• You can now help support WorldwideDX when you shop on Amazon at no additional cost to you! Simply follow this Shop on Amazon link first and a portion of any purchase is sent to WorldwideDX to help with site costs.
  • Click here to find out how to win free radios from Retevis!

Lets talk about amp biasing and building. Bring in the experts.

Yes both class C.

The most common "AB" biasing you see in CB amps is the resistor / diode feeding the input transformer. Not very good as far as biasing goes, but is easy to spot. The other biasing methods are generally pretty easy to see as well.

A huge amount of CB amps are just class C though, and if advertised as class AB I always assume resistor / diode until I see the circuit / schematic.
 
Yes both class C.

The most common "AB" biasing you see in CB amps is the resistor / diode feeding the input transformer. Not very good as far as biasing goes, but is easy to spot. The other biasing methods are generally pretty easy to see as well.

A huge amount of CB amps are just class C though, and if advertised as class AB I always assume resistor / diode until I see the circuit / schematic.
Ill have to check the one i have here that says class ab
 
Ill have to check the one i have here that says class ab

Never hurts to have a peak inside!

And I should note that MOSFET amps will not have the same style biasing circuits the bipolar / "pill" amps have, so going to look a little different should you crack one of those open.
 
Never hurts to have a peak inside!

And I should note that MOSFET amps will not have the same style biasing circuits the bipolar / "pill" amps have, so going to look a little different should you crack one of those open.
I've never had or dealt with any. I would geuss the rm Italy ones?
Im mostly trying to wrap my head around how things work, i fix stuff here and there. Mostly ham stuff. If i had enough stuff laying around it would be fun to build a big amp like something 1kw plus. I have a few 3-500z amps already but i hate tubes. They will kill ya if you make a mistake
 
Never hurts to have a peak inside!

And I should note that MOSFET amps will not have the same style biasing circuits the bipolar / "pill" amps have, so going to look a little different should you crack one of those open.
Looks pretty nice in comparison but weird design for the matching transformers

Ive used this one alot mobile when i had a magnum 257. Worked great on ssb except the relay sucks. Rx drops out. Gotta smack it

20201209_152630.jpg 20201209_152252.jpg 20201209_152258.jpg 20201209_152316.jpg
 
How can a amp like this easily be made class ab?

This is what gets people into trouble...

Look at the Datasheet please!
upload_2020-12-10_18-5-47.png
It isn't set up for BIAS​

The reason why T/S and others flock to this part, is for the very reason why you DON'T see Bias. For it doesn't - it doesn't need it - and isn't truly designed for AB bias, Linear operation - AM or SSB - because it's will flop if you do try to engage a Bias into it - you'd be lucky to achieve 30 watts versus the "rated" 60 watts - or even the 100 watts these things can do.

Because it's not built for AB use, it's built for Class C, D or above (E,F e&c)

I'm not saying you CAN'T - but you won't like the results if you do try - to get maximum efficiency - you'll need to redo the input and output to obtain (re-obtain Gain) that 60 watts.

Many could use some help or lessons learning how to understand the datasheet and what the part is for.
(...Including me - so thank you for this trip down Memory Lane...)

T/S designed it for FM use, they then took out :L2 - which makes this part TRULY Class C - and then reworked the changed Conjugate - so you can drop in an MRF455, but not a true Class AB SSB part like the MRF421 - simply due to the conjugate matching you make for one transistor; in this example - the MRF455.

To attain those figures of merit
  • - it has a highly inductive input when driven in Class C
  • - but what you get in return when you raise the BIAS up off the Floor (Read this as - NOT GROUNDED GRID) you change the way the transistor looks or sees the RF input
  • - so again refer to the Datasheets
  • - you can attain Class C operation
    • but if you change the BIAS in the circuit - you change (you lose really) how much power the device can accept.

I'm no expert but...
upload_2020-12-10_18-19-46.png

It's not one of the "Best performers" but works for the job of you paying good $$$ to Texas Star for an amp and Motorola then partakes in the Booty - all is Fair in Love and War...

Green = 100mA Bias But you lose a Working Factor for Voltage limitations are limited (50% of MAX)
Amber = 50mA Bias But you lose a Working Factor for Voltage limitations are limited MAX rating 36V only attain 55% Efficiency (Class C)

...Although...
HFE at DC is only 10 - Proper Conjugate gets GPE your RF Gain value ŋ =13 only 3dB more!
Again, refer to the Datasheet - no where on it does it tell you to BIAS in AB - Note the DUT Schematic shows, provides and proves the DC grounding of Base is thru L2 to attain that 60W figure - it's conservative, but no where does it tell you to bias AB!

It's your money.

You can Bias it, but you won't like it's new power curve.

Better to refer to the 421 and build that input and output network - to sub in those parts to replace the 455.

Either way, to Bias in AB, you will lose power in both PEP and Average/RMS - simply due to the new input Bias and how it affects the output - changes the way the part amplifies the RF power.

The only thing I'm trying to stress here is when you try to Bias these amps into Class AB to make them work for SSB, you lose a lot in the process

  • To help
    • - best to look at other Datasheets in how they Bias
    • - with their Support Parts and designs
    • - Changes have to be made to the support chain in order to regain your power curve
    • By adding BIAS - you are changing how the transistor looks for "work" so you are also changing the way the power is admitted and vectored to attain output.
Just some stuff to think about.
 

Attachments

  • MRF455_M.pdf
    85.1 KB · Views: 11
  • Like
Reactions: Shadetree Mechanic
This is what gets people into trouble...

Look at the Datasheet please!
View attachment 41944
It isn't set up for BIAS​

The reason why T/S and others flock to this part, is for the very reason why you DON'T see Bias. For it doesn't - it doesn't need it - and isn't truly designed for AB bias, Linear operation - AM or SSB - because it's will flop if you do try to engage a Bias into it - you'd be lucky to achieve 30 watts versus the "rated" 60 watts - or even the 100 watts these things can do.

Because it's not built for AB use, it's built for Class C, D or above (E,F e&c)

I'm not saying you CAN'T - but you won't like the results if you do try - to get maximum efficiency - you'll need to redo the input and output to obtain (re-obtain Gain) that 60 watts.

Many could use some help or lessons learning how to understand the datasheet and what the part is for.
(...Including me - so thank you for this trip down Memory Lane...)

T/S designed it for FM use, they then took out :L2 - which makes this part TRULY Class C - and then reworked the changed Conjugate - so you can drop in an MRF455, but not a true Class AB SSB part like the MRF421 - simply due to the conjugate matching you make for one transistor; in this example - the MRF455.

To attain those figures of merit
  • - it has a highly inductive input when driven in Class C
  • - but what you get in return when you raise the BIAS up off the Floor (Read this as - NOT GROUNDED GRID) you change the way the transistor looks or sees the RF input
  • - so again refer to the Datasheets
  • - you can attain Class C operation
    • but if you change the BIAS in the circuit - you change (you lose really) how much power the device can accept.

I'm no expert but...
View attachment 41946

It's not one of the "Best performers" but works for the job of you paying good $$$ to Texas Star for an amp and Motorola then partakes in the Booty - all is Fair in Love and War...

Green = 100mA Bias But you lose a Working Factor for Voltage limitations are limited (50% of MAX)
Amber = 50mA Bias But you lose a Working Factor for Voltage limitations are limited MAX rating 36V only attain 55% Efficiency (Class C)

...Although...
HFE at DC is only 10 - Proper Conjugate gets GPE your RF Gain value ŋ =13 only 3dB more!
Again, refer to the Datasheet - no where on it does it tell you to BIAS in AB - Note the DUT Schematic shows, provides and proves the DC grounding of Base is thru L2 to attain that 60W figure - it's conservative, but no where does it tell you to bias AB!

It's your money.

You can Bias it, but you won't like it's new power curve.

Better to refer to the 421 and build that input and output network - to sub in those parts to replace the 455.

Either way, to Bias in AB, you will lose power in both PEP and Average/RMS - simply due to the new input Bias and how it affects the output - changes the way the part amplifies the RF power.

The only thing I'm trying to stress here is when you try to Bias these amps into Class AB to make them work for SSB, you lose a lot in the process

  • To help
    • - best to look at other Datasheets in how they Bias
    • - with their Support Parts and designs
    • - Changes have to be made to the support chain in order to regain your power curve
    • By adding BIAS - you are changing how the transistor looks for "work" so you are also changing the way the power is admitted and vectored to attain output.
Just some stuff to think about.
So its just the nature of the 455. Are others like this? Whats a cheap alternative to 28979s
 
So its just the nature of the 455. Are others like this? Whats a cheap alternative to 28979s


I can start a flame fest from this but; It has to be mentioned because of Logistics, Economics - Manufacturing viability - and profitability.

There really is not a lot of differences amongst the MRF line in that package, they all use similar dies, I said SIMILAR - doesn't mean EXACT - so they had to design the die to make it mass production-able.

So that means many of them are nearly identical - expect for substrate - mountings and perhaps a special doping requirement for Mil-Spec operations we simply can't get into...

They are pretty much the same die in the Motorola - Signetics, Thompson and Toshiba and various others, differentiate their dies to enhance layers or aspects of performance for their benefit - but all are the same underneath that cap - only layout and the structure of it is different - but not by much.

So in light of performances between the die layouts - obviously many prefer the ones that are able to perform at twice the values they are rated for.

Well, that's great, but how is that obtained?

Bandwidth mostly.

Note the DUT schematics, all show one way or another the layout equals to the Frequency of Interest - usually the MUF (Maximum Useable Frequency) - so you see their "peaking" or test bench results at the 30MHz - or perhaps the 1MHz if it's needs a type of reproducible disastrous event or SWR measurements to attain their Conjugate values

Narrow the bandwidth, you can then idealize for the MAIN frequency you use it for - so now you know why they tend to "tweak" amps to attain those Gigawatt values Scotty always Save Kirks' Butt from those "Aliens" by having saved some type of power source in some console somewhere...

upload_2020-12-10_20-16-48.png

But I digress...

