No RF gear here?

Administrator
Joined 2004
Paid Member
Hi w5jag,
Sounds like you're winning with that project. This should spur you on to complete it.

Just went to a Hamfest yesterday in Milton, Ontario. Smaller turnout of dealers and guys who sell from their trunks, but lot's of good stuff that spans the categories from junk, old computers and bits, test gear (some relics) and even some HAM related stuff. I scored a book on antennas, a Heathkit IT-18 transistor checker in pristine condition
and also an M-Audio Pre-USB mic preamp. Lots of things I would love to buy, but the cash just wasn't there. There's always next year!

-Chris
 
Member
Joined 2016
Paid Member
Your project is in the finest tradition of amateur radio - you've experimented a lot, and learned a lot! And, I've lurked but not commented before, but I think many of us are enjoying the process! Keep up the enthusiasm and good work! Definitely the Right Stuff!
 
Thanks for the words of encouragement - I thought this project would be drama free, especially considering that I have a pre built 9 MHz back end.

It's my own fault for digging my heels in with that seriously flawed VFO. I must have unknowingly damaged the firmware implementing the fixed LO output. Since I cannot rule out that I am the one that botched the thing, I have not shared my observations with the designer / vendor.

With my "antenna" connected, the background noise increase at 14 MHz is slight, so I added an MMIC between the crude front end filter and the input to the mixer. This definitely increased the gain - it's hard to say for sure that it increased the sensitivity. This NEC upc1651 is not a low noise device to start with, and it's not implemented correctly. It's unmatched at the output ( I was hoping that would lower the gain ), and I haven't got the voltage right. The data sheet says best NF is between 5.5 volts and 6 volts ( max for the device ). I can't get it dialed in with the SMD 1206 zeners I have - a 5.6 volt device is giving me about 5.1 at the chip, and a 6.X device put me at 6.6 - over the max spec. It makes a difference - according to the sheet, NF goes up 2.5 dB between 5.5 and 5.1 volts ...... I find rework to be really difficult on this type of board with my present tools. You really need to get it right the first time.

This is not the front end filter that will actually be used, so fooling with it is a dead end. The MMIC's are interesting though - it looks to be next to impossible to make one oscillate, even when badly implemented.

Win W5JAG
 

Attachments

  • IMG_20180716_202131.jpg
    IMG_20180716_202131.jpg
    510.6 KB · Views: 230
As part of the SMD thing, I've had these HP IAM 81008 chips laying around, so I tore down the 2N3823 mixer, and substituted one of these in it's place.

Not much to it - 50 ohm in and out, so I just hooked one end to the input filter, and the other end to the FT-817. A couple of bypass caps, blocking caps, the usual decoupling from the supply line, that's pretty much it.

Seems to work okay; the battery was low on the FT-817 so I didn't get much time to play with it, but it heard the local 25 watt LF NDB, and 50 watt BCB station in the next town over, quite well. CW on 40 meters was good, also.

These chips are cheap on eBay, and apparently genuine. Datasheet attached. There is an old patent for it, #5379497, but my copy is too big to upload

Win W5JAG
 

Attachments

  • IMG_20180717_205340.jpg
    IMG_20180717_205340.jpg
    391.9 KB · Views: 211
  • IAM81008.pdf
    58.4 KB · Views: 35
The IAM 81008 looks like it could have some potential.

It is a doubly balanced mixer, so here are some FFT captures of the port to port isolation. I am taking the IF at 45 MHz, and was listening to CW on 20, so the VFO is running about 59 MHz.

I am running the VFO flat out, so this could be right at, or slightly over, the spec for LO input. The chip is at 8 volts, again the maximum recommended operating spec for this device. Max, period, is 10 VDC.

Capture 1 is the VFO amplitude at the LO port of the mixer. I put the pip at the horizontal mark to make it easy.

Capture 2 is at the IF port - the LO is suppressed about 25 dB.

Capture 3 is at the RF port - LO suppression here is only about 17 dB, and some spurious response is showing up.

I haven't looked at the rules since last summer, but I doubt that this meets the rule for spurious emissions, so something more is needed here. I do recall that the rule was that spur readings have to be taken at the antenna port - any suppression from a narrow bandwidth antenna can't be counted in meeting the rules.

