EMRFD Message Archive 12259

Message Date From Subject
12259 2016-02-02 11:45:51 Ken Chase Hycas-Cascode amp

Hi All


I am experimenting with Wes' cascode amp to use as a 7 MHZ preamp. The circuit is in the link below.




Looking at figure 3 the input network is for 9 MHZ, so I need to scale this for 7 MHZ. The 5.8 uH on a T50-6 is 328.0 ohms. So for 7 MHZ at 328.0 ohms I get 5 turns on a FT37-43.


Am I doing this correctly?





12260 2016-02-02 11:54:54 kerrypwr Re: Hycas-Cascode amp
#43 ferrite is not suitable for this use; what you have described will be an RF choke as, for example, L2 & L3 in Fig. 4.

Stay with an iron-powder core.

Kerry VK2TIL.
12262 2016-02-02 13:23:44 Tayloe, Dan (Noki... Re: Hycas-Cascode amp

You don’t want to use FT37 -43 for this as the Q is poor.


I would just scale the inductance to make it larger.  5.8 uH * 9 MHz (the source frequency)/ 7 MHz (the target frequency)


That is about 7.45 uH as a target.  That is 43 turns on a T37-2 (use # 32 wire) or 39 turns on a T50-2 using # 30 wire.


That is a lot of turns and you might not have wire that fine.  If you want, you can stack two T37-2 cores and use formulas for a single T37-2 using half the inductance.  That gives 27 turns, which means you could use #26 or # 28 gauge wire.


-          Dan, N7VE


12264 2016-02-02 15:27:45 Gene W5DOR Re: Hycas-Cascode amp

Hey Ya'll.  I have some spare HYCAAS pcb's if anyone needs one.  I have 5 and I only want to keep 2 of them.


Gene, W5DOR




12265 2016-02-02 19:47:54 Ken Chase Re: Hycas-Cascode amp
Thanks Kerry.

73 Ken VA3ABN

12266 2016-02-02 20:05:56 Ken Chase Re: Hycas-Cascode amp
Thanks Dan.

I have all sizes #20-#32 on hand but if I can wind less turns, the better.
Now in Fig 3 can I use the -43 for the op transformer.



12267 2016-02-02 21:32:44 kerrypwr Re: Hycas-Cascode amp
Yes; the Fig 3 schematic shows an FT37-43.  Note, though, that T1 in Fig 4 (the final design) is a little different, a 16:4 ratio on an FB43-2401.  I think that the FB43-2401 has roughly twice the cross-sectional area of the FT37-43 so the 4 turns on the FB will have much greater inductance than the 2 turns on the FT.

A design rule for this kind of transformer is that the shortest winding should have a reactance of four times the impedance that it "sees"; I'm not sure what impedance the 4-turn winding "sees" in the Fig 4 circuit but, if it's 50-ohms, an FT37-43 requires four turns to reach a 7-MHz reactance of a little over 200 ohms.

Perhaps that's why the core size was changed in the final design.

As Dan said, you can "stack" two FT37-43s; that will, I think, approximate an FB43-2401.

We have just had a good discussion on this kind of "conventional" transformer;



My tests, described in that topic, indicated that interleaved windings are better.

Kerry VK2TIL.
12268 2016-02-02 22:53:15 vasilyivanenko Re: Hycas-Cascode amp
Good show Ken

Some great advice shared by the group!

I’ve built 4 of these wonderful stages and it remains a go-to design . On the last version, built with Ugly Construction, I plied a FT37-61 ferrite core for L1.

Some 61 material wound coils in my collection measured a QuL of 135 - 161 @ 1 MHz – enough for many single frequency network match schemes without excessive losses. The only advantage over powdered Fe is you can make a coil such as a 7.1 uH job with just ~10-11 turns of wire. This is especially helpful for tuned circuit coils > than 10 uH  -- more so for us older guys with presbyopia  and shakey hands who dislike winding a meter of wire (or so) on a little powdered Fe donut.  They're not for VFOs however.

Wes’ L1 /C2   L -network Q is ~8 or so, but for me the ferrite inductor worked OK --- and seemed much easier to wind with only 10 or so windings.  In-situ adjustment is needed with thumb and forefinger to get reasonable precision with respect to the L value since you have only ~10 turns of wire.

