|
All,
Looking to order some spare output finals for the ubitx I have on order.
In looking around the IRF520 appears to have higher ratings than the
IRF510 with a lower Rds.
The 510 does have much lower input and output capacitance.
There is hardly any difference in price.
Any recommendations on which to order?
tim ab0wr
|
Tim:
Given the druthers, stick with the IRF510.
Yes, the IRF520 and IRF530 have better voltage stats,
but much higher input gate capacitance. This DOES
make a difference. Remember, these devices were
designed for low frequency switching operations. They
were designed for on/off operation.
Adapting them to RF is another ball game entirely.
Input and output capacitance do matter...a great deal.
But the IRF520 has been adapted successfully to PA
operation (they are in a lot of Chinese amps). So it
can be done...but adjustments have to be made, usually
by increases negative feedback. The IRF530 is MUCH
harder to work with. Yeah, it can be done but keeping it clean is a real pain...
Operating the IRF510 on 18-24v (or more up to 48v)
will give very good results provided it is done right and attention
is paid to heat generation and impedance.
john
AD5YE
|
Or the Mitsubishi RF RD16HHF1 MOSFET designed for 2-30MHz RF applications.? Under $5 each.? Or the RD16HVF1 only pennies more and good through 2M.?
With a 12VDC to 24/28VDC DC to DC converters (about $4) and the Mitsu parts can be bolted directly to a big heatsink attached to the chassis, no insulator required. Probably good for 20W through 10M although run off batteries a bit shorter battery life.? At real 5W QRP they should be happy forever.
Dave WI6R
On Sun, Jan 14, 2018 at 4:33 PM, Tim Gorman <tgorman2@...> wrote: All,
Looking to order some spare output finals for the ubitx I have on order.
In looking around the IRF520 appears to have higher ratings than the
IRF510 with a lower Rds.
The 510 does have much lower input and output capacitance.
There is hardly any difference in price.
Any recommendations on which to order?
tim ab0wr
|
You can find a post using RD06HVF1? for HF too.
It would surely supass 2m and really means QRP regards ?sarma ?vu3zmv
On Mon, Jan 15, 2018 at 12:21 PM, Dave Bottom <ars.kd6az@...> wrote: Or the Mitsubishi RF RD16HHF1 MOSFET designed for 2-30MHz RF applications.? Under $5 each.? Or the RD16HVF1 only pennies more and good through 2M.?
With a 12VDC to 24/28VDC DC to DC converters (about $4) and the Mitsu parts can be bolted directly to a big heatsink attached to the chassis, no insulator required. Probably good for 20W through 10M although run off batteries a bit shorter battery life.? At real 5W QRP they should be happy forever.
Dave WI6R
|
Dave, Would this be a satisfactory drop-in replacement? I can't find any info on input or output capacitance to compare it to an irf510. From the Smith chart on the datasheet I'm guessing the input capacitance is about 60pf (been a long time since I've used a Smith Chart) at 30Mhz, much less than the irf510. Would that impact the negative feedback used on the amp significantly (0.1uf in series with a 220ohm resistor)? By the same token the output capacitance would appear to be much higher than the irf510. I suspect that will impact the design of the bandpass filters but I'm not sure by how much without doing a lot of calculations. I'm just looking for something that would provide more reliability under adverse conditions (e.g. high swr) than the irf510. Don't really care if it provides higher output or not - but that might come later <grin>. On Sun, 14 Jan 2018 22:51:06 -0800 "Dave Bottom" <ars.kd6az@...> wrote: Or the Mitsubishi RF RD16HHF1 MOSFET designed for 2-30MHz RF
applications. Under $5 each. Or the RD16HVF1 only pennies more and
good through 2M.
With a 12VDC to 24/28VDC DC to DC converters (about $4) and the Mitsu
parts can be bolted directly to a big heatsink attached to the
chassis, no insulator required.
Probably good for 20W through 10M although run off batteries a bit
shorter battery life. At real 5W QRP they should be happy forever.
Dave WI6R
On Sun, Jan 14, 2018 at 4:33 PM, Tim Gorman <tgorman2@...> wrote:
All,
Looking to order some spare output finals for the ubitx I have on
order.
In looking around the IRF520 appears to have higher ratings than the
IRF510 with a lower Rds.
The 510 does have much lower input and output capacitance.
There is hardly any difference in price.
Any recommendations on which to order?
tim ab0wr
|
Thanks for the reply, John. I'll stock up on irf510's! tim ab0wr On Sun, 14 Jan 2018 16:55:50 -0800 "John Backo" <jabac@...> wrote: Tim:
Given the druthers, stick with the IRF510.
Yes, the IRF520 and IRF530 have better voltage stats,
but much higher input gate capacitance. This DOES
make a difference. Remember, these devices were
designed for low frequency switching operations. They
were designed for on/off operation.
Adapting them to RF is another ball game entirely.
Input and output capacitance do matter...a great deal.
But the IRF520 has been adapted successfully to PA
operation (they are in a lot of Chinese amps). So it
can be done...but adjustments have to be made, usually
by increases negative feedback. The IRF530 is MUCH
harder to work with. Yeah, it can be done but keeping it clean is a
real pain...
Operating the IRF510 on 18-24v (or more up to 48v)
will give very good results provided it is done right and attention
is paid to heat generation and impedance.
john
AD5YE
|
I think I mentioned it before, but anyone seeking to drop in a Mitsubishi RDMos part should be aware that?
a) They're designed for 12V operation.? You won't want to run them at 18V/etc b) They have a different pinout, you'll have to solder them in with the leads mangled/crossed.? IRF510/20/30's are Gate-Drain-Source.? The RDMos series are Gate-Source-Drain.? This is a mod I still intend to try at some point, but have other projects on the plate first. c) They have a different bias voltage.? Turn down the bias voltage before you key up, and then re-bias appropriately. If your IRF510's are biased at something like 4V when the bias voltage for the RDmos's are 2V (I'm making total wild assumptions, it could be the other way around), you'll instantly turn the devices on full-current and potentially blow something up.?? d) The feedback network will want to be different.? These have much more gain at high frequencies.? You may need more negative feedback. e) You can use an uninsulated grounded heatsink, as the tabs of the these devices is the source and not drain.
I'm looking forward to hearing from someone who's done the mod and tried it out.
On Mon, Jan 15, 2018 at 10:52 AM, Tim Gorman <tgorman2@...> wrote: Dave,
Would this be a satisfactory drop-in replacement? I can't find any info
on input or output capacitance to compare it to an irf510.
From the Smith chart on the datasheet I'm guessing the input
capacitance is about 60pf (been a long time since I've used a Smith
Chart) at 30Mhz, much less than the irf510. Would that impact the
negative feedback used on the amp significantly (0.1uf in series with a
220ohm resistor)?
By the same token the output capacitance would appear to be much higher
than the irf510. I suspect that will impact the design of the bandpass
filters but I'm not sure by how much without doing a lot of
calculations.
I'm just looking for something that would provide more reliability
under adverse conditions (e.g. high swr) than the irf510. Don't really
care if it provides higher output or not - but that might come later
<grin>.
On Sun, 14 Jan 2018 22:51:06 -0800
"Dave Bottom" <ars.kd6az@...> wrote:
> Or the Mitsubishi RF RD16HHF1 MOSFET designed for 2-30MHz RF
> applications. Under $5 each.? Or the RD16HVF1 only pennies more and
> good through 2M.
>
> With a 12VDC to 24/28VDC DC to DC converters (about $4) and the Mitsu
> parts can be bolted directly to a big heatsink attached to the
> chassis, no insulator required.
> Probably good for 20W through 10M although run off batteries a bit
> shorter battery life.? At real 5W QRP they should be happy forever.
>
> Dave WI6R
>
> On Sun, Jan 14, 2018 at 4:33 PM, Tim Gorman <tgorman2@...> wrote:
>
> > All,
> >
> > Looking to order some spare output finals for the ubitx I have on
> > order.
> >
> > In looking around the IRF520 appears to have higher ratings than the
> > IRF510 with a lower Rds.
> >
> > The 510 does have much lower input and output capacitance.
> >
> > There is hardly any difference in price.
> >
> > Any recommendations on which to order?
> >
> > tim ab0wr
> >
> >
> >
> >
>
>
-- Martin Held - AE7EU
-------------------------------------------------------------------------------
If there aren't any questions, then what is there to learn?
|
Of course the upside of using RD16HVF1 is that it is a real RF transistor... not a motor switch like the IRF510. ?What you get for that is actually hidden in your comment about gain at high frequency. In fact, the gate capacitance with frequency is far better behaved with the RD device so complicated compensation networks to keep gain constant is not needed. ?Hans Summers has a second set of pads on his 5 watt PA just for the use of these parts.?
Dr.?William J. Schmidt - K9HZ J68HZ 8P6HK ZF2HZ PJ4/K9HZ VP5/K9HZ PJ2/K9HZ ? Owner - Operator Big Signal Ranch – K9ZC Staunton, Illinois ? Owner – Operator Villa Grand Piton - J68HZ Soufriere, St. Lucia W.I. Rent it:
email:??bill@... ?
toggle quoted message
Show quoted text
On Jan 15, 2018, at 1:39 PM, Diver Martin < diver.martin@...> wrote: I think I mentioned it before, but anyone seeking to drop in a Mitsubishi RDMos part should be aware that?
a) They're designed for 12V operation.? You won't want to run them at 18V/etc b) They have a different pinout, you'll have to solder them in with the leads mangled/crossed.? IRF510/20/30's are Gate-Drain-Source.? The RDMos series are Gate-Source-Drain.? This is a mod I still intend to try at some point, but have other projects on the plate first. c) They have a different bias voltage.? Turn down the bias voltage before you key up, and then re-bias appropriately. If your IRF510's are biased at something like 4V when the bias voltage for the RDmos's are 2V (I'm making total wild assumptions, it could be the other way around), you'll instantly turn the devices on full-current and potentially blow something up.?? d) The feedback network will want to be different.? These have much more gain at high frequencies.? You may need more negative feedback. e) You can use an uninsulated grounded heatsink, as the tabs of the these devices is the source and not drain.
I'm looking forward to hearing from someone who's done the mod and tried it out.
-- Martin Held - AE7EU
-------------------------------------------------------------------------------
If there aren't any questions, then what is there to learn?
|
Thanks, Diver. I think I'll just stick with the irf510's for now. They are cheap if I play too hard with them! tim ab0wr On Mon, 15 Jan 2018 11:39:36 -0800 "Diver Martin" <diver.martin@...> wrote: I think I mentioned it before, but anyone seeking to drop in a
Mitsubishi RDMos part should be aware that
a) They're designed for 12V operation. You won't want to run them at
18V/etc
b) They have a different pinout, you'll have to solder them in with
the leads mangled/crossed. IRF510/20/30's are Gate-Drain-Source.
The RDMos series are Gate-Source-Drain. This is a mod I still intend
to try at some point, but have other projects on the plate first.
c) They have a different bias voltage. Turn down the bias voltage
before you key up, and then re-bias appropriately. If your IRF510's
are biased at something like 4V when the bias voltage for the RDmos's
are 2V (I'm making total wild assumptions, it could be the other way
around), you'll instantly turn the devices on full-current and
potentially blow something up. d) The feedback network will want to
be different. These have much more gain at high frequencies. You
may need more negative feedback. e) You can use an uninsulated
grounded heatsink, as the tabs of the these devices is the source and
not drain.
I'm looking forward to hearing from someone who's done the mod and
tried it out.
On Mon, Jan 15, 2018 at 10:52 AM, Tim Gorman <tgorman2@...> wrote:
Dave,
Would this be a satisfactory drop-in replacement? I can't find any
info on input or output capacitance to compare it to an irf510.
From the Smith chart on the datasheet I'm guessing the input
capacitance is about 60pf (been a long time since I've used a Smith
Chart) at 30Mhz, much less than the irf510. Would that impact the
negative feedback used on the amp significantly (0.1uf in series
with a 220ohm resistor)?
By the same token the output capacitance would appear to be much
higher than the irf510. I suspect that will impact the design of
the bandpass filters but I'm not sure by how much without doing a
lot of calculations.
I'm just looking for something that would provide more reliability
under adverse conditions (e.g. high swr) than the irf510. Don't
really care if it provides higher output or not - but that might
come later <grin>.
On Sun, 14 Jan 2018 22:51:06 -0800
"Dave Bottom" <ars.kd6az@...> wrote:
Or the Mitsubishi RF RD16HHF1 MOSFET designed for 2-30MHz RF
applications. Under $5 each. Or the RD16HVF1 only pennies more
and good through 2M.
With a 12VDC to 24/28VDC DC to DC converters (about $4) and the
Mitsu parts can be bolted directly to a big heatsink attached to
the chassis, no insulator required.
Probably good for 20W through 10M although run off batteries a bit
shorter battery life. At real 5W QRP they should be happy
forever.
Dave WI6R
On Sun, Jan 14, 2018 at 4:33 PM, Tim Gorman <tgorman2@...>
wrote:
All,
Looking to order some spare output finals for the ubitx I have
on order.
In looking around the IRF520 appears to have higher ratings
than the IRF510 with a lower Rds.
The 510 does have much lower input and output capacitance.
There is hardly any difference in price.
Any recommendations on which to order?
tim ab0wr
|
?So Dr. Schmidt --- are you saying that the severe dropoff in power output from the ubitx / bitx40 would be ameliorated?
If so, I might be interested!? ? It would be quite nice to have a bit more power on 30 and 20 meters.....
comment?
gordon
toggle quoted message
Show quoted text
Of course the upside of using RD16HVF1 is that it is a real RF transistor... not a motor switch like the IRF510. ?What you get for that is actually hidden in your comment about gain at high frequency. In fact, the gate capacitance with frequency is far
better behaved with the RD device so complicated compensation networks to keep gain constant is not needed. ?Hans Summers has a second set of pads on his 5 watt PA just for the use of these parts.?
Dr.?William J. Schmidt - K9HZ J68HZ 8P6HK ZF2HZ PJ4/K9HZ VP5/K9HZ PJ2/K9HZ
?
Owner - Operator
Big Signal Ranch – K9ZC
Staunton, Illinois
?
Owner – Operator
Villa Grand Piton - J68HZ
Soufriere, St. Lucia W.I.
Rent it:
email:??bill@...
?
I think I mentioned it before, but anyone seeking to drop in a Mitsubishi RDMos part should be aware that?
a) They're designed for 12V operation.? You won't want to run them at 18V/etc
b) They have a different pinout, you'll have to solder them in with the leads mangled/crossed.? IRF510/20/30's are Gate-Drain-Source.? The RDMos series are Gate-Source-Drain.? This is a mod I still intend to try at some point, but have other projects on
the plate first.
c) They have a different bias voltage.? Turn down the bias voltage before you key up, and then re-bias appropriately. If your IRF510's are biased at something like 4V when the bias voltage for the RDmos's are 2V (I'm making total wild assumptions, it could
be the other way around), you'll instantly turn the devices on full-current and potentially blow something up.??
d) The feedback network will want to be different.? These have much more gain at high frequencies.? You may need more negative feedback.
e) You can use an uninsulated grounded heatsink, as the tabs of the these devices is the source and not drain.
I'm looking forward to hearing from someone who's done the mod and tried it out.
