Building A VHF Amplifier For Valve Sets

[G0HZU_JMR]

I agree with Paul, the main threat to stability is an unterminated output because the device gain will be huge without a regular 75 ohm output load. Having a 390R shunting the output terminals (all of the time) might make it more tolerant of there being no output load and it wouldn't dampen the gain very much. The initial bias resistors looked about right to me so I'd advise putting it back to original but with a 75R termination resistor at the output. Then check the dc operating point. If it is OK, there should be something like 4 to 6V dc at the collector.

[G0HZU_JMR]

I do not think that a common emitter amplifier will have the gain or impedance matching suited to the intended application.

The input match will probably be quite poor but this could be improved with a suitable L match at the input without degrading the noise figure very much. I don't think gain will be a problem, I simulated the amplifier using a model for a similar BJT and it produced about 20dB gain across the 88-108MHz FM band with a noise figure just under 2dB. It might not be quite this good with a BF199 but I doubt it will be that far off.

[G0HZU_JMR]

Running at just a few mA Ic with the collector looking straight into a 75 ohm output will mean that this amplifier won't have very good signal handling compared to a modern alternative so some care would be needed if you use it in an area with large signals nearby. A lot also depends on the (noise and signal handling) performance of the receiver used after it. But for now I think the aim is to get this basic little amplifier working as it should.

[paulsherwin]

Originally Posted by G0HZU_JMR

The input match will probably be quite poor but this could be improved with a suitable L match at the input without degrading the noise figure very much. I don't think gain will be a problem, I simulated the amplifier using a model for a similar BJT and it produced about 20dB gain across the 88-108MHz FM band with a noise figure just under 2dB. It might not be quite this good with a BF199 but I doubt it will be that far off.

Simple single transistor common emitter broadband VHF/UHF preamps normally give in the region of 12-20dB gain, so your simulation sounds plausible.

[Chris55000]

Hi!

If you're looking for something you can try from plans, you'll get an idea from "Practical Television Circuits", by R.E.F. Strest, 1968 edition – you can download this from the Internet Archive.

Have you thought about using a cascode circuit? You can make one up by using a second transistor of the same type, you feed the signal input into the emitter circuit of the BF199, connected in common–base with a 470 ohm emitter resistor to 0V and a series coupling capacitor of about 100p from the signal input.

The base of the common–base BF199 can go to the slider of a 4k7 preset connected to a 22k resistance between the top (cw) end of the pot and the +12V supply, the bottom of the 4k7 pot going to 0V. Decouple the base of the first BF199 to 0V with a 1000p low loss disc ceramic capacitor.

The collector of the first BF199 connects directly to the emitter of a second BF199 used in common–emitter mode, with no components connected to 1st BF199 collector/2nd BF199 emitter junction.

Make a top transistor base potential divider of two resistors across the +12V supply, use a 4k7 for the +12V resistor and a 10k resistor for the bottom 0V resistor. Decouple the junction of the two resistors to 0V with a second low–loss 1000p disc ceramic cap to 0V and connect the base of the "top" BF199 to the junction of the two resistors at the mid–point of the base divider and the 1000p cap.

That leaves only the collector of the "top" BF199, connect this to +12V with a 1k resistor, then use a second 100p capacitor to take the output signal from the "top" BF199 collector.

The 4k7 preset is adjusted on a 12V supply to give between 0.2 to 1.5V across the 470 ohm emitter return resistor for the "bottom" common–base BF199.

A collector voltage of about 7.5 to 8.5V on the "top" BF199 is a useful suggested working point to begin with.

Make all connections as short and direct as possible on a piece of copper–clad board, returning all the 0V points described above to the copper foil, with all non–earthed connections spaced about 1–2 cm above the copper.

The amplifier can be tested using a standard v.h.f. broadcast receiver tuned to a weak local f.m. station.

Chris Williams

[Herald1360]

Sounds like a useful circuit but it's not a cascode nor is the top transistor operating in common emitter. It looks like a directly coupled cascade of two common base stages. Does it have any significant advantage over a simple common base single stage amplifier?

A cascode is a common emitter driving a common base arrangement. (Originally common cathode driving common grid.)

[G0HZU_JMR]

I marked up the schematic with the component values as below and I would suggest the first test would be to temporarily remove the 100k resistor R2 and measure the voltage at TP1 in the diagram below. TP1 is the collector of the BF199 and the voltage here should rise to about 12V with R2 removed. There is close to zero chance of the circuit oscillating when R2 is removed and the voltage at TP1 should go up to be the same as the supply voltage at about 12V.

The reason I'm suggesting this basic test is because of the strange voltages measured in post #6.

I've double checked the connections for the transistor and it is in correctly.

B: 180mV

C: 1V

E: 148mV

Ic: 9mA

I suppose this could be due to instability but it could also be because the BF199 is faulty or maybe it isn't really a BF199.

If the BF199 is actually genuine and the pinouts are correct then it could be that the device has failed and there is some kind of breakdown failure inside the device. Do you have another BF199 to try?

If the device isn't a genuine BF199 and pinouts are different and the base of the transistor was somehow fitted to TP1 and the collector was fitted to the R1 and R2 junction then I think you would see quite similar voltages to those strange readings above.

That's why I think it's worth doing the test at TP1 with R2 removed. A healthy transistor will show 12V here.

If it passes this test then I'd suggest fitting a 75R termination across the output from C2 OUT to the 0V Ground connection and then refit the 100k resistor R2. If the circuit isn't oscillating then it should be possible to adjust the value of R2 such that you see about 5V at TP1. This will mean that (with a 12V power supply) there is about 5mA collector current.

Don't try too hard to get exactly 5V at TP1 as anything from about 3.5V to 8V would be fine for some initial health checks.

[Chris55000]

Hi!

Sounds like a useful circuit but it's not a cascode nor is the top transistor operating in common emitter. It looks like a directly coupled cascade of two common base stages. Does it have any significant advantage over a simple common base single stage amplifier?

A cascode is a common emitter driving a common base arrangement. (Originally common cathode driving common grid.

Thanks for pointing out the mistake I made in my description – I had mentally put together a common base V.H.F. r.f. amp, as in an f.m. broadcast receiver, with a common–base top stage – the common–base output half I had intended but I described the input wrong!

The correct way is to use a 22k resistor and 4k7 bias pot as before with the wiper of the bias pot connected to the "bottom" BF199 base via an additional 10k series resistor, with the signal input connected direct to base via a 100p capacitor. The emitter of the "bottom" BF199 goes to earth via 470 ohm and a 1000p in parallel with no other connections, and the collector of the "bottom" BF199 going directly to the emitter of the "top" BF199 as I described it before!

Me thinks I ought to draw it out first before describing it in words!..

Chris Williams

[G0HZU_JMR]

If it helps, I've posted up my earlier simulation below. I don't have a BF199 here or a decent model for it but I do have a reasonable model for the classic 2SC2026 in a similar TO-92 package.

This transistor is a little faster than the BF199 and it is suitable for VHF and UHF preamps. I chose it because it boasts a sub 3dB noise figure at collector currents of a few mA and the noise figure doesn't degrade much even with 6 or 7mA collector current. So this transistor would be close to ideal for this design as long as a tight PCB layout was used.

I simulated it for 75R test ports with the equivalent of a tight PCB layout using conventional leaded parts on a groundplane. So there is a bit of self capacitance and several nanohenries of lead inductance iin the resistor models.

It shows about 20dB gain and a noise figure just below 2dB. However, the two tone plot shows that the output intercept is only about +10dBm and this is really quite puny for an amplifier with 20dB gain. This would be OK in a rural location 30 years ago but would probably suffer a lot of intermodulation distortion in a modern urban environment with lots of big signals across the various VHF bands.

I also tried modelling slower devices and the gain and noise figure degrades. The gain dips to about 15dB and the noise figure varies from 2.5dB to 3.5dB depending on the transistor model I choose. I'd expect the BF199 to fall into this category. For these devices the layout would need to change over to SMD I think in order to minimise the inductance in the R and C components at the emitter. This would boost the gain slightly.

The red trace in the plot below is the gain (LHS scale) and this shows about 20dB gain.

The green trace is the noise figure in dB (