UK Vintage Radio Repair and Restoration Powered By Google Custom Search Vintage Radio and TV Service Data

Go Back   UK Vintage Radio Repair and Restoration Discussion Forum > General Vintage Technology > Components and Circuits

Notices

Components and Circuits For discussions about component types, alternatives and availability, circuit configurations and modifications etc. Discussions here should be of a general nature and not about specific sets.

Closed Thread
 
Thread Tools
Old 28th Jan 2014, 1:14 pm   #81
G8HQP Dave
Rest in Peace
 
Join Date: Sep 2008
Location: Solihull, West Midlands, UK.
Posts: 4,872
Default Re: FET Questions

Quote:
Originally Posted by G6Tanuki
[2] To overcome mixer-noise [old multi-electrode mixers like the ECH35 and 6SA7 had equivalent-noise-resistances of 250,000 Ohms-plus and you needed significant RF amplification to get to a point where first-RF-tuned-circuit noise overcame mixer-noise]
I suspect this is a widely-believed myth. A voltage gain of 10 from RF grid to mixer grid would be sufficient to allow RF valve noise to dominate over mixer noise. Old receivers were noisy because of the loose antenna coupling required to give good tracking for random antennas, and poor L/C ratios at higher frequencies because of the need for general coverage over the whole HF band.
G8HQP Dave is offline  
Old 31st Jan 2014, 3:13 am   #82
Synchrodyne
Nonode
 
Join Date: Jan 2009
Location: Papamoa Beach, Bay of Plenty, New Zealand
Posts: 2,943
Default Re: FET Questions

Quote:
Originally Posted by Radio Wrangler View Post
Lots of RF gain in a supertuner is OK just so long as you have a supermixer to handle it.

Very few of these firms were into careful design at the system level. A standard failing was that there were hundreds of articles and books written on planning the frequency conversions of a receiver and combatting VFO drift, and getting noise figure of a stage down, BUT there was a general lack of info on how to design a system for good dynamic range.

Consequently there are tuners, receivers etc where the gain/loss distribution and the distribution of selectivity aren't poorly thought out, they weren't thought out at all. A lot of choices were done by copying someone else's and maybe tweaking something a bit.

For planning such systems, there was a lot of tedious arithmetic needed, so it was fertile ground for what to do with that new computer the department had bought.

Gain distribution (and variation by agc) is an interesting subject, but there does not seem to be too much information readily available relative to specific receivers. Some of the FET-rich Japanese HF receivers of the 1980s have published level diagrams as previously mentioned. I have attached one or two others that I have come across. The Mullard valve TV diagram provides quite a bit of information, including gain reduction with agc and the agc voltages required. The GEC BRT400 table simply gives stage gains, presumably maxima; it would be interesting to know what were the reductions under agc and what the required agc voltages were.

Cheers,
Attached Thumbnails
Click image for larger version

Name:	Mullard Gain & Level.jpg
Views:	616
Size:	43.9 KB
ID:	88452   Click image for larger version

Name:	Benson & Whitaker p.13.94.jpg
Views:	755
Size:	39.0 KB
ID:	88453   Click image for larger version

Name:	GEC BRT400 Gain Distribution.JPG
Views:	477
Size:	31.8 KB
ID:	88454  
Synchrodyne is offline  
Old 31st Jan 2014, 3:49 am   #83
Synchrodyne
Nonode
 
Join Date: Jan 2009
Location: Papamoa Beach, Bay of Plenty, New Zealand
Posts: 2,943
Default Re: FET Questions

Quote:
Originally Posted by G8HQP Dave View Post
I suspect this is a widely-believed myth. A voltage gain of 10 from RF grid to mixer grid would be sufficient to allow RF valve noise to dominate over mixer noise. Old receivers were noisy because of the loose antenna coupling required to give good tracking for random antennas, and poor L/C ratios at higher frequencies because of the need for general coverage over the whole HF band.
That would seem to hold even for valve VHF TV tuners, which, for Band III at least, seem to have been an extreme case (see the attached excerpt from Fisher) basis the perverse use of noisy pentode mixers. The Mullard level diagram attached to the previous post shows a Band III gain of 7 for the PCC84 RF stage.

For VHF TV, IFs were high enough (from the mid-1950s, anyway) that image rejection seemed not to be a big problem, and sufficient could be obtained with a single-tuned aerial circuit and a double-tuned interstage. Thus there was never any need for more than one RF stage.

Cheers,
Attached Thumbnails
Click image for larger version

Name:	Fisher p.50,51.jpg
Views:	595
Size:	108.1 KB
ID:	88455  
Synchrodyne is offline  
Old 31st Jan 2014, 7:05 am   #84
Synchrodyne
Nonode
 
Join Date: Jan 2009
Location: Papamoa Beach, Bay of Plenty, New Zealand
Posts: 2,943
Default Re: FET Questions

Quote:
Originally Posted by G6Tanuki View Post
From the traditional valved HF communications-receiver perspective the "two RF-amps" thing had several purposes:

[1] To allow three loosely-coupled RF tuned-circuits ahead of the mixer. This helped reduce image response [which despite 3 tuned-circuits was still a problem at 25+ MHz if your IF was only 465KHz]
Even with an IF around 1.6 MHz, three tuned circuits were probably desirable for good image rejection at 30 MHz.

Quote:
Originally Posted by G6Tanuki View Post
[3]An extension of [1] - the additional pre-mixer selectivity could give reduced cross-modulation effects in the mixer [but paradoxically the extra pre-mixer amplification could increase the risk of it!]
It is very much a trade-off then. In the FM tuner/receiver case, intermodulation is said to be the main bugbear, as evidenced by the half-IF response. Reducing this was the rationale given by Ambit for using two dual-gate mosfet RF stages and five tuned circuits ahead of the mixer.

Quote:
Originally Posted by G6Tanuki View Post
[4] Pre-mixer selectivity provided greater isolation of the LO signal from the antenna - potentially important in some military/commercial applications where multiple receivers were fed from the same antenna [or if you were at risk of the enemy DFing your LO signal].
I wonder then if this is why the Eddystone 880 had two RF stages, despite its highish 1st IF which would probably have allowed a single stage simply from an image rejection viewpoint. Apparently the 880 was designed for very low radiation, for use in embassies, etc. (As a corollary, was the RA17 somewhat the other way in radiation terms, with its harmonic generator, etc.)

Quote:
Originally Posted by G6Tanuki View Post
In the context of FET or bipolar VHF/UHF TV tuners, image interference [where extra tuned-circuits ahead of the mixer would be a Good Thing] never seemed to be a big issue - not in the UK anyway. Bands IV and V had national/regional station-planning so that you rarely had a significantly-powerful local transmitter on the image-frequency of another station serving your locality: though I must admit to having experienced image-reception of aircraft on a Band-II receiver.
Even where two RF stages were used in bipolar UHF TV tuners, such as in the Mullard ELC1043, the first stage had a broadband input (presumably for low noise) and the first interstage was single-tuned. So there was no, or little RF selectivity gain (possibly the fist interstage did slightly better through being sandwiched between two amplifiers).

