Topics

Want to increase audio gain in Q70 stage of Ubitx - ideas?

David Feldman
 

I'm working on a homebrew version of ubitx derived from v2. As my project is aimed at weak signal VHF work, wide-open gain at zero signal input is important and as previously confirmed in a post on this subject, ubitx (at least v2 that I am working with) has by design modest receive gain at wide-open.

I've started experimenting with the single audio stage around Q70 but so far to little effect. I'd like to increase the gain there by some dB (ideally about 10 dB), but I'm having trouble understanding this very simple circuit's operation (for example, why bias Q70 base with R50 between base and collector, when other class A bipolar amp schematics I've looked at have a thevenin-style bias network with resistors tied to ground and + supply rail?)

My ubitx v2 build is largely identical to the original designer version, except that I'm using 9V power supply as I wish to power from battery (I'm thinking this would reduce max audio output by reducing excursion on the audio PA chip, but not reduce overall receive audio gain - perhaps I'm wrong.) In looking at ubitx v4 schematic, the circuitry around Q70 is identical.

I measure about 400 mV pk-pk receive noise at speaker at max audio gain (VOL control max.) Q70 appears to be running about 8.6 uA base current and 1.55 mA collector current (based on making voltage measurements and deriving from the values of R50 and R51.

I tried changing Q70 from MMBT3904 (2N3904; I'm building with SMD) to MMBT5089 (2N5089), which has higher hFE spec (2-3x higher). Bias points changed and I changed R51 experimentally to 2.4K from 4.7K, but in the end I see only slight increase in mV pk-pk at the speaker.

I'd like to work exclusively around Q70 if possible, as I'm using a homebrew PCB and adding an additional gain stage following Q70 would be less desirable than increasing gain at Q70, given my modest need for additional gain.

Thanks for any advice/comment,

Dave

Ashhar Farhan
 

There are a number of things you will have to do to lower the noise figure and increase the gain.
1. Change all the IF transistors to BFR93 or some other UHF transistor with low noise figure.
2. Add an additional IF amplifier at the other side of the 45 Mhz filter.
3. Change the audio preamplifier to something like what is used in the Progressive Receiver designed by W7ZOI. It is also used in the DC40 receiver of mine in http://phonestack.com/farhan

- f



On Sat 10 Aug, 2019, 8:39 PM David Feldman via Groups.Io, <wb0gaz=yahoo.com@groups.io> wrote:
I'm working on a homebrew version of ubitx derived from v2. As my project is aimed at weak signal VHF work, wide-open gain at zero signal input is important and as previously confirmed in a post on this subject, ubitx (at least v2 that I am working with) has by design modest receive gain at wide-open.

I've started experimenting with the single audio stage around Q70 but so far to little effect. I'd like to increase the gain there by some dB (ideally about 10 dB), but I'm having trouble understanding this very simple circuit's operation (for example, why bias Q70 base with R50 between base and collector, when other class A bipolar amp schematics I've looked at have a thevenin-style bias network with resistors tied to ground and + supply rail?)

My ubitx v2 build is largely identical to the original designer version, except that I'm using 9V power supply as I wish to power from battery (I'm thinking this would reduce max audio output by reducing excursion on the audio PA chip, but not reduce overall receive audio gain - perhaps I'm wrong.) In looking at ubitx v4 schematic, the circuitry around Q70 is identical.

I measure about 400 mV pk-pk receive noise at speaker at max audio gain (VOL control max.) Q70 appears to be running about 8.6 uA base current and 1.55 mA collector current (based on making voltage measurements and deriving from the values of R50 and R51.

I tried changing Q70 from MMBT3904 (2N3904; I'm building with SMD) to MMBT5089 (2N5089), which has higher hFE spec (2-3x higher). Bias points changed and I changed R51 experimentally to 2.4K from 4.7K, but in the end I see only slight increase in mV pk-pk at the speaker.

I'd like to work exclusively around Q70 if possible, as I'm using a homebrew PCB and adding an additional gain stage following Q70 would be less desirable than increasing gain at Q70, given my modest need for additional gain.

Thanks for any advice/comment,

Dave



David Feldman
 

Thanks very much for the suggestions and quick reply - much appreciated!

I am having trouble with intuitive understanding the concept of Q70 amplifier as shown in ubitx - it is a Class A common emitter amplifier, but I do not understand how to analyze or calculate expected gain (which must be a function mostly of R50 and Q70, maybe also R51?)

I have previously moved 2nd IF to 33 MHz with a very low-loss 2-pole crystal filter with 50 ohm port matching, and that provided some gain improvement, getting farther from roll-off of MMBT3904.