I connected a return loss bridge to the input and squished or expanded the coils while tweaking the trimmer to get a proper match.  -- Just 1 option to consider on the bench for some situations --

Further, the #61 materials can make excellent RF chokes.  The 43 material with a Q <= 10 often works just a little better than a decoupling resistor for DC line filtration purpose @ some frequencies -- although without the DC voltage drop of a resistor.  The boosted Q in the #61 coil can increase the attenuation and bandwidth in a properly designed RF choke network.  All verifiable in a sweep system.



12269 2016-02-02 23:46:15 kerrypwr Re: Hycas-Cascode amp
I also thought about #61 instead of iron-powder;Fair-Rite data shows that the loss component of #61 is just beginning to be apparent at about 7MHz.  #61, or something like it, is commonly used for ferrite antennas in broadcast-band radios as it has reasonable Q in, and a little above, the AM broadcast band.

I wound 12 turns of 0.4mm wire on an FT37-61; that gave about 7.5uH as Dan calculated.

A quick & dirty sweep showed that the resistance was about 6 ohms; the reactance is about 350 ohms so the Q is about 60.

That is roughly 20% of the Q that might be expected from an iron-powder core in similar circumstances.

If striving for perfection, iron-powder should be used in this particular application; if a smaller size of inductor and less winding turns is a consideration and greater loss is acceptable, #61 would do.

Engineering, in its many branches, is a game of trade-offs.  :)

Kerry VK2TIL.
12270 2016-02-03 03:06:00 Dennis Anderson Re: Hycas-Cascode amp
He Gene, will you ship one to uk if any left. 


12271 2016-02-03 04:00:55 Roelof Bakker Re: Hycas-Cascode amp
Hello all,

Further in regard to ferrite types for rf transformers, for wideband
use. Accidentally, I found that a Magnetics 0W40705TC core, AL=8350,
with 12 bifilar windings is good (-3dB) from 4 kHz (!) to 200 MHz.
The inductance for each winding is 1.2 mH.

This material is intended for LF use and the only drawback is that
it is electrically conductive. So a layer of clear lacquer is on

Used as a rf isolating transformer with two three turn windings
opposite each other on the core, it is good from 50 kHz - 50 MHz.
In this case the loss is 1 dB, which is no problem when your
reception antenna delivers enough signal.

12272 2016-02-03 09:11:37 Gene W5DOR Re: Hycas-Cascode amp

I'm sorry but all the extra HYCAS boards are gone.


Gene, W5DOR





12273 2016-02-03 09:19:55 KENNETH CHASE Re: Hycas-Cascode amp
Thanks Gene



12274 2016-02-03 09:56:23 Alberto I2PHD Re: Hycas-Cascode amp
12275 2016-02-03 10:13:57 Scott Nolde Re: Hycas-Cascode amp
Hi Gene,
I would like to have one of the boards if I may.

Scott Nolde
cell:    2052295892
office: 2056633332

12276 2016-02-03 12:02:18 Vern Re: Hycas-Cascode amp
You can find board info info at:


Sent via the Samsung Galaxy S®6 active, an AT&T 4G LTE smartphone
-------- Original message --------
12277 2016-02-03 13:57:06 Alberto I2PHD Re: Hycas-Cascode amp
12278 2016-02-03 16:33:06 vasilyivanenko Re: Hycas-Cascode amp
Thanks Kerry

The first kit version [ supplied by Roger ] I built in ~ 2008 ran a 6.8 uH molded choke for
L1. Anything with a Q >= 50 worked fine --- in fact, it eliminated some parasitic oscillations associated with
a Fe toroid L1 version experienced by some of the early builders of this stage.

Best T/V
12279 2016-02-03 17:44:45 Ken Chase Re: Hycas-Cascode amp
I think I will give both cores a try, just to compare. I will wind a FT37-61 and a dual T37-2.



12282 2016-02-04 19:30:40 w0gk Re: Hycas-Cascode amp


  I’ve built this circuit and used an 8.2uH molded choke in the input matching network with good results at 8 MHz. These chokes typically don’t have very high Q (I didn’t measure it), but that’s OK in this usage.

  If you have a return loss bridge you can tweak the trimmer (C2) in the input network for best match. I was able to get a bit over 22dB input return loss.

  Note that Figure 3 and Figure 4 show different values for the gate resistor; Rin=2.2k on Figure 3 and R7 = 3.3k

12326 2016-02-09 18:48:42 Ken Chase Re: Hycas-Cascode amp
Hi All

I got the Hycas amp working using two stacked T37-2 with 30 turns #26. The gain control works nicely. One problem is the o/p signal is distorted on the positive part of the signal. I think is because I am driving the amp too hard.

My source is a SI5351 through a step attenuator. But f I attenuate too much so the o/p is not distorted, I can't get a scope signal strong enough.



12329 2016-02-09 22:07:03 vasilyivanenko Re: Hycas-Cascode amp

Hi Ken, although I built my last version 8 years ago, here’s how I put it together and tested it so at the end– no surprises; it worked as advertised.   In the statements, I’m assuming a 9 MHz IF, although that can be any IF.   Further, I’m going from memory, so some of the voltages might be off --- but you will get the idea.


Build the first hycas stage and terminate the Q1 collector via a series 10n cap and a 1K resistor shunt to ground. Put a temporary BNC or SMA connector on the input. Hook up the string of 3 diodes to the FET source.

Temporarily hook up a 10K pot to allow you to manually adjust the “AGC” voltage – the DC voltage applied to the common base amp in the hycas stack(s). Middle lug to the Q1 base via a snubber or isolation resistor; another to ground and the other to the 9 VDC supply. You may want to also put a 10K resistor between the DC supply .and the hot lug to limit the maximum DC voltage to ~ 4.5 VDC ( ~½ VCC) assuming a 9V DC supply


Connect a return loss bridge to the input and match the L-network.  I got ~ 23-26 dB by squishing the coils and trimming the trim cap. Once you got the 9 MHz match, remove the RLB set up.

Connect a 50 ohm resistor --- or just us a 50 ohm BNC/SMA terminator across the input. Put your 10X  ‘scope probe  across the shunt 1K output resistor. Tweak the 10K bias pot up and down. Changing your time scale on your ‘scope (or DSO) look for oscillations at low frequency, HF and if possible UHF. If you see oscillations you will need to stop them and if so, they will most likely happen with a higher DC bias.

As aforementioned, choosing a high Q  L1 coil with lots of wire/windings might increase the tendency to turn this IF amp into an oscillator --– especially when you have all 3 hycas stages hooked up -- and throttled up with lots of DC bias on the BJTs. With the #61 ferrite [ and my small selection were hot parts with way higher than normal  QuL measured carefully on a freshly calibrated HP-Agilent network analyzer ] or low Q molded choke   I found no parasitic oscillations in my 4 builds of this IF amplifier.

Now , remove the 50 ohm  input termination and connect up a 9 MHz signal at say 0 dBm or so to the input . Tweak the 10K bias pot. In your ‘scope you should see the gain increase along with the DC bias on the common base amp in the hycas stack.  Measure the DC bias with your DMM and watch it go 0 to ~~ 4.5 VDC maximum. [Going above ~ 1/2 VCC bias will just make distortion] You should not see parasitic oscillations, nor distortion of the AC signal in your ‘scope. Measure your DC voltages and stage DC current at min and maximum bias.  Does everything make sense?

Next build stage 2 (Q3-Q4) and connect it to first hycas stage per the schematic and terminate it with a 2K2 or so resistor – AC coupled; shunt to ground. Connect the 10K bias pot wiper to the isolation resistor (think it was 220 ohms or something) going to the 2nd stage BJT base -- so you are now biasing up both hycas common base amps.  Connect the diode string to that FET source. Run all the AC tests again except for return loss measures.

....Then add the third stage……. You can even calculate maximum gain at maximum bias and most importantly determine of your amps are working sweetly, are stable and ready for the AGC circuitry. Remove the temp 10K bias pot when done.

I built and each logical part of the AGC circuitry and tested at DC (and AC as possible). The old build a stage-test it and then move on when things are correct story. By the way, I absolutely love this AGC --  Wes on fire!

Repeat the AC tests with the AGC circuitry connected….

