|
There are many posts here about problems with the BS-170. , such a misfortune happened to my QDX. Everything was in order, the antenna is good tuned, the reasons are not clear.
There was no BS-170, I didn¡¯t want to wait for delivery, so I put the IR-510. The transceiver is working. The current consumption during transmission increased to 2 Amps, I lowered the supply voltage to 8 Volts. Now the current during transmission is 1.8 A. I also had to adjust the band-pass filter. Reduced the number of turns by two turns of the inductor. While everything is OK. On load from QRP-LAB about 20 Volts. There is no way to accurately measure power. More than 300 contacts were made on 20 meters. 52 countries including PY and VK. The antenna is a dipole at a height of 3 meters. On other bands so far QRT. There are no antennas.
On IRF - 510 I installed small aluminum radiators from a beer can :-)
Perhaps my solution is not beautiful, but allowed me to restore the transceiver
-- 73, Oleh, I2/UY2ZA
|
Take a reading at this article!
and watch for the page stating:?¡°Bigger is not Better¡±
Have fun, de dl2arl
|
The driver will most likely not have enough power to drive the IRF510 FETs hard enough. Especially on higher bands the gates of the IRF510 FETS have too high capacity to open to saturation quickly or at all.
Vojtech OK1IAK
|
The IRF510 has virtue but not for this.
THe inpput (gate) capacitance is about 150PF about 3x the BS170s!
The gate threshold voltage is 2-4V (most about 2.7V) and the ACT08 peak is 5V
except for the higher input C of the IRF510 will likely depress that (and make it run hot).
That means the actual mode is likely closer to class B (more current, heat, less efficient).
That it works at all is pretty amazing.? Its still a bad solution.
Allison
------------------
Post online only, please no email.
|
The capacitance issue isn't quite as bad as that 3x number suggests. It appears that Oleh has replaced each parallel pair of BS170s with a single IRF510, so the total capacitance is only 1.5x as high.
On Sat, Oct 8, 2022 at 10:20 AM ajparent1/kb1gmx < kb1gmx@...> wrote: The IRF510 has virtue but not for this.
THe inpput (gate) capacitance is about 150PF about 3x the BS170s!
The gate threshold voltage is 2-4V (most about 2.7V) and the ACT08 peak is 5V
except for the higher input C of the IRF510 will likely depress that (and make it run hot).
That means the actual mode is likely closer to class B (more current, heat, less efficient).
That it works at all is pretty amazing.? Its still a bad solution.
Allison
------------------
Post online only, please no email.
|
Actually its actually worse as we are dealing with dynamic waveforms.
BS170 s 24 (max 40) PF per device input capacitance.? times 2 devices.
Also the Crss or reverse transfer capacitance is 7pf (10 max).
IRF510? the Cin is 135PF (for some can be as high as 180pf).? X1 device
The reverse capacitance? is 20pf? (for that vendor, others are higher)
You have both input capacitance (gate to source) and what is known
as Miller capacitance which is essentially negative feedback
(the actual Crss capacitance) multiplied by the output voltage swing
(simple description).? So with IRF510 your really about? four times
greater load to drive with a tiny CMOS gate.
With all that, it makes the IRF510 much harder to drive than BS179s
in parallel.? How much difference?? BS170s are saturated RDS_on is
the limiting factor for power.? For Irf510 the only reason power out
happens is for drive above he gate threshold it has a gain of 1S
so for 5V drive you get about 2-2.5V gate drive not allowing for
capacitive load and you can maybe get to 2A drain current but
it is not saturated power its linear.? If it were saturated the .54ohm
RDS_on would make for a lot of power.?
FYI the current drive needed to hard switch a IRF510 at 13.55 mhz
is about 8A peak for?6 volt (measured) drive. I needed a 10V pulse
but no devices existed.? I was doing class D power conversion.?
Where does that power (current) go?? ?Charging that gate cap
then discharging it while the Miller coupling is working against you.
Allison
------------------
Post online only, please no email.
|
Is it an end in itself to achieve the transistor operation mode in class E? For me, no. And if you approach strictly, then class E in this version of the amplifier cannot be. A push-pull output stage will work well without a charge-discharge cycle of the capacitor circuit through the inductance and load.?
I do not impose on anyone to change transistors. Most likely I would put BS-170 again, but I don't have them. I wrote:
Perhaps my solution is not beautiful, but it allowed me to restore the transceiver
-- 73, Oleh, I2/UY2ZA
|
I believe that the PA is class D, not class E in QDX.
Gary
W9TD
|
For those that are interested in learning more about? this subject, Paul Harden NA5N wrote a couple of very good articles way back for QRPp.? ?
Rolf
K9DZT
|
Thanks for sharing, Oleh! Good job adapting the parts you have on hand to get your QDX back on the air. 99% of operators wouldn't take the time or energy to fix their radio like this.
