A New Line of Transceivers ~ DifX
Transceiver Architecture 2.115~04~2017 1St Transmitter Tests of the LM373 Transceiver. Pure Joy.
A Way Forward is to Look Back.
Soon it will be time for another Solder Smoke Podcast (#196) and Bill, N2CQR and I hope to spend some time on "Old Time Homebrew Transceivers". Not necessarily that the transceivers themselves are old for they are both new; but that they involve older technology. BTW this is the 3rd year that I have been doing the podcasts which started with #161 in May of 2014.
[I am taking a short detour from the dual conversion DifX as I await my bag of twenty five 11.5 MHz crystal to arrive and then I will share (using Method #4) how I built the filter.]
In SS #195 I casually mentioned that my first solid state SSB transceiver used an amazing single IC from National Semiconductor (now a part of Texas Instruments) called the LM373 and that was in the early 1970's. My rig was on 20 Meters. My first introduction to this device was from an article in a 1970's QST by Tom Sowden, W6KAN and today a friend of mine. His project used a 9.0 MHz crystal filter and covered two bands with a single VFO -- the old 20/80 Meter trick. If you take the sum frequency of a 5.0 MHz VFO and the IF you get 14 MHz, and of course the difference 9 - 5 = 4 which is 75 Meters. In one case the VFO tunes forward and in the latter case backwards -- but hey two bands. That receiver circuit formed the basis of my transceiver project. My transceiver worked as did one built by my then next door neighbor, Bill K6ACJ. I think Bill still has his rig.
Below is the LM373 datasheet. It was good through 30 MHz which of course is 10 Meters.
But a real uptown version of an LM373, 20 Meter SSB Transceiver was built by Charles Hill W5BAA as detailed in a November 1973 Ham Radio Article.
It should be mentioned that I contacted Charley, W5BAA at the time this article came out and he sent me some additional data not in the article. Today Charley and I are still in regular contact and I even bought my first SDR kit from him which was the subject of a QRP Quarterly Article.
Charley's transceiver was absolutely amazing. His rig used three LM373's as this versatile IC had many applications. The primary use was as the main circuit element of the transceiver where it was used as the IF amplifier wrapped around a crystal filter and as the product detector on receive and as the transmit mixer on transmit. A second LM373 was wired as the Receiver Mixer stage. In W6KAN's receiver it used a 40673 dual gate MOSFET for the mixer stage. The third LM373, in W5BAA'a rig was used as the Balanced Modulator.
In most of my rigs I use double balanced mixers such as the SBL-1 or, TUF-1 or even the ADE-1L for the front end Rx Tx mixer stage and a second DBM is used as the Product Detector on receive and as the Balanced Modulator on transmit. This approach facilitates making a singular connection of the LO to one DBM and the BFO to the second DBM. So there is no switching of these signals as you go from receive to transmit. This saves some wiring and the need to do switching of signals. What we have is the bilateral approach where the signals are sent left to right in the IF for receive and right to left on transmit.
But another non-bilateral method is to send signals a single way through the rig and at either end you change the elements. On receive the input side has the receiver mixer and on the back end is the product detector. In transmit the front end has the balanced modulator and the back end the transmit mixer and RF chain. In this case circuit elements must be steered to the front and back ends of the IF block depending whether you are in receive or transmit.
In W5BAA'a rig his signal steering involved the use of a property of diodes to act as switches. When diodes are biased properly they appear as a closed switch. Change that bias and it is like an open circuit. In the receive mode the VFO is "steered" to the LM373 acting as the Receiver Mixer and the BFO is steered to the product detector port on the main LM373. In the transmit mode the BFO is now steered to the 3rd LM373 that acts as the Balanced Modulator and the VFO is now steered to the port where the BFO was connected. Interestingly the output pin that connected to the audio chain on receive is now connected to the transmit circuitry in the transmit mode. [An Italian ham built a vacuum tube transceiver ( I think it had five tubes) and the audio output tube became the final RF amp stage on transmit]. Charley's rig put out a whopping 5 watts --so certainly QRP.
So based on my utterance on SS#195 I decided to recreate my first solid state transceiver circa 1970 using the LM373. I thought this would be blast although my first easy decision: NO Analog VFO and No Crystal BFO. Going back in time is fun; but using an Analog VFO is like doing brain surgery with a rusty spoon!
I also decided that I would use SBL-1's for the receiver mixer and the balanced modulator. It wasn't until I had finished cutting a PC Board and began installing parts that I realized that I really only needed one SBL-1 as it could have functioned as the Receiver Mixer on receive and as the Balanced Modulator on transmit. Should have noodled more on this one. But true to Charley's design I have used diode steering. Well if I ever build a version #2 that would be a change.
I also decide to use two separate Band Pass Filters as this simplified switching. My board at this time has only the receive one installed but the second unit is on today's work list. The Rx RF amp is a single 2N3904 as will be the transmit pre-driver stage.
So where is my project now? Well after wiring it up I discovered it didn't work and then right in the middle of doing some trouble shooting my Hantek 200 MHz Digital Storage Scope shot craps and it was toast! I always thought it acted flaky. So I had to get a new scope which I settled on a Rigol 100 MHz DSO --good move and works much better (Model #DS1102E).
Initially I had chosen to build the LM373 rig using a mix of mostly surface mount parts and a few leaded parts and my board layout was designed for that purpose. I should mention that I had a small stash of LM373's left over from 1970 and some were the 10 Pin TO-5 style while others were the 14 Pin DIP. 10 is less than 14 and so the board was designed using the 10 Pin can style. When the rig didn't work, my first thoughts were bad devices and so I made an adapter board so that the 14 Pin Dip style (using sockets) was retrofitted to the board. Still didn't work --thus had to wait on the new scope arrival.
Well 5 minutes with the new scope convinced me the problem was N6QW. What I hadn't mentioned was this was the first time I used a new soldering iron and what I had done is applied too much heat to many of the SMD parts and they were burnt "open". If you put the scope probe on one side of a cap, say on the LO side, you get a nice signal and plenty of juice. Put the Scope on the other side and you get almost nothing. So step one: change out all of the SMD coupling caps (where I found several others that were open) to leaded versions and boom it worked. At this point just the receiver is working as I am awaiting some parts (coming with the bag of crystals) to finish up the transmitter stages and then we can have another go with my 1970's SSB transceiver. Maybe before the next podcast
Here for your listening pleasure is the LM373 SSB transceiver vintage 1970's
As was stated the way forward is to look back -- signal steering would not be my 1st choice today; but in 1970 it was the right thing to do. The LM373 has a built AGC circuit and it can even be made adjustable --that certainly is a desirable feature in today's rigs. I have no good explanation why there was not more widespread use of the LM373. But for a low part count rig --it is perfect.
Oh you will have a terrible time finding LM373's (I happen to have some in the bin's) so for those who want to replicate this rig -- just buy a Bitx40 kit and move on.
When I get the transmit side working I will share that with you. Lets have a cheer for W6KAN and W5BAA and the amazing 1970's --remember the even / odd gas days.