The low pass filter at the antenna input should take care of this, so I took a look at it, and determined it's not working properly - at the antenna terminal, I am only seeing an additional 3 dB attenuation of the LO at 59 MHz. So, if I go with a wideband front end, that will have to be sorted out. The MW high pass filter looks to be working as intended.

I will probably follow up with a double conversion down to 9 Mhz to see if these chips have any merit, or not.

Win W5JAG
 

Attachments

  • pic_59_1.jpg
    pic_59_1.jpg
    48.7 KB · Views: 191
  • pic_59_2.jpg
    pic_59_2.jpg
    47.4 KB · Views: 199
  • pic_59_3.jpg
    pic_59_3.jpg
    48.1 KB · Views: 198
Last edited:
So, maybe someone more experienced at reading data sheets can look at this Toko sheet, which appears to be the only slug tuned SMD coil in the Mouser catalog.

I made a narrowband front end with the slug set to midpoint, and an appropriate resonating cap, only to find it was impervious to 14 MHz signals, and no amount of diddling would change that. I swept it with the Feel Tech function generator while looking at the FFT on the Hantek, and resonance was well into the low VHF range. I pulled the coil, measured it as best I could, and found it was nominally about 0.23 uH, and can be varied from about 0.2 to 0.3 uH. Thinking I had damaged it, I checked another and got the same result. Maybe not coincidentally, 11 pf and 0.23 uH is resonant at the 100 MHz test frequency specified.

Looking at the sheet, I thought I was getting a slug that would go from 0.05 to 2.7 uH.

So, I backed up and made a single tuned circuit with a toroid and a link, that worked OK, but looked too fragile for the rough roads around here.

I switched to a fixed inductor, and instead of a link, rigged up a series capacitive divider to get a Lo-Z point for a 50 ohm antenna, as well as a Hi-Z point for the random wire I am testing with. This looks like it might work, and the trimmer can be substituted with a tuning diode that can be tuned remotely if I stay with just one pole, or multiple poles can be stagger tuned to make a fixed front end.

I probably need to diddle with the capacitor values in the divider some, this is just for a test of concept.

I was listening last night with this set up hooked to the FT-817 as a 9 Mhz IF, RF amp off, and 20 dB attenuator turned on to make the FT-817 as deaf as possible, and heard a local guy about two towns north working phone on 20. From about 11:00 PM local to midnight, he worked Nebraska, Michigan, Indiana, Minnesota, and Wisconsin, and, with the random wire, this set up could Q5 copy every station he worked.

I ordered some more of those HP mixer chips, but haven't had a chance to do anything more with the ones I have.

The board looks just flat out awful at this point. It is not apparent from the picture, but I am starting to mix in some 0805 with the 1206. I have 0603 also, but haven't tried any 0603 yet.

Win W5JAG
 

Attachments

  • tuned circuit.jpg
    tuned circuit.jpg
    278.8 KB · Views: 53
  • IMG_20180730_084307.jpg
    IMG_20180730_084307.jpg
    465.2 KB · Views: 59
  • toko smd coil.pdf
    247.3 KB · Views: 41
I played with this circuit for an hour or so last night.

I set up the FY6060 as a signal source at 14 MHz, injected into the "Lo-Z" series point, hooked the scope up to the gate, and peaked the trimmer cap for maximum response.

I then swept it with the FFT, and found it was actually resonant at about 12.5 MHz, the trimmer was just going to minimum capacitance. I changed the lower cap to 130 pf in parallel with the trimmer, and this brought it to resonance in the twenty meter band.

It doesn't behave like other single tuned circuits I've played with / used. Instead of a sharp response on the resonant frequency with a steep slope, if the Hantek FFT is to be believed, and I am testing this correctly, it has a very flat response across the resonant frequency, with very shallow skirts. Shallow enough that I only saw 6 dB difference between the original resonant point and the bottom of the twenty meter band. Once modified, I didn't have any trouble finding a point where it was ruler flat across the twenty meter band, together with a good 75 Khz either side of it.

I"m aware of Q, but won't pretend that I understand it. I came across some slug tuned coils in my junkbox that would work here, and eliminate the need for the trimmer cap, and that might have a sharper response than this molded inductor, but I'm not sure the broad bandwidth is a bad feature. A few poles of this would sharpen the filter, and it should not need to be peaked from the front panel. I just want to keep what I don't need out of the weak mixer.