UHF channel allocation generally seemed to be planned on the basis of specific receiver image rejection performance and IFs. In the USA, the FCC UHF planning (including the UHF “taboos”) assumed a receiver IF (vision) of 45.75 MHz, and there was concern when some receiver makers were slow to adopt it.

Quote:
Originally Posted by G6Tanuki View Post
I can see the logic in having two MOSFET RF amps in a broadcast-band tuner: the first one essentially being noise-figure-optimised [with the AGC delayed quite a bit], the second one being deliberately designed to provide relatively low-gain (tap the gate-connectors well down the tuned-circuits so you get the highest possible tuned-circuit 'Q') 'selective-matching device with a bit of gain' then into the mixer.
Yes, even given the downsides, if two RF stages were used in the valve era and also for solid-state FM front ends, then why not for mid-range solid-state HF receivers with low 1st IFs, although these were probably few in number. But it seems not to have been done.

The data provided by Zenith for its VHF TV tuner (and used as a generalized example in the literature) showed that a dual-gate mosfet was somewhat better in terms of its cross-modulation performance than a typical triode-pentode valve. So it would not appear that mosfets mixers were less tolerant of RF gain than valve mixers. At HF heptode mixers were commonly used; their cross-modulation threshold might be higher than for TV pentodes; the answer to that question is not readily apparent. But even if so, I should not think that the difference would be huge. Incidentally, the Zenith data show that jfet mixer was better in cross-modulation terms than the mosfet, although not used in TV tuners because its low gain required the use of another IF amplifier stage. But that may explain why the Japanese makers often used jfet mixers, usually push-pull, in their FET-based mid-range HF receivers. A jfet mixer thus might be a better choice in FM front ends that have two mosfet RF stages, as well; in a “supertuner” an extra IF stage to make up the gain is unlikely to be a problem.

Cheers,
Attached Thumbnails
Click image for larger version

Name:	Mullard ELC1043.jpg
Views:	1409
Size:	63.7 KB
ID:	88456   Click image for larger version

Name:	Zenith VHF TV Mixer.jpg
Views:	951
Size:	49.1 KB
ID:	88457  
Synchrodyne is offline  
Old 3rd Apr 2014, 3:49 am   #85
Synchrodyne
Nonode
 
Join Date: Jan 2009
Location: Papamoa Beach, Bay of Plenty, New Zealand
Posts: 2,943
Default Re: FET Questions

Variations on a Theme?

In one of the TI papers on the use of dual-gate mosfets in TV front ends and IF strips, it was noted that an early developmental device, the SFB8970, was commercialized as the 3N201, 3N202 and 3N203. The differentiation appears to have been by specific application. Thus the 3N201 was described as a VHF TV RF amplifier, the 3N202 as a VHF TV mixer, and the 3N203 as a TV IF amplifier. I am not sure how the functional separation was done; was the basic device “tweaked” along specific vectors to optimize functional performance in each case, or did normal production scatter result in a mix from which the specific variants were selected?

The same functional relativity applied to an improved series, namely 3N204, 3N205 and 3N206.

Further improvements came with the 3N211, 3N212 and 3N213. Here the 3N211 was used as a VHF TV RF amplifier and a TV 1st and 2nd IF amplifier, the 3N212 was the VHF TV mixer and the 3N213 was a TV final IF amplifier, intended to drive a diode demodulator. This was noted as having a higher breakdown voltage than the others, so it may have been a physically different device. Hitherto three-stage IF strips offered by TI had used mosfets in the first two stages, but a bipolar device as the final IF in order to drive a diode demodulator at a sufficiently high level.

The SFB8970 was also the basis for TI’s early work on FM front ends, used as RF and mixer. Whether there were separate type numbers for the commercial derivatives I do not know.

RCA referred to developmental device TA7151 being used as an FM mixer in its early paper on radio applications. I suspect that when commercialized, this became the 40603 (FM RF amplifier) and 40604 (FM mixer). Actually, it would not be too surprising if the 40600 (VHF TV RF amplifier), 40601 (VHF TV mixer) and 40602 (TV IF amplifier) were also derivatives of the TA7151.

The 4060n series were not gate-protected devices, but the apparently later 4082n series were. The latter comprised the 40820 (VHF TV RF amplifier), 40821 (VHF TV mixer), 40822 (FM RF amplifier) and 40823 (FM mixer).

The previously mentioned RCA paper also referred to the TA7150 as an AM RF amplifier. What that became when commercialized I cannot ascertain. But the 40841 might have been a later and similar gate-protected version. For example, RCA showed the 40841 as AM RF amplifier ahead of its CA3088 AM radio IC.

Another possible case of variations on a theme were the Signetics SD307 and SD308, respectively intended for use as UHF mixer and UHF RF amplifier. I am not sure whether they were intended for use in UHF TV tuners, but if so, then when the SD307 was used as a mixer, it would represent a departure from the previously noted and apparently “normal” combination of a dual-gate mosfet RF amplifier with a diode mixer.

Something else that is evident is that there was a narrow time window early in the FET age when non-gate protected devices were used, presumably ahead of the availability of the gate-protected types. One example is the initial version of the Revox A76 FM tuner, which used the RCA 40603 and 40604 pair. Leak used a 40468A (single-gate) and a 40603 in its Stereofetic FM tuner of 1969. Of interest is that the 40603, designated as an FM RF amplifier, was used as a mixer by Leak; one may infer that the differences between the 40603 and 40604 were not inimical to their cross-use. The Rogers Ravensbrook (also of 1969) had a 40603 RF amplifier (with a bipolar mixer); the slightly earlier Ravensbourne (of 1968) had non-gate protected dual-gate mosfets in the RF and mixer positions, but I do not know their types. The earliest use of gate-protected devices that I can find from the information on hand is in the Eddystone EC958 HF receiver, announced I think around the beginning of 1969. This used the RCA 40673 in the 1st, 2nd and 3rd mixer positions. I suspect that the 40673 preceded the 4082n series, and it appears to have been aimed primarily at military, commercial and industrial applications. Its main point of differentiation appears to have been its capability of operating up to 400 MHz, whereas the upper limit for the 40820/1 was stated as 250 MHz, and that for the 40821/2 150 MHz.

Cheers,
Synchrodyne is offline  
Old 3rd Apr 2014, 10:54 am   #86
Andrew2
Nonode
 
Andrew2's Avatar
 
Join Date: Mar 2010
Location: Dukinfield, Cheshire, UK.
Posts: 2,034
Default Re: FET Questions

I'm sure I'm not alone in finding this stuff fascinating, so thanks for going to the trouble of unearthing it and posting it.
I had a quantity of 40673's and I used them in RF stages, mixers and IF's in my homebrew transceivers and transverters for many years with excellent results.
When they ran out I was lucky enough to spot a member of this forum flogging a bag of 3N140 devices which I snapped up.
These work well too, but they seem not to have gate protection diodes as they come with a wee spring clip that is removed after the FET has been safely soldered into the circuit.
__________________
Andy G1HBE.
Andrew2 is offline  
Old 3rd Apr 2014, 11:10 pm   #87
Synchrodyne
Nonode
 
Join Date: Jan 2009
Location: Papamoa Beach, Bay of Plenty, New Zealand
Posts: 2,943
Default Re: FET Questions

It does seem that the 40673 was a standard general-purpose device, quite widely used.

