I just found this:
This is described by A****f as a "high quality power amplifier for Band II". They claim that it does 30 Watts - it probably does, but on lots of frequencies!

OK - who can spot all the (very obvious) errors in basic design and the construction of this interference generator? A packet of Spangles for whoever comes up with the most!

I'm no expert by any means but the first thing I'd say is the coils should be at right angles to each other.


Sent from my iPhone using Tapatalk

Apart from the obvious idiot syndrome with the inductors and the (SIX!) very expensive variable capacitors, what's with the half turn to ground on the output after the 3 parallel through-hole capacitors?

There's quite a lot of wasted board space. The layout isn't very efficient.

Are all those vias really necessary at ~100 MHz? This is pretty low frequency kit. Just looking at the placement, regardless of the spacing between vias and/or the operating frequency, why are they not a uniform distance away from the input trace when it turns the (right-hand) corner?

I don't understand the solder mask tenting around the transistor input ground planes, given that there is a lack of tenting around the input trace...

I can't see an example of a requirement for the dual-layer board construction, apart from the old-fashioned through-hole components. It must be dual-layer as there would be no vias if it were single-sided. Soldering would also be difficult. Why increase cost by making the underside a ground plane, plus the through-hole plating cost, when you can just adjust your microstrip geometry to suit single-layer construction and spend less on leadless components?

Fuses make me laugh. They're nowhere near fast enough to save the day.

And what the hell is that, a resistor with fine enamelled wire wrapped around it?!

wrong filter and must be screened apart and about the resistor maybe they dont have or dont listen about the vk200 chock coil and after all that they dont ever listen about chip or clad capacitor.you have also to buye a pll kit tool to trim all these onces u failed with one you will have surprises.

if also anyone could guet a picture from this brand you will notice a combiner wilkinson with two board using ordinary carbon resistor instead of rf flange ones.i see this before but i forget where is it so if anyone found post it you will also laugh highly but to be honest i like the flat veronica 1w pll sound.

if also anyone could guet a picture from this brand you will notice a combiner wilkinson with two board using ordinary carbon resistor instead of rf flange ones.i see this before but i forget where is it so if anyone found post it you will also laugh highly but to be honest i like the flat veronica 1w pll sound.

Here's the combiner its on their Facebook page, it looks like it's a resistive combiner too.

looks like a cool thing to hook up to an itrip or tv modulator maybe even mess up 4g because they seem to do that with tv, its like will just give people filters for it, anyway its been replaced by 5g anyway

This PA is being sold by an Australian guy on eBay, so I don't think it's anything to do with a****f - although it does look like his overpriced shit.

Here is the eBay item number - 282100420485

The thing is by the time it reaches the resistors there's probably only about 5w going into them, so they can handle it!

After writing half a page on this I noticed it could be construed as a criticism of the engineering that's gone into this amp. Given that I've now spent almost exactly 16 years forcing myself to bite my tongue on this topic, I've decided to keep it to myself and Albert you can keep the Spangles (again). They're probably as old as the technology on show here.

What are Spangles anyway, some kind-of drugs?

thewisepranker wrote: Are all those vias really necessary at ~100 MHz?

Yes.

Personally, I like to put some many f**king vias on my PCBs that board ends up like a Crunchie bar and can fall to pieces at any moment. I also enjoy the idea of them getting through loads of drill bits at the PCB fab place, it feels like I'm getting better value for money somehow.

shuffy wrote:After writing half a page on this I noticed it could be construed as a criticism of the engineering that's gone into this amp. Given that I've now spent almost exactly 16 years forcing myself to bite my tongue on this topic, I've decided to keep it to myself and Albert you can keep the Spangles (again). They're probably as old as the technology on show here.

Shuffy - don't hold back: say what you think!

I find it particularly funny that they think that a whole load of 5-turn (poorly wound) coils in series, tapped with 65p trimmers could constitute any kind of useful filter and the ½-turn from the collector to deck is just bizarre. I think Paul Hollings got a blurry photo of a PA from somewhere, and this was his attempt at trying to copy it.

Incidentally, you missed a couple of quotation marks around the word "engineering".

Also - we found that "Spangles" are still being made in the Central African Republic, so LB brought me a few packets when he was over there last month. They really don't taste like I remember!

Just for the benefit of the kids around here - "Spangles" (made by the same company that makes "Mars" Bars) were square, individually-wrapped fruit-flavoured boiled sweets. They looked quite like "Tunes" (the menthol sweets for people with blocked noses), if that helps!

Well, OK then as I've had a good afternoon! But this is winding my neck in!

There are a few things I don't like and I think they're getting away with it because it's only a 30W amp.

I don't like the layout. Seems to accommodate the matching, which doesn't look appropriate and I'm sure could be simplified. This looks like a 2 stage amplifier, and for 30W out that's not particularly efficient. I suspect it's part of the attempt to make it broadband. I don't like the use of a pot as part of the input attenuator (that's an interesting attenuator architecture by the way, I expect it's not flat but you might get away with it over the 20MHz of band II). On the subject of the matching (in particular the output) it looks like, at some stage along the line, there's been an attempt to widen the match. This will be out of the window as soon as some Herbert starts messing with the trimmers on the filter (which beggars belief, others have listed the obvious problems). I would like the designer of this to explain the theory.

You say you think the coils are poorly wound. I know it's a promo shot but my initial assumption from looking at the picture was that the coils have been stretched - most of the trimmers are full mesh and therefore haven't been bothered with. It implied to me that someone had already tried to "set it up" in situ. So, that's another reason I don't "get" the trimmers. Due to the filter layout, I expect its characteristic will change with output power and operating frequency, so could this be the reason for the trimmers? Why bother with an attempt at a broadband match and then plonk that lot on the end of it? For this amp architecture wouldn't it be cheaper, safer and ultimately more sensible to use a designed, fixed output filter and a single trimmer in the output match?

Personally if I was building a 30W amp, I'd use a single FET with proper matching. I would expect this to be cheaper, smaller, more efficient, and achieve the stated specification every time without intervention.

That said, I'm sure this amp performs perfectly well as described and is extremely clean and will run for many years without maintenance.

Oh, and I forgot to mention the Spangles - they were OK if there was nothing else on offer!

Albert H wrote: Incidentally, you missed a couple of quotation marks around the word "engineering".

In fairness, I think you started it Albert. You used the word design in your original post, without quotation marks.
Not much I can say, which hasn't already been said. I remember being slightly gobsmacked - when I saw that "attenuator design" in other a****f products - as I was when I first saw the schematic for the "Easytune", and the approach to "broadband" operation.

This screenshot, is from a youtube video - looking at one of their 100W products. Second prize (ie two packets of Spangles), for guessing what's under those round black protective caps (near the "LPF" inductors). It's a bit of a swizz, because I don't actually have the Spangles, and I haven't seen one of these uncapped, but I thought it might be amusing to speculate.

Regarding the half turn inductor. I guess it may be intended as some form of lightning protection ?

Hey Shedbuilt - you're right! I missed the quotation marks around "designed".

Incidentally - have you noticed that Hollings' rip-off of Stephen's oscillator is fundamentally crippled? It can't ever work as designed (when Stephen came up with the circuit back in the late 70s). The whole principle of the "self-doubling push-push oscillator" is that it MUST be truly balanced, The coils must be identical, the oscillator transistors must have the same Hfe and Vcesat, and the loading on the oscillator must be symmetrical.

The brain-dead copier - Paul "I'll sue you" Hollings - doesn't understand the first thing about the circuits he ripped off from Stephen. He's tried to "simplify" the circuit by removing the balancing load components from one side of the oscillator. He's also "cut down" on the "unnecessary" decoupling and filtering components. He's also run the driver and output stages into saturation to make it appear to be "broadband". What he's done is turn a piece of brilliant 1980s design into a complete turd. The crap that this board puts out cannot be filtered to cleanliness!

You can be certain that the components beneath the dipole end caps (for that's surely what they are) will be the usual yellow trimmers - carefully, precisely set in a vain attempt to filter the crap produced by the cheap SMPSU, the "amazing" exciter board and the completely mismatched and incorrectly terminated driver and output transistor. No possible combination of adjustments could ever make this thing in any way clean.

