开云体育

here's a? design which I believe to be be mature that will take 3mW or so up to? 270 mW or so to drive RD15HH or similar mosfets.

No originality? claimed but? it's potentially useful to many ? well not many Qmx users ,? but hell I'ms sure Joe Biden or even? the CIA could use one.
Or indeed as a building block for the QSX.?

This OP chip from Tx has been around for quite a while and has been yanked into a? number of other? ham radio designs - which makes me suggest it is the only
one about with the HF performance and power output that can do this duty . Second source ?? prolly none.

The transformers appear to be from? Mini circuits but peoples can wind their own on bihole ferrites (generally type 43 for the IP - output side could be type 61.

NB I have some questions on how this chap has come up with the Av for the? first? chip. I get similar but not the same.? However I have to rush out will lealve for some other time.? NB nbt two - he makes claims for? Av only ,not power gain.? His claim that a given transformer is? lossless looks like Drek to me.? Trust be Verify...

Now there's a UKR outfit that is selling something like this? but with the RHS? Video amp condensed out and a cut and shut done.? Their one? is for? HF only and they claim? 10 W out at the? mosfets with? 12V Vcc and? 2mW of drive at? 50R .? I have huge reservations about that claim.? THe UKRs give no circuit details on ebay. They do give photographs of the component layout? ?and some of them concern me. I'll drop them in a later post, if any interest shown (and if their wsite is not shut down).

It appears that the? 5 plus 5? R? power resistors in the OP circuit are required to protect the OP when it is run into a nominal open circuit . Something similar appears on the? standard design cct? from Tx instruments for the intended use as a high tech communications driver? (modem duty).

I purchased? two of these chips from an Ebay seller at about? 7USD each.? That's about the going rate for suppliers like? Mouser too last time I looked.

in order to use the chip you need a Frit? to hold it and convert it to DIP or you have to design up a SM PCB.? Frits are available from likes of? Futurlec.com which is an Aussie outfit that pretends to be US based.? That's a statement not a recommendation.

Is there any other? Video amp that offers the same performance in a single chip -? does anyone know ?

Ok need to jump aboard my broom and flit off into the night....call sign for the design originator is on the lower RHS of the drafting sheet .? Looks Americano.

TEF

Questions / statements on this design.

First one is how the? drafter of the schematic - not the original designer came to? his Av? results

For instance the? input transformer has a voltage gain of? 2 and? a power gain of slightly less than one.

so as a? log ratio surely the? voltage gain is? +3 dbB ? But he uses? 20 log? V2/ V1 type formula I guess and he's getting? 6 dB

THe reason this is intriguing? is I want to see where his? Av of? 13.4 for the first video op amp comes from.

I 'd say that? the way to analyse this is? to replace teh 180 R feed back resistor with two 90Rs and a? virtual tap to GND.? That's because
it is a fully balanced circuit.??

So Av would? be? 332 /90? ? +1, using the std form for an inverting in put? FB amplifier.? THis gives a voltage gain of 4.7

I we use the Incorrect formula of? ? 20 Log? 4.7? ?we get 13.4? as a the? Av as a logarithmic gain. This aligns with his table.

This query is more about? the correct way of analysing the? performance of the RF amp than picking holes in whatever formulation he?
used for the? logarithmic voltage gain.??

any comments. ?
The other comment which I wills save is those ballasting resistors on the output? - they seem to be required as per the TI? datasheet for this chip.

Generally...

video/RF amplifiers are measured the same.

Power output (undistorted and t 1db comprssion)?
Power gain (also votlage gain)
Distortion IMD, IP3, harmonics
Noise, noise figure of the amplifier.

You can use voltage gain or power gain so long as the input and output load
impedance are known.? For power the common use standard is 50 ohms and
DBW? or DBM.? Votage gain is sometimes used as well and if the load
impedance is know the conversion is easy math.??

That is an opamp amplifier so gain is the ratio of the feedback plus or minus 1.
The latter parts sign reflects if the amplifier is inverting or noninverting.

The series output resistors are for stability as most opamps are rather fussy
about reactive loads.? Fussy translated to they become unstable or can oscillate.

--
Allison
------------------
Post online only,?
direct email will go to a bit bucket.

The main question Allison is do you agree with the splitting of the second feedback resistor as I did to do the maths.? ?

Indeed for the? AD 8009 the datasheet? recommends? about a? 50R resistor in the output drive line to? reduce chance of instability in the amplifier.?