From the RCA data one might infer that the 40673 and 4082n were variations on the same device. One wonders if the 4082n series were favourably priced to equipment makers who used them in large quantities for their designated applications. Perhaps for the home constructor the price differences were small-to-negligible, so that the 40673 was the version of choice.

At HF, rather than VHF, it may not have mattered so much, either. Wireless World 1969 July included an article (first part thereof) on an advanced amateur band HF receiver, by D.R. Bowman. (It might be noted that gain distribution was a key factor in its design, too.) For the RF amplifier, a dual-gate mosfet was chosen. In respect of device choice, the author made the comment:

“R.C.A. have recently marketed an integral cascode device which has the advantage of a somewhat lower h.t. requirement than separate devices, as well as a very low reverse transfer admittance value.

“These devices are marketed under an assortment of code numbers and vary in price from about 7s to 14s. The author tested the following types and at up to 30 MHz could find very little difference between them:- 3N140, 3N141, TA7149 and 40500. (Since writing the MEM 564C has become available and is to be recommended since gate protection is incorporated.”


Also, that tends to confirm 1969 as the time when protected-gate devices became generally available.

An odd reference to the 40673 was made in WW 1975 October, in the first part of an article by D.C. Read on a high quality TV tuner (with video flat to 5.5. MHz, group delay corrected to 5 MHz, and split sound.) After describing this choice of vision IF amplifier (emitter follower feeding a bipolar shunt cascode circuit, between an LC block filter and an MC1330 IC vision demodulator), the author goes on to make the comment:

“Similar results can be obtained from a dual-gate f.e.t. (e.g. the RCA 40673), but these components are fragile, and the more robust circuit used here was chosen in preference.”

So concerns about mosfet fragility had not disappeared even several years after their widespread use. But choosing the 40673 as an example suggest that this was perceived as being amongst the best known devices of its type.

Cheers,
Synchrodyne is offline  
Old 19th May 2014, 2:37 am   #88
Synchrodyne
Nonode
 
Join Date: Jan 2009
Location: Papamoa Beach, Bay of Plenty, New Zealand
Posts: 2,943
Default Re: FET Questions

Early on in this topic – post #15 - Radio Wrangler mentioned Rafuse’ use of quad DMOS fets in a ring mixer, and Racal’s subsequent adoption of this approach in its epoch-marking RA1772 professional HF receiver of 1974, thereby lifting its performance (IP3, etc.) well above the level previously prevailing.

The empirical evidence though is that for quite some time, this technology was too costly and/or too difficult to use in consumer/amateur HF receivers, which sold at prices that were say an order of magnitude below those of the professional types. The earliest example I can find of a not-too-obscure consumer/amateur HF receiver using a DMOS quad 1st mixer is the AOR AR7030 of circa 1996. For this, excellent strong signal performance was a key design desideratum and the leading feature in its sales literature. (I had forgotten about the AR7030 until recently, when I discovered a brochure, etc., amongst my data collection.) Its 1st mixer is in the form of an SD5400 integrated circuit.

I wonder whether the DMOS devices themselves were initially too costly for use in consumer/amateur receivers, or whether it was the surrounding circuitry that made them so. I do not have the full Rafuse paper, just the digest, but in that the DMOS quad mixer is shown as requiring a local oscillator input of 2 to 3 watts, which seems very high, and probably was not easy to do in a clean way. That might well have inhibited its use outside of the top professional receivers.

The AR7030 was otherwise not a fet-rich circuit in terms of the signal path. There was a push-pull jfet 1st IF amplifier ahead of the 45 MHz 1st IF crystal filters, but thereafter it was mostly bipolar IC, including a consumer IC, the TDA1572, that provided the 455 kHz IF gain. The RF amplifier was a bipolar RF transistor with feedback, an approach borrowed I think from professional receiver techniques, although seen before in consumer/amateur HF receivers in the McKay Dymek DR22 et al from circa 1977. (These did have signal paths that were fet-rich.)

Nevertheless, the DMOS high-level mixer was a technique that was developed early in the FET era, circa 1968. So whilst small-signal FETs were enabling a transition from valves without the performance penalties commonly associated with small-signal bipolar devices, including in consumer/amateur equipment the other end of the spectrum a specific variety allowed a step change improvement in professional HF receiver performance.

Cheers,
Synchrodyne is offline  
Old 19th May 2014, 8:19 am   #89
Radio Wrangler
Moderator
 
Radio Wrangler's Avatar
 
Join Date: Mar 2012
Location: Fife, Scotland, UK.
Posts: 22,800
Default Re: FET Questions

Rafuse's FET ring mixer really did need that amount of LO drive. His MOSFETS needed rather a lot of volts to turn them on into saturation. The LO feed wasn't tuned but was resistor-loaded. It needed a broadband QRP transmitter to drive it. But in the early 70s that was OK. Some good bipolars were appearing targeted at CATV systems. The 2N5109 and 2N3866 for example. Racal picked it up at this point and used a quad of discrete MOSFETs

Cost-wise 4 FETs cost more than one, and the lower on-resistance parts have a larger die than low-current types. Consumer equipment design often hinges on differences of a penny, so there were two clearly differentiated paths of development. THe dual gate MOSFET got consumer market sized turnover and the price in bulk was pennies. The low resistance single-gate parts and the devices needed to drive them stayed in the 50p region. Add in the transformer and the mixer was priced significantly out of consumer gear. Oh and did I mention that being a lossy mixer, it would force an extra amplifier stage over the dual gate MOSFET design?

When quad DMOS chips became available, they were pitched as conveniently matched analogue switches and their prices had to be competitive, so they were in the couple of quid region. But on top there still were drivers, transformers and gain costs. This was the era of the SD5000. Ed Oxner had Siliconix do a ready-connected ring of the same basic DMOS devices, the Si8901. Racal stuck to four transistors.

The gate on-voltage of the SD5000 was an improvement which kept the performance and reduced the LO drive needed - a bit.

The thing that has always interested me is that Rafuse missed a trick. A big trick. He'd found an interesting new switch device, yet he stuck with a circuit that was dedicated to handling the issues of separating LO and signal flows for 2-terminal devices. He'd now got a good 3-terminal device. What he did was good progressive development, but he stopped at what should have been an intermediate step, and failed to take the next one.

I'm not sure who took that next step. Colin Horrabin was the first to publish, and Pat Hawker gave it space in Tech Topics (Sadly missed, both Pat and his column). Other people were also active in that field and I was merely one of them. Ian B, a chap with a background in audio and low frequency analogue suggested the grounded-source topology to me when we were working on a receiver and I had one of those 'Oh Crikey' moments when I saw what he'd said in conjunction with the Rafuse paper which I'd had in my head for many years.

The threshold of the SD5000 meant that reasonable saturation could be obtained with only an HC series CMOS gate driving the DMOS gates. Fourteen years earlier,I'd been working on a seriously state of the art receiver and had run into the issue of 'why bother to balance a mixer so carefully if you're going to hit it with a poor mark:space ratio LO drive?' My solution then was to redo my synthesiserfor twice the frequency I'd thought of, and to integrate a fast ECL divide-by-two into the mixer (diode ring in this case) LO drive chain, and keep everything balanced from then on. For the first grounded source DMOS mixer I built, I used a 74HC74 flip flop straight onto the gates, and then later an AC version.

This breadboard mixer with some quickly wound transformers was giving over +40dBm third order intercept point, about 6dB insertion loss and the flip flop was using about 10mA DC from a 5v supply.

I was awash with high level mixers at this time. I'd spotted an ad from John Birkett in SPRAT, where he had some RAY-3 minicircuits mixers for sale. I phoned him and asked if there was an error in his price. He knew they were high performance ones, but said that the amateur radio market was for the MD108, and anything else would only sell for a lower price than a real MD108, and that people would find high LO power difficult. I suppose he was right, but it seemed sad. At least I knew that John wasn't missing out, so I bought a dozen with a clear conscience.

So, with very high dynamic range mixers available, the spotlight gets directed elsewhere. David Norton had been busy patenting transformer-feedback amplifiers and one firm had patented the use of directional couplers in feedback. There were good amplifiers to go with the mixers. The limiting factor fell on crystal filters. So we can make receiver structures where the intermod performance is set by the crystal filters. The best high-level filters I ever came across were made by C.E.P.E. Part of the Thomson-CSF empire of old. These are no longer available and they were bespoke units at terrifying prices back in their day.

The quartz limitation is still true today, and a modern receiver design is an interesting tradeoff between Quartz and ADC.

I'm now going to run my typing fingers under the tap

Cheers,
David
__________________
Can't afford the volcanic island yet, but the plans for my monorail and the goons' uniforms are done
Radio Wrangler is offline  
Old 19th May 2014, 2:06 pm   #90
Radio Wrangler
Moderator
 
Radio Wrangler's Avatar
 
Join Date: Mar 2012
Location: Fife, Scotland, UK.
Posts: 22,800
Default Re: FET Questions

The firm with the patented directional-coupler feedback amplifiers was Q-Bit. It just took a while to surface in memory,

They also did a stepper motor driven automatic tuning tool with ends to suit most trimmer capacitors and tuning slugs!

David
__________________
Can't afford the volcanic island yet, but the plans for my monorail and the goons' uniforms are done
Radio Wrangler is offline  
Old 22nd May 2014, 6:10 am   #91
Synchrodyne
Nonode
 
Join Date: Jan 2009
Location: Papamoa Beach, Bay of Plenty, New Zealand
Posts: 2,943
Default Re: FET Questions

Thanks very much for the detailed commentary, David.

I imagine that apart from cost, a de facto QRP transmitter to provide mixer drive would have been difficult for consumer/amateur receivers because of the difficulty of stopping the LO signal from escaping from what was often a single-box construction, as compared with the multiple screened boxes that were used in some professional receivers. If I have this right, the MD108 DBM required something like +7dBm LO drive, which seems modest in comparison.

It’s interesting that the 2N5109 and 2N3866 should be mentioned in the QRP context. The AOR AR7030 had the 2N3866 as its RF amplifier, and either could be used as the RF amplifier in the McKay Dymek receivers. One might say that they were examples where the right kind of bipolar device (low power RF) in the right kind of circuit (feedback) outperformed small-signal fets. The Redifon R550/R551 commercial/marine receiver had a push-pull pair of 2N3866 as RF amplifier. It seemed to have been a tour-de-force of push-pull and feedback bipolar circuitry.

The SD5400 mixer in the AR7030 was driven by a 74AC86 gate, so apparently did not need a QRP transmitter to drive it. Anyway, there was fet technology in both the mixer and its driver.

Cheers,
Synchrodyne is offline  
Old 22nd May 2014, 8:14 am   #92
Radio Wrangler
Moderator
 
Radio Wrangler's Avatar
 
Join Date: Mar 2012
Location: Fife, Scotland, UK.
Posts: 22,800
Default Re: FET Questions

Back in ye olde days receivers used, as you've already said, antennae of almost any impedance lightly coupled into a first tuned circuit, with an antenna tune capacitor to allow a fine twiddle of the tracking of the resonator. This well suits the high input impedances of valve RF amplifier stages. On the whole it worked well enough.

Since then, HF radio in homes has changed. Off the peg transceivers have replaced war surplus receivers with homebrew transmitters. Controlled-impedance antennae and coax cable have become the most common arrangement to keep the transceivers happy on transmit. So amateur-owned receivers have joined the professional ones with 50 Ohm antenna input circuits, and if you use a different antenna, then an external ATU is needed.

Inside the transceivers, the trend was away from gang-tuned front ends and towards switched arrays of bandpass filters. There was no-longer a resonator tuned to the signal frequency. There was no-longer a convenient impedance transformation for the gate of an FET or the grid of a valve. This has opened up a vacancy where a bipolar device seems to be the best fit. Transistors intended for a few watts work well, especially in feedback circuits. Similar RF MOSFETS in the few watts region are also possibilities (VMP4 was good). So the receiver uses a QRP transmitting device as its RF amplifier. If it has one at all.

Recognition that intermodulation was the main bugbear menat the evolution of a number of receivers with no gain before the mixer. Probably the G3PDM design which can be found in RSGB handbooks of around 1970 vintage was one of the early appearances of such concepts in amateur circles. Large signal handling was the flaw in the RA17/117 and the RA1217 with the introduction of transistors into low power wideband circuits was worse not better. Racal found this out and the result was the RA1772.

For over 30 years we've been able to do designs with good intermodulation performance, and that has shown up phase noise as the next battleground. Sosin knew about this in the sixties and put measures into his H2900 to improve it. Racal knew too. There is a crystal filter in the LO feed which is inserted to stop climbing phase noise spoiling the sensitivity at LF/VLF.

It's an interesting field. I have an RA1217 in my collection, and though it's easily overloaded it's a cute receiver and rather well made. It's also their last Wadley system. I'd like to get an RA1772 to keep it company. I also have an RA1792 which is the result of a cost-cutting exercise. Definitely built down to a price.

I had an H2900 prototype, but with thumbwheel switch tuning and a slow-settling synthesiser I was never going to use it, so I passed it on. Similarly, the GEC RC410 was a disappointing performer and I didn't like it enough to keep it. I enjoyed my RA117, I'd bought it new, unused from storage and it shone inside and out, but in the end I wanted space in the shack for operating, so I replaced it with an Icom 765 24 years ago.

