How to perform the 2sc2999 and Schottky diode swap

[Robb]

Then get a radio and play with it and see for yourself.

You simply won't believe it any other way.

Additionally, using your own logic on re-biasing the 1674 for better results can equally be applied to the 2999e as well.

 

[G0HZU]

One more time on the noise blanker diodes...

The noise blanker detector typically uses unbiased germanium diodes. An unbiased germanium diode detector will still work reasonably well for tiny signals.
However, swapping these diodes for Schottky may well actually make the detector worse because Schottky diode detectors really do need to be slightly biased in order to become good low level detectors.
In reality (for a noise blanker detector) I suspect it would make little or no difference on car ignition interference either way anyway

The NB detector is there to detect a large and sudden pulse of noise energy and it will blank the receiver (i.e. you don't hear any output) for a time period that is slightly fatter than the pulse of ignition noise from the car.
So you are actually listening to a soundtrack with tiny pulsed gaps in it that line up with the ignition events in the car.

In order for the noise blanker to be able to detect and capture the fast pulse of ignition noise 'quickly' it listens for large noise events over a bandwidth of several hundred kHz at the first IF in a typical Uniden CB. i.e. maybe across 20 or 30 channels.

This can't work with fuzzy hiss noise you hear on your chosen channel and there won't be 'louder' CB signals with this mod. Period.

As for the Schottky diode mod in the AM detector I am pretty much telling you it won't make any difference because the detector uses a forward bias current that is optimised for the 1N60 germanium diode.
If you swapped in Schottky diodes the bias current would still be pretty close to ideal but there wouldn't be any difference in signal to noise ratio as both detectors will have similar efficiency.

Note that if you put the Schottky diodes in a radio with an AM detector that used silicon diodes eg 1S2075 diodes then I would expect that the detector efficiency would become quite poor because the detector circuit would have a higher forward bias to suit the silicon diodes.

This would bias the Schottky diodes to a point well away from the ideal bias point and would kill the detector efficiency
If you rebiased the detector to suit the Schottky diodes then you may well make the detector perform better but you would also have to make sure the detector didn't then overdrive the audio stage after it.

In short, each AM detector circuit in the CB will be biased to suit the original diode type fitted and the load impedance and drive level from the detector will be chosen to match the diodes.

Just swapping the diodes might mean you make the radio worse.

 

[Robb]

You have gone to great lengths to prove that this mod cannot possibly work. Yet for those who had radios - in the real world - that have less than comparable receive to another radio of the same model find improvements with this mod.

You cite every possible fact that explains why it shouldn't work. Yet it does because you fail to take into account that cheaply mfd radios are anything but consistent.

I don't sell this mod nor do I distribute them. You act as though that I am the proximate cause for your elaborations to be necessary.

You fail to buy a cheap $50 Cobra 148 GTL or Galaxy 949 and test this mod yourself. You claim that you don't need to as though you are above it, as if you might be proven wrong. For which you would.

I have heard for myself that which you refuse to accept as fact. I don't need to prove a thing; yet you insist upon proof which you have no need to seek for yourself. That isn't scientific; it becomes a stalemate of an argument that is simply going nowhere fast.

Many hundreds - if not thousands of CB radio operators have used this mod and are both glad and successful in their endeavor. Maybe it is because you hold a fierce dislike for the CB crowd? But I won't speculate to that, your reasoning is less than credible because you failed to find out for yourself. Had you gone the distance and experimented in a true scientific fashion, your claims might carry more weight.

Of course, what has been said in this argument is moot. Many of us who have done this mod time and again know that you are wrong. There might be only a few that had less than expected results. But they are the exception; not the rule.
Because of that fact, your arguments are oh so useless.

 

[G0HZU]

Maybe it is because you hold a fierce dislike for the CB crowd? But I won't speculate to that

But you already have speculated that because it's in text above for all to see?

BTW I do have several CBs here.
eg the mk2 Cobra 148GTL-DX export model.

I'm also aware of what contributes to the noise floor of the 148 receiver because I've owned a 148 since I was a student and I've tested the radio many times and examined the circuits.

I've also modified mine to make it more sensitive across all of its channels but I've done it the scientific way with test equipment to verify the results.

Also, the AM detector diodes in it are 1N60 and they work just fine for the type of detector.

I'm tempted to swap some schottky diodes into the AM detector just to prove there will be no improvement but I know it will be a waste of time because you won't believe me or you will say the diodes are fakes

 

[Robb]

No, your contentious behavior has done its own damage on your own behalf.

