Sanity Check Request: 2200m & 1750m to AM Broadcast Upconverter / General Science and Electronics / Forums

Forums

4hv.org :: Forums :: General Science and Electronics

« Previous topic | Next topic »

Sanity Check Request: 2200m & 1750m to AM Broadcast Upconverter

1 2 next

Move Thread

LAN_403

Sigurthr

Tue Feb 23 2016, 02:26AM

Registered Member #4463 Joined: Wed Apr 18 2012, 08:08AM
Location: MI's Upper Peninsula
Posts: 597

EDIT: I've changed the schematic some: 6 pole IF filter instead of 10 pole, and widened the passbands to 300KHz from 200KHz. Now it will upconvert 100-400KHz. (cost was ludicrous). I've also changed values listed to the nearest available values.

3cWaLMO

Schematic:

RF Bandpass Sweep:

IF Bandpass Sweep:

-------------------------------------------------- -

I nearly posted this in the EMR section, but really none of the matter at hand has anything to do with EM, it's all small signal stuff; hence my request for help, hah. Feel free to move the thread if you deem it so needed, Mods.

Original Schematic:

I'm trying to homebrew a design for a simple upconverter to shift a band between 150KHz and 350KHz (FWHM) up by 1MHz (to 1.15MHz - 1.35MHz) so that it can be received by a simple household AM radio. The reason behind the project is threefold; I'm an Amateur Radio operator and I could definitely use the experience making an upconverter, the band of interest is the experimental "LowFER" band, and both of my CWSSTCs have fundamentals within this band and I've love to actually see their transmitting potential.

I've designed the upconverter, which consists of a bandpass filter for the band of interest, an RF preamp, a dual balanced mixer, a local oscillator of 1MHz (square wave, 5Vpk), and an Intermediate Frequency bandpass filter. While I've made plenty of RC and LC lowpass and highpass filters, I've never actually made a bandpass filter before. I've used online calculators (from WA3DSY) for the filters, so values are nonstandard but I'll use the closest standard values, and from the initial construction of a BOM the standard parts are within normal tolerances of the ideal values, so it should work fine I think. Anyway, I'd really appreciate a good once-over on the design to check for errors.

One matter I'm not clear on is the amplitude required by the mixer. The datasheet recommends an input level of 5-10dBm. Now, my studies stated that dBm is a decibel referenced to 1mW, as in power; not a particular voltage. However, through some research online I found a few RF experts mentioning that "7dBm is about 0.5Vrms". I extrapolated from this that 0.8Vrms should be between 7-10dBm. From this I concluded that I'd need -6dB attenuation of a 5Vpk 50% duty 1MHz square-wave to bring it around to 0.8Vrms; thus the Pi Pad attenuator. I'm not confident of this because of the extrapolation and unconfirmed dBm-to-Vrms conversion, and I REALLY don't want to blow up my one and only $25 mixer. So any advice here would be really appreciated.

FWIW; the BOM is up to about $25 now of purely passive/reactive components. If you add in the cost of the mixer it's already over $50. I'm starting to wish I just went the cheap chinese eBay SDR route, as some have upconverters built in already and they cover 100KHz to several hundred MHz. The downsides to that route is having to be tethered to a computer, and of course the SDR's front end is wide open so RFI is a major problem. This isn't only about bang for buck though, I really do feel like I need the experience with mixers to advance my studies in RF.

Hazmatt_(The Underdog)

Tue Feb 23 2016, 03:13AM

Registered Member #135 Joined: Sat Feb 11 2006, 12:06AM
Location: Anywhere is fine
Posts: 1735

Hey there,

Not really sure what you are trying to do here, it looks to me more like a downconverter than an upconverter.

But more to the point, you are mixing a signal with a LO, this is going to cause a frequency shift in the mixer proportional to the modulated input, hence FM not AM.

If you want AM modulation you need to modulate the bias applied to the final PA that is tuned to your desired frequency.
So your basic RF PA there is being fed the LO of 1.1 MHz, then you modulate the bias on the transistor at the base through a capacitor, or at the emitter of the transistor (assuming you have a 100 ohm resistor on the emitter). OR you do it like the old days and use an audio interstage transformer to do both coupling and modulation (modulation transformer).

Sigurthr

Tue Feb 23 2016, 03:40AM

Registered Member #4463 Joined: Wed Apr 18 2012, 08:08AM
Location: MI's Upper Peninsula
Posts: 597

I think you might be misreading and misunderstanding it, Hazmatt; it's nearly identical to the input stage of any superheterodyne AM receiver. I simply am using a prefabbed AM radio (and fixed XO for the LO) in place of a detector and AF amp down stream. Quite literally this is a pretty direct (but BPF and LO altered) rip off of upconverters sold for use with RTL-SDR dongles so they can receive 100Khz-100Mhz within their normal >25MHz reception. Mine is just much, much narrower bandwidth for a specific band of interest (and to circumvent all the "spurious" mixer products that plague the SDRs).

As for the frequency translated signal, it follows any modulation scheme present at the input as there is no modulation, detection, or demodulation present in my schematic. It would thus work with all modes (but I'm only planning on using AM or FM via slope detection). The LO being fixed causes a static linear shift of all signals present within the input passband.