It isn't just a hobby, it's the day-job as well.

David
__________________
Can't afford the volcanic island yet, but the plans for my monorail and the goons' uniforms are done
Radio Wrangler is offline  
Old 26th Jun 2014, 6:09 am   #93
Synchrodyne
Nonode
 
Join Date: Jan 2009
Location: Papamoa Beach, Bay of Plenty, New Zealand
Posts: 2,943
Default Re: FET Questions

Quote:
Originally Posted by Radio Wrangler View Post
Another quirk is that Gallium Arsenide devices have worse noise figures than can be achieved with silicon devices below about 300-500MHz.
The Agilent Noise Figure Analysers have one of each type as a preamp, and switch between them at 500MHz
The same seems to have been true of the AOR AR2002 VHF-UHF receiver of the second half of the 1980s, a contemporary of the ICOM R7000. It had NEC IC RF amplifiers, an MC-5800 for the 25 to 550 MHz range, and an MC-5805 for the 800 to 1300 MHz range.

I’d guess that the MC-5800 was silicon-based, and that the MC-5805 was GaAs based, although I cannot find information to confirm that. Also, I think that back then, GaAs was still synonymous with fets, thus making the MC-5805 pertinent here.

Cheers,
Synchrodyne is offline  
Old 23rd Aug 2014, 11:17 pm   #94
Synchrodyne
Nonode
 
Join Date: Jan 2009
Location: Papamoa Beach, Bay of Plenty, New Zealand
Posts: 2,943
Default Re: FET Questions

Back in post #40 I mentioned the jfet/bipolar cascode that was used as an RF amplifier in some MF AM receiving equipment, including car radios and hi-fi tuners.

I have recently discovered that GM held the patent for this combination, US4277757 A, filed 1979 December 05. Presumably the development work was done by GM’s Delco division, hence its use in Delco car radios.

A quote from the “Background” section of the patent provides a good overview.

“This invention relates to radio frequency amplifiers and particularly to those utilizing automatic gain control. Such amplifiers are commonly used in the initial amplification of a received radio frequency signal, particularly in AM receivers.

“An RF amplifier which has been utilized in a significant number of radio receivers since its introduction is the dual gate MOSFET, in which the RF signal may be applied to the lower gate and a gain control voltage applied to the upper gate. The gain controlled output is obtained from the drain of the device. However, although useful, the dual gate MOSFET RF amplifier has limitations in AGC range and overload characteristics.


“An improved RF amplifier which may replace a dual gate MOSFET in many circuit applications is a cascode combination of a JFET and bipolar transistor. The drain of the JFET is connected to the emitter of the bipolar transistor; the RF signal is applied to the gate of the JFET; the AGC voltage is applied to the base of the bipolar transistor; and the output is taken from the collector of the bipolar transistor. This amplifier shows superior AGC range and overload characteristics under applied AGC; and, in addition, provides simpler biasing requirement and lower noise than a dual gate MOSFET. However, the bipolar transistor of the combination requires a base current for operation which is considerably greater than the leakage gate current of the upper gate of a dual gate MOSFET. Therefore, if this cascode amplifier is to be substituted directly for a dual gate MOSFET in a circuit application, a larger current drive must be supplied for the bipolar transistor.

“In addition, it may be desirable, particularly in some varactor tuned applications, to provide a second stage of RF amplification in order to overcome losses in the tuning circuit without overloading the tuning varactor diodes.”


Basis the RCA work of the late 1960s, the dual-gate mosfet was an improvement over bipolar transistors in the RF stage of AM receivers. But evidently even better was desirable. The comment about varactor overload is interesting. I wonder if that was why in some cases a broadband input circuit was used, with a varactor-tuned bandpass pair following the cascode RF stage. Presumably AGC applied to the latter, with its wide agc range, would result in signal reduction for the largest RF inputs, so protecting the varactors from overload.

Cheers,
Attached Thumbnails
Click image for larger version

Name:	US4277757-1.jpg
Views:	1286
Size:	21.9 KB
ID:	96640  
Synchrodyne is offline  
Old 13th Sep 2014, 1:05 am   #95
Synchrodyne
Nonode
 
Join Date: Jan 2009
Location: Papamoa Beach, Bay of Plenty, New Zealand
Posts: 2,943
Default Re: FET Questions

As previously noted in this thread, it is a fair generalization in respect of hi-fi FM tuners to say that from the late 1960s, FETs became dominant in front ends, and ICs became dominant in IF strips.

Early IC IF strip implementations typically used two or more ICs, and embraced the whole IF section from front end to demodulator. Example combinations were 4 x µA703, CA3005 (later CA3028) + CA3014; CA3012 + CA3014; CA3028 (or similar) + TA661 (or similar), MC1350 + MC1357, and so on.

Perhaps oddly though, following the arrival of the CA3089 in 1971 and the subsequent widespread use of it and its progeny, one could say that whilst the main IF strip was integrated, what might be called the pre-IF section, embodying selectivity and any additional gain required, was left very much to individual designer choice. Therein could be found FETs, discrete bipolar devices, ICs and admixtures thereof. The attached compilation of excerpts about FM IF strips from Ambit Catalogues #1, 2 and 3 are illustrative. So a reasonable inference is that satisfactory results were obtainable with a range of technologies, and unlike the front end and main IF cases, none had a clear lead. Interestingly, the RCA application notes for the CA3089 and CA3189 both show discrete bipolar pre-IF stages, which is perhaps a little surprising as from 1966s RCA was advocating all-IC based FM IF strips, partly predicated on the superior limiting qualities of integrated differential amplifiers.

In the early 1980s or thereabouts, the National LM1865 IC effectively integrated the pre-IF stage with the “CA3089” functions in a single IC, which also provided RF agc that was either wideband or narrow band according to incoming signal levels and distribution. But I don’t think that this established a new industry norm. Thus in the 1980s could be found FM tuners that used FETs in their pre-IF sections.

Cheers,
Attached Files
File Type: pdf Ambit #1-3 FM IF Circuits.pdf (1.60 MB, 725 views)
Synchrodyne is offline  
Old 28th Nov 2014, 4:28 am   #96
Synchrodyne
Nonode
 
Join Date: Jan 2009
Location: Papamoa Beach, Bay of Plenty, New Zealand
Posts: 2,943
Default Re: FET Questions

My reference to the “late 1960s” in my immediately previous post needs amending to “mid-late 1960s”, basis early examples that I have come across in searching for additional information on this topic.

Very early – perhaps even first - to use FETs in its FM tuners was H.H. Scott. One of Scott’s designs was described in “Electronics World” (EW) 1965 December. In this case it was a 3-gang front end that used a pair of jfets as a shunt-cascode RF amplifier, followed by a common gate mixer with oscillator injection at the gate. Then by mid-1966, Scott was advertising its use of FETs in the AM circuitry of its FM-AM hi-fi receivers In 1967 came the use of ICs (in this case the Fairchild µA703) in FM IF strips, as recorded in EW 1967 April.

Another early adopter was Heath, in its Heathkit AR-15 hi-fi receiver. This combined a jfet-based FM front end with an IC-based (RCA CA3012) FM IF strip. It was described in EW 1967 January. It was claimed to be the first use of an IC in the hi-fi component field. That would have made it the first “fetic” consumer FM item.

Fisher introduced a jfet-based FM front end in 1966, in this case also using a PIN-diode RF attenuator, controlled by an agc loop. It claimed a 120 dB dynamic range, although how this was measured was not stated in the associated article, in Electronics 1966 September 22.

So, jfets arrived in consumer FM (and AM) equipment in 1965-66.

Cheers,
Synchrodyne is offline  
Old 6th Dec 2014, 10:40 pm   #97
Synchrodyne
Nonode
 
Join Date: Jan 2009
Location: Papamoa Beach, Bay of Plenty, New Zealand
Posts: 2,943
Default Re: FET Questions

As mentioned in the posting above, Heathkit (USA) was an early adopter of the jfet for FM tuner front ends, and also of ICs for FM IF strips. Discovering that led to a little research to see what it might have done in respect of mosfets in the early days.

It turns out that Heathkit was also a relatively early adopter of the dual-gate mosfet. It published an article in Electronics World for 1970 March comparing communications receiver front end devices in respect of cross-modulation and intermodulation performance. The conclusion was that over the agc range, dual-gate mosfets were better than the other devices tested, namely jfets, (small signal) discrete bipolar and IC bipolar. Fairly clearly, this comparison was limited to the kinds of devices that could be used in the relatively modestly priced consumer-type equipment that Heath made, so did not include the more esoteric kinds that might be used in professional equipment. Thus for example, the 2N3866 bipolar RF power transistor was probably out-of-bounds for this kind of equipment. (Also, a high-dissipation transistor – compared with small signal devices – might not have been too welcome in relatively compact, non-compartmentalized consumer equipment insofar as temperature rise and its effects on oscillators was concerned.)

This article actually followed the release of the Heathkit GR-78 communications receiver. This was advertised at least as early as 1969 October and described in an article in Popular Electronics (PE) 1970 June. It was relatively simple, single-conversion to 18 MHz and double-conversion over the 18 to 30 MHz range. It included three dual-gate mosfets in its front end, of the ubiquitous RCA 40673 type. One was the RF amplifier, the second was the 1st mixer, and the third was the 1st IF amplifier on the lower bands, and the 2nd mixer on the highest band. The rest of the IF amplifier, following the block filter, was discrete bipolar. The local oscillator used a single-gate mosfet. Another 40673 dual-gate mosfet was used as a product demodulator.

The PE article on the GR-78 noted that the mosfets were of the protected gate type. The inference is that Heath waited until such were available before using mosfets in a production model. If so, then it was fairly quick off the mark, as 1969 was probably the first year in which gate-protected mosfets were available in production quantities. The EW article on front end devices showed the mosfet tested as being the RCA 40603. This was an earlier type without gate protection.

The SB-303 amateur band HF receiver was advertised from 1971 February. This was a double-conversion unit that was more-or-less a solid-state version of the earlier valve-based SB-301. In the SB-303, dual-gate mosfets were used in the RF amplifier, 1st and 2nd mixer and 1st IF stages. The 2nd and 3rd IF stages were discrete bipolar.

In both cases it appears that starting at the “front”, mosfets were used for as far along the RF/IF chain as they offered performance advantages that were justified by their higher unit cost as compared with bipolar devices, correlated with the receiver price points. In the case of the GR-78, the dual-gate mosfet made for a simple product demodulator in a receiver whose overall level probably did not justify the more elaborate four-diode unit as used in the SB-303.

The GR-78 was a contemporary of the Drake SPR-4 HF receiver, which seems to have been advertised from 1969 November. The SPR-4 was at a noticeably higher price and performance point though, and employed fets throughout the full RF and IF signal path.

Heath also used mosfets in AM (MF) receiving equipment. The AR-29 hi-fi AM-FM receiver, advertised from 1969 December, was a less powerful companion to the previously mentioned AR-15. It had a high-quality AM section that used 40673 dual-gate mosfets in the RF and mixer stages, along with a jfet oscillator. The two-stage IF strip was discrete bipolar.

The FM side though looked more like a scaled back version of that in the AR-15. There was a jfet 1st RF stage followed by bipolar 2nd RF and mixer stages. That looks to have been a case of using a single jfet to best advantage. The IF strip retained the CA3012 IC pair as used in the AR-15, but also included a µA703 IC pre-amplifier ahead of an LC block filter, which was new. (The µA703 with CA3012 may well have been an unusual admixture) It also had an IC-based (MC1305) stereo decoder. I am not sure when Heath adopted mosfets for its FM front ends. The AR-1500, successor to the AR-15, was advertised from 1971 November. It had adopted the AR-29 type AM circuitry and FM IF and decoder circuitry. The FM front end was said to have been improved; I suspect that it was based upon dual-gate mosfets, but I have not found confirmation of this. The Radio-Electronics review (1971 December issue) included an interesting comment: “Unlike most solid state FM receivers tested in this location (including the AJ-15), there was no difficulty with overloading or cross modulation any of the stronger stations located on the Empire State Building. Apparently Heath has overcome this problem very nicely with the AR-1500.” The AJ-15 was the tuner-only derivative of the AR-15. So there does seem to have been a non-trivial step-up in performance from the AR-15 to the AR-1500. A change from jfets to mosfets might have been part of that, as might have been additional tuned circuits ahead of the mixer.

The AR-29 was a contemporary of the Scott 386, advertised from 1969 October, which also had a high quality AM section. This combined the jfet AM front end from 1966 with an IF strip that include an L-C block filter, then a jfet amplifier followed by an IC (MC1350) amplifier. (This circuitry was described in Electronics World 1970 June.)

And with its TV receivers, Heath introduced a dual-gate mosfet RF amplifier in the VHF tuner of its GR-270 and 370 models advertised from 1970 September. The mixer was bipolar, as was the IF strip. Later, with the GR-2000 advertised from 1974 April, the VHF tuner had a dual-gate mosfet mixer as well as an RF amplifier, and an IC-based IF strip with block L-C filter.

The Heath case is probably reasonably illustrative of the initial and judicious incursion of fets (and ICs) into consumer products, particularly interesting because it involved quite a wide range of product types, each with its own trade-off point as to just how far down the signal chain fets could be justified.

Cheers,
Synchrodyne is offline  
Old 17th Dec 2014, 3:47 am   #98
Synchrodyne
Nonode
 
Join Date: Jan 2009
Location: Papamoa Beach, Bay of Plenty, New Zealand
Posts: 2,943
Default Re: FET Questions

By way of a little more on the FET timeline, apparently RCA released the 3N98 HF single-gate mosfet in 1964. It was followed by the 3N128 VHF device. The big step – according to RCA – was its development of the dual-gate mosfet, first offered in 1965 as the 3N140. Minimizing feedback capacitance was a primary goal, and better cross-modulation performance under agc was a benefit. RCA conceded though that the development of gate-protected devices was necessary for general market acceptance. In 1971 RCA was working on UHF (1 GHz) dual-gate mosfets. The then-current devices covered VHF and low UHF.

As to the use of FETs in RCA consumer products, the CTC-49 TV chassis was new in 1971. Its main feature was its extensive use of ICs, including the CA3068 vision IF amplifier and demodulator. (It has been said that in this case, the division handling consumer IC development rather pressed the consumer products division to use its new range of TV ICs. The CA3068 was significant because it was the first IC to contain a complete vision IF strip and demodulator (diode-type), something which others in the industry were sceptical about because of potential IF feedback problems.) The VHF tuner of the CTC-49 was said to have followed previous practice, being of the 4-gang type with a dual-gate mosfet RF amplifier and a cascode (bipolar) mixer. So evidently RCA had adopted the dual-gate mosfet in its TV receiver VHF tuners some time before 1971.

RCA’s next major TV chassis iteration was with the CTC-74 and CTC-81 in 1975. These had a VHF tuner that included a dual-gate mosfet mixer as well as a dual-gate mosfet RF amplifier. It was claimed that the complete mosfet front end allowed greater RF agc delay and so an improved signal-to-noise ratio without causing overload that would result in cross-modulation. Or in other words the mosfet mixer could handle more incoming signal than the bipolar cascode. That RCA converted to FET-based VHF tuners in two steps suggests that in the early days, device cost was an inhibiting factor. But a significant change was that the vision IF strip in these receivers used mosfets (3 of), not an IC, with the claim that this also contributed to a better signal-to-noise ratio.

Some corroboration of RCA’s move from an IC-based to a mosfet-based TV IF strip is provided by the fact that at around the same time, Motorola proposed a low-noise IF strip that comprised a pair of 3N203 dual-gate mosfets, each with agc, and an MC1331 demodulator IC. That was against the background where Motorola had been the industry leader in respect of integrating TV IF systems, initially with its MC1550 (which was not well accepted for the role) and then with its MC1350 and MC1352, which were widely used. Motorola’s MC1349 (whose release date I have not been able to pin down) was an improved version of the MC1350 with a better noise vs. agc curve, so perhaps the semiconductor folks were inspired to do better when they saw the mosfet-based IF strip.

Cheers,
Synchrodyne is offline  
Old 9th Jan 2015, 11:24 pm   #99
Synchrodyne
Nonode
 
Join Date: Jan 2009
Location: Papamoa Beach, Bay of Plenty, New Zealand
Posts: 2,943
Default Re: FET Questions

As noted earlier in this thread, the Japanese makers of amateur/domestic class HF receivers (of the non-portable type) adopted substantially fet or all-fet based RF and IF signal paths from the early 1970s onwards, if not before.

Dual-gate mosfets featured prominently, but an interesting facet was the use of push-pull jfets (and sometimes push-pull mosfets) as 1st mixers (and sometimes 2nd and 3rd mixers) from about the mid-1970s. Actually, a very early user of this configuration was Drake, who used a push-pull jfet 1st mixer in its SPR-4 receiver of 1969.

Evidently this configuration, which was still a nominal quadratic mixer not a switched mixer, must have provided what was considered to be a worthwhile decrement in spurious cubic (and higher order) intermodulation products, and with the presumably not too precise balancing that could reasonably be obtained in mass production.

Being quadratic mixers, the push-pull fets still required sufficient RF selectivity ahead of them in order to avoid the generation of unwanted in-band 2nd order products. At least for receivers with lower 1st IFs, that was anyway usually covered by the RF selectivity required for good image rejection.

The previously-mentioned Drake SPR-4 receiver was to some extent a solid-state derivative of the earlier valved R-4 and SW-4 receivers, and had broadly similar topology. Again as noted earlier in the thread, the use of dual-gate mosfets in particular allowed matching of, but not materially exceeding valve performance in the RF, mixer and IF stages. The empirical evidence is that the push-pull mixer was evidently something more easily done in solid state than with valves, at least with this class of receiver, so it possibly represented an achievable “gain” for the solid-state case, as compared with typical valve receiver practice.

Most examples of push-pull fet mixers that I have found appear to have in common that they were used in receivers with separate RF (i.e. preselector) tuning or with switched sub-octave RF filters. As some of these would have broader RF tuning than was typically the case for receivers with closely tracked RF tuning, cubic mixer products would have been more of a potential problem. So to some extent the push-pull performance increment may have been used to cover this.

Mostly jfets seem to have been used in push-pull mixers. Kenwood used push-pull dual-gate mosfets from the R-820 through the R-2000, but switched to push-pull jfets for the R-5000. This is an application where a dual-gate mosfet advantage over jfets is not readily apparent, given that the push-pull layout allowed separate signal and mixer input ports, even with single-gate devices. Kenwood had a prior history with single dual-gate mosfet mixers, so perhaps familiarity was a reason for staying with this kind of device in its early push-pull mixer efforts. Unusually, it also used a dual-gate mosfet (in source follower mode, I think) as a buffer stage between the RF amplifier (another dual-gate mosfet with single-tuned input and bandpass-tuned output) and the 1st mixer of its R-820 receiver.

Less frequently, but notably in some of the JRC HF receivers, push-pull jfet RF amplifiers were found, presumably with the same objective of reducing cubic products.

Of course, and as already recorded upthread, whilst all of this was happening at one level, at a significantly higher level, professional receivers had moved beyond the world of small-signal fets to high-level switching mixers and with commensurate RF amplifiers using bipolar power devices in feedback circuits. This had also applied to one or two examples at the upper end of the amateur/domestic field, such as the Drake R-7 and the McKay Dymek DR22 et seq. Some aspects of professional practice found their way into the Japanese HF group. ICOM used DBM 1st mixers in its R-70 and R-71 receivers, and Yaesu graduated to the DBM late on (in the period under consideration) with its FRG-8800, having previously used push-pull jfets. All of these, though, had RF amplifiers based upon small-signal jfets in push-pull (in fact 2SK125 pairs in each case).

Looking back against that background, The Eddystone EC958 series HF receiver of 1968-69, being strictly a professional unit, must have been quite unusual in that it had a signal path based largely upon small-signal fets, including three dual-gate mosfet mixers in succession. Although its topology was soon outmoded, and perhaps even on the edge of obsolescence when it was designed, it stayed in production for quite some time and seems to have been well-regarded at its price-performance point. I assume that it would have been noticeably cheaper than say a Racal RA1772, but I do not know that for sure. Sutton of Eddystone provided something of a post facto rationalization for the EC958 in the second part of his paper “Eddystone Radio: A Short History of Receiver Development from 1965-1995”, where the primary topic was dynamic range and the development of the Eddystone 1650. Essentially the good performance of the EC958 despite its use of active devices of limited dynamic range was attributed to its use of tracking RF selectivity ahead of the active devices.

Mosfets also trickled down to entry-level HF receivers. Below the Heathkit GR-78 was the SW-717 of 1971, which was a very simple MF/HF unit with but a two-gang front end, and so no RF amplifier. It had as its input stage a 40673 dual-gate mosfet mixer, fed by a jfet oscillator. The IF strip, following a block ceramic filter, was bipolar, but included an additional stage for amplified agc. The SW-717 followed a sequence of similarly very simple valve receivers, the last of which was the GR-64, with a 12BE6 self-oscillating mixer as its input stage. I imagine that the SW-717, with its separate oscillator, would have offered somewhat improved performance. So even down in the “bargain basement”, some effort was made to get above the performance delivered by the typical bipolar self-oscillating level mixer input stage.

Cheers,
Synchrodyne is offline  
Old 15th Jan 2015, 7:41 am   #100
Synchrodyne
Nonode
 
Join Date: Jan 2009
Location: Papamoa Beach, Bay of Plenty, New Zealand
Posts: 2,943
Default Re: FET Questions

In my previous post I said:
Quote:
Originally Posted by Synchrodyne View Post
Some aspects of professional practice found their way into the Japanese HF group. ICOM used DBM 1st mixers in its R-70 and R-71 receivers, and Yaesu graduated to the DBM late on (in the period under consideration) with its FRG-8800, having previously used push-pull jfets. All of these, though, had RF amplifiers based upon small-signal jfets in push-pull (in fact 2SK125 pairs in each case).
That was not fully correct. Forum member Orbanp1 has kindly let me know that the ICOM R-71 in fact had a balanced (push-pull) jfet mixer using a pair of 2SK125. Orbanp1 also pointed me to this excellent site: http://www.qsl.net/va3iul/Low_Noise_...or_SSB_TRX.htm

Thus one may see that, to quote Radio Wrangler’s quasi-eponymous comment from the Quasi-Synchronous demodulation thread, there are “horses for courses”, and for amateur/domestic receivers that are not usually fed from large antenna farms, there are valid reasons for choosing balanced push-pull fet mixers over DBMs.

I mentioned in my previous post that the Eddystone EC958 had three dual-gate mosfet mixers in succession, these following a cascode jfet/single-gate mosfet RF stage. A reasonable inference is that Eddystone saw no advantage, in situ, of moving up to push-pull fet mixers. A factor here would appear to be that the EC958 had “tight” selectivity from throughout the signal path. The RF input was of the bandpass type (on the HF bands at least) with a single-tuned interstage. So notionally it had a four-gang front end. As well as tracked tuning, it also had an RF peaking control. And the tunable 1st IF had a triple tuned bandpass filter between the 1st and 2nd mixers.

More on early fet applications, I happened upon the attached items which are early 1969 advertisements for Vanguard Laboratories amateur ancillary equipment. These are self-explanatory. Vanguard extols the virtues of dual-gate mosfets as compared with bipolar devices, and also as compared with jfets and singe-gate mosfets. The RF pre-amplifier was available for frequencies up to 450 MHz. Some of the early dual-gate mosfets, such as the well-known RCA 40673, were rated for use up to 500 MHz, so providing low-UHF coverage. UHF coverage towards 1 GHz, as required for UHF TV front ends, was still in the future though, as previously recorded.

In the hi-fi field, the American McIntosh MR 73 tuner dated from 1969 and might have been its first solid-state design. Not unexpectedly the FM front-end was fet-based. It had a cascode RF amplifier using a pair of two jfets, and a single jfet mixer. It was four-gang, with a bandpass input and a single interstage. Bandpass inputs seemed to feature on some early solid-state FM tuners, but later on single-tuned inputs with bandpass interstages seemed to be the norm, perhaps because there was a noise advantage from a less tightly tuned input. Anyway, the FM front end spoke of earlyish FM fet practice, before dual-gate mosfets largely superseded jfets. On the other hand the AM section was very highly specified, with two dual-gate mosfets, one each in the RF amplifier and mixer positions, and with two single-gate mosfets in the IF strip.

In the MR 74 of 1972, McIntosh had changed the FM front end to have a two-stage RF amplifier using dual-gate mosfets in each position, followed by a jfet mixer. It was still four-gang, with single-tuned input and interstages. But the AM stage had regressed to being a simple and all-bipolar circuit. At least in part this apparently backward move may be explained by looking at the periodic peaks and troughs in the way AM radio was treated in American hi-fi equipment over the years.

Fets were found in consumer-type active aerials as well. Since I had looked out the McKay Dymek materials in respect of a contribution to the “Indoor AM Aerial” thread (https://www.vintage-radio.net/forum/...d.php?t=112116), I also looked at the circuitry. The DA5 (MF) and DA7 (LF/MF) indoor tuned, screened ferrite loops had a 40822 dual-gate mosfet following the loop tuning capacitor, this feeding a TIS74 jfet source-follower which provided the output.

The DA100 (LF-to-HF) outdoor active whip had in its head-end a jfet input stage amplifier feeding a jfet source-follower, which in turn drove a 2N5109 (or 2N3866) bipolar RF power transistor with feedback from the latter’s collector to the input jfet source.

Both aerials were mid-1970s products. They seem to have fitted the mould, so as to speak. The empirical evidence is that the extremely high impedance dual-gate mosfet was typically used in narrowband circuits, with tuned inputs and usually with tuned outputs. In the DA5/DA7, the dual-gate mosfet had a tuned input, but an untuned output, the jfet source follower presumably providing an appropriate high-impedance load. On the other hand jfets were used in both tuned narrowband and untuned wideband circuits. So in the wideband DA100, the jfet provided a very high impedance load for whip, whilst the bipolar power transistor with feedback provided wideband RF amplification with minimal cubic (and other) spurii production.

Cheers,
Attached Thumbnails
Click image for larger version

Name:	AR73 196902 p.59 Vanguard Labs.jpg
Views:	562
Size:	72.3 KB
ID:	102714   Click image for larger version

Name:	AR73 196902 p.27 Vanguard Labs.jpg
Views:	536
Size:	73.1 KB
ID:	102715   Click image for larger version

Name:	McIntosh MR 73 FM-AM Tuner Section.jpg
Views:	3018
Size:	64.9 KB
ID:	102716  
Synchrodyne is offline  
Closed Thread

Thread Tools



All times are GMT +1. The time now is 12:22 am.


All information and advice on this forum is subject to the WARNING AND DISCLAIMER located at https://www.vintage-radio.net/rules.html.
Failure to heed this warning may result in death or serious injury to yourself and/or others.


Powered by vBulletin®
Copyright ©2000 - 2024, vBulletin Solutions, Inc.
Copyright ©2002 - 2023, Paul Stenning.