[Amps] Solid State HF amps and kits (VERY LONG)

[Steve Katz]

Jim,

I read all this but still don't really understand your position, or your
point.

CB amplifiers are illegal, period, and anyone who builds and sells them (at
least in the U.S.) should be cited, charged and convicted.  Is this a
difficult concept?

Regarding the PCB prices and costs, your references have nothing to do with
what CCI is selling, which is double-sided, tinned copper, holes drilled and
plated through, transistor mounts punched,  with silkscreening, sheared and
tested.  This is far from the $2 boards you were claiming would be
equivalent.

I know Accutrace, et al, and have used them before.  If I provide them with
Gerbers, they'll make an AR347 amplifier board for about $50 each, as I
previously stated would be the case from Circuit Express.  That includes
shearing, drilling, punching, plating, silk screening and placing each board
in a 2 mil poly bag, just like CCI does.  How does a $2 board figure into
this, anywhere?

It sounds to me like you're just someone with a gripe, and although I fully
understand gripes and frequently have my own, I don't get this one.

73 de Steve, WB2WIK/6

"Each success only buys an admission ticket to a more difficult problem." --
Henry Kissinger

> -----Original Message-----
> From:	Jim Strohm [SMTP:jstrohm@texas.net]
> Sent:	Friday, April 19, 2002 9:24 AM
> To:	Steve Katz
> Subject:	RE: [Amps] Solid State HF amps and kits  (VERY LONG)
> 
> >Jim, I have some comments and questions contained below.  If you'd answer
> >them, perhaps I and other readers might have a better understanding of
> what
> >you're trying to say.
> >
> >73 de Steve, WB2WIK/6
> >
> >
> >> Do you know what fabricated PCBs cost, even in short runs?
> 
> >	[Steve Katz]  I think I do.  I use Circuit XPress for short runs,
> >and after paying the $500 setup charge per layout for Gerber file
> transfer,
> >a PCB the size and quality of the AR347 amp board is about $50/each for
> the
> >first 10 or so, diminishing to about $25/each in 100+ lot runs, assuming
> >high grade FR4 and 2 oz copper, tinned, plus silk screening.
> 
> 
> Some sample prices from EE Times 15 Apr 2002.  Keep in mind that these ARE
> low-ball teaser rates.
> 
> Imagineering (pcbnet.com)
> 2 layers, 90 sq. in.  $25 each, free tooling and silkscreen, first web
> order only, price not verified
> 
> Looks like you could get four amp boards from each $25 board ==> $6.25 per
> board
> 
> 
> Accutrace (pcb4u.com) (real bid!!)
> 20 x 14 (280 sq. in.), 2-sided, 2 oz copper FR-4, no drill, no mask, no
> screen, 4 week delivery
> 
> This should produce 12 pcs of MOT EB-63 (est. 20 sq.in. per amp board) --
> I
> have NOT optimized for production.
> 100 pcs. $48.01 ea. = $4800 for 1200 pcs amp board. ==> $4 per amp board
> 10 pcs. $147. ea = $1470. for 120 pcs amp board ==> ~$12 per amp board
> 
> 1 week delivery is the same for 10 pc., $59.66 per board for 100 pc.
> ==>4.97 per amp board
> 
> Oh yeah -- 4 week, 2500 pcs. -- $18.65 ==> $46625 ==>30,000 amp boards ==>
> $1.55 per board.
> 
> Now, these numbers are TOTALLY without any layout optimization and do not
> reflect any price negotiation.  They also assume customer board shearing,
> drilling, and cut-outs.  I'm confident that with optimization and
> shopping,
> the board costs would be a lot lower than this, and the customer labor
> would be limited to shearing, punching cut-outs for the transistors, and
> drilling or punching four mounting holes.
> 
> And $46K would pay for a PCB prototype mill, lots of tooling, lots of
> blank
> boards, and lots of programming.
> 
> 
> >> With the art work in hand, I could go to a custom board house and have
> >> one-off boards custom-made for less than what CCI is asking.  In a
> >> quantity
> >> run, the price would fall well below two dollars a board very quickly.
> 
> 
> >	[Steve Katz]  If you can get me these boards for $2/each, seriously,
> >I'd like to place an order with you today for 1000 pcs.
> 
> How about 30,000 pcs?  If you can live with a 4-week lead time, we can do
> some business, but you can get a better deal working on your own.
> 
> 
> 
> >> And -- did you happen to check the price of the active devices for that
> >> 1000-watt CCI amp?  You probably will not believe me when I tell you
> that
> >> it's over $900.
> 
> >	[Steve Katz]  If you buy all the parts one-off, I'm sure you're
> >right.  But CCI didn't design the amp, Motorola did, and I think they
> sell
> >the BOM pretty reasonably.
> 
> In million-unit quantities, perhaps.  With buyer negotiation, perhaps.
> But
> the fact with MOT is -- they have never been interested in small
> customers.
> Unless I wanted to buy a million MRF151Gs, I'd probably have to call
> Richard$on and pay whatever those robbers want to charge.  (Actually, I'd
> use a different transistor manufacturer.)
> 
> 
> >> On the other hand -- most of the RFE amps use the 2SC2879 device, which
> >> sells in small-handful quantities at about $15. each.  Properly biased
> and
> >> heatsinked, these are good for 100 watts each all day long.
> >>
> >> So on the one hand you've got CCI selling amps made from expensive,
> >> hard-to-power, hard-to-cool Motorola devices, and on the other hand
> you've
> >> got RFE selling amps made from cheap, easy-to-power, easy-to-cool
> Japanese
> >> devices.
> 
> >	[Steve Katz]  I don't understand the "hard to cool" comment.  Is the
> >device Rthj-c higher?  What's the Rthj-c of the 2SC2879?
> 
> Without looking JUST at thermal resistance of the package, look at the
> _footprint_ of the set of devices required to deliver a kilowatt-class
> amp.
> (this only looks right with a mono-spaced font)
> 
> Motorola "Gemini" footprint:      [==]
> 
> 
> 2SC2879 x 8 footprint:           *     *
>                                  *     *
>                                  *     *
>                                  *     *
> 
> Both sets of devices will generate comparable amounts of waste heat.  The
> thermal resistance of the package divided by the total contact area of the
> sets of devices is probably comparable -- within a factor of 2 or less.
> How much more critical is it to properly affix one 1000-watt device to a
> copper heat spreader on a heatsink, or two sets of four 150-watt devices
> to
> two separate copper heat spreaders on a heatsink, with mechanical
> tolerances built into the multi-spreader design?
> 
> 
> >> The RFE amp goes straight out the door and under the dash of a semi for
> >> about 20% of the cost of the CCI amp -- which requires some relatively
> >> sophisticated work to power up.
> 
> >	[Steve Katz]  All I can say is the obvious, which is: "If it's so
> >great, why don't they just apply for certification?"  It's not expensive
> to
> >do so.  They'd have to lock out 22+ MHz per CFR47 requirements, but could
> >include a retrofit kit to put 12/10m back in for licensed amateurs, same
> way
> >everybody else does.  I don't get their gripe.
> 
> 
> Certification does not require merely negligible gain above 22 MHz.  It
> also specifies a minimum power input level to produce usable gain.
> 
> Fixing the >22 MHz gain issue is simple -- increase the inductance of the
> windings broadband input and output transformers so that they serve as a
> low-pass filter with rolloff starting at 20 MHz.  The sleeve-core balun
> design was (as far as I can tell) selected because it was easy to optimize
> for the HF transistors developed in the early 1970s, and it's easy to
> reporduce cheaply.  With two or three turns on the low-impedance side,
> larger cores, and equivalent turns-count adjustment on the high-impedance
> side, it should be easy to roll your own at home to fit this design
> requirement.
> 
> Now that I have a cheap inductance meter and a cheap accurate HF sweep
> generator, I could do this at home -- in my son's bedroom this time, now
> that he's living with mommie in Oregon.
> 
> 
> >> Are you starting to get a clue about the political and societal
> >> ramifications of the RFE decision?**
> 
> >	[Steve Katz]  Not really.
> >
> 
> 
> OK, here's a clue -- Motorola is an old-line, respected American
> communications manufacturer with significant connections inside the
> Federal
> government and most state and local governments as well, primarily in law
> enforcement.
> 