As for calling it an up or downconverter there's no real clear naming convention that I can find other than what appears to be if Rx only it follows the signal chain (in this case 250KHz being shifted up to 1.25MHz), if Rx and Tx, or TX only the naming is based on the final transmitted band (if this was Tx as well it would be part of a downconverter). This could be wrong though, again I haven't seen anything official, but it's just semantics.

johnf

Tue Feb 23 2016, 04:01AM

Registered Member #230 Joined: Tue Feb 21 2006, 08:01PM
Location: Gracefield lower Hutt
Posts: 284

Sigurthr
All looks pretty good It is best drive the mixer at its rated power or above ie you cannot underdrive the mixer --its performance falls off extremely quickly if you do.
I would however consider a bigger transistor for the pre-amp to stop intermod from strong signals 2N5179 BSX20 or even 2N3866 or even better still a Jfet J310. You should not need a preamp as manmade noise dominates 5kHz to 30MHz

Sulaiman

Tue Feb 23 2016, 07:03AM

Registered Member #162 Joined: Mon Feb 13 2006, 10:25AM
Location: United Kingdom
Posts: 3141

Your antenna is going to be a small fraction of a wavelength long so it will be a high impedance source
so the input filter is not suitable
Some type of AMU / ATU is required
and since the source impedance is high I would treat the front end more like an active antenna front end
I'd use a j310 jFET or a BF964 dual gate FET, because that's what I have in stock, use something similar, or whatever is used in published designs.

I would also consider a tuneable multi-turn loop in the vertical plane
as they are less sensitive to noise at these frequencies and can be rotated to null out the worst interference

P.S. radio that can receive L.W. a.m. would be sufficient ... very cheap
you could consider a ferrite rod antenna too.

Sigurthr

Tue Feb 23 2016, 06:46PM

Registered Member #4463 Joined: Wed Apr 18 2012, 08:08AM
Location: MI's Upper Peninsula
Posts: 597

Thank you Johnf and Sulaiman.

I actually couldn't find any published designs for this kind of circuit. In fact, looking at similar applications (such as the 136KHz 2200m band receivers) all I could find were designs from ~20years ago which use obsolete and unavailable parts. That's why I took to designing this from the ground up instead of just modifying the input of an off the shelf radio.

Speaking of the 136KHz band. I'm debating changing the filter values to increase the bandwidth enough to include this band in reception. As it stands now there's about a 5dB attenuation for 136KHz. It's only another 50KHz bandwidth to add, so I don't think it will be pushing things too much. I haven't done the tedious process of finding available components which closely match the computed values yet though, and this really determines viability. This change would let the completed upconverter be useful for more than just listening to my SSTCs and the 175KHz LowFER band though, so it does have merit.

Re: Imput Impedance
I am planning on using a vertical loop antenna (actual loop dimensions yet unplanned, will be experimented with) and a ATU made specifically for matching loop antennas to 50-ohm coax. I also have a random-wire ATU which can match extremely high impedance antennas to 50-ohm coax which I can try as well.

Re: RF Preamp
I am actually completely unfamiliar with the application of JFETs or Dual Gate FETs. I've seen the latter used as mixers once or twice, and only have sparse recollection of the former as being "difficult/sensitive to drive".

I would be exceptionally grateful if either one of you (or anyone) could do a quick schematic of a drop in replacement for my NPN BJT using the J310 jFET or any other suitable amplifier stage. My background lies primarily with Class-D and Class-A systems, so please include specifics about biasing as it is extremely weak area of mine. Thank you.

Bored Chemist

Tue Feb 23 2016, 09:29PM

Registered Member #193 Joined: Fri Feb 17 2006, 07:04AM
Location: sheffield
Posts: 1022

Not my field, but is the o/p of your 1MHz osc a square wave or a sine wave. If it's square I'd put some sort of LP filter in there to take the corners off it before the mixer.

Sigurthr

Tue Feb 23 2016, 10:01PM

Registered Member #4463 Joined: Wed Apr 18 2012, 08:08AM
Location: MI's Upper Peninsula
Posts: 597

It's square. I initially thought the same thing too, but then upon reading into the double balanced mixer I found out that even a sine introduces the same harmonics as a square, as there is a deadtime when there is no conduction of any of the schottky diodes which effectively squares the output, thus reintroducing any odd harmonics you might have filtered out. Thus the only change having an input low pass filter has is less requirements on the output bandpass filter. My IF filter is very steep in the transitions and quite strong attenuation of the stopbands, so I'm not expecting any issues from the odd harmonics, their images, or any of their reflected mixer products. That's why I went with a 10 poly chebyshev instead of a simple low pass.

Edit: related to the output of the LO though, the datasheet isn't perfectly clear whether it is 5Vpk, 4.5Vpk, or 3.9Vpk. If it is >4.5V the shown 6dB Pi Pad attenuator should work fine, but if it is a lower peak voltage it won't meet the 10dBm requirement of the mixer and I'll have to switch to a 4dBm attenuator or use an Emitter Follower buffer amp.

Edit 2; Updated OP with changes.

Blackcurrant

Wed Feb 24 2016, 02:13PM

Registered Member #2989 Joined: Sun Jul 11 2010, 12:01AM
Location: UK
Posts: 94

Hi,
I thought the signal from a SSTC might be rather large. Would an attenuator be better at the input of the mixer.
for your mixer LO +7dBm @50ohm, 5mW =700mV peak to peak

Recommend this book, (it might be a bit out of date now)
Mastering Radio Frequency Circuits Through Projects and Experiments

johnf

Wed Feb 24 2016, 06:21PM

Registered Member #230 Joined: Tue Feb 21 2006, 08:01PM
Location: Gracefield lower Hutt
Posts: 284

Sigurthr
square wave is fine for the mixer -its called hard switching
Try and find an old Siliconix applications book or app noted on the U310 which is the same as the J310

1 2 next

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob