How to perform the 2sc2999 and Schottky diode swap

[skybolt58]

I was away from home when I made my last post, and got the figures for the 2SC2999 wrong. Actuallly, the Hfe range is 40-200, and the current gain ranges for the C, D, and E versions are 40-80, 60-100, and 100-200.

The 2SC1764L which is standard equipment in most high end CBs and Exports is rated Hfe = 90-180. So... If your rig happens to have a really good 1764L, and you replace it with a so-so 2999E, there could actually be a decrease in performance, because decreased gain = decreased sensitivity. The best solution is to use a VOM with a built-in transistor tester and see what you got. The 2SC1730L in my 858 rig measured an Hfe of 96. My distributor here let me go through his bin of 2999Es and cherry-pick the 50 best. Most were in the 160-180 range. Of the 200 or so I checked, the best one measured 183 on my meter.


BTW... There is only a 1db difference in Noise Figure between the 1764L (3db) and the 2999E (2db). This is barely discernible. So just make sure the transistor that's in there has really high gain. If you have a 1764 with a Hfe of 130 or better leave it alone and just do the schottky diode mod. Anyone out there REALLY know why the schottky diode mod works??
- 399

SO Like I said in my first post here.





DON'T WORK

--------------------------------------------------------------------------------

I just did this receive upgrade to my texas ranger tr696f ssb and my galaxy DX959 and I actually think it made the receive worse, and I got the good part too, NOT FROM EBAY. I also did a re alignment with still no better.

sad

I have a OLD sears road talker 40 ssb slantface and it receives up to 4 s units that the other 2 don't even pick up.

 

[skybolt58]

Woops

 

[G0HZU]

Judging the (CB amplifier) performance of the 2SC2999 based on Hfe is daft.

The operating point of the transistor when in a typical CB RF amplifier will be set by the resistor networks around it. By operating point I mean the collector current obtained from the biasing nework around the device.

Hfe will have such little impact in this it isn't worth considering. You would have been better off selecting the transistors from a batch for lowest noise.

Even that would be a waste of time as the spread would be insignificant for a CB.
Also, just slapping in a higher Ft transistor could result in instability, excess gain and sometimes the noise can be worse even if the device has a lower NF on the spec sheet.

I'm not saying this mod doesn't work, just that all the guys judging these transistors by Hfe grouping are misinformed.

 

[G0HZU]

Also, if anyone thinks excess gain is a good thing then they should spare a thought for the mixer that follows the RF amplifier. If this mod did provide 6dB more gain then that will degrade the dynamic range of the radio.

If you don't know what this means then just imagine a world where everyone else that is NOT on your channel is suddenly running 4 times as much power and causing more overload to your receiver. That's what 6dB increase in input level does to the mixer.

Sadly, you don't get a 6dB increase in signal to noise ratio on your tuned channel. you might get 0.5dB to 1dB on test equipment but if there is any band noise then this benefit is instantly lost and you just get the overload from the 6dB excess gain.

Does that sound good?

 

[unit_399]

Judging the (CB amplifier) performance of the 2SC2999 based on Hfe is daft.

GOHZU -

For an rf circuit, the biasing resistors are there to set the no-signal collector current at the mid point of the linear portion of the device's operating curve. A transistor is a current in/current out device. Hfe is the ratio of current out/current in. If the circuit uses a transistor with an hfe spec of 40 - 200, then for a given input current, the output of one with an hfe of 40 will be 5 TIMES LESS then the output of one with an hfe of 200 . You say this isn't even worth considering ??? Get a clue.

If a circuit is designed around a device that has a max hfe of 180, then any replacement should have an hfe as close to 180 as possible for best circuit performance. Typically, front end amp gain can be increases by 10% over stock without causing mixer overload problems.

Whoever claimed that this mod will result in a 6 db gain from the rf amp is REALLY stretching it. Most receiver front end amps have 10-15 db gain stock, and an additional 6db would make the mixer section beg for mercy.

Next time you're going to call someone "daft", first make sure that you know what you're talking about.

- 399

 

[G0HZU]

GOHZU -

For an rf circuit, the biasing resistors are there to set the no-signal collector current at the mid point of the linear portion of the device's operating curve. A transistor is a current in/current out device. Hfe is the ratio of current out/current in. If the circuit uses a transistor with an hfe spec of 40 - 200, then for a given input current, the output of one with an hfe of 40 will be 5 TIMES LESS then the output of one with an hfe of 200 . You say this isn't even worth considering ??? Get a clue.

If a circuit is designed around a device that has a max hfe of 180, then any replacement should have an hfe as close to 180 as possible for best circuit performance. Typically, front end amp gain can be increases by 10% over stock without causing mixer overload problems.

Whoever claimed that this mod will result in a 6 db gain from the rf amp is REALLY stretching it. Most receiver front end amps have 10-15 db gain stock, and an additional 6db would make the mixer section beg for mercy.

Next time you're going to call someone "daft", first make sure that you know what you're talking about.

- 399

The collector current in a typical amplifier is always is set such that the operating point is dominated by the biasing resistors and not HFe. A correct design will make this operating point stable wrt changes in HFe.

Hfe is really only relevant when designing the transistor to operate as a switch.

