its fun, and I like playing... and I have a packet of 20 spare 510s... and the priceless look on the XYLs face when one does go off will be hillarious.

73, Nick, VK4PLN

I have the?6.7" x 5.1" x 3.1" blue metal case from ebay.?

https://www.ebay.com/itm/6-7-x-5-1-x-3-1-Blue-Metal-Enclosure-Project-Case-DIY-Junction-Box-F8Z7/252898313469

It barely fits my bitx40, and I mean barely. Like 1mm or less of clearance after I cut out some plastic bits.?

Definitely will not fit the ubitx.

--
David K4DBZ
Unofficial bitx chatroom:?https://discord.gg/CrHvWFc

This is what I went with:

I believe you need at lest 6" x 6" x 2.5"