What you would like to say and have everyone believe is that the Emperor is naked.
In truth, it is the other way around. We already know it works. You don't. At this point, it would be suspicious if you suddenly claimed that you tested it and said it doesn't work. Listen once again: There are many people who have done this mod and can hear the difference. And you have done nothing at all except to use smoke and mirrors. No proof; no substance. And yes; at this point no one will believe you after I have urged you to try it and you then you make claims that you have. Thousands will disagree with you; they already know it works.

At this point, I won't contend with you any longer. You have lost this argument before you started. I have repeated ad nauseum and no longer going to entertain your need for a pointless argument.

 

[RatsoW8]

Although the CB radio is a product of science, many of the mods, peak and tune info and general misinformation and street logic about what works and what doesn't is definately religious in nature.

I doubt Darwin ever threw down on the subject with the Pope.

Who knows...on an older radio along with an alignment?? Magic.

 

[G0HZU]

What you would like to say and have everyone believe is that the Emperor is naked.

For the noise blanker mod as claimed by the OP, yes.

We will also help out the noise blanker section of the radio for quieter operation. This will not change the signal you receive on your meter but it will make a 3 s-unit signal as loud as a 4.5 s-unit signal

The claim above is utter BS. He's effectively claiming it will make the signal several dB louder.

The signal you receive doesn't even go through the noise blanker path so how can the diodes make it louder? Total BS.

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

This is so wrong it's funny.
TR14 is the first RF amplifier.
More naked emporer BS.

You can now replace the diodes at D1, D2, D21, and D22 with the new 1N6263 diodes. These are low noise diodes that help quiet the natural hiss of the radio.

D1 and D2 are the noise blanker detector so this is more BS as the noise blanker can't deal with hiss noise. It only activates on impulse noise.

In D21 and D22 you have an AM detector with biased germanium diodes. This is a peak detect and hold circuit so faithfully tracks across each RF carrier peak and produces a demodulated output.

Any noise contribution from the diode (germanium or Schottky) will be tiny (microvolts) compared to the size of the modulated RF carrier it is detecting. eg 150mV of RF at the detector even on a tiny signal at the antenna. This is because there is a LOT of RF gain in the receiver. So the tiny detector noise is negligible compared to the massively amplified signal and noise that is hitting it from the receiver IF output.

So another naked emporer.

 

[G0HZU]

Although the CB radio is a product of science, many of the mods, peak and tune info and general misinformation and street logic about what works and what doesn't is definately religious in nature.

Exactly. That's very well put

I'm actually trying to help people like Robb spot the BS but he really wants to keep up with the beliefs and reject any attempt at scientific analysis. His post #95 sounds a bit like someone defending a religion based on numbers of followers. They can't all be wrong can they?

 

[G0HZU]

It's late here but I had a go at fitting some Schottky diodes to a 148GTL-DX on the AM detector.

I don't have any 1N6263 diodes but I did find some old HP 2800 diodes and these are very similar.

The results are quite interesting and there's probably something for everyone here

First of all I tested with a decent AM signal generator for S/N with a 30% 1kHz tone from -50dBm right down to -120dBm and looked at S/N ratio on an accurate meter.

As expected, both diode types were the same in terms of signal to noise ratio across the whole range right down to 2dB signal to noise. The 1N60 appears slightly better with weak signals but the difference is tiny. Less than 1dB.

Then I tried looking at using two modulation tones 1kHz and 1.6kHz at 30% each.
I redid the tests and got the same S/N for both diodes with the 1N60 just shading it.

However, what I did notice was that the 1N60 diode detector became less linear at very small signal levels with the two tone modulation test. These are at tiny signal levels with just 3 or 4 dB signal to noise.

This detector non linearity actually makes the radio appear to have LESS background hiss if the background signal is very low, i.e. with just a dummy load fitted. In other words the demodulated output falls more than 1 dB for a 1dB change in signal generator level. However the signal to noise ratio is the same if you turn up the volume.

Technically, this characteristic with the 1N60 diodes isn't a good thing but I think I can see why some people don't like the schottky mod to the AM detector. There is definitely no improvement in signal to noise ratio on weak signals but the detector is more linear in action right down at the threshold of sensitivity.

This will mean the standard radio can sometimes appear quieter during pauses between stations transmitting (assuming band conditions are extremely quiet with very low band noise)

Some people may prefer this with medium to strong signals on a quiet band although on very weak signals (so weak they are barely intelligible) you would have to use more volume.

I swapped the diodes back and forth several times to double check all tests.

Despite the differences listed above I would expect that under normal conditions there would be little difference as both detectors appear identical if the signal is more than just a few dB above the minimum it can detect.

 

[loosecannon]

G0HZU, i appreciate the testing that you have done here. i dont think ive read of such an intricate test being done on this mod before.

you definitely sound like you know what you are talking about when it comes to receiver design.

i am one of the many who have done this mod on many occasions.
i continue to do it because i continue to enjoy the results.

i do not have the knowledge and experience to debate the merits of the mod with you on a technical level. i can only say that every radio i have done the diode changes to has been more enjoyable to listen to when on an empty channel.

the terms i would use to describe the after effects would be "smoother" and "less poppy".

i do not do the 2SC2999 transistor change anymore because i just couldnt tell the difference between it and the 2SC1674.

i have done enough rudimentary testing to see that an S-5 signal before the mod is still an S-5 signal after the mod. these tests were done with my signal generator and a dummy load; not on-air with another station.

maybe you can explain why this is?
i am not being facetious, i am honestly hoping you can explain why changing the diodes in the AM detector circuit and the noise blanker circuit from germanium 1N60 diodes to 1N6263 schottky barrier diodes makes for a more pleasant sounding receive when no signals are present.


since it seems clear that you are not much of a fan of this mod, i am also hoping that you might help further the hobby by coming up with something better.
as i said, it seems that you know receiver design quite well.

this mod is popular not because it is the best thing you can do for your receive, but because its the only thing out there that seems to show some improvement.

im sure if you were able to show us a better way, that your mod would become equally popular if not more.

i believe this chassis would be the best example to use, as its similar to your 148GTL-DX and the mod could be adapted for other chassis quite easily:
http://www.cbtricks.com/radios/galaxy/dx959/graphics/dx959_sch.pdf

this chassis is nearly identical to those found in other CB radios being manufactured these days, and they all have the same issue: excessive white noise when no signals are present.
i have personally been told by a pretty knowledgeable source that the copper used on the PC boards is too thin to provide much of a groundplane at 27mhz, and that this is the majority of the problem. (just so you can keep that in mind when looking for improvements)

i sincerely hope that you will take a little time and try to come up with some ideas.
i assure you that you will find no shortage of "beta testers" willing to try what you come up with.

the CB hobby is severely lacking in the "elmer" dept. it seems that those with the knowledge cant be bothered to take the time to apply it, and that is why you see mods that fall short when put to the test becoming so popular.
we take what we can get!


thank you for taking the time to read this, and i hope it inspires you to stick around a while.
LC

 

[G0HZU]

Hi Robb I really appreciate your reply

I guess we all share a common cause here in that early Uniden CBs are not as sensitive as they could be.

Over here in the UK the SSB CB of choice was always the Cobra148GTL-DX and it's the radio I have the most fondness for.

It often gets modded to span 200 channels and the receiver front was never designed to cover this range so it goes very deaf on the outer channels.

There are ways to improve this and some are quite simple although to get this radio to be as sensitive as some of the later CBs would require quite a few changes.

I suppose the 'good' news is the bands are much noisier these days so a lightly modded 148 won't be missing out because the noise level on the band is so high.


On the mk2 148GTL-DX on AM the schottky diode mod does have some merit because it has better linearity right down at the noise floor.

I did experiment with the detector bias on the 1n60 and could improve its linearity in the noise floor to be about the same as the schottky but not sure if this is of much significance really.


However, I can list the pros and cons of what I think a typical CBer would notice with this mod on a mk2 148 on AM.

Advantages
1/ Late at night on flat band conditions when listening to very weak AM signals 'in the noise' (noisy enough you can't get more than maybe 75% of the words) the schottky detector will maintain the volume of the weak signal better if it has any fading.

Basically, very weak signals will hold their volume better as they fade.

2/ This is related to the above. If you are in a three way conversation (late at night, quiet flat band) and one station is so weak they are hard to copy

and the other station is maybe S1 then you won't have to adjust the volume as much between them to get a good copy.


Disadvantages

1/ Because the schottky AM detector is more linear close to the noise it will make the background noise appear louder to the user than before. i.e. late at night or on flat band conditions the radio will show a higher noise level between transmissions. This isn't a failing of the detector, it's a symptom of its better linearity.

Some people might not like this higher noise volume if they like to have the radio monitoring with no squelch.

The simplest way of describing the above differences is that the standard radio (as in my 148GTL-DX) effectively develops a lower volume level for very weak signal levels (including weak noise).

i.e. very weak signals lose volume quicker on the standard radio (but maintain signal to noise compared to the schottky) as the signals get smaller.

Technically, this is bad. But some users may actually prefer it for the reasons given above.

the terms i would use to describe the after effects would be "smoother" and "less poppy".

You are doing the mod to a different radio so we may see some differences anyway.

However, the poppy effect may be due to the non linearity of the standard detector. We are talking tiny differences here but the standard 1N60 detector has this non linear gain characteristic where it gets quieter by more than 1dB for every 1dB drop in signal when very close to the noise limit.

So presumably if there was a +1dB pop or crack in noise level then it would go up more than 1dB.

This would make the pop appear louder?

This would only really apply for very weak signals on a flat band where the person's voice was well in the noise level.


Note that I can't realistically test this 'on air' here because the noise levels are too high and AM doesn't get used here in the UK anyway.

 

[G0HZU]

i do not do the 2SC2999 transistor change anymore because i just couldnt tell the difference between it and the 2SC1674.

Oh no, I've just ordered 2x 2SC2999 to do a comparison on :laugh:


Seriously though I don't expect to find a significant difference. The common base amplifier used in the 148GTL-DX has fairly predictable performance at 27MHz and the gain is fairly well defined by the components around the transistor.

The noise figure is going to be dominated by the common base topology as well and will be most affected by the setting of the transformer at the input.


I did do a quick youtube video about the 148 receiver sensitivity and how to ge the best alignment from an unmodified radio.

I'm afraid I have a poor voice for presentations and I do tend to waffle a bit but you may find this two part video to be of interest.

Cobra 148GTL-DX Receiver Alignment and Sensitivity Part 1 (of 2) - YouTube

I do plan to do another one with some minor mods on but I can't say when I will be able to do it

 

[wa4gch]

Oh no, I've just ordered 2x 2SC2999 to do a comparison on :laugh:


Seriously though I don't expect to find a significant difference. The common base amplifier used in the 148GTL-DX has fairly predictable performance at 27MHz and the gain is fairly well defined by the components around the transistor.

The noise figure is going to be dominated by the common base topology as well and will be most affected by the setting of the transformer at the input.


I did do a quick youtube video about the 148 receiver sensitivity and how to ge the best alignment from an unmodified radio.

I'm afraid I have a poor voice for presentations and I do tend to waffle a bit but you may find this two part video to be of interest.

Cobra 148GTL-DX Receiver Alignment and Sensitivity Part 1 (of 2) - YouTube

I do plan to do another one with some minor mods on but I can't say when I will be able to do it

INTERESTING .....

Radio here is a GRANT LT
SIGNAL GEN IFR 1200 S using the SINAD jack hooked to speaker ( AM ) set for about 6DB S/N.

The 2SC1674 was good to about .7UV about -110 DBM ....
The 2SC2999 was good t9 about .4UV about -115DBM ....

I DID NOT try the diodes however what you say makes good sence

Bottom line is it may or may not work on any given radio ....

Now this is MY results with THIS radio a radio that is hot out of the box may or may not have reacted the same.

 

[G0HZU]

Hi Wa4gch

Your tests show a huge 5dB difference in sensitivity between the two transistors and I think something surely isn't right there.

I have actually built the front end of a 148 receiver on a separate PCB and tested it for noise figure. i.e. I tested the input amplifier stage.
I tried it with the AGC diodes and without and also with and without the input L match and L4.
I always got consistent results showing the transistor in this circuit was giving about a 4dB noise figure.
I also tried putting loads of different transistors in the front end amp in the actual radio and always got about the same sensitivity.

Your 5dB improvement seems a lot. Is there any way your radio has been modified in some way or maybe the original transistor was duff?

You can see by my test results in my video that the standard radio can lose 4dB of gain at the top/bottom channels because of the L8, L9 BPF but the S/N only degrades by about 2dB.

This tells me the gain of the first stage in the centre of the band is already good enough to pretty much overcome the noise contribution of the rest of the radio path. So the only way to get better sensitivity would be to lower the front end NF somehow (rather than just pile on lots more amplifier gain).

There's no way you can get the NF 5dB lower just swapping out the transistor. I'd be amazed if it could be 1dB lower NF with just a transistor swap.
However you CAN get a very big change in NF if the input transformer is not set to the best setting for the before and after comparisons.
i.e. it is possible to get poor signal to noise by doing nothing other than adjusting the input transformer L7 slightly to the noisy side of the peak.

i.e. I can make the same radio look bad even though nothing has changed apart from the L7 setting is now 'noisy'

Can I ask, are you testing in the middle of the band for your 5dB improvement or at the outer band edges?

 

[moose]

REALLY glad to see this thread go from "you're wrong, no you're wrong" :bdh: back to discussion and testing. I have been on the fence for a while as to whether to take my Galaxy77 (EPT3600) under the knife and do the diode swap. Not so much the 2SC2999 as it will only see base station use for te rest of it's forseeable life.

Thanks guys! Please carry on... I'm glued!!! op:

73's