WA7MLH

My shack in Beaverton 2018

This page will have a lot of edits during 2019

I have left this page as a long one due to lack of time & effort to create linked pages.










My first homebrew xtal filters in this 20m CW/SSB TXVR using 7.5 MHz IF.
I got the xtals at the Rickreall Hamfest (OR) back in the late 1980s.
Filters are Cohn ladder topology (equal coupling coefficients with 5 xtals).
I am using them mainly due to the nostalgia of them being my first ever xtal filters.
Their performance is a bit lacking compared to proper Chebycheff design with more xtals.
The chassis was a hamfest carcass with speaker grill and S meter already set up on the front.
It was years later I realized it was silly to worry about the cost of new xtals from Mouser/Digikey.
Buying 50 or 100 at about $0.32 each should never have been an issue since I would live with
and use the filters for the rest of my life. And I could build filters with performance far beyond
anything that was available as a commercial one, or one from a typical amateur "appliance".




Hartley VFO at lower right at 6.5 MHz
RX LNA is on far left, then xtal filters, then 1st IF RX diode ring mixer and post mixer broad band amp.




Hybrid cascode IF/AGC amp. Thank you KA7EXM for the PCB design!




IF/AGC amp on the right in a shielded Bud chassis.
RX audio section in the middle.




RX BFO. I like tuneable BFO for RX over fixed xtal.





7/7 TX & RX boards for QRP CW TXVR. These are for the W7ZOI design in EMRFD.
These were the first ones, so RX xtal filter is only 3 xtals.(As in original)
Next version added 2 more xtals so that the filter design would be easier in XLAD (from EMRFD).
XLAD has bugs for 3 & 4 xtals, requiring a manual over-ride for k & q values.
I am not a proponent of board stuffing and then calling it home brew.
Stuffed PCBs are hard to debug and do not lend themselves to modifications the way ugly does.
Once an ugly version has been assembled and is up and running properly, then a PCB is fair game.
The one benefit of a PCB is that it can be stuffed and soldered quickly, and can be replicated.
Not so with ugly method.




Took me decades to realize that a roller inductor in an antenna tuner has many drawbacks.
Qu of the roller inductor is easily degraded by poor contact mechanisms for the variable "tap".
Most of the surplus rollers I have seen at hamfests and ones I have purchased had, in hind sight,
really bad electrical contact mechanics between the actual contact wheel and the wire on the solenoid.
Rollers tend to be large and heavy.
Clip leads on a solenoid inductor are easy to move and implement.
Plus, the plexiglass front panel allows me to see exactly where I am tapping.
Also, the AirDux does not need a turns counter!
And the AirDux and caps are quite cool looking IMHO.
#10 solid house wire wound on a cylindrical piece of wood or plastic would be a cheap and high Q inductor.
And the coil does not need to be cylindrical...a rectangular form will work just as well.
Unit on top is a 30W Wheatstone bridge for isolating the TX during tune up and measuring VSWR.




The CLC AirDux tuner ready for antenna matching. The wood base allows the caps to "float" in series.




Wheatstone bridge schematic.
Lots of flexibility here in that one simply needs to get the ladders on each side set up as equal value resistors.
The left divider value does not have to equal the right divider.
Values are not critical. I paralleled up a bunch of 82 ohm 2W resistors I got at a hamfest for next to nothing.
The paralleled sets were then soldered together as series strings and measured with DMM to make sure they were matched.
The nice feature of this bridge scheme is that under worst case loads of open or short, the TX nevers sees worse than
about 2:1 VSWR so it is always protected regardless of tuning. I use a DPDT switch to switch the bridge in or out.
In hind sight, I think a directional coupler and high power dummy load is the better strategy. WAY easier, and not
much to fuss with to get it up and running. See EMRFD for details. Thanks W7ZOI for this one!





This is a rebuild of a 9 MHz IF HF CW/SSB TXVR I designed back in the late 1980s.
It uses a 1/1 analog PLL with 5-5.5 MHz master VFO and 12 slaved VCOs for high side LO injection
on both RX & TX paths.
Bummer is that 9 MHz sucks as an IF for 17m band, so 17m will not happen in this one.
I had all of the RX path up and running when I accidentally had a power supply alligator clip
land on one of the KVG xtal filter pins! That fried the SSB filter.
It was a horrible sense of loss until, by sheer accident, I discovered that Mouser carries 9.000 MHz xtals!
That allowed me to keep all of the VCOs and xtal offset oscillators and build a very serious set
of filters that significntly outperformed the KVG filter by a lot!
Now this rig has 8 xtals at 3.2 KHz BW for SSB with approximately .1dB ripple Chebycheff response.
Good luck affording that kind of selectivity in a commercial filter, even if you could find it!
Now the rig has a pair of CW xtal filters of Gaussian to -6dB response for 220 Hz & 470 Hz BW.
The ugly board is the PLL down mix with 5.5 MHz LPF and some gain before going in to the phase/freq det PCB




I opted for the convenience of using a cheap LCD freq counter to count the VCOs
and simply ignore the digits above the decimal.




The KVG SSB filter response before I zorched it.
I am not big fan of 2.6 KHz SSB filters because they lack the audio quality of a wider BW.




The new home brew SSB filter. This was mesh tuned in XLAD.
I used end section matching via LC networks because ALL of the filters I have built using
discrete transformers wound up with unpredictable "ears" on the skirts close in near the -6dB area.
I never use transformers on my xtal filters any more. Also, one can make the argument that
the extra L & C improve the skirt performance.(which they do)




9 MHz .1dB Chebycheff at approx 3.26 kHz. Design target was 3.0 kHz.




Amazing response for the new home brew SSB filter.
2dB insertion loss vs 6dB for the KVG!.
Not a big deal on TX but will help on improving system NF on RX.




Comparison of insertion loss between KVG & homebrew xtal filter.
I need to get a few more data points to fill in the gap.
Then the KVG will look even uglier.




This CW filter came out to 220 Hz for Gaussian to -6dB response.




The Hartley VFO.




RX tuneable BFO.




I have switched to RCA phono jacks for shielded interconnect in this rig.
They offer great RF performance in to UHF in spite of them being considered as audio connectors.
I learned that to my surprise by measuring them on an 8753B vector network analyzer.
And they are much cheaper than BNC and weigh less for protable rigs and take up less space than BNC.




CW TX carrier xtal osc & LSB xtal BFO & USB xtal BFO.
Freq adjust happens through a hole in the lid for access to the tuning trimmer caps.
IF/AGC is in chassis under the plate.




PIN diode switched DTC band pass filters for RX front end & TX mixed band select.




PIN diode DPDT switch for xtal filter path.
Diodes are 400V Si rectifiers at approx 30mA.




The ugly version of the digital phase/freq detector.




MC1496 Jones cell balanced modulator.
Some readers might be interested in more of the history behind the 4 quadrant multiplier at
Analog RF History




My xtal characterization set up.
See EMRFD chapter (4?) on G3URR method.
Another good source for xtal characterization is the QEX paper by G3OTK at
An Automated Method for Measuring Quartz Crystals




Digital phase/frequency detector as a dedicated PCB.
I will probably use this one since it takes up less space than the original ugly version.
This is a shared (public) design at OSHPark




SMD 4 section Sallen-Key audio filter.
This was inspired by coming acrosss about 100 cheap (free?) OP074 SOIC op amps.
This is a shared (public) design at OSHPark SMD Audio LPF




Diode ringe balanced modulator for generating DSB.
This is a shared (public) design at OSHPark SMD Audio LPF