--
Martin Held - AE7EU
-------------------------------------------------------------------------------
If there aren't any questions, then what is there to learn?
|
Name is Bill… ? There are considerations in the amplification stages of the uBITx leading up to the PA too … the bandwidth of the output and interstage transformers T9, T10, and T11.? I have not scanned them to know… or measured the drive from the earlier stages. ?Theoretically if they were designed right, the PA with an RD15HVF1 would generate about 18 watts output from 1-54 MHz (60% efficiency).? The biasing is at a different point but what is on the board should work. ??Maybe someone has already measured the output of the last driver stage before the PA on all bands with a scope.. and knows if there is change in available drive level with frequency?? You wouldn’t want to go through the work of replacing the 510 with the RD device just to find out the other stages have a severe gain variance with frequency. ? I always replace the 510’s with RD parts whenever I can… ? ? Dr. William J. Schmidt - K9HZ J68HZ 8P6HK ZF2HZ PJ4/K9HZ VP5/K9HZ PJ2/K9HZ ? Owner - Operator Big Signal Ranch – K9ZC Staunton, Illinois ? Owner – Operator Villa Grand Piton – J68HZ Soufriere, St. Lucia W.I. Rent it: Like us on Facebook! ? ? email:? bill@... ?
toggle quoted message
Show quoted text
From: [email protected] [mailto:[email protected]] On Behalf Of Gordon Gibby
Sent: Monday, January 15, 2018 7:20 PM
To: [email protected]
Subject: Re: [BITX20] irf520 vs irf510? ?So Dr. Schmidt --- are you saying that the severe dropoff in power output from the ubitx / bitx40 would be ameliorated? ? If so, I might be interested!? ? It would be quite nice to have a bit more power on 30 and 20 meters..... ? comment? ? gordon ?
Of course the upside of using RD16HVF1 is that it is a real RF transistor... not a motor switch like the IRF510. ?What you get for that is actually hidden in your comment about gain at high frequency. In fact, the gate capacitance with frequency is far better behaved with the RD device so complicated compensation networks to keep gain constant is not needed. ?Hans Summers has a second set of pads on his 5 watt PA just for the use of these parts.? ? Dr.?William J. Schmidt - K9HZ J68HZ 8P6HK ZF2HZ PJ4/K9HZ VP5/K9HZ PJ2/K9HZ ? Owner - Operator Big Signal Ranch – K9ZC Staunton, Illinois ? Owner – Operator Villa Grand Piton - J68HZ Soufriere, St. Lucia W.I. Rent it: ? email:??bill@... ? I think I mentioned it before, but anyone seeking to drop in a Mitsubishi RDMos part should be aware that? a) They're designed for 12V operation.? You won't want to run them at 18V/etc b) They have a different pinout, you'll have to solder them in with the leads mangled/crossed.? IRF510/20/30's are Gate-Drain-Source.? The RDMos series are Gate-Source-Drain.? This is a mod I still intend to try at some point, but have other projects on the plate first. c) They have a different bias voltage.? Turn down the bias voltage before you key up, and then re-bias appropriately. If your IRF510's are biased at something like 4V when the bias voltage for the RDmos's are 2V (I'm making total wild assumptions, it could be the other way around), you'll instantly turn the devices on full-current and potentially blow something up.?? d) The feedback network will want to be different.? These have much more gain at high frequencies.? You may need more negative feedback. e) You can use an uninsulated grounded heatsink, as the tabs of the these devices is the source and not drain.
I'm looking forward to hearing from someone who's done the mod and tried it out.
? On Mon, Jan 15, 2018 at 10:52 AM, Tim Gorman <tgorman2@...> wrote: Dave,
Would this be a satisfactory drop-in replacement? I can't find any info
on input or output capacitance to compare it to an irf510.
From the Smith chart on the datasheet I'm guessing the input
capacitance is about 60pf (been a long time since I've used a Smith
Chart) at 30Mhz, much less than the irf510. Would that impact the
negative feedback used on the amp significantly (0.1uf in series with a
220ohm resistor)?
By the same token the output capacitance would appear to be much higher
than the irf510. I suspect that will impact the design of the bandpass
filters but I'm not sure by how much without doing a lot of
calculations.
I'm just looking for something that would provide more reliability
under adverse conditions (e.g. high swr) than the irf510. Don't really
care if it provides higher output or not - but that might come later
<grin>.
On Sun, 14 Jan 2018 22:51:06 -0800
"Dave Bottom" <ars.kd6az@...> wrote:
> Or the Mitsubishi RF RD16HHF1 MOSFET designed for 2-30MHz RF
> applications. Under $5 each.? Or the RD16HVF1 only pennies more and
> good through 2M.
>
> With a 12VDC to 24/28VDC DC to DC converters (about $4) and the Mitsu
> parts can be bolted directly to a big heatsink attached to the
> chassis, no insulator required.
> Probably good for 20W through 10M although run off batteries a bit
> shorter battery life.? At real 5W QRP they should be happy forever.
>
> Dave WI6R
>
> On Sun, Jan 14, 2018 at 4:33 PM, Tim Gorman <tgorman2@...> wrote:
>
> > All,
> >
> > Looking to order some spare output finals for the ubitx I have on
> > order.
> >
> > In looking around the IRF520 appears to have higher ratings than the
> > IRF510 with a lower Rds.
> >
> > The 510 does have much lower input and output capacitance.
> >
> > There is hardly any difference in price.
> >
> > Any recommendations on which to order?
> >
> > tim ab0wr
> >
> >
> >
> >
>
>
--
-------------------------------------------------------------------------------
If there aren't any questions, then what is there to learn?
|
All:
As noted before, the Mitsubishi parts and the IRFxxx power MOSFETS
are totally different. The Mitsubishi parts are designed to be RF amplifiers,
not power switches.