Win W5JAG
 
Looking at the sheet, I thought I was getting a slug that would go from 0.05 to 2.7 uH.
That seems like a really wide range (50 to 1) for a slug-tuned variable inductor. I wonder if it's a misprint. Do other models from that manufacturer have (or claim to have) similar ranges?

There's something called a variometer (google it and see how it works) that can go over a wide range of inductance (two coils, one on an axis, they could add or cancel depending on the angle of one in relation to the other), but it seems doubtful that a slug-tuned coil could do that.
 
There's something called a variometer (google it and see how it works) that can go over a wide range of inductance (two coils, one on an axis, they could add or cancel depending on the angle of one in relation to the other), but it seems doubtful that a slug-tuned coil could do that.

Also -- Collins R-390 used a "Permeability Tuned Oscillator" which was quite linear.
 
They are indeed. Everyone should own an R-390, R-390A at least once, if your back is in good shape. Sadly, I am old enough to remember when you could get those Collins PTO's from mil rigs as cheap surplus.

I think these Toko SMD coils are the only shielded SMD part in production. Relevant to this trainwreck, these coils are not totally useless - they might make an okay 45 MHz IF transformer. Would take a lot of C, but it would probably be okay for amateur grade stuff. A lot easier than winding a toroid.

There is probably surplus stuff on eBay that would be better for HF.

Win W5JAG
 
That's AM radio in style. If you can find ( or make ) a loop antenna, it helps a lot with BCB reception.

My last remaining Collins general coverage receiver - I don't know what agency specified this green finish, but I like it - reminds me of elementary school! Swapped an extra Drake R-7 even up for the 51S-1.

Win W5JAG
 

Attachments

  • Palomar Loop Antenna.jpg
    Palomar Loop Antenna.jpg
    48.7 KB · Views: 149
  • Collins 51S-1.jpg
    Collins 51S-1.jpg
    85.1 KB · Views: 162
I once made a loop antenna from a copper tube, about 2 meter diameter, with matching anode coil of the same inductance, to tube both by a single 2-gang cap, with 12AT7 cascode. It was pretty cool! I listened to hams at 3.5 mHz talking about sound quality of their transmitters and microphones. :)
The receiver I used then, was a BC-348, and a Fisher stereo as a second converter and IF. :)
AFAIR,it was tuned to 1600 KHz - IF of BC-348.
 
Looking at the sheet, I thought I was getting a slug that would go from 0.05 to 2.7 uH.

So I went and looked:
https://www.mouser.com/datasheet/2/281/5cceg-606175.pdf

Yeah, that could be read a few ways, but it appears what it's really saying is "we have versions of this part with max inductance ranging from of 0.05uH to 2.7uH."

I see four part numbers, but I don't see any inductance listed. Looking in the Mouser online catalog shows something about capacitance, but the inductance column just has a dash (-). What's going on here?


ETA: At least the Coilcraft parts list inductance:
Variable Inductors | Mouser
 
Last edited:
Mouser only carries the -0001 Toko part. The -0002 or -0003 would be more useful, since they have a center tap on the primary.

The CoilCraft parts are fine, but they are leaded, and they don't have a secondary. AFAIK that Toko 5CCEG part is the only slug tuned coil that is shielded, has a secondary, and is SMD.

I looked at DigiKey and they have it:

Adjustable Inductors | Inductors, Coils, Chokes | DigiKey

but they list the 50 nH part which is way worse, if correct.

Win W5JAG

edit: real IF XFMRS are here, but 455 KHz types are all that are available, as far as I can tell:

Xicon IF Audio Transformers / Signal Transformers | Mouser
 
Last edited:
From some Murata literature:

Murata Variable Coils for Signal Circuit are inductors ideal for high-density mounting in a wide range of compact equipment. With a frequency ranging from 50MHz to a high of 120MHz, the series offers an inductance range of 0.05µH to 2.7µH and a Q of approximately 70 (at 100MHz).

Note the words "the series offers an inductance range of 0.05µH to 2.7µH"

It is rare for a single variable inductor to have more than a 50% inductance range, especially at 100 MHz.


Note this on the data sheet:

"5CCEG type is custamized item. Please feel free to contact us for details."

These are made to order by someone ordering big volume. You got a "standard" part designed for operation at 100 MHz, probably the one given out by sales reps for engineers to stick in their component measuring equipment or test circuit for correlation with the factory specs.

When I did RF design for Motorola I got Coilcraft to make me whatever I wanted and often the custom part arrived in a day or two. I also had a large cabinet full of every design kit they had. I might still have a few in a box somewhere.
 
Yes, it appears I got the generic part that resonates with the specified 11.4 pF at 100 MHz. With 20 / 20 hindsight, the clues are all there.

As usual, my mistake is even dumber than it looks on the face of it. I've actually ordered these parts twice - a few at first just to see if they were something I thought I could work with on this style protoboard. Did I bother to verify the inductance when I actually had some in hand?

No. No I did not.

So I got some more on my next order, and only measured them when they didn't work like I expected.

The generic part is by no means ham diy useless - as is, they should be okay for a 45 MHz IF XFMR, or front end coils for a six or two meter rig, or an FM BCB tuner. I suppose a couple of them could be put in series to get more inductance. Might be able to be useful at 14 MHz that way. That would also make a center tapped coil, and the slugs could be tweaked for balance for push pull stuff? Might be interesting to play with them at some point in the future.

I will probably get some more on my next order in case they go extinct.

Win W5JAG
 
....The low pass filter at the antenna input should take care of this, so I took a look at it, and determined it's not working properly - at the antenna terminal, I am only seeing an additional 3 dB attenuation of the LO at 59 MHz .....

So, it occurred to me that attaching a random wire antenna to a 50 ohm filter port might not accurately (duh) indicate whether or not the filter was working properly.

I put a 47 ohm resistor to ground on the input of the hi pass filter, and another 47 ohm resistor to ground at the output of the low pass filter, and retested the filter.

The signal generator is the cheap FeelTech FY6600 waveform generator. The measuring device is the Hantek 70 MHz DSO. Test frequencies were 30, 40, 50, and 60 (upper limit of the FeelTech) MHz. Signal amplitude was 200 mv p-p. At each test frequency, I checked the FeelTech output, and reset the level to 200 mv p-p, so the filter saw a constant amplitude across the measurement range.

The first FFT is at 30 MHz, and in order, 40, 50, and 60 MHz. Horizontal divisions are 15 MHz - I re-centered it on 75 MHz so the 12.5 MHz displayed by the scope tag is not applicable. Vertical scale is 10 dB per division

The low pass portion is five poles, intended to begin attenuation at a bit past 30 MHz. It appears to be working properly, possibly a few dB better than predicted, considering this is pretty mediocre test equipment.

The 5 pole BCB high pass filter is a virtual brick wall. At 2 MHz, the test signal is there - at 1.75 MHz, it is basically just gone. These measurements are consistent with the observed performance. Even the strongest local BCB stations (5 kw daytime) are undetectable unless the high pass filter is bypassed.

Win W5JAG
 

Attachments

  • 30 MHz reference FFT.JPG
    30 MHz reference FFT.JPG
    49.5 KB · Views: 118
  • 40 MHz reference FFT.JPG
    40 MHz reference FFT.JPG
    48.1 KB · Views: 110
  • 50 MHz reference FFT.JPG
    50 MHz reference FFT.JPG
    52.4 KB · Views: 108
  • 60 MHz reference FFT.JPG
    60 MHz reference FFT.JPG
    46.1 KB · Views: 52
  • 2 MHz reference.JPG
    2 MHz reference.JPG
    52.9 KB · Views: 50
  • where did it go.JPG
    where did it go.JPG
    48.5 KB · Views: 54
Last edited:
A cautionary tale for the reflexive capacitor changers.

Over the weekend, I changed out the electrolytic caps on the SG-9. The SN76514 (AKA TL442CN) IC is date coded 1978, so that sets the vintage of the parts. There were eighteen or so electrolytics, no tolerance marked on the parts, but I presume they were the +80 - 20 typical low cost part.

After changing them, I checked the value of the pulled parts on a DMM, and of the bunch two were spot on the marked value, one was at the - 20% tolerance, and the rest were about 25-50% over the marked value, so, likely all were still completely servicable, even at the advanced age. The newer parts are quite a bit smaller - will they still be good after forty years? Chances are I won't be around to know.

To get the sideband right on twenty meters, I also changed the 9.0015 BFO crystal to one of the new 8.9985 MHz crystals Mouser is now carrying. I bought the seventy nine cent, lower tolerance part, and it adjusted to it's marked frequency without issue.

Win W5JAG