If the PA achieves anything like 100 Watts (which it might - on a wide choice of frequencies, all at the same time), the cheap yellow trimmers will start to fail. They will start to arc through their polyester dielectrics and they will begin to burn. The output device will see a dead short soon enough, and will die. I've seen these "PAs" with tracks burnt off, with the cap blown off the output device, and with charred and melted capacitors.

A few years ago, I suggested that some guys who'd been fooled into buying one of these abortions went round to Hollings' place in Queensbury to demand their money back and to suggest that he might like to make recompense for their lost earnings caused by his utter ineptitude. These guys were ex-squaddies......

Hollings ran away to the Dominican Republic a week or so later!

@ Albert
Lol look here
http://www.a****f.net/2016/10/21/fixing ... ansmitter/

RF-Head wrote:@ Albert
Lol look here
http://www.a****f.net/2016/10/21/fixing ... ansmitter/

Look at the file name of the image
http://www.a****f.net/wp-content/upload ... tter-1.jpg

REPAIRING MELTED TRIMMERS !!!!!!!!!!!!!!

Maybe if you didnt use yellow trimmers which are not rated at 100w they wouldnt burn out haha.
Even the plastic ends that he has placed over the trimmers have melted from the heat.

I once lived in Peurto Banus in costa del sol, I was lucky enough to test Kiss FM's backup transmitter that was located on a gold course. The site had a built in analyser at the base of the tower and you could sample each transmitter there. We tested the a****f setup which was actually a 400w setup compromising of amps combined. It had the main carrier then about 2 mhz either side it had spurs, then 2 mhz lower than that it had another 2 spurs and they went down in the shape of a christmas tree. Somehow this passed its test by the engineer I was with but it just shows how a****f are getting away with it.

Im only an amateur and learning and tbh ive no clue but id never do some of the things Ive seen on a****f's newly 'designed' exciter. The 6 coils in a straight line in oscillator, the tracks must run down the whole length of that being prone to picking up RF from amps nearby... The extra poles on the filter haha. It looks aweful. You would think he would add some high value smoothing cas across the switchmode also to try and filter some of the high freq switching.

Is a****f an approved offcom dealer?

Nope. Not even Mrs Hollings approved of our Paul these days.

Your description of the outputs from a Hollings rig looking "like a Christmas Tree" is quite apt. The one I saw recently (a "30 Watt" one) had almost 2 whole Watts on the frequency it was meant to be on, and the rest appeared as a general mush from about 47MHz (the ½f was 52.1MHz) up to over 150MHz. It was radiating slightly under a Watt on 121.5MHz, which would really piss off the local airport. There was even some output at 243MHz so even the military would be interested in getting this shut down!

The owner of this "high quality stereo transmitter" had paid over £850 for this thing. I took pity on him, and knocked up a little 30 Watt FET job, with a linear supply, nice stereo coder and a proper limiter. This one only appears on the frequency he wants - 104.2MHz - and nowhere else. The 2nd is at -77dBc and the third is in the grass. It actually delivers 35 Watts exactly (but can be turned down if necessary).

The whole thing is in a 2U 19" case, and even has a nice front panel with a few status LEDs - green for OK, red for out of lock, RF off, bad SWR, or too much modulation. It even has a bargraph audio input level indicator. The RF out is through an "N" connector, and the input audio is through two XLR sockets - the input is 600Ω balanced. Mains goes in through the usual IEC "kettle" lead, and there's a couple of outputs through a 9-way socket for silence detect and an alarm output. There's the facility internally to add an RDS board later if he wants. It's good enough to conform to any specification that's thrown at it. He's paid me slightly less than my usual rate (I did it as a favour to get the guy out of trouble) and it only took a few hours to put together out of stuff I had in stock.

The owner phoned tonight to order a back-up rig (at full price!), and said that he can actually be heard in his target area now and the owners of the station are very happy with the results. He's using stacked Yagis to point the signal into a city, and his ERP is close to 180 Watts! It's surprising how good low power coverage can be when the noise floor is low (like it used to be in London).