The IRF510 is good to about 17m. Thereafter, it falls off considerably in
normal PA applications. The Mitsuishi parts are good to at least 10m,
and many go beyond that.
The advantages of the IRFxxx series is 1)they are cheap. 2} they can use
voltages up to 48v or so. The Mitsubishi parts are designed to be operated
at 12v. Driving them with a higher voltage will damage them. Because of this
they have a more or less fixed power output.
Both have found their place among hams. They require different approaches.
john
AD5YE
|
Bill—. Thanks!!!
Sent from my iPhone
On Jan 16, 2018, at 00:12, K9HZ < bill@...> wrote:
toggle quoted message
Show quoted text
Name is Bill…
?
There are considerations in the amplification stages of the uBITx leading up to the PA too … the bandwidth of the output and interstage transformers T9, T10,
and T11.? I have not scanned them to know… or measured the drive from the earlier stages. ?Theoretically if they were designed right, the PA with an RD15HVF1 would generate about 18 watts output from 1-54 MHz (60% efficiency).? The biasing is at a different
point but what is on the board should work. ??Maybe someone has already measured the output of the last driver stage before the PA on all bands with a scope.. and knows if there is change in available drive level with frequency?? You wouldn’t want to go through
the work of replacing the 510 with the RD device just to find out the other stages have a severe gain variance with frequency.
?
I always replace the 510’s with RD parts whenever I can…
?
?
Dr.
William J. Schmidt - K9HZ J68HZ 8P6HK ZF2HZ PJ4/K9HZ VP5/K9HZ PJ2/K9HZ
?
Owner - Operator
Big Signal Ranch – K9ZC
Staunton, Illinois
?
Owner – Operator
Villa Grand Piton – J68HZ
Soufriere, St. Lucia W.I.
Rent it:
Like us on Facebook!
?
?
email:?
bill@...
?
?
?So Dr. Schmidt --- are you saying that the severe dropoff in power output from the ubitx / bitx40 would be ameliorated?
?
If so, I might be interested!? ? It would be quite nice to have a bit more power on 30 and 20 meters.....
?
comment?
?
gordon
?
Of course the upside of using RD16HVF1 is that it is a real RF transistor... not a motor switch like the IRF510. ?What you get for that is actually
hidden in your comment about gain at high frequency. In fact, the gate capacitance with frequency is far better behaved with the RD device so complicated compensation networks to keep gain constant is not needed. ?Hans Summers has a second set of pads on his
5 watt PA just for the use of these parts.?
?
Dr.?William J. Schmidt - K9HZ J68HZ 8P6HK ZF2HZ PJ4/K9HZ VP5/K9HZ PJ2/K9HZ
?
Owner - Operator
Big Signal Ranch – K9ZC
Staunton, Illinois
?
Owner – Operator
Villa Grand Piton - J68HZ
Soufriere, St. Lucia W.I.
Rent it:
?
email:??bill@...
?
I think I mentioned it before, but anyone seeking to drop in a Mitsubishi RDMos part should be aware that?
a) They're designed for 12V operation.? You won't want to run them at 18V/etc
b) They have a different pinout, you'll have to solder them in with the leads mangled/crossed.? IRF510/20/30's are Gate-Drain-Source.? The RDMos series are Gate-Source-Drain.?
This is a mod I still intend to try at some point, but have other projects on the plate first.
c) They have a different bias voltage.? Turn down the bias voltage before you key up, and then re-bias appropriately. If your IRF510's are biased at something like 4V when the
bias voltage for the RDmos's are 2V (I'm making total wild assumptions, it could be the other way around), you'll instantly turn the devices on full-current and potentially blow something up.??
d) The feedback network will want to be different.? These have much more gain at high frequencies.? You may need more negative feedback.
e) You can use an uninsulated grounded heatsink, as the tabs of the these devices is the source and not drain.
I'm looking forward to hearing from someone who's done the mod and tried it out.
?
On Mon, Jan 15, 2018 at 10:52 AM, Tim Gorman <tgorman2@...> wrote:
Dave,
Would this be a satisfactory drop-in replacement? I can't find any info
on input or output capacitance to compare it to an irf510.
From the Smith chart on the datasheet I'm guessing the input
capacitance is about 60pf (been a long time since I've used a Smith
Chart) at 30Mhz, much less than the irf510. Would that impact the
negative feedback used on the amp significantly (0.1uf in series with a
220ohm resistor)?
By the same token the output capacitance would appear to be much higher
than the irf510. I suspect that will impact the design of the bandpass
filters but I'm not sure by how much without doing a lot of
calculations.
I'm just looking for something that would provide more reliability
under adverse conditions (e.g. high swr) than the irf510. Don't really
care if it provides higher output or not - but that might come later
<grin>.
On Sun, 14 Jan 2018 22:51:06 -0800
"Dave Bottom" <ars.kd6az@...> wrote:
> Or the Mitsubishi RF RD16HHF1 MOSFET designed for 2-30MHz RF
> applications. Under $5 each.? Or the RD16HVF1 only pennies more and
> good through 2M.
>
> With a 12VDC to 24/28VDC DC to DC converters (about $4) and the Mitsu
> parts can be bolted directly to a big heatsink attached to the
> chassis, no insulator required.