THe datasheet is a bit coy on the conditions that lead to such instability.? Yes I refer to the? 8009.? ?IN the case in question? the OP resistors have been put in? parallel to? ensure that can dissipate? up to? 0.5? W? that is what makes me suspect that those reisistors are there not only for stability reasons but also as? a form of power dump if the? RHS transformer is left unterminated.?

Are there any? dual? Video amps of this type (two to a package )? that you know of? which? can do a similar duty at HF? to? 50 Mhz drive? I get the impression that this OPA is? pretty much unique in this regard.?

I will include a? photo of a commercial application that uses this type of? Tx chip but I have some considerable reservations about what is shown.? Specifically there appear to be be cappies in the lines from the? driver transformer to the OPA . This would mean that? ?there is no way that the signal bias can get to the op amp.. Clearly there could be components under the PCb but given there is a heat sink there , that seems unlikely.? The alternate is that the UKR Mafr has published a false photo and what is supplied does work and is subtly? different from the photo. This design has zero filtering post the? RF power amp mosfets.? ?I include it mainly because it shows how the transformers? were implemented as hand wound designs and not? commericial designs as per schematic supplied (or a completely different application).

Regards,? TEF

?

?

Why do you say there is no input bias?? There appears to be a divider from +V on pin 8, via the 1kR under the chip to the second 1kR and bypass cap to the left of the chip.? Pins 3 and 5 look like they are tied at that junction under the chip.

/Don W3MW

>>>The main question Allison is do you agree with the splitting of the second feedback resistor as I did to do the maths.? ?

It is not my analysis beyond eyeball.? Not a debate for me.

>>Indeed for the? AD 8009 the datasheet? recommends? about a? 50R resistor in the output drive line to? reduce chance of instability in the amplifier.?

>THe datasheet is a bit coy on the conditions that lead to such instability.? Yes I refer to the? 8009.? ?IN the case in question? the OP resistors have been put in? parallel to? ensure that can dissipate? up to? >0.5? W? that is what makes me suspect that those reisistors are there not only for stability reasons but also as? a form of power dump if the? RHS transformer is left unterminated.?

Unterminated transformer has a reflected condition that with dependencies will be reactive and op amps dislike reactive loads.?
Adding resistance makes it less so.? IF your driving a mosfet reactive in the normal case.
?
>Are there any? dual? Video amps of this type (two to a package )? that you know of? which? can do a similar duty at HF? to? 50 Mhz drive? I get the impression that this OPA is? pretty much unique in this >regard.?

NOt studied and likely the LT1254 series (dual 1253) that are fairly old likely qualify.? That was the driver use by
Elecraft in the K2.

>I will include a? photo of a commercial application that uses this type of? Tx chip but I have some considerable reservations about what is shown.? Specifically there appear to be be cappies in the lines >from the? driver transformer to the OPA . This would mean that? ?there is no way that the signal bias can get to the op amp.. Clearly there could be components under the PCb but given there is a heat >sink there , that seems unlikely.? The alternate is that the UKR Mafr has published a false photo and what is supplied does work and is subtly? different from the photo. This design has zero filtering post >the? RF power amp mosfets.? ?I include it mainly because it shows how the transformers? were implemented as hand wound designs and not? commericial designs as per schematic supplied (or a completely different application).

As to filtering the board is too small and a wideband design so that would be on the input and output sides
as needed.

The picture does not obviously agree with what you say.? The backside of the board and whats under the part is not
shown. heatsink is not a limitation.? ?Is it more than a two layer board?? ?I find when I do my own work I
get better answers.? I always find a copy of a bad copy is generally bad.? ?Best case for that is many of
the Motorola apnotes are not a predictable nor reliable when copies are made and subtle changes added.
As a note generally those are at best guides or advertising.? I prefer to look at all with a critical eye.

Generally I do not use those opamps due to other factors like generally high power
consumption and higher noise figures.

--
Allison
------------------
Post online only,?
direct email will go to a bit bucket.

Doug, there might be? an H topology arrangement in place with 2x 1 K resistors forming the voltage splitter and then two resistors under the PCB forming the other part of? the H which provides a? say 200 R termination to the transformer and at the same time places? 6V on both legs of the driver copper? post the coupling capacitors. This assumes components and Vias that I cannot see but makes some sense.? Why are the coupl capacitors there at all ? Perhaps they are there to attempt to null out parasitic capacitance? from the RF input transformer.?

