NXP

[Handy Andy]

Hey! Welcome to the club - sometimes we use them in repairs - like "WHY DO THEY DO THAT" Wham - Wham Wham ...

No, er - well I guess it doesn't hurt to let others know how you feel about these things... but still - GRRR...

I remember the older times where Motorola used to show why they did certain things.

Was it you that mentioned the IMD THD issues that Single Pill designs they produced took care of?

There was a thread or two about how one pill can produce just as much as a two pill - although much of the ruckus over it - there were provable points amongst the efforts - that showed real world results - however used in Bandwidth limitations, narrow as they may be - the sky was the limit - but in the Broadbanding of the platform - a lot of the effort was then taken to matching networks for the conjugate match. The device loses it's glory when it comes to wide dynamic range range and band width performance - one or the other, but not both.

I am not sure if I caught it right, but in everyone's effort to find the right matching network design - the Motorola "test jigs" showed that a single pill design - if done correctly - did not need a filtering Pi network because their Die design as well as the Test Jig itself - took care of the IMD / THD problems - but it narrowed the bandwidth the device needed to perform in. This process of the test jig performance design - took / demonstrated the devices singular ability to perform cleanly.

Well I just wanted to kinda extend on that; for things that I'm noticing We aren't doing anymore that somehow either went away because they didn't think IMD and THD were such issues, or they just got lost in the Translation somehow.

But with these new Pills arriving on our shores - knockoffs and copy-catters - I think they feel that were are responsible for their garbage they send to us and we have to produce the results - not the other way around.

So since I'm losing access to those BJT's of yesteryear (read TO-220 style stuff not the others) - I've had no choice but to segue over to the MOSFET designs and even then, their graphical design overlays and "slice thru" structure and "open platform architecture" the developers were willing to produce in a plethora-show of die designs and overlays even on their websites - are now wrapped up in secrecy - strange world we live in now...

You're not alone in your endeavors - it's been interesting here too...

 

[Shockwave]

Hey! Welcome to the club - sometimes we use them in repairs - like "WHY DO THEY DO THAT" Wham - Wham Wham ...

No, er - well I guess it doesn't hurt to let others know how you feel about these things... but still - GRRR...

I remember the older times where Motorola used to show why they did certain things.

Was it you that mentioned the IMD THD issues that Single Pill designs they produced took care of?

There was a thread or two about how one pill can produce just as much as a two pill - although much of the ruckus over it - there were provable points amongst the efforts - that showed real world results - however used in Bandwidth limitations, narrow as they may be - the sky was the limit - but in the Broadbanding of the platform - a lot of the effort was then taken to matching networks for the conjugate match. The device loses it's glory when it comes to wide dynamic range range and band width performance - one or the other, but not both.

I am not sure if I caught it right, but in everyone's effort to find the right matching network design - the Motorola "test jigs" showed that a single pill design - if done correctly - did not need a filtering Pi network because their Die design as well as the Test Jig itself - took care of the IMD / THD problems - but it narrowed the bandwidth the device needed to perform in. This process of the test jig performance design - took / demonstrated the devices singular ability to perform cleanly.

Well I just wanted to kinda extend on that; for things that I'm noticing We aren't doing anymore that somehow either went away because they didn't think IMD and THD were such issues, or they just got lost in the Translation somehow.

But with these new Pills arriving on our shores - knockoffs and copy-catters - I think they feel that were are responsible for their garbage they send to us and we have to produce the results - not the other way around.

So since I'm losing access to those BJT's of yesteryear (read TO-220 style stuff not the others) - I've had no choice but to segue over to the MOSFET designs and even then, their graphical design overlays and "slice thru" structure and "open platform architecture" the developers were willing to produce in a plethora-show of die designs and overlays even on their websites - are now wrapped up in secrecy - strange world we live in now...

You're not alone in your endeavors - it's been interesting here too...

I have mentioned that the one transistor tuned input and output design is clean like the tank output in a tube circuit. It's also been tested several times that one 2SC2879 can produce as much output as two in push-pull because a pair can never draw anything close to 50 amps.

I've learned that only held true for low voltage, high power transistors that were already operating at or close to their full rated input current. In this case, a 25 amp fuse will almost never burn open before it can protect one, or TWO 2SC2879's. As soon as voltage increases and the operating currents are more reasonable, twice the power output can be expected from a pair of transistors in push-pull because the full dissipation of both devices can be reached before the instantaneous current in either device, exceeds maximum ratings.

 

[kopcicle]

I have mentioned

Insistently, repeatedly, redundantly, consistently, predictably...

that the one transistor tuned input and output design is clean

One, it's not a transistor. Second, a single ended design doesn't do near as good a job of suppressing even harmonics. Only then define "clean" as something more than opinion.

like the tank output in a tube circuit.

It's nothing like a tube type tank circuit. A tube type tank circuit, Pi , or Pi-L, is used for translating a higher impedance to a lower impedance. I can allow for the analogy but it's not a tube.

It's also been tested several times that one 2SC2879 can produce as much output as two in push-pull because a pair can never draw anything close to 50 amps.

Just stop it.

I've learned that only held true for low voltage, high power transistors that were already operating at or close to their full rated input current. In this case, a 25 amp fuse will almost never burn open before it can protect one, or TWO 2SC2879's. As soon as voltage increases and the operating currents are more reasonable, twice the power output can be expected from a pair of transistors in push-pull because the full dissipation of both devices can be reached before the instantaneous current in either device, exceeds maximum ratings.

I don't even know how to respond to this anymore, but I'll try.
This has less than nothing to do with 2SC2879's. This is about LDMOS. The only thing the two have in common is that they are amplifiers.
If your understanding of radio frequency amplification is limited to tubes and bipolar junction transistors then so be it. Create your own thread promoting your own beliefs about tubes and bipolar junction transistors and quit attempting to convince me, in my thread , of something I care so little about.

 

[Shockwave]

I'm not going there Kop, enjoy the holiday weekend.

 

[Handy Andy]

Just stop it.
I don't even know how to respond to this anymore, but I'll try.
This has less than nothing to do with 2SC2879's. This is about LDMOS. The only thing the two have in common is that they are amplifiers.

Now just settle down young man!

You're right, it has nothing to do with 2879's but ...

It has everything to do with the NEWTORKS the MOSFET as a drop in, has to interface with...

THE WHOLE POINT of my post was to segue into a point being made in certain datasheets...

ST Micro doesn't show this as an event but does issue a footnote about RESISTIVE loading...Which if you're into Dimmers; Triacs, Diacs and SCR's - you already know of the limitations of Pulse width when it comes to Inductive loading...

Here's the examples...

Think of it as Device Under TEST - DUT - test Circuits...

Download the PDF's attached to this for more information and those Graphs...

It becomes important when we are to design the Tank circuit to utilize the most effective conversion and translation in/as the Conjugate. That being the output of the MOSFET being the best it can be, without adding IMD and THD artifact's (boxcars, flat topping the like et.al.) because of the changes from the BJT's wide power curve dynamics onto the MOSFET relatively 2 dots - as plotted on a smith chart.

A secondary event if Rise and Fall Times are now being "blamed" on the Leads to hook up the device.

The statement from VISHAY is stating the Rise and Fall times are more INDUCTIVE per the circuit design than the Part itself - for now the UPPER FREQUENCY LIMITS are now placed back to the BOARD design.​

There is one small problem - the MOSFET is so good at what it does, we may need to re-think how we approach the Complexity of the Conjugate of the BJT's part, and design networks that can ACCEPT in ADMITTANCE of the output of the MOSFET into the PI-Filter Network.

• READ: We may need to Rethink about putting the Cart before the Horse.
• 
• Place stronger Emphasis on Using NFB network support designs within the various stages to limit the stages amplification so that the cascade of one stages over amplification or error - is not propagated into the PI-filter or into pass bands the PI-Filter cannot truly remove and keep out of as an Emission TVI/RFI/EMI problem.

Sure these things may be indestructible - but they also produce - in Pronounced effects, Spurries, and Appearances of the Garbage In - Garbage Out quandary...

This was the point of the post, to interface the experiences on the BJT and IMD/THD problems of it (LEGACY), into the MOSFET striipline they use now to replace the BJT's throughout the Transmitter output strip (Future).

We need to work together guys, not flame each other from the past...

 

[Shockwave]

One area that could be politely addressed is how we may view the output of a tuned single ended solid state circuit, as a typical tank circuit in reverse.

It's nothing like a tube type tank circuit. A tube type tank circuit, Pi , or Pi-L, is used for translating a higher impedance to a lower impedance. I can allow for the analogy but it's not a tube.

Since you know that the tube tank circuit has a higher input impedance that is transformed down to a lower one, what stops you from turning that same circuit around and using it in reverse with respect to its input and output, in a solid state circuit? In this case the circuit behaves just like an RF transformer with its primary and secondary reversed, other than the fact its resonance limits bandwidth while also acting as a 3 pole, low pass filter.

While this basic 3 pole filter does not "cancel" even harmonics like a push-pull circuit, it does provide some attenuation to all spurious emissions outside of its tuned bandwidth. The push-pull circuit allows everything to pass without attenuation in the output stage, other than the even harmonic content. Granted, neither one of these solid state circuits will meet current FCC regulations for harmonic attenuation without additional band filtering.

 

[Crawdad]

Edgar however has not yet come out of hiding

Keep a wok handy in case he shows up