> Probably good for 20W through 10M although run off batteries a bit
> shorter battery life.? At real 5W QRP they should be happy forever.
>
> Dave WI6R
>
> On Sun, Jan 14, 2018 at 4:33 PM, Tim Gorman <tgorman2@...> wrote:
>
> > All,
> >
> > Looking to order some spare output finals for the ubitx I have on
> > order.
> >
> > In looking around the IRF520 appears to have higher ratings than the
> > IRF510 with a lower Rds.
> >
> > The 510 does have much lower input and output capacitance.
> >
> > There is hardly any difference in price.
> >
> > Any recommendations on which to order?
> >
> > tim ab0wr
> >
> >
> >
> >
>
>
--
-------------------------------------------------------------------------------
If there aren't any questions, then what is there to learn?
|
uBITX PA
I will be trying out the Mitsubishi parts after I get all of the other things sorted out with my uBITX, but it is a very pleasing rig.? My uBITX starts out at 12W on 80M, 7.5W on 40M, 2W on 20M, and is down at 0.4W on 10M. so I will be checking the driver stages as well.? I expect they are not flat over the 80-10M bands either.? ?I will take it step by step and I have a good spectrum analyzer to keep things clean as I work my way through that.?? I'll post the results when I get to this down the road...??
Ahead of the potential PA mods are: 1) I just received the neat little CW Keying Adapter PCBs from the QRPGuys.? Nice job, but I'll need some smaller standoffs (4-40 or smaller mounting it looks like). 2) AGC is up after that using the unused Audio AGC amplifier to control ether IF or RF or both for AGC. 3) Finally I have an annoying high frequency tone at medium volume leaking through somewhere on receive to locate and get rid of.?
Keep the soldering irons warm...
Dave WI6R
On Tue, Jan 16, 2018 at 12:55 AM, Gordon Gibby <ggibby@...> wrote:
Bill—. Thanks!!!
Sent from my iPhone
On Jan 16, 2018, at 00:12, K9HZ < bill@...> wrote:
Name is Bill…
?
There are considerations in the amplification stages of the uBITx leading up to the PA too … the bandwidth of the output and interstage transformers T9, T10,
and T11.? I have not scanned them to know… or measured the drive from the earlier stages.? Theoretically if they were designed right, the PA with an RD15HVF1 would generate about 18 watts output from 1-54 MHz (60% efficiency).? The biasing is at a different
point but what is on the board should work. ??Maybe someone has already measured the output of the last driver stage before the PA on all bands with a scope.. and knows if there is change in available drive level with frequency?? You wouldn’t want to go through
the work of replacing the 510 with the RD device just to find out the other stages have a severe gain variance with frequency.
?
I always replace the 510’s with RD parts whenever I can…
?
?
Dr.
William J. Schmidt - K9HZ J68HZ 8P6HK ZF2HZ PJ4/K9HZ VP5/K9HZ PJ2/K9HZ
?
Owner - Operator
Big Signal Ranch – K9ZC
Staunton, Illinois
?
Owner – Operator
Villa Grand Piton – J68HZ
Soufriere, St. Lucia W.I.
Rent it:
Like us on Facebook!
?
?
email:?
bill@...
?
?
?So Dr. Schmidt --- are you saying that the severe dropoff in power output from the ubitx / bitx40 would be ameliorated?
?
If so, I might be interested!? ? It would be quite nice to have a bit more power on 30 and 20 meters.....
?
comment?
?
gordon
?
Of course the upside of using RD16HVF1 is that it is a real RF transistor... not a motor switch like the IRF510.? What you get for that is actually
hidden in your comment about gain at high frequency. In fact, the gate capacitance with frequency is far better behaved with the RD device so complicated compensation networks to keep gain constant is not needed.? Hans Summers has a second set of pads on his
5 watt PA just for the use of these parts.?
?
Dr.?William J. Schmidt - K9HZ J68HZ 8P6HK ZF2HZ PJ4/K9HZ VP5/K9HZ PJ2/K9HZ
?
Owner - Operator
Big Signal Ranch – K9ZC
Staunton, Illinois
?
Owner – Operator
Villa Grand Piton - J68HZ
Soufriere, St. Lucia W.I.
Rent it:
?
email:??bill@...
?
I think I mentioned it before, but anyone seeking to drop in a Mitsubishi RDMos part should be aware that?
a) They're designed for 12V operation.? You won't want to run them at 18V/etc
b) They have a different pinout, you'll have to solder them in with the leads mangled/crossed.? IRF510/20/30's are Gate-Drain-Source.? The RDMos series are Gate-Source-Drain.?
This is a mod I still intend to try at some point, but have other projects on the plate first.
c) They have a different bias voltage.? Turn down the bias voltage before you key up, and then re-bias appropriately. If your IRF510's are biased at something like 4V when the
bias voltage for the RDmos's are 2V (I'm making total wild assumptions, it could be the other way around), you'll instantly turn the devices on full-current and potentially blow something up.??
d) The feedback network will want to be different.? These have much more gain at high frequencies.? You may need more negative feedback.
e) You can use an uninsulated grounded heatsink, as the tabs of the these devices is the source and not drain.
I'm looking forward to hearing from someone who's done the mod and tried it out.
?
On Mon, Jan 15, 2018 at 10:52 AM, Tim Gorman <tgorman2@...> wrote:
Dave,
Would this be a satisfactory drop-in replacement? I can't find any info
on input or output capacitance to compare it to an irf510.
From the Smith chart on the datasheet I'm guessing the input
capacitance is about 60pf (been a long time since I've used a Smith
Chart) at 30Mhz, much less than the irf510. Would that impact the
negative feedback used on the amp significantly (0.1uf in series with a
220ohm resistor)?
By the same token the output capacitance would appear to be much higher
than the irf510. I suspect that will impact the design of the bandpass
filters but I'm not sure by how much without doing a lot of
calculations.
I'm just looking for something that would provide more reliability
under adverse conditions (e.g. high swr) than the irf510. Don't really
care if it provides higher output or not - but that might come later
<grin>.
On Sun, 14 Jan 2018 22:51:06 -0800
"Dave Bottom" <ars.kd6az@...> wrote:
> Or the Mitsubishi RF RD16HHF1 MOSFET designed for 2-30MHz RF
> applications. Under $5 each.? Or the RD16HVF1 only pennies more and
> good through 2M.
>
> With a 12VDC to 24/28VDC DC to DC converters (about $4) and the Mitsu
> parts can be bolted directly to a big heatsink attached to the
> chassis, no insulator required.
> Probably good for 20W through 10M although run off batteries a bit
> shorter battery life.? At real 5W QRP they should be happy forever.
>
> Dave WI6R
>
> On Sun, Jan 14, 2018 at 4:33 PM, Tim Gorman <tgorman2@...> wrote:
>
> > All,
> >
> > Looking to order some spare output finals for the ubitx I have on
> > order.
> >
> > In looking around the IRF520 appears to have higher ratings than the
> > IRF510 with a lower Rds.
> >
> > The 510 does have much lower input and output capacitance.
> >
> > There is hardly any difference in price.
> >
> > Any recommendations on which to order?
> >
> > tim ab0wr
> >
> >
> >
> >
>
>
--
-------------------------------------------------------------------------------
If there aren't any questions, then what is there to learn?
|
Thats great Dave. ?I have not followed through myself by testing each of the stages before the PA and recording the drive levels but it should be enlightening. ?Ive been focused on the antenna tuner.?
Ive thought through the AGC problem and i think there is a very good solution to replace one of the early IF amplifiers with a rebiased dual gate fet... and let the second gate be voltage controlled by the AGC circuit. ?AGC should have an RF and Audio components in logarithmic function with a low pass filter. ?That takes it to almost late Collins 75A4 level.?
Dr.?William J. Schmidt - K9HZ J68HZ 8P6HK ZF2HZ PJ4/K9HZ VP5/K9HZ PJ2/K9HZ ? Owner - Operator Big Signal Ranch – K9ZC Staunton, Illinois ? Owner – Operator Villa Grand Piton - J68HZ Soufriere, St. Lucia W.I. Rent it:
email:??bill@... ?
toggle quoted message
Show quoted text
On Jan 17, 2018, at 10:16 PM, Dave Bottom < ars.kd6az@...> wrote: uBITX PA
I will be trying out the Mitsubishi parts after I get all of the other things sorted out with my uBITX, but it is a very pleasing rig.? My uBITX starts out at 12W on 80M, 7.5W on 40M, 2W on 20M, and is down at 0.4W on 10M. so I will be checking the driver stages as well.? I expect they are not flat over the 80-10M bands either.? ?I will take it step by step and I have a good spectrum analyzer to keep things clean as I work my way through that.?? I'll post the results when I get to this down the road...??
Ahead of the potential PA mods are: 1) I just received the neat little CW Keying Adapter PCBs from the QRPGuys.? Nice job, but I'll need some smaller standoffs (4-40 or smaller mounting it looks like). 2) AGC is up after that using the unused Audio AGC amplifier to control ether IF or RF or both for AGC. 3) Finally I have an annoying high frequency tone at medium volume leaking through somewhere on receive to locate and get rid of.?
Keep the soldering irons warm...
Dave WI6R
|
Bill,
Yes, I think that sort of idea is needed.? Right now I can't tell you how much the uBITX reminds me of my SBE SB-33 SF-1. Even audio AGC doesn't do it? I'm hoping that the IF stage is enough or then the need to add an RF stage (very low gain, mainly an attenuator stage). Possibly using the second channel of the existing Audio amplifier connected ahead of the Volume control with it's own detector to feed the AGC system, maybe a real S-Meter (such as they are on superhets, spoiled by SDRs with accurate signal level display.)??
AGC is what separates a good SSB rig from a GREAT SSB Rig.? Collins did it right early on (although I've changed attack and decay timing to suit) and Heathkit SB series rigs always had great if simple AGC. Even the lowly HW-12/22/32 tube type or HW-104 solid state were superb to listen to even on a round table with so many varying signal strengths.??
My Alda 103 also had superb superb AGC as well as the first rig I ever saw that specified IMD intercept point with two signal inside the pass band (e.g. 2 KHz spacing, better than +10 dBm) 1979!? ?Decades before Elecraft did with their K3.? The Alda service manual; is available on line and it has some very interesting circuit ideas including using a high power RF transistor for the receiver's RF preamp.
I wish I had another Alda to install a Raduino for VFO, HFO and BFO.? Fun with a neat digital display and how much it could simplify this rig along with modern relay band switching of filters.??