TEF

Allison? - comment on LT 1254 noted.

I also have some things to say about the Motorola high power application notes and they aren't all good.??

I'll write the? out in due course and perhaps save some people from a? bit of heart ache.?

You might note that I have come to the conclusion that? the resistors for? the voltage divider and? loading resistor for the driver might?
be in H configuration with the CT take to ground via an RF bypass.? I cover that with a separate note. That would assume that? some of those?
resistors are under-board? mounted.?

?

TEF

Allison? re? - substitutes for the OPA2674.? and other important matters?

Interestingly Mouser lists? most video amps under? High speed operational amp category.? . If you go to? Video amplifiers you find it very bare.

There are now only two major manufacturers being? AD and? Texas? There are about three other minor manufacturers?

These latter have? Iout max figures of down around 110 mA typically so they are not much chop for the application.

OPA2674. - there is no substitute from? AD since the OPA can source up to 500mA? RF energy? at the output?
The nearest two channel device I can find from AD is the? AD? 8008? which has a 180 mA drive capacity. Threshold frequency might be slightly higher than Texas.

The highest? Iout I can find from an AD device is 175mA from the AD 8009. This is a single channel device.? Frequency specs are somewhat better than the? OPA2674

All three device have? Vmax of? 12 V when operated in single supply mode and generally have? a? +/-5V swing at maximum drive.

Ergo - designing in the? OPA into any bit of? RF? ?kit is fraught with danger commercially? as there is no compatible/comparable substitute. .
=================

I'll leave my critique of? H Granberg's best efforts in terms of what he left out and what he covered by ex Cathedra statements for another time? ?viz.? Motorola application notes.? I have a full bound set on rice paper that I managed to pick up for cheap a long time back.? Full as in Motorola said? they were the most important ones !! Not just Granberg -all sorts of other clowns, many oft them seemingly much more talented.? ?I will stick that critique nder the? ?RF Preamp header for those who may be interested.?

?

Allison and anyone? else who may be looking? -?

I have a chicken scratch list for the issues? that Motorola/ Granberg touch on and then leaves alone or give in consistent information for -

I'll write it up properly and put it on the board for all who are interested as long as Hans is willing to present it? (ie gets past the moderator).

Typically -?

- Rational design basis for phasing chokes for? Mosfet? PP amplifiers never stated by? Granberg.? Just implemented as a fait accompli

- input impedance of unloaded? RF Amp? OPut? transformer? design stated but all over the place and inconsistent.

- Magic? Silver mica capacitors on primary of? the OPut transformers used everywhere including on? centre fed designs but never discussed.

This really does seem an empirical enhancement but? I've never seen it analysed.? I 'll present one? idea on what it might be doing .

Mosfets -? (which admittedly were less available in low cost RF versions when Granberg was writing his material).

- EB -104 "Get 600 W from four off? power? FETs"? ?(EB is engineering bulletin as opposed to? application note )? is the only Granberg article I can find? covering? MOsfet? RF amps and it's along the lines -?
Isn't it wonderful - aren't I smart , with not a lot of discussion of the design basis. Moreover it 's not scalable down because of the schematic
design.? No other Motorola engineers seem to cover PPull? RF mosfet? amplifier design? where an engineering model of a practical say 50-100 W amplifier
is presented.?

I'll get around to this in the next 72 hrs or so.?

Just a note that this issue of? some of Granberg's pronouncements and inconsistencies is still being looked at.

1.? John Z's comment about? Adsl? Line drivers applies to some? Texas chips . I think the part numbers start MHD.? Anways they can be driven from 24 V and?
the indications are they would produce about? 6W rf out up to 30 Mhz.? The 30 Mhz is from the Tx industries'? datasheet.

2. Some basic work on the? phasing choke used in the Kees Talen? 16 W RF amp? (13.8 V supply) shows that the? ?BN 43-302 appears to only just work.? That's to say? it appears to be near saturation and that would cause the impedance to fall away badly.? This? would lead to power from the primary of the? OP transformer being dumped into the series impedance of the RF choke.? ? I need to do some LTspice simulation? on this aspect (but bear in mind LTS treats magnetic components as completely linear ). H? Granberg? ?(EB and AN series from Motorola) provided no discussion of? Pt choke design anywhere I can see - eg impedance requirements versus the parasitic impedance of the transformer primary it was driving or any criterion like that.?

Will report in due course.? TEF