Thanks again,

Dave

MadRadioModder
 

An old designers trick: When I'm not set on the exact amount of gain I need for a stage that operates at less than 1 MHz, I pop in an LM324 op amp with a variable feedback resistor and adjust it to where it needs to be to make the next stage work properly. I then can calculate the gain that works (measure R1,R2) and design a transistor circuit around it. Of course, you can just leave the LM324 in place too...

-----Original Message-----
From: BITX20@groups.io [mailto:BITX20@groups.io] On Behalf Of David Feldman via Groups.Io
Sent: Saturday, August 10, 2019 11:00 AM
To: BITX20@groups.io
Subject: Re: [BITX20] Want to increase audio gain in Q70 stage of Ubitx - ideas?

Thanks very much for the suggestions and quick reply - much appreciated!

I am having trouble with intuitive understanding the concept of Q70 amplifier as shown in ubitx - it is a Class A common emitter amplifier, but I do not understand how to analyze or calculate expected gain (which must be a function mostly of R50 and Q70, maybe also R51?)

I have previously moved 2nd IF to 33 MHz with a very low-loss 2-pole crystal filter with 50 ohm port matching, and that provided some gain improvement, getting farther from roll-off of MMBT3904.

Thanks again,

Dave





---
This email has been checked for viruses by AVG.
https://www.avg.com



--
…_. _._

David Feldman
 

That's a cool idea - that would give me a way of determining the existing and desired amount of gain, which takes me back to the desire to reconfigure Q70 stage gain...

I suspect Q70 and it's two biasing resistors (one in C-B path and one in C-Vdd path) have gain controlled by the two resistors (when I changed to a higher-gain transistor type, the circuit gain barely changed), but I've been searching and so far haven't found a model/analysis of this sort of amplifier stage, even though it's really only 3 parts (!)

Dave

Gordon Gibby
 

Transistor Q 70 is operated in a very simple low level amplifier configuration.  That’s a very simple biasing arrangement, identical to what we used in our VOX circuitry in our home brew Signalink  equivalent  digital interface


The input current to the base will be equal to the input voltage divided by the equivalent input impedance of the stage.  

The collector current will be equal to the HFE of the transistor times the input current.

The output voltage will be equal to the collector resistor times the collector  current.

So the output voltage is equal to
Rc times hfe times vin / Rin

Gain of the stage is equal to output voltage divided by input  voltage and therefore is equal to

Collector resistor times HFE, divided by input resistance.

And in general it’s going to be pretty high!  I don’t know the input impedance but you could guess maybe 1000 ohms  or actually measure it by putting a 1K resistor in series and seeing how much the output voltage drops and doing a little math.

It’s a simple biasing circuit, if the collector voltage rises a bit then more bass current will be injected to bring the collector voltage lower.   It’s not as fancy as a voltage resistor divider, but it works if you are using very very small signals and therefore of the biasing point of the collector can be off a volt or  two. and not matter

Here are some links to educational systems I created to teach this kind of stuff to our local ARES group:

Gordon




image1.png


On Aug 10, 2019, at 14:14, David Feldman via Groups.Io <wb0gaz@...> wrote:

That's a cool idea - that would give me a way of determining the existing and desired amount of gain, which takes me back to the desire to reconfigure Q70 stage gain...

I suspect Q70 and it's two biasing resistors (one in C-B path and one in C-Vdd path) have gain controlled by the two resistors (when I changed to a higher-gain transistor type, the circuit gain barely changed), but I've been searching and so far haven't found a model/analysis of this sort of amplifier stage, even though it's really only 3 parts (!)

Dave



MadRadioModder
 

This is all very good… IF… you know the gain needed for the stage.  If not, it’s trial and error until you figure that out.

 

 

From: BITX20@groups.io [mailto:BITX20@groups.io] On Behalf Of Gordon Gibby
Sent: Saturday, August 10, 2019 1:46 PM
To: BITX20@groups.io
Subject: Re: [BITX20] Want to increase audio gain in Q70 stage of Ubitx - ideas?

 

Transistor Q 70 is operated in a very simple low level amplifier configuration.  That’s a very simple biasing arrangement, identical to what we used in our VOX circuitry in our home brew Signalink  equivalent  digital interface

 

 

The input current to the base will be equal to the input voltage divided by the equivalent input impedance of the stage.  

 

The collector current will be equal to the HFE of the transistor times the input current.

 

The output voltage will be equal to the collector resistor times the collector  current.