SSB 100 Years old (1917 patent Bell Laboratories).? How far we have come...
Dave WI6R
On Wed, Jan 17, 2018 at 11:50 PM, K9HZ <bill@...> wrote: Thats great Dave.? I have not followed through myself by testing each of the stages before the PA and recording the drive levels but it should be enlightening.? Ive been focused on the antenna tuner.?
Ive thought through the AGC problem and i think there is a very good solution to replace one of the early IF amplifiers with a rebiased dual gate fet... and let the second gate be voltage controlled by the AGC circuit.? AGC should have an RF and Audio components in logarithmic function with a low pass filter.? That takes it to almost late Collins 75A4 level.?
Dr.?William J. Schmidt - K9HZ J68HZ 8P6HK ZF2HZ PJ4/K9HZ VP5/K9HZ PJ2/K9HZ ? Owner - Operator Big Signal Ranch – K9ZC Staunton, Illinois ? Owner – Operator Villa Grand Piton - J68HZ Soufriere, St. Lucia W.I. Rent it:
email:??bill@... ? uBITX PA
I will be trying out the Mitsubishi parts after I get all of the other things sorted out with my uBITX, but it is a very pleasing rig.? My uBITX starts out at 12W on 80M, 7.5W on 40M, 2W on 20M, and is down at 0.4W on 10M. so I will be checking the driver stages as well.? I expect they are not flat over the 80-10M bands either.? ?I will take it step by step and I have a good spectrum analyzer to keep things clean as I work my way through that.?? I'll post the results when I get to this down the road...??
Ahead of the potential PA mods are: 1) I just received the neat little CW Keying Adapter PCBs from the QRPGuys.? Nice job, but I'll need some smaller standoffs (4-40 or smaller mounting it looks like). 2) AGC is up after that using the unused Audio AGC amplifier to control ether IF or RF or both for AGC. 3) Finally I have an annoying high frequency tone at medium volume leaking through somewhere on receive to locate and get rid of.?
Keep the soldering irons warm...
Dave WI6R
|
You may not need such drastic surgery. Attached is a cobbled up template using side 2 of the 2822 and a pnp transistor to generate a control voltage. By lifting R10 and R11 and feeding this control voltage to Q10 as its bias voltage you can control the gain of the amp section. Or you could use Q30 instead. I left off the biasing of the 2822 and the values of the RC filter after the diode, those values would impact the attack and release timing. The pot in the collector lead of the PNP transistor would be used to set the control level being sent to Q1. All this would have to be calculated. You could even tap into the output of Q32 instead of tapping the audio line. But you would have to use a different AGC amplifier section. Something to think about. Simple and easy to build and easy to reverse. tim ab0wr On Thu, 18 Jan 2018 01:50:58 -0600 "K9HZ" <bill@...> wrote: Thats great Dave. I have not followed through myself by testing each
of the stages before the PA and recording the drive levels but it
should be enlightening. Ive been focused on the antenna tuner.
Ive thought through the AGC problem and i think there is a very good
solution to replace one of the early IF amplifiers with a rebiased
dual gate fet... and let the second gate be voltage controlled by the
AGC circuit. AGC should have an RF and Audio components in
logarithmic function with a low pass filter. That takes it to almost
late Collins 75A4 level.
Dr. William J. Schmidt - K9HZ J68HZ 8P6HK ZF2HZ PJ4/K9HZ VP5/K9HZ
PJ2/K9HZ
Owner - Operator
Big Signal Ranch – K9ZC
Staunton, Illinois
Owner – Operator
Villa Grand Piton - J68HZ
Soufriere, St. Lucia W.I.
Rent it: www.VillaGrandPiton.com
email: bill@...
|
I forgot the attachment! tim ab0wr On Thu, 18 Jan 2018 01:50:58 -0600 "K9HZ" <bill@...> wrote: Thats great Dave. I have not followed through myself by testing each
of the stages before the PA and recording the drive levels but it
should be enlightening. Ive been focused on the antenna tuner.
Ive thought through the AGC problem and i think there is a very good
solution to replace one of the early IF amplifiers with a rebiased
dual gate fet... and let the second gate be voltage controlled by the
AGC circuit. AGC should have an RF and Audio components in
logarithmic function with a low pass filter. That takes it to almost
late Collins 75A4 level.
Dr. William J. Schmidt - K9HZ J68HZ 8P6HK ZF2HZ PJ4/K9HZ VP5/K9HZ
PJ2/K9HZ
Owner - Operator
Big Signal Ranch – K9ZC
Staunton, Illinois
Owner – Operator
Villa Grand Piton - J68HZ
Soufriere, St. Lucia W.I.
Rent it: www.VillaGrandPiton.com
email: bill@...
|
"pot in the collector lead of the PNP transistor"
I assume you have the emitter and collector wrong in this diagram.
Brian VK4BAP
|
Yep, got it backwards. Emitter should be toward + and collector toward the pot! There should also be an emitter resistor and a resistor in the AGC lead toward the controlled transistor. I'll have to work on the actual layout when I have time. tim ab0wr On Thu, 18 Jan 2018 14:28:00 -0800 "Brian" <vk4bap@...> wrote: "pot in the collector lead of the PNP transistor"
I assume you have the emitter and collector wrong in this diagram.
Brian VK4BAP
|
K9HZ