If you narrow the Bandwidth, you can improve the peak performance to get back some of that lost gain values from their (READ Maker) adaptation to make the Best of a Class C but try to make it Linear.

So review some of the changes in values - the circuit isn't much different, aside from missing the Bias - but look over the PARTS LIST's between and see the changes - they applied many to make the FM-based amp to work in AM - but to ask it to be SSB is possibly asking too much from that design.

Let's keep it simple.

upload_2020-12-10_20-28-14.png
This circuit looks VASTLY different than the 455
But study it - note there are similarities in functional "blocks"
They simply put in a tuning "tweak" support to help the end user
develop the proper platform and make the DUT work for their application.
  • R1 isn't 10 ohms - it's 5.6 (5R6) 1/2W
  • R2 and R3 are a divider - the VBB is set per your design specs - Diode is your business - they only want the VBB to be able to keep the device in AB mode - if it falls into C - it's only going to produce more IMD and heat - but if your ducks are in a row, it doesn't stay in the "C" Class all the time.
  • The VBB setup is unique, you only need to set the CURRENT and note the "breakdown" Base voltage - so you can apply 4 volts (or slightly less) to attain the 50mA drive and leave it alone. If you use the Resistor values specified, and your Bias is less than 4V - you have set the part to drive for SSB mode. (Idle Current)
    • If any of you have been paying attention, you don't need tons of AMPS to drive the bias, you should focus more on the input drive levels to make sure you don't cause the "Bias" to bottom out and force the Part into Class C.
      • Why This Emphasis on Amps?
      • Forcing your part to stay in Class AB, with more amperage
      • - will actually PUSH the device into Class A operation
      • and with more distortion and power dissipation as the result
      • - remember what was said earlier about the RF rectification occurring at the Base of the DUT. It will make the BIAS climb - in this design, you now see why they provided a divider instead of a Diode - you need to drain the RF leaking in, as well as any DC level BIAS that results from the overdriven input. (Your Input Window drive level is key here - too much and BOOM on any device and any platform)
      • In keeping the BIAS the same mA effort, your device; when it's exposed to higher levels of RF at the INPUT - will make the device operate more into Class C (Bias starvation) and all it's distortion and power dissipation - but the moment you lower the input values, Bias gets Restored - the admittance works and you obtain your Gain, clean up your signal and get less IMD from your Bias arrangement.
IF you decide to Go cheap, you may have to pay a price on the Pills and their Prescription costs - but if you need to use a Diode - then go ahead, you are simply applying that which is your business to offer some form of Regulation for the Bias supply in the first place. So Temperature compensation would be needed - might as well use it for that purpose - but also know, the DIODE's intrinsic impedance to the flow of power across it's junction is and will be different than Room Temperature.

So if you want to COMPENSATE - you may want to add in a 1W ZENER in, (3.2 ~ 3.6V) using a typical Power Regulator using a Power transistor capable of 2A. You can also physically place the Zener close to the parts - BEFORE R2 - so it arrives to R2/R3 (Divider) as compensated. So include it in the circuit, but not going to ground - instead have it in the VBB Regulator SUPPLYING power to the DUT - like this...

upload_2020-12-10_21-19-13.png

Idle current would only be affected by the Voltage Drop of the Diode back at the Regulator - but thermally using the DUT or heat sink of the amp itself - as the thermal feedback.

Now the above example shows a NPN - but if you're in a pinch, and need something to handle this - using a spare parts radio, you can obtain MOST of the needed parts - if they work - to install your Regulator
  • - Got a Cobra 29 or PC78 laying around?
upload_2020-12-11_9-0-22.png

  • Regulator Design is from that radio...use it - just change some values to meet the needs...
upload_2020-12-11_9-18-52.png

R104? 22 ohm 1W
R105? 1.2 ~ 1.5K 1/2W

You can see an article and read about this Regulator more at...
https://e2e.ti.com/blogs_/b/powerho...ies-linear-regulator-with-discrete-components
Just D22 needs to be Thermal (physically) contact with the Heat sink to offer some form of Temperature compensation.
Now, it's not going to be a huge honkin' power regulator - it's only to provide some mA control you still have to fix the problem with input admittance, tuning and the PEP limitations

Again C1 and C2 provide some "window" as well as the R1 values can be altered to offset the recurrent problem of rectified voltages that affect the Base Bias.

You can then use the R2/R3 values from the 2290 datasheet, if and as you need - the Zener will compensate better than trying to press fit diodes into the Flanges of the parts.
 

Attachments

  • 2sc2290.pdf
    121.4 KB · Views: 5
Last edited:

dxChat
Help Users
  • No one is chatting at the moment.