Power transistors in receiver front ends.

[Biggles]

Way back in the 'eighties, I remember being on a training course for (I think, Marconi?) main site receivers and was surprised to find the circuit used a bipolar power transistor as the RF amplifier in the first stage of the radio. Apparently this was to improve the intermod/overload performance, as the receiver was designed to be co-sited with some powerful transmitters. This is the only instance I have come across where this technique is used. I can understand that running the front end transistor at higher currents could improve immunity to overloading, but it doesn't appear to be common practice, unless any forum members know otherwise? It would be interesting to find out.
Alan.

[ParcGwyn]

I have always wondered if noise performance could bve improved using power transistors in front ends. 1/f noise may be lower in these devices due to a larger base area.

Dave

GW7ONS

[turretslug]

I think the famous AR7030 receiver uses something like a 2N5109 in a low-gain, high feedback configuration as switchable RF amp? This was a set pitched at the "upper enthusiast" level of the consumer market, making me feel that this technique would surely have been found in the sign-here-Mr. Taxpayer arena of "serious" radios. I get the feeling that JFETs/dual-gate MOSFETS had looked attractive in theoretical isolation with their favourable characteristics including gentle input loading and low noise but being pitched into the rude world of HF congestion highlighted their limited signal handling.

Perhaps, the design goal with many transistor sets of offering straightforward battery operation had hobbled front-end capability? The instant and obvious attraction when transistors came on the scene was their capability of operation at far lower voltage and consequent power- say 10V vs. 200V with typical front ends. A headline point of 12V vehicle/battery usage was obviously attractive, but perhaps if semiconductor front-ends had been designed around 25V, 40V, whatever supply there would have been less obvious disappointment compared to classical valve receivers. Admittedly, many RF transistors have a rather limited voltage capability, though, particularly with early devices. I wonder if any early "pro" front ends adopted the AF118? Maybe nowadays, the final generation of CRT video driver silicon transistors would offer possibilities.

[Argus25]

Transistor noise has a lot to do with the the source impedance driving the transistor (BJT). Low source impedances are better due to the their lower input voltage noise, which is also minimized by choosing a transistor with a low base spreading resistance (rbb) and operating a higher range collector current, provided the hfe remains high. This requires a higher power rated transistor for any collector-emitter voltage. So you might get to see one on the input of a low noise amplifier, looking like it doesn't belong.

With high source impedances the current noise can be reduced by operating at a lower collector current, but as pointed out by Horowitz & Hill, its best to use a Fet when the source impedances are high, not a BJT.

This topic is well covered in Horowitz & Hill.

[G8HQP Dave]

For HF receivers noise is not a major issue, as most noise comes in from the antenna. Linearity is the problem. To get linearity from bipolars you need massive negative feedback, which means high gain. To get high gain you need high current. To get high current you need power transistors.

JFETs and dual-gate MOSFETs were more used at VHF and UHF, where noise matters and signals are generally smaller. However, a JFET could be used at HF if lowish power consumption is a requirement. I seem to recall that MOSFETs had quite high 1/f noise corners so no good for the lower HF range.

[Biggles]

Apologies but I should have said that the receiver in question was an AM VHF High Band fixed main site type.
Alan.

[Bazz4CQJ]

There's an extremly popular design for an RF pre-amp known as the 'W7IUV Beverage Pre-Amplifier' (many hits on Google) which uses a 2N5109. I think that its very high linearity is a key factor. The 2N5109 is discussed in some detail in the the recent thread on magnetic loop aerials on this Forum, with some concern that some of the RCA 5109's on sale in the UK are low spec.

The same thread cites an article on magnetic loop aerials by a Bulgarian amateur, which in turn cites an article looking at the linearity of various transistors. Curiously, the 'cheap as chips' PN2222 turns out to be extremely linear, and seems to be a popular choice for use in mag loop amps.

B

[short wave]

May be to try and protect from the E.M.P. threat at the time?

[Biggles]

It could have been for improved EMP protection, but at that time, site equipment was often fitted with EMP/discharge protection in line on the antenna distribution, as were some vehicles. I remember seeing boxes full of them, all designed around the standard die cast box that seemed to house everything at the time. Whether they would have done much good is of course the question.
Alan.

[ionburn]

I think (if I remember rightly) there are a few circuits which suggest use the 2N5109 for RF amplifiers in some of the radio ham books. I got some many years ago with the express purpose of using them for RF pre-amplifiers. It was a while ago now so I don't remember details.

[Synchrodyne]

The use of RF power transistors in HF receiver front ends was mentioned by Pat Hawker in his Wireless World (WW) 1970 June article on communications receivers. To quote:

“Where bipolar transistors are used in r.f. amplifiers a useful extension of dynamic range can often be achieved by the use of r.f. overlay power transistors, an approach found in some recent Redifon receivers, which also make use of voltage-controlled diode attenuators in the input circuits.”

This I think would have been a reference to the Redifon R499, R550 and R551 series of receivers. Pat Hawker briefly had mentioned the R499 (and its use of RF power transistors) in Technical Topics (TT) for 1968 April, and the R499 was announced in WW 1968 June. The R550 was announced in WW 1969 March, and the R551 in WW 1970 January.

So Redifon was evidently an early user of RF power transistors for receiver RF amplifiers, and I’d guess that it probably used the 2N3866 or 2N5109 device. The latter was announced in Radio-Electronics for 1967 October, and was described as being designed to provide high dynamic range, low noise and low distortion in cable TV (CATV) applications (wideband) at frequencies up to 1 GHz. The 2N3866 was I think similar but earlier. So these devices had VHF-UHF as well as HF credentials.

One may see their utility for HF receivers. Small-signal bipolar transistors had notoriously poor signal-handling capabilities as compared with valves. Dual-gate mosfets could about match valve performance when used in tracking-tuned RF amplifiers. But the emerging need was for wider-band, non-tracking tuned RF amplifiers whose out-of-band protection might be at best sub-octave filters, if any at all. And here, RF power transistors in feedback circuits did the job, particularly for professional receivers that were typically fed from rather large antenna systems.

The Racal RA1772 had a bypassable wideband RF amplifier, but I do not know what form it took. The RA1792 used a 2N3866 and 2N 5160 in a feedback circuit.

Another user of this technique was McKay Dymek, in its upper consumer level and professional HF receivers of the late 1970s, DR22 et seq. It used a 2N3866 or 2N5109 in a feedback circuit. The same was also used as the main amplifying element for its DA100 broadband active antenna. On the other hand, its tuned loop MF active antenna designs used a dual-gate mosfet as the main amplifying device.

Nonetheless, dual-gate mosfets were used in tuned RF amplifiers (and mixers) in HF receivers, and indeed down to LF and VLF. At the consumer level, Heathkit and Drake were early users (1969) in HF receivers that went down to LF. They were also used in quality AM (MF) tuner front ends (e.g. McIntosh and Heathkit 1968-69). At the professional level, Marconi used a dual-gate mosfet RF amplifier in its Apollo marine receiver, which tuned down to 15 kHz. There was also some use of jfets. The GEC RC410-R had a pair of cascode jfet RF amplifiers. The Marconi Hydrus ISB receiver had a cascode jfet RF amplifier followed by a cascode jfet 1st mixer. The Eddystone EC958, which tuned down to 10 kHz, had a hybrid cascode RF amplifier, a single-gate mosfet over a jfet.

The decade from the mid-1960s to the mid-1970s was one of rapid change, where the goal seemed to move from “how do we match valve performance with semiconductors using basically similar topology” to “what do we need to do to lift performance well above that which has gone before”. The use of RF power transistor-based RF amplifiers was evidently an important part of the answer to the second question, so Redifon provided a major stepping stone c.1968. In an holistic sense it all came together with the epoch-marking/paradigm shifting Racal RA1772 of c.1972.


Cheers,

[G0HZU_JMR]

At work we dabbled with the classic Norton 'lossless feedback' amplifier topology up at VHF using BJTs. This was mainly to get low noise figure and decent IP3 performance. But this was a long time ago. Nearly 30 years ago. The PCB photoplots were still being designed with sticky on tracks and traces in those days! But we did get very good performance. I think the Norton amplifier is now used a lot by short wave listeners who like to make high performance receiver front ends using a pair of 2N5109 BJTs as a push pull Norton amplifier.

A modern solution up at UHF is to use a pair of decent low noise HEMT devices in a balanced amplifier. i.e. combine two HEMTs using -3dB 90deg hybrids to make up a balanced LNA. The noise figure can be typically <1.5dB for the overall balanced LNA even with a low loss BPF ahead of it. This configuration can give very good IP3 performance and we have used this topology for amplifiers designed to be used near transmitter masts.

In theory at least, HEMT devices rated up at 5W (RF) can be used to make LNAs up in the V/UHF band although I've never tried designing an LNA using a meaty device like this. I don't think I've seen one used in a real application either. But it should be possible to get sub 2.5dB NF over a huge bandwidth and high IP3 performance too.

[Radio Wrangler]

I'e used a 2-Watt HEMT as a stage in a GHz transmitter (later stages take things to 300W so it's quite fun!)

I wondered about using one as a high dynamic range HF preamp. The quoted noise figure is attractive, but I found problems. The input Z looks like a few Ohms in series with severl pf... nothing scary until you try to match it at low frequencies. The gain goes up like a ski-jump at lower frequencies and both the noise match and the gain match go horribly up in Z.

To keep the thing stable you need to present something healthily resistive to the input and the match opposes this. They hoot!

There is so much gain you need feedback around the thing... Grrr!

David

[Argus25]

Here is an interesting article on low noise amplifier design, but the amplifiers appear to be for fairly low frequency spectrum use.Perhaps they were striving for low noise in audio applications. In any case the author makes mention of the fact that transistors with higher collector current ratings have lower rbb values.

http://www.janascard.cz/PDF/Design%2...amplifiers.pdf

['LIVEWIRE?']

Excuse my ignorance, but what is a HEMT? I know the UJT= Unijunction transistor, so assume that BJT is BiJunction Transistor (wrongly??), but have never knowingly come across a HEMT.

[Radio Wrangler]

High Electron Mobility Transistor

and there's also the p-HEMT and the little case p isn't for the polarity of the semiconductor its "Pseudomorphic High Electron Mobility Transistor"

BJT is Bipolar Junction Transistor... our friends the PNP and NPN jobs with two junctions insde them. And they've got a fair number of weird types like gold-doped ones for faster switching.

There's quite a menagerie all told. Could be useful in scrabble - where they aren't acronyms.

There's no ignorance involved, most people never have to work with these things, though they may well have one or two in their pocket. They offer high gain and fairly low noise at GHz and upwards frequencies.

David

[G0HZU_JMR]

At work we often use GaN HEMT devices for the PA stage in ultra wideband transmit amplifiers. A popular part is the CGH40010F as in the link below:

http://www.wolfspeed.com/downloads/d...7/cgh40010.pdf

We have used these to make amplifiers at frequencies up to about 6GHz. When these devices were first released by Cree I used a couple of these devices to make a wideband lab 'test' (push pull) amplifier that can produce 15-20W across about 6MHz to 1300MHz. This amplifier has been used and abused for many years by various engineers at work for testing limiters and damage protection for various radios. It has never ever failed despite all this mismatch abuse. I have a homebrew copy of this amplifier here at home and I use it quite often to test circuits at high power levels. However, a more typical design might be a single ended 10W amplifier covering 1GHz to >3GHz.

They are depletion mode devices so great care is needed when powering them up as the bias supply and drain supply have to be sequenced correctly or the device can be damaged very quickly and easily.

[G6Tanuki]

Originally Posted by Biggles

Apologies but I should have said that the receiver in question was an AM VHF High Band fixed main site type.
Alan.

A classic example of where good strong-signal-handling is needed: your receive antenna may be sharing a mast with a whole bunch of transmitters, everything from Kilowatt police base-stations to equally-murderously-powered paging transmitters and probably a local-radio FM transmitter or a TV 'infill' repeater too.

There's only so much tuned-cavity ftont end filters can do: your RF amp needs to be able to handle a plethora of signals as well as what you're wanting to amplify: power transistors (often those originally intended for use in cable-TV distribution amplifiers) were the traditional way ahead here.

Plenty of volts HT and tens of milliamps standing current. it can surprise some people to see a finned heatsink fitted to the first RF-amplifier transistor, but believe me they can need it.

One trick I'm surprised wasn't used more to get low noise as well as good signal-handling ability was that of wiring multiple classic junction-transistors in parallel. You get the benefits of increased power dissipation, and noise falls by 1/n where n is the number of transistors.

['LIVEWIRE?']

Thanks for the explanations, David & Jeremy. I've never come across thesee devices, despite working with electronics for almost 50 years, but most of that time has been spent repairing audio gear, AM/FM car and domestic radios. Construction-wise, I've built Valved TRF radios, Transistor (AM) Superhets, including a multi-band one, FET based AM RF sig. generator, etc. None of these things would, of course, incorporate HEMTs.

[Radio Wrangler]

Originally Posted by G6Tanuki

One trick I'm surprised wasn't used more to get low noise as well as good signal-handling ability was that of wiring multiple classic junction-transistors in parallel. You get the benefits of increased power dissipation, and noise falls by 1/n where n is the number of transistors.

That trick crops up in moving-coil record playing pre-amps.

It's functionally equivalent to having one larger area device. It doesn't so much reduce the amount of noise being generated as it reduces the optimum source Z for best noise figure. It's a good trick if your source has a low output impedance and you need an amplifier to suit it.

David