> When Motorola introduced cheap, powerful bipolar RF transistors that were
> easy to use in the early 1970s, they uncorked one hell of a genie.  Now,
> anybody with a hacksaw and a soldering iron in their garage could build a
> simple, rugged solid-state HF amp that could be driven by 5 watts.  (I
> built my first one in my bedroom in 1981.)  That genie will NEVER go back
> in the bottle.
> 
> And "mass production" of CB linears became a very profitable cottage
> industry, especially in impoverished parts of the country where minimum
> wage work was "good work, if you can get it."  And the finished product
> appealed to the workers -- before Henry Ford, an auto maker could never
> afford a hand-built Cadillac; before Helge Granberg, an electronics
> assembler could never afford a hand-built Collins 30S-1.
> 
> And face it -- it's FUN to have the biggest radio on the block.  Right?
> 
> Japanese manufacturers were able to produce their own transistors that
> were
> "close enough" that the CB amp builders could plug in the Jap parts at
> will.  The other American manufacturers (e.g., TRW and Raytheon) making
> power RF devices left the market; after surrendering this market niche to
> the Japanese, Motorola sold its manufacturing rights to MA/COM.
> 
> Motorola shifted its marketing focus to military, industrial, broadcast,
> and mobile comm markets, where the price point is not a factor in
> transistors -- the cost to the end user reflects the engineering content,
> not the component-cost content.
> 
> So now the American semiconductor manufacturers and Fortune 100 comm gear
> builders are effectively shut out of the marketplace for CB amps by the
> commoditization of components AND labor, as well as by their (reasonable)
> choice not to serve the bootleg marketplace.  Meanwhile, the CB radio
> genie
> that got uncorked in the early 1970s drove a market for maybe two or three
> million bootleg, illegal amps per year, not counting the tremendous black
> market for these radios south of the border.  Since the entire marketplace
> for CB amps is the black market, just how many of the cottage
> manufacturers
> do you think are going to report or pay ANY taxes, or cobform to ANY
> government regulation?
> 
> While it may be unfair to characterize ALL CBers as undereducated yokels
> in
> pickup trucks and semis who hate the government and mostly live in places
> like Mississippi and South Carolina and Texas, fact is -- a lot of us DO
> resemble that characterization.  I live in Texas, used to be a CBer, used
> to drive a semi, and have owned two pickup trucks.  And a couple of times,
> I have sipped some Georgia 'shine.
> 
> So there's quite a bit of political pressure to clamp down on tax-evading,
> law-breaking, ratchet-jawed, chicken-choking radio users who'd be
> delighted
> if the government would just leave them alone so they could buy Japanese
> semiconductors and sell a few radios a week to feed their family and put a
> few bucks in their local economy.  The problem with this sympathetic
> picture of the cottage-industry amp builder is obvious -- ALL radio users
> suffer EVERY time a radio is used improperly.  But with all the genies out
> of all their bottles, what can the FCC do?   Stomp on every amp builder
> they can find?  Engage in grass-roots enforcement measures at every ham
> swapfest Riley Hollingsworth attends?  That's a start.  RFE got stomped.
> And Riley's famous for bitching out  licensed hams at swapfests who were
> displaying CB amps for sale, and demanding that they lock them back up in
> their cars.
> 
> My only issues with the present state of enforcement -- CCI appears to be
> "in bed" with Motorola over the marketing of its products.  And Texas Star
> can freely sell inoperable 4-watt 10 meter CW transmitters with 1600 watts
> worth of finals in them.  I don't exactly understand how all that works,
> but fundamentally we still reap some minimal benefits from products and
> technologies like these being available to us.
> 
> And -- a Texas Star 667 has a street price under $400.  That's kind of an
> attractive price for that particular QRP transmitter.
> 
> Jim N6OTQ
>