And to Allison - as my mother used to say "If you don't have anything nice to say, don't say anything at all". In the time it took you to write out those messages you could have easily dropped some replacement BS170's in the mail for the old man...
|
Come on friends. ?Discussion (like confession) is good for the hobby (and soul). ?Those who are more knowledgeable than me are welcome here in my book. ?Again, this is a hobby and for those who really want to know there is so much information (much open to different interpretation) to study and learn from. ?Keep the information and correction coming. ?My feeling are not damaged. ?Again, one may be using the incorrect terms but if one is happy with how it works and learns from it, it works! My wife corrects me many times but I still do things the same way, now with better understanding. We are both happy with the outcome.?
Dave K8WPE since 1960 and still open to correction and learning new things every day. David J. Wilcox¡¯s iPad
toggle quoted message
Show quoted text
On Oct 8, 2022, at 10:01 PM, Luke / AD0KI <ad0ki@...> wrote:
?Thanks for sharing, Oleh! Good job adapting the parts you have on hand to get your QDX back on the air. 99% of operators wouldn't take the time or energy to fix their radio like this.
And to Allison - as my mother used to say "If you don't have anything nice to say, don't say anything at all". In the time it took you to write out those messages you could have easily dropped some replacement BS170's in the mail for the old man...
|
Luke,
Never confuse technical conversation for insult? and taking offense.?
Your comment is offensive.? I'm old?enough to be your mother.? My
comment is if you can't add anything useful, just don't.? Chill please.
As to the comment I could have sent parts, well why don't you as its far
from trivial to send even a few small parts overseas.??
Further:
As to class or D or E its a pedantic issue the IRF510 will not do
that either with ACT08 as driver.? ?Been there did that have the scars.
Both modes are dependent on saturating the device.? As I commented
its barely class B, marginal Class C.? The devices are not switching.?
I do know Harden's work well.? Have you read Yang and Rooij of
?EPC-CO.com as they took it to the next level.? I have used the
GaN devices for class D power conversion (wireless power transfer)
as part of a project.? We met several times. I discovered that TI has
the only suitable driver for any frequency above 7Mhz.
Its nice for Oleg that it puts RF out.? That he had to reduce supply
voltage to keep current down is an indicator of inefficiency.? He also
had to add cooling fins.? Excellent experiment. HE demonstrated
why we can't just drop in a higher power device and expect the same
or improved performance.
While its interesting to me it works I do wonder how hot the act08 is
getting.
Allison
------------------
Post online only, please no email.
|
If you look at the handbook for the QDX, you will see that the PA transistors are not involved in the circuit for the band-pass filter. The PA transistors are used for transmit and the bandpass filter is used on receive.
Altering the number of turns on L12 will have no effect on the PA.
|
Geoff,
Yes, L12 is RX only.??
Whose question are you replying to?
Allison
------------------
Post online only, please no email.
|
I think that if the driver wanted to break down, he would have done it already. I promise to write when it is.
I lowered the voltage to reduce the current. 9 volts gives an adjustable stabilizer with a maximum current of 2 amperes.
I screwed on the radiators because there is room. Worked without them. The transistors were hot, now they are warm.
L12 did not touch. I had to pick up the inductance L10.
I fixed the midpoint of T1 by placing a capacitor on the common wire. You can see it in the photo.
Hans is a talented designer. I am sure that he will overcome the problem with the reliability of the output stage and everything will be fine.
Thanks for the discussion. That's why we communicate.
-- 73, Oleh, I2/UY2ZA
|
I can't see anything at all offensive in Allison's posts.
I've tried, I looked, and I looked again.
Nope, still can't find anything.
10-4 ?
--
- 73 de Andy -
|
Oleah,
IF the act08 is not running hot? (surface temp under 40-50C) and the finals are
at an acceptable temp (anything under 50C is fine)? Go make contacts and enjoy.
Solid state parts can run wam and most if you keep the silicon inside under 100C
will do so for a long time.??
The need to adjust L10 is understandable,? Its a filter and if the terminating impedance
changes the filter performance changes.? That will likely happen with L6 and L8
for the other bands.? The change is a result of using the IRF510s, neither good
or bad just the interactions that come from such a change.
Allison
------------------
Post online only, please no email.
|
A bit OT but, reading this thread, I feel necessary to thank the very knowledgable and patient people in this forum for their outstanding? teaching.
Vr
Giuliano, IV3DLW
|
Replying to Allison, from the original poster.
"I also had to adjust the band-pass filter. Reduced the number of turns by two turns of the inductor."
Geoff
|
Yes. I was wrong. Sorry
--
73, Oleh, UY2ZA
|
DL2ARL
Shirley Dulcey KE1L
Gary W9TD
David Wilcox K8WPE