 

So the output voltage is equal to

Rc times hfe times vin / Rin

 

Gain of the stage is equal to output voltage divided by input  voltage and therefore is equal to

 

Collector resistor times HFE, divided by input resistance.

 

And in general it’s going to be pretty high!  I don’t know the input impedance but you could guess maybe 1000 ohms  or actually measure it by putting a 1K resistor in series and seeing how much the output voltage drops and doing a little math.

 

It’s a simple biasing circuit, if the collector voltage rises a bit then more bass current will be injected to bring the collector voltage lower.   It’s not as fancy as a voltage resistor divider, but it works if you are using very very small signals and therefore of the biasing point of the collector can be off a volt or  two. and not matter

 

Here are some links to educational systems I created to teach this kind of stuff to our local ARES group:

 

Gordon

 

 

 

 

image1.png


On Aug 10, 2019, at 14:14, David Feldman via Groups.Io <wb0gaz@...> wrote:

That's a cool idea - that would give me a way of determining the existing and desired amount of gain, which takes me back to the desire to reconfigure Q70 stage gain...

I suspect Q70 and it's two biasing resistors (one in C-B path and one in C-Vdd path) have gain controlled by the two resistors (when I changed to a higher-gain transistor type, the circuit gain barely changed), but I've been searching and so far haven't found a model/analysis of this sort of amplifier stage, even though it's really only 3 parts (!)

Dave



Virus-free. www.avg.com

--

…_. _._

Tom, wb6b
 

Hi,

Usually I'm bothered when I see a transistor amplifier without an emitter resistor. The gain should be set by the ratio of the collector resistor to the emitter resistor and be set significantly below the gain of the transistor. This is so things will be stable and you can do large production runs of products and not have to stop production to determine how to fix the product (or add adjustable pots that will cost labor time to adjust) every other time you get parts from a different supplier or batch. The other important thing the emitter resistor provides in stability of the bias point of the transistor. As the emitter provides a negative feedback based on how much current flowing through the collector. This works to keep the voltage around 1/2 the Vcc with wide variations in the gain of the transistors.

In the case of Q70 the next stop for the audio is a volume control, so if the gain is higher or lower in various batches, so be it. Likely nobody will notice. In the case of the "wide open" Q70 it is important to note that the bias resistor is connected from the collector to the base. That is an alternative form of negative feedback for keeping the bias level from going out of bounds.

If a batch of transistors with higher gain were used, the gain of the transistor would tend to want to conduct more and drive the collector voltage down. Maybe so close to ground, the amplifier would not work. But, the collector voltage goes down, also reducing the bias current. That limits how far the collector voltage can go off from the optimal middle point. Low gain transistors would be the same idea but the reverse would happen, keeping the collector voltage from going too high from center. The impedance of the audio feeding Q70 from the mixer is low enough that the 100K bias resistor should not provide much negative feedback at audio frequencies and lower the audio gain.

It is surprising that higher gain transistors did not make a noticeable difference. Although they would need to be more than twice the gain to be noticeable to the ear, per common wisdom. Maybe changing Q70 to a darlington and adding an emitter resistor to tame it down some might work. 

Another thing that has been discovered by folks that have had low transmit power, or trying to boost it further, is the supply voltage (especially below 12/13 volts) has a major effect on the obtainable transmitter power. A large contributor to that seems to be that the gain of the IF chain drops significantly below these voltages. As the same IF chain is used for transmit and receive, running those circuits at 9 volts may be a problem. Your measurements on your radio will let you know if this is an issue.

Tom, wb6b

Tom, wb6b
 

On Sat, Aug 10, 2019 at 11:46 AM, Gordon Gibby wrote:
It’s a simple biasing circuit, if the collector voltage rises a bit then more base current will be injected to bring the collector voltage lower.
Gordon covered the Q70 biasing arrangement very well, too.

Tom, wb6b

David Feldman
 

After a short discussion on EMRFD regarding the "minimalist DC receiver" (which uses the same gain stage design) resulting in a detailed explanation from W7ZOI (see there if interested), and previously more experimentation on my bench around different devices and bias settings, I've realized that I cannot solve the problem in the single NPN stage as originally envisioned.

As an alternative approach, I am taking the output of the AGC driver from the TDA2822 (U1-3 that ends up delivering output across R72), then coupling that signal into the audio gain pot which in turn drives U1-7. This provides a surplus of gain in this path (it puts the two stages of the TDA2822 in cascade) so that the audio gain control is now useful and can easily drive the output section of TDA2822 into clipping on no-signal noise (my original goal).

Some further work will be needed to ensure that the AGC detector (D10 etc.) does not adversely influence the audio at that point.

Dave