This is a nice compact broad band amp PCB designed around the TO-92 MPSH10 NPN.
MPSH10 has fT=650 MHz and is relatively cheap and easy to find.
With 350mW power dissipation and Ic=50mA (max) it is a great device for most HF applications.
Bias stabilization at DC is accomplished with emitter degeneration and AC coupled emitters
to parallel the devices at RF. This allows a lot of flexibility on the board for loading 1 to 4
transistors depending on how much power out one needs or an IMD agenda.
MPSH10s in this board work well out through 6m. Note that MPSH10 pin out is not the same as most NPN in TO-92 pkg.
4 devices at 250mW each will provide >+20dBm useful Pout for class AB SSB driver applications
or a superhet RX front end mixer post amp. The benefit of running at 250mW or less is not needing any form of heat sink!.
Another option for this board is to bias for class AB, maybe 5mA per stage (min) and use it as a QRP final at 1W out.
This design has a wide range of options for gain and power out depending on feedback and emitter resistors
and bias setting on the base.
This is shared (public) at OSHPark BBAMPx4





This is a 6 MHz SSB filter designed with 2 primary agendas.

One was to go through the exercise of extending the XLAD xtal filter software out beyond it's
default limit of 10 xtals. A bit of ASCII & notepad copy & paste allowed me to extend out to 14 xtals!!!

Two, these 6 MHz xtals I found for next to nothing at SEAPAC hamfest (Oregon) turned out to have absolutely stellar
unloaded Qs, so they begged for becoming an amazing filter.

6 Mhz would not be my first choice for IF freq, yet it is quite OK for a lot of bands on TX,
especially with high side LO injection via 1/1 analog PLL.
This RX is now a working superhet on 75m with a 9.5-10 MHz Hartley VFO. I will rebuild with 1/1 loop
and add TX to it for a dedicated 80m/75m CW/SSB TXVR with low drift master VFO running 1.5-2 MHz





Ugly build up of 15m RX front CW TXVR.
I like to build up on the bench and verify everything is working before commiting to a chassis.
IF is 16 MHz (approx) with 5 xtals for Gaussian to -6dB response and 5 MHz VFO.
I found a lot of the 16.xxx MHz xtals at SEAPAC hamfest for super cheap.
They had the feature of being extremely tight on frequency, so I did not have to do any sorting.
Their Qu was not stellar, yet quite OK for a portable superhet RX.





80m Field Day CW TXVR. This is a cascode of 2N4416 JFETs at around 3mA drain current.
The 3mA is a compromise for lowered battery drain on rx vs front end dynamic range.
I have rarely found IMD to be a limiting issue in cw contests.




80m Field Day CW TXVR front end.




80m FDay Colpitts RX BFO tuneable from the front panel.
Series inductors with the xtal were needed to pull the osc freq lower than where it wanted to be.
In hindsight, much higher Colpitts C values would have likley eliminated the need for the series inductance.






80m FDay active audio LPF. 4 sections of Sallen-Key low pass at 800 Hz




80m FDay LM386 audio amp.




80m FDay TX chain for 5W variable Pout from an IRF-510.




80m FDay TX chain PA LPF and RX LC tap off.




80m FDay VFO and external PI LPF.



radio buttons next to pictures have audio narration


WA7MLH Set 1

WA7MLH Set 2

WA7MLH Set 3

WA7MLH Set 4

Schematics

Field Day 2001

Shared OSHPark RF PCBs

Shared OSHPark Modular RF PCBs



Salem OR Shack 2008

80m/40m Band Imaging RX

80m/40m Band Imaging RX2

80m/40m Band Imaging RX3

Rickreall Hamfest 2009

2nd Rickreall Hamfest 2009

GPLA Double Tuned Circuit

GPLA 20m Double Tuned Circuit Tutorial

Homebrew HF Power Amps

EMRFD DTC Filter Design Tutorial

15W HF LDMOS Linear Amp


Updates: November 2009

Rickreall Feb 2009 Hamfest

Rickreall 2003 Hamfest

Rickreall October 2003 Hamfest

Rickreall October 2004 Hamfest

Rickreall February 2006 Hamfest

SEAPAC 2007 Hamfest

Rickreall October 2008 Hamfest

Rickreall October 2009 Hamfest



return to neoanderthal Main Page