... 1SS272 ... and ... with ... Is there ... think ... Pete uses the 1N5767 axial lead PIN diode with good results. He does not recall that his former DX-394Message 1 of 6 , May 1, 2005View SourceSynopsis of comments from Pete, KE9OA:
--- In RADIOSHACKDX394@yahoogroups.com, "Tom Holden"
> I'm wondering what one might do to improve its strong signal
> intermod performance.
> It uses HSU277 diodes to switch among 3 SW preselector bands and
> diodes for the LW/MW bands. I have nearly doubled the reverse biasand
> forward current by switching from the 7V supply to the 13.8V supplywith
> little obvious benefit. At best, I can shove 10-12 mA through them.Is there
> any point in changing the diodes? If so, what would you suggest? Ithink
> SMD's are preferable so as not to degrade stray coupling.Pete uses the 1N5767 axial lead PIN diode with good results. He does
not recall that his former DX-394 suffered badly from intermod and on
reviewing the specs for the stock diodes suggested that it might be
good to leave the unit as-is. However, he suggested I look at the
Icom IC-R75 for another reason and I found that it uses PIN diodes on
the inputs of its band preselectors and diodes similar to the DX-
394's on their outputs. That's intriguing - there must be a reason.
ARRL Labs reported that the DX-394 Two-tone, third-order IMD dynamic
range: 14 MHz, CW1 position, 84 dB at 100-kHz spacing. For the IC-
R75, 86-91dB, depending on choice of preamp (none, #1 or #2) but the
spacing was not shown in the extract available at
http://groups.yahoo.com/group/icomr75/files/Information/ titled arrl-
comparisons.pdf. Elsewhere in the extracts, it says the ARRL Lab
standard spacing is 20kHz but it also used 50kHz spacing for one
radio. There was no 3rd order or 2nd order intercept measurement on
the DX-394 to compare against these other radios.
> The RF preamp is a 3SK195 and the 1st mixer uses a pair of these
> for RF and driven in parallel by the LO. Is there any point inparalleling a
> second 3SK195 (piggyback style) on the preamp? on the mixer? Orchanging out
> the transistor type? How would one determine/set the correctoperating
> point?Pete suggested I look at the IC-R75. Indeed, it uses two 2SN2171
FET's, paralleled pin-to-pin, in the 1st RF stage. However,
it's a grounded-gate JFET configuration while the DX-394's 3SK195 is
a Dual insulated Gate MOSFET common source (unless I have the source
and drain mixed up). The 3SK195 is described as being for TV Tuner,
VHF RF Amplifier Applications, FM Tuner Applications and exhibits
Superior cross modulation performance, Low reverse transfer
capacitance: Crss = 0.015 pF (typ.), and Low noise figure: NF = 1.1dB
I think piggybacking a parallel 3SK195 on Q1 could be worth trying. I
did it once without soldering and was not aware of any effect but it
would be necessary to cut the resistors R36 and R38 in half to get
the same current flow through each as the current through Q1 alone.
Another idea coming from the IC-R75 is a control to turn off/bypass
the RF preamp stage. It's probably not needed below 10mHz and
intermod performance is improved in the R75 by running without it.
Any comments, suggestions, advice???
73, Tom VE3MEO
Got the Dec 94 issue of QST out of the stacks at the Metro Reference Library. Interesting reading and raises more questions as well as suggesting a course ofMessage 1 of 6 , May 5, 2005View SourceGot the Dec'94 issue of QST out of the stacks at the Metro Reference
Library. Interesting reading and raises more questions as well as
suggesting a course of action. W0IVJ's tests demonstrate that PIN
diodes intended for RF switching are superior to PN switching diodes,
especially at low bias. Surprisingly, the common 1N4007 1A rectifier
which has a PIN structure equals or outperforms the MPN3700 and BAR17
RF PIN diodes for intermod at 5-10mA bias but its reverse bias
isolation would require two of them in series to match these others.
The DX-394 preselector band switching diodes are not described as PIN
but they are intended for RF switching - does this mean that they may
be as good as PIN diodes or might they suffer from higher intermod?
The 1N4153 PN diode tested had excellent isolation but lousy
intermod, great for turning something off but poor when turned on!
The author measured a 10-12 dB improvement just by raising the
forward bias from 5mA to 10 mA, as I have done in an earlier mod.
What was especially revealing in this article was the testing of a
varactor diode as the capacitor in a tuned preselector circuit
contrasted to a fixed capacitor switched by a PIN diode. The author
concluded that "using tuning diodes to tune high impedance circuits
in the presence of strong signals is not a good idea". I'll have to
go back to the library because a footnote to this sentence was that
the same issue of QST had an article by Ulrich Rohde, the highly
respected communications engineer, on a diode-tuned filter designed
to minimise IMD. Interestingly, Pete KE9OA tells me, "One
modification I did do [to his long-gone DX-394] was to replace the
LW/MW loopstick with a low-pass for the LW band and a bandpass for
the MW band. It was a bit of work, but the results were pretty good."
What this suggests is that the bandswitching diodes and the RF preamp
may not be the main determinant of intermod performance but that
diode D8 could well be. An interesting experiment would be to switch
between D8 and a conventional capacitor. Perhaps even replace it with
a tunable cap - might have to be pretty big - probably a dual 365pF
with both halves used for LW/MW and one half for SW. Alternatively,
use an outboard preselector and disable the internal one so that none
of the diodes are in circuit.
Anybody want to try?
73, Tom VE3MEO
--- In RADIOSHACKDX394@yahoogroups.com, "Tom Holden"
> Does anyone have QST December, 1994? If this article is not toolong, could
> you scan it to me?Switching and
> Tom Thompson W0IVJ, Exploring Intermodulation Distortion in RF
> Tuning Diodes, QST December 1994, pp 25-27suggests
> It's referenced in the Kiwa article "Tips for Improving Receiver
> Performance" at http://www.kiwa.com/rxtips.html . The PIN diodes he
> for preselector switching have ON resistance an order of magnitudeabove
> that of the stock diodes but "they also exhibit very high immunityto
> distortion". It was this article that prompted me to increase theforward
> and reverse bias on the stock diodes to see if I could get thetarget 10 mA
> he cites as a good operating point for the PIN diodes. That mod was20Data%20Sheets/
> described in msg #601 and preceding at
> I've posted the spec sheets for the stock preselector diodes under
> (watch the wrap!).Intermod
> 73, Tom VE3MEO
> ----- Original Message -----
> From: "Tom Holden" <holden_family@s...>
> To: <radIOSHACKDX394@yahoogroups.com>
> Sent: Sunday, April 24, 2005 5:59 PM
> Subject: [RADIOSHACKDX394] Ideas for improved Strong Signal
> > It uses HSU277 diodes to switch among 3 SW preselector bands and
> > diodes for the LW/MW bands. I have nearly doubled the reversebias and
> > forward current by switching from the 7V supply to the 13.8Vsupply with
> > little obvious benefit. At best, I can shove 10-12 mA throughthem. Is
> > thereI think
> > any point in changing the diodes? If so, what would you suggest?
> > SMD's are preferable so as not to degrade stray coupling.