Baffled by What I See!
August 16, 2020
I started this by asking what should you believe and one blog reader suggested I look at a couple of plots that measured the V out divided by the V in. This measure tells much about your circuit but has left me with some yet more questions about what to believe. My earlier plots looked only at output form the circuit not a measure of how the circuit responds (V out) in comparison to the input (V in).
So I did that (interesting results):
Wow the ratio of the output to the input is FLAT and the V in appears to be affected by circuit loading versus frequency. So this is not termination insensitive although the relative measure of the ratio is consistently flat. Now I ask the question that what really counts is what is coming out of the stage to drive the next stage looking only at the output. We next look at charts where we measure the Ratio, V out and V input.
There is close to a 3db Drop in output at 10 Meters with the 100nF. Now we repeat with the 4.7nF
The output drops by about 1.75 dB at 10 Meters so a bit better. But the price is paid on the front end.
The Long Story Short: For IF Applications leave C1 = 100nF. If using this amp as a general amp stage as in a broad band RF amp consider changing the 100nF to 4.7nF.
********
Long standing there has always been a belief that if anything is ever published in EMRFD then it has to be Gospel and all things for some techie types must have their roots traced to some chapter or Figure in EMRFD. There is even a cult of illuminati who spend endless hours on blogs and websites scouring circuits and then commenting "well in EMRFD this is how it is done". Actually EMRFD serves me well as a bookend.
There are about two or three circuits in that publication that have caught my eye and up front I have used them. One in particular has been the Plessey circuit of Fig 6.110 which is a simple two transistor amplifier (2N3904 and 2N3906) that is bilateral. I have used it in many transceivers as the signal reversal merely involves changing one connection from ground to a positive source -- easily done with a localized relay. Below is the schematic of two such stages wrapped around a 9 MHz Filter and used in The Simple SSB Transceiver which can be found at www.n6qw.com
The first use of this arrangement was in my 20M shirt pocket SSB Transceiver.
As you can see this got shrunk down to a very small size.[Sitting behind that amplifier board is a homebrew 4.9152 MHz Crystal Filter.]
But not leaving well enough alone I then created a SMD layout for this circuit so I could use the CNC Mill. I have labeled the components so you can readily identify what is what. This is for the EMRFD illuminati who can't read schematics!
But not leaving well enough alone I then created a SMD layout for this circuit so I could use the CNC Mill. I have labeled the components so you can readily identify what is what. This is for the EMRFD illuminati who can't read schematics!
So how does this circuit really work? Well for starters the 2N3904 is used to send signals from LEFT to RIGHT where typically if ahead of the filter would be to send the output to the Crystal filter and most often as in Receive.
The output of the second stage following the filter is to the Product Detector. Now "throwing the switch" the signal path reverses and the circuit works RIGHT to LEFT such as you would have in the Transmit sequence. The PD is now a Balanced Modulator and the signal goes through the IF Module in the opposite direction.
Close examination of the 2N3904 has the 100 Ohm Emitter resistor grounded in Receive and so it is operating and sends the signal from LEFT to RIGHT. Simultaneously the 100 Ohm Emitter resistor on the 2N3906 is Grounded SO IT IS NOT AMPLIFYING! Now we engage the Relay and the 2N3904 Emitter is being fed +8 VDC so it is NOT amplifying but now the Emitter of the 2N3906 (PNP) is supplied +8 VDC is now the active element. Pretty cool beans!
Since this circuit came from EMRFD and obviously blessed by the illuminati, I never had cause to look at a simulation of the circuit. Of particular interest was the PNP portion. I must confess when I think about amplifier stages never do I think PNP. So now was the time!
My results leave me with a huge knot in my stomach. Here is the PNP circuit redrawn in LT Spice.
So to those pseudo illuminati who read this blog please check that I have accurately reproduced just the PNP circuit. The load is 50 Ohms as shown as R6.
Now the "AGHAST PART" -- the Plot of this PNP Amp above.
This circuit would work well on the lower frequencies but sucks beyond 30M. So a huge RED FLAG --while it amplifies it drops by 3 dB over the HF band.
So now lets "diddle a bit" with circuit values. As to the why I looked at this one area in particular I can only suggest --been here before and this change usually affects frequency repsonse. C1 was Changed from 100nF to 4.7nF.
This small circuit change appears to have flattened the response out to 30MHz with a bit of a lower gain on 80M. I think less gain on 80M is a far better trade off to having a flatter gain across the rest of the HF band. Same circuit with only one capacitor change appears to make a marked improvement in overall frequency response.
So what is wrong here? This circuit is in EMRFD, it is in the bible and is read as the Holy Gospel by the radio illuminati. Is there something wrong here?
Wow surprise, no emails on this but there is information in the data which may shed some light on why no one has jumped on this before.
C'mon EMRFD illuminati you may have missed this. But the Plessey circuit is being used as an IF amplifier stage where typically the IF Frequencies are less than 13 MHz. (Look at the Bitx series at 12 MHz). So as long as there is reasonable gain up to 12 MHz or so, then having the gain drop off beyond that may be of no consequence.
But just suppose this stage was used elsewhere such as a bi-directional Receiver RF Amplifier and Transmit Pre-Driver say on 20 Meters and above, then the capacitor change is critical. Look at the plots above.
Wow surprise, no emails on this but there is information in the data which may shed some light on why no one has jumped on this before.
C'mon EMRFD illuminati you may have missed this. But the Plessey circuit is being used as an IF amplifier stage where typically the IF Frequencies are less than 13 MHz. (Look at the Bitx series at 12 MHz). So as long as there is reasonable gain up to 12 MHz or so, then having the gain drop off beyond that may be of no consequence.
But just suppose this stage was used elsewhere such as a bi-directional Receiver RF Amplifier and Transmit Pre-Driver say on 20 Meters and above, then the capacitor change is critical. Look at the plots above.
The circuit was developed by Plessey for use in military equipment. History tells us a lot of military gear operated in the 2 to 12 MHz range (often with a 16MHz IF) so that is a possible explanation. Another is that circuit simulation techniques may have been limited at the time this was developed. Still a possibility I did something wrong.
But more importantly I did something right! For this circuit as I have drawn it -- one capacitor change dramatically improves the over all wide band frequency response The bonus: this modified circuit likely would make a good Receiver RF Amp stage in a homebrew receiver -- note no tuned circuits! The Double Bonus --- I have a large jar of PNP transistors that I often said too bad these are not NPN. Well maybe I now have some very new opportunities and challenges to use the PNP devices. Just thinking -- an all PNP SSB transceiver.
For the keen observer(s) -- the Fig 6.110 uses 6 VDC with an explanation as to why. I have moved that voltage up to 8 VDC with no issues.
Again Baffled by what I see in the plot. The EMRFD authors say the circuit is good for about 17 dB through 40MHz. Given the losses in the filter such a module could have an overall gain in the 25 dB range. 17dB X2 = 34 dB and assuming a 10 dB Filter Insertion loss results in the around 25 dB gain number.
Now there are some things that could impact the results; but I am dubious of such an outcome. The PNP circuit was taken in isolation as if the NPN part were absent. True, based on how Ground or Voltage is applied only one portion is ON at any one time. But there are other components connected to the circuit. Inactive components I would think would actually degrade performance and not enhance it.
So this exercise with the Plessey Circuit is an extension of what good ham friend from Austin, Texas inscribes on all of the shematics he generates. TRUST BUT VERIFY!
I have TRUSTED the Plessey Circuit for many years and the results have been excellent mostly becuase I have used it with IF Frequencies in the range of 3 to 10 MHz, where obvioulsy it works. Recently I had an interest in using this amplifier for applications other than an IF stage which caused me to VERIFY its wideband performance. Now I can see where use in others applications such as a wide band bilateral RF ampilfier would require some tweaking such as the capacitor change.
73's
Pete N6QW