Look again at a CB amplifer design and the bias resistors.

eg on a Cobra 148GTL-DX that uses a typical common base amplifier the source impedance providing the base bias is less than 1k ohm. (2k2 ohm from 8V source with 820R shunt)

The resistance in the emitter is about 1k ohm at full RF gain.

The source to bias it is about 8V. So you get maybe 2.2V at the base and 1.5V at the emitter.

The collector current is therefore about 1.5mA. (ballpark from just looking at a circuit rather than measuring it)

So as you claim you are are 'not daft' why don't you work out the spread in collector current across devices with HFE ranging 80 to 220?

How far different will that 1.5mA collector current be?

I suggest the deviation from 1.5mA will be vanishingly small. (I'd guess ballpark 1% change but you can work it out more formally if you like?)

You seem to think there will be a big difference?

Get a clue.

Next time you're going to call someone "daft", first make sure that you know what you're talking about.

Keep it coming. Answer my stuff above and below and lets see who is daftest?

 

[G0HZU]

As you may need a hint:

Work out the voltage drop in those bias resistors for the change in required base current for 1.5mA collector current according to Hfe across 80-220.

This will give you the change in emitter voltage.
Then you can work out the change in collector current. The result will be ballpark 1% change. i.e insignificant.

Then redo the sums applying a 5% tolerance on the resistors.

See how the resistors are more significant wrt changing the bias current than the Hfe? That's the sign of a good design.

You'd have more success selling people 1% tolerance bias resistors to better define the operating point compared to all this daftness about hfe :laugh:

Also note that with a common base amplifier the gain is proportional to the collector current because the collector current defines the input impedance = 26/Ic(mA) for a simplistic model.

So you'd get a bigger swing in gain if you simply fished out 5% tolerance resistors from a resistor box and measured the gain each time you tried a different 1k ohm resistor in the emitter. Same for the other bias resistors.

If you compared worst case resistor tolerance to best you'd see a much bigger spread in collector current/amplifier gain compared to the hfe difference

if you don't believe any of this then look at the circuit for a Cobra 148GTL-DX and see how the RF gain control works?

It increases the emitter resistance and therefore reduces collector current and therefore reduces the amplifier gain.

 

[G0HZU]

For an rf circuit, the biasing resistors are there to set the no-signal collector current at the mid point of the linear portion of the device's operating curve.

I would suggest that Uniden bias their common base RF amplifiers to provide a compromise between gain and signal handling and stability and (maybe) noise figure. The collector current affects noise figure and so there's a tradeoff with NF and gain and signal handling.

You can get more RF gain from the common base 2SC1674L in a 148GTL-DX by simply adding a shunt resistor across the 1k resistor in the emitter to increase collector current but if you overdo it the gain will go too high and the NF will start to creep up and the amplifier may also go unstable especially with different antenna impedances presented at its input.

 

[N5IT]

Just thought you might care to know that the main board in the current production DX-2547 is now an EPT069611A, and that Galaxy has replaced Q17 with a 2SC4043S. I believe the NF for this device is 3.5 dB @ 500 MHz. I also believe that they are no longer using 1N4148's for the detectors, but I will have to verify that. The 2SC1969 PA has been replaced with an IRF-520 MOSFET.

Dave, N5IT

 

[N5IT]

The detectors are still of the 1N60P variety, not 1N4148.

Dave, N5IT

 

[G0HZU]

People keep rating transistors based on the NF on a spec sheet when really each amplifier design should be looked at from the perspective of the components and impedance driving it and the device operating point.

The front end common base amps in old Uniden CBs are not ideal designs but if you want to improve the noise figure of the complete amplifier then really this should be done either by trial and error with real test equipment or (much better) the design should be simulated on a computer and the biasing and input components optimised for source impedance and operating point to get best noise figure. I'd guess you could get a useful improvement in noise figure just by keeping the 2SC1674L and swapping the parts around it. I'd be very surprised if the input transformer LA279 was ideally optimised for NF for the 148GTL-DX radio.

I've not tried the 2SC2999 swap but I'd also be very surprised if people saw much benefit unless other parts were swapped to optimise the design as above.

Also, there's some really dodgy information being handed out on this thread about suitable diodes for the front end TR/AGC diodes without any appreciation of how these parts were spec'd by the manufacturer.

 

[Cozy]

Well... B&D are out of stock for the 2SC2999E-SPA...

Have to find another place....

 

[RatsoW8]

Well... B&D are out of stock for the 2SC2999E-SPA...

Have to find another place....

I bought a few of the 2SC299E high gain transistors a while ago. How many do you need? I also have a bunch of the correct schottky diodes too. PM me.

 

[unit_399]

Before doing the 2SC2999 swap, it's a good idea to look at the specs of the transistor you're replacing. For instance, in the Uniden 858ssb chassis, replacing the 2SC1730 transistor with the 2999 will actually degrade performance because the 1730 has a lower noise figure than the 2999. In most cases, the schottky diode mod is a good thing.

- 399

 

[G0HZU]

OH yea, that guy in PA is gonna re do everything, take those diodes out and I guess put the 1n60's back in, he told me on the phone to never do those on line mods.. what can I say. for me it didn't work. and I made sure I got the 3.00 sanyo transistor.

I guess you now know not to listen to techno 'kids' on forums that don't have any RF design experience.

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

:laugh::laugh::laugh: