How to perform the 2sc2999 and Schottky diode swap

[Klondike Mike]

The best bits for me are their ridiculous technical explanations as to why these mods work.

Get off your high horse and quit looking down your nose at "these mods put out by some high school kid in their basement". Well, that's the impression you put across.

This mod was intended for and works for certain types of radios. No, it will not work in the ICOM 7800. I am not an RF engineer but I have implemented the mod and it works for the radio/radio's it was intended for.

You are good at spewing out numbers and you have not EVEN implemented it! Typical theorist babble. Your attitude is one of some engineer/designer who needs to get out of their closet and see how their designs/creations are working in the real world.

 

[Robb]

Get off your high horse and quit looking down your nose at "these mods put out by some high school kid in their basement". Well, that's the impression you put across.

This mod was intended for and works for certain types of radios. No, it will not work in the ICOM 7800. I am not an RF engineer but I have implemented the mod and it works for the radio/radio's it was intended for.

You are good at spewing out numbers and you have not EVEN implemented it! Typical theorist babble. Your attitude is one of some engineer/designer who needs to get out of their closet and see how their designs/creations are working in the real world.

X 2

 

[G0HZU]

Get off your high horse and quit looking down your nose at "these mods put out by some high school kid in their basement". Well, that's the impression you put across.

This mod was intended for and works for certain types of radios. No, it will not work in the ICOM 7800. I am not an RF engineer but I have implemented the mod and it works for the radio/radio's it was intended for.

You are good at spewing out numbers and you have not EVEN implemented it! Typical theorist babble. Your attitude is one of some engineer/designer who needs to get out of their closet and see how their designs/creations are working in the real world.

Look, I'm not the one spewing out false info like:

I like to start with the transistor at position TR-14, this is the Am Detector.

TR14 is NOT the AM detector. Not even close!

You are good at spewing out numbers and you have not EVEN implemented it! Typical theorist babble.

I've certainly looked at the front end circuit design of similar CBs and examined what causes the most noise.

A better tip would be to tell people how to adjust the transformer ahead of TR14 on the radio ahead of RF amplifier TR14 in the example CB (eg Cobra 2000GTL) in the first post.

http://www.cbtricks.com/radios/cobra/2000gtl/graphics/cobra_2000gtl_om_main_sch_11x17.pdf


This will impact the receiver noise figure much MUCH more than the choice of transistor at TR14. Eg, set this slightly wrong and you can make the radio noisier by several dB (but little change on the S meter) This is because the noise figure of the amplifier stage TR14 will be affected by its source impedance and transformer L4 affects this impedance. L4 needs to be set to the lowest noise position.

Also, the AM detector diodes and the SSB product detector are a long way down the receiver chain and any noise contributed by these parts will be miniscule compared to the level of signal they see following all the signal gain in the receiver.

The noise blanker can only deal with low duty cycle impulse noise rather than hiss so changing the diodes D1 and D2 will do nothing for hiss noise. i.e. the noise blanker 'works' by killing the receiver gain during the brief activity period of the impulse noise. i.e. the type of regular pulsed noise you get from a vehicle ignition system.

I don't need to test this because I know how the circuit works and I've characterised the Uniden noise blanker for how well it works for pulse repetition frequency and pulse width etc. It won't remove hiss noise.


Also, unless you use reasonably accurate test gear to evaluate any mods then it isn't possible to judge them by ear using tests on air with an antenna. Too many variables at work here. Also, even if the 2SC2999 reduces the noise by 0.5 to 1dB it simply isn't possible for a human to detect this change in S/N unless you could flick between 'before and after' instantly and under very strict test conditions. Even then the difference would be tiny.

So tiny you wouldn't be able to tell which was which if you left the room and came back in again and listened to just one version.

 

[G0HZU]

The other thing is I just downloaded the datasheets for the 2SC2999 and the 2SC1674L and the test conditions for determining noise figure are very different.

In the case of the 2SC1674L the device test fixture is driven directly from a 50 ohm source into the transistor base and uses a 22k resistor for bias within the RF test fixture.

In the case of the 2SC2999 the test jig for noise figure on the datasheet shows a very elaborate design with feedback and also a tapped tuned circuit on the input. This steps up the impedance into the base and this will affect the noise figure of the overall test fixture. It also keeps any bias resistors outside the RF area and this is good for keeping noise low.

So I can't see how you can directly compare the noise figure data for these two devices based on the info in the datasheets.

This is because the figure in the datasheet is the noise figure for the overall TEST JIG CIRCUIT and not some magic figure burned into the transistor itself. The noise contribution from the transistor will vary and will depend on the components around it and the source impedance driving it and also its collector current.