Gotta go to bed – hope this makes some sense



12330 2016-02-10 03:19:50 Ken Chase Re: Hycas-Cascode amp
Hi V

Thank you for the informative testing procedure. I guess its time to build a RLB.

A few things I must mention. Firstly I am using the amp as a preamp for an RX and it will be for 40 M or 7 MHZ. Also I am NOT building the entire amp as shown in Figure 4 in the link below.


I am actually building figure 3 in the link, a single stage.

The input network was scaled to 7 MHZ with L being 30 turns on 2 stacked T37-2. I don't have a RLB, so I peaked the signal at the gate of J310 with my scope. Yes I should invest in a RLB.

My DC control voltage for AGC is setup so the voltage range is approximatley 1 to 4 volts.

As I suspected, I was driving the amp too hard. Adding more attenuation with my step attenuator, the o/p signal cleaned up nicely. It is now a nice sine wave.

I have to remember that a S9 signal is only 50 uvolts, not the 200 mvolts or so I was using.

My SI5351 o/p signal is aproximately 200 mvolts unloaded and square wave. Loaded it is a little under 200 mvolts and the signal is distorted, not square.

On the o/p of the network loaded, the gate of J310, the signal is sinusoidal and approximately 700 mvolts.

On the amp o/p, secondary side of FT37-43 transformer, the signal is sinusoidal and varies with the AGC dc level. The signal is no longer distorted.

Unfortunately I can not see the signal on my scope when I turn down the AGC voltage. It disappears at approximately30 mvolt level.



12332 2016-02-10 07:25:22 Bill Carver Re: Hycas-Cascode amp
I'm using HYCAS as preamp/attenuator ahead of J310 regenerative detector
on both 80 and 40 meters. HYCAS gain controlled by potentiometer for an
RF GAIN control so regenerative detector is not overloaded by big
signals. Yet can bring up DX on 7003 in the evening. Great combination.

I'm not using an L network for at input, though. Instead have a link
coupled parallel tuned circuit with the gate of the J310 tapped down
with a capacitive divider. That is, on 40m, gate of J310 has 27K in
parallel with 330 pF to ground and 75 pF to the top of a T50-2 with
about 30t of #28 and a small variable to resonate it to the RX freq.
Bottom of coil grounded. The antenna goes to a 4t link winding on the
toroid. This is NOT impedance matching, just bringing the antenna
voltage up a bit and giving the J310 some protection from off-40m signals.

HYCAS collector output has a one turn link on the coil of the regen

12335 2016-02-10 16:12:20 vasilyivanenko Re: Hycas-Cascode amp
Clever. Thanks Bill
12338 2016-02-11 03:40:33 Ken Chase Re: Hycas-Cascode amp
What is the proper way to test the gain of this amp?

Do I measure i/p signal at the source or measure at the gate of the J310?

I am referring to the amp in the link below FIGURE 3.




12343 2016-02-11 14:16:12 John Harding Re: Hycas-Cascode amp
Glad this device has raised some questions - I plan on using one at 455kHZ following 2 selectable xtal filters with, I believe,  500 and 600 ohm impedances. Has anyone used the hycas-cascade amp at 455kHz??
I can lose some gain without a problem. I could just replace R7 per the app notes with 550 Ohm. I could used the pg 4 circuit of the app note with 2 matching L -networks via a 50 ohm link, however room is getting tight....if anyone used the amp at 455kHz the same situation probably was probably already addressed. I have lots of selectivity ahead of this and probably too much gain (dual conversion)

12348 2016-02-11 20:23:35 vasilyivanenko Re: Hycas-Cascode amp
At the source.  

That's your available power --- from a 50 ohm output impedance sine wave signal source
12350 2016-02-11 23:41:23 Gene W5DOR Re: Hycas-Cascode amp

Hey ya'll.  I have one more HYCAS pcb I can part with.  It has been modified by me to include R38 which is an  18 or 22 ohm resistor in series with C12 between emitters of Q7 and Q8.  This mod has been mentioned in the papers about  this circuit but wasn't included in the original pcb layout.


First come first serve.  Email me off group.


Gene, W5DOR



12351 2016-02-12 03:05:56 Ken Chase Re: Hycas-Cascode amp
Can you ship to Canada? Contact me off group if you can.