 

[Robb]

I think I can see your dilemma G0HZU. Your only real failure here is to take into account the fact that the radio mfr - lets use Cobra as the example - doesn't buy and use the text sheet version of the 1674. They buy in lots and do not pick and choose each device that would exhibit the specifications you quote.

They just buy in huge lots and use the whole lot. The radio design may be sound; but the device quality is anything but consistent to that spec sheet. So it is no small wonder that the hand picked 2999e device that has been screened for the higher spec is coming up with the changes that are claimed by the use of the better substitute.

In essence, it is the difference between the spec sheet and the real world application of somewhat substandard devices being replaced with a different device of known and tested character and quality.

Since the other major factor in this claimed mod is the Shottky device, you can realize a change that you can perceive by ear if you have been used to using the very same radio. I do not say this out of theory; but from several actual experiments with this upgrade.

I would ask that you spend just a few quid or less and try this modification yourself, and you can see with little discernment that the amount of change from two or three different radios. The change may well be slight to noticeable, even if the radios tested aren't your standard radio of use.

This mod isn't earth-shattering; it is a simple change that yields modest to mild results in favor of less noise and a change in positive signal strength. When the 2999e/Shottky are used together, the difference in mass production radio components and hand picked devices can be appreciated.

Just try it with a Cobra, Ranger, or Galaxy radio.
You will no doubt hear a difference.

 

[G0HZU]

Hi Robb
I'd expect that there wouldn't be any significant difference (<1dB gain) if you (or the CB manufacturer) bought a bucketful of 2SC1674L devices and tried each one in turn assuming none of the devices were faulty. This is because of the nature of a common base amplifier like TR14.

The gain is dictated by the ratio of load to source impedances and transistor Hfe has very little impact on this on a regular design like the one in the Cobra 2000GTL.

Put another way, you would see a MUCH bigger spread in gain if you simply bought a bucketful of the biasing resistors around TR14 (5% tolerance?) and fitted all of these these in turn. This is because the biasing affects the input impedance of the transistor and hence its gain.

As for the schottky diode mod the only place I can see that having a possible benefit is at the AM detector as some CBs use silicon diodes here. However, there may be implications with how the detector is biased in its original form and simply swapping out diodes might not be ideal. This is because the detector will be forward biased to suit the original diode characteristic for weak signals.

The place I think it is VERY unwise to use a schottky diode is at the front end, i.e. the AGC diodes.

Also, changing the diodes at the noise blanker will not reduce hiss noise.
It might alter the characteristic of the blanking pulses for a given interference waveform from a car ignition system but that isn't the same thing.

Put another way the noise blanker does NOTHING unless it detects pulse type interference then it springs to life and 'blanks' the receiver gain for very short bursts that coincide with the pulsed interference (eg from a car ignition event)

The noise blanker cannot deal with hiss noise. Not even close. Hiss noise is totally the wrong type of interference.

 

[Robb]

Theory and specifications lose in the real world application.
Try it.

 

[G0HZU]

Theory and specifications lose in the real world application.
Try it.

But you are happy to quote specs for Hfe to justify your purchase of 2SC2999E devices?

Despite the fact that even elementary analysis of the circuit shows that typical Hfe values will have miniscule impact on the gain of TR14?

Much less impact than the effect of bias resistor tolerance and certainly too small for a human to notice without test gear?

Also, getting a tiny amount of extra gain doesn't automatically yield more sensitivity. i.e. gain does not equal sensitivity.


The noise figure test circuit on the 2SC1674L datasheet looks to be less optimal for low noise compared to the circuit used for the 2SC2999.

This is because it is being driven with a less than ideal (too low) 50 ohm source impedance whereas the 50 ohm test source on the 2SC2999 test jig gets stepped up using a tuned circuit. This has been done in order to get higher gain and lower noise from the test jig.

Also note that TR14 is configured as a common base amplifier which isn't the same as the NF test jig on the two datasheets.

So apples are not being compared with apples. However, I guess you will ignore this and cling to that 2.2dB figure and 'E grade' Hfe stamped on the 2SC2999 datasheet.

Try it

Quantify what test method you used to prove that the 2SC2999 works better.
Just simply saying you tried it and the radio sounded better with an antenna won't impress me.

You need to prove that the receiver really is better. Maybe it is better but from where I am looking at things I doubt any improvement will be more than a fraction of 1dB and it could always prove to be worse.

 

[Robb]

With so much variation from radio to radio - no real world mass production radio can be a definitive test subject. Given the variations due to all of the devices within the radio. Mass production tolerances don't exist for the common CB radio. You will find variations throughout the entire radio in each different circuit.

I recommend for testing a few different radios of the same model and cite the various differences. I think you need to experiment and see for yourself. If you want to hear me say that your evidence is useless, that is only true due to poor device quality from the mfr. In a perfect world where the mfr cared about the device quality, you would be right.

You also keep 'running home to mamma' when it comes to the db change issue. The truth is is that I have done this mod to over a dozen radios, it cost so very little to do. The amount of change in each radio - sometimes the same model - does vary. I have done the real work - and yes - the results are subjective to me. But I can say that; you cannot. As you haven't gone the distance. It is too easy for you to continue your line of argument of 'who is on first?'. You have failed to be subject to your own findings. So your conclusions mean something; but lacks evidence.

 

[G0HZU]

So basically you have nothing. This thread is several years old and there are no credible tests done to prove this 2SC2999 mod works.

Just a lot of very bad science trying to justify that it DOES work and that any failures are due to fakes.

Also a lot of crap about the importance of Hfe and some dodgy noise figure interpretations from manufacturers datasheets by non technical screwdriver experts.

 

[G0HZU]

Just so you understand where I'm coming from on this, I have a lot of test gear here for RF design etc.

But no way, NO WAY could I walk into a room, switch on a CB and tell the receiver sensitivity by simply plugging it into an aerial and tuning across the band and listening to other signals and noise on the band.

That would be a ridiculous test method for lots of reasons. Totally uncontrolled and affected by band noise etc. Even if you gave me an accurate signal generator and I tested it directly (rather than with an aerial) I would struggle to tell the S/N ratio on a fairly weak signal without further test gear. This is despite having spent all my working life designing receivers and transmitters and testing the results under controlled conditions.

So how come you can manage it?

 

[Robb]

I am not stuck. Not like you are. As you refuse to prove me wrong by allowing yourself to be subject of your own line of reasoning by testing it for yourself. When you do and you find that this mod does have significance, there is no shame in saying so. You won't be the first and doubtful you will be the last.

I can give many reasons why a prime rib is an excellent dinner, and better than filet mignon. Now, if you only had filet mignon and had never tried prime rib, you would have every right to convince me that filet mignon has less cholesterol and fat, is leaner and tastier, as well as a better buy. In some ways, you would be right. However, you have tied yourself to a fixed position. As Gen. George Patton once quoted: "any fixed position is a monument to the stupidity of mankind".

I have done the tests, and that verdict is in.
I have gone overboard to tell you as much so many times.
Have a prime rib, a beer, and cheers . . .

 

[G0HZU]

The schottky diode mod on the noise blanker diodes won't help with hiss noise. I don't 'need' to test that because I know how the circuit works. It might help with very weak ignition noise from a car but that isn't the same thing as hiss noise from the receiver.

As for using the schottky diodes in the AM detector then that might have merit on detectors that use ordinary silicon diodes but I can 't be sure.

However, most CBs use a pair of germanium diodes like the 1N60 for the AM detector and the detector will be weakly forward biased with a few microamps.

If you swapped the 1N60s for Schottky diodes I would expect to see no measurable difference whatsoever in AM detector performance because the detector has a weak forward bias to suit the 1N60. Do I need to do this test? No I don't think so. (by measurable I would say less than 1dB either way right from high S/N right down to signals near the noise level)

Obviously you won't accept this.

As for the 2SC2999 mod I doubt it would make more than 0.5dB difference in NF (could even make it worse) and this is based on experience of designing common base amplifiers in the VHF region using various NPN transistors.

If you think a difference of 0.5dB or even 1dB will be enough for you to regard this as a significant change then you are deluding yourself and every sucker who lets you fit one of your (you bought 60!) 2SC2999 transistors you shelled out your $ on

 

[Robb]

I can accept the fact that you refuse to try it for yourself.

You can talk about the spec sheets and your qualifications all you like, too.

However, you are looking at all of this from a flat piece of paper and categorically refuse to take in the variables presented by real radios in the real world on the market that have all assembled with parts that are lacking from those same spec sheets.

The mass production of radios that don't fall under what you expect due to the mfr's willingness to sell a radio with less than desirable parts keeps you blind to the spec sheets.

Sheesh!

Tell me; what can you not understand about that?
You cannot believe that Cobra or Galaxy or Ranger sells radios for profit and short change the buyer with a product that is substandard to the initial design?

It is all about the money they save (production) and the profit they make.
$$$

 

[G0HZU]

The mass production of radios that don't fall under what you expect due to the mfr's willingness to sell a radio with less than desirable parts keeps you blind to the spec sheets.

Sheesh!

Tell me; what can you not understand about that?
You cannot believe that Cobra of Galaxy or Ranger sells radios for profit and short change the buyer with a product that is substandard to the initial design?

It is all about the money they save (production) and the profit they make.
$$$

What is hilarious is that you are saying this and then 'promoting' changes that won't make any significant difference!

If you wanted more gain from TR14 then bias it slightly harder and the gain will go up. All it will cost is a 2 cent resistor.

Tell me; what can you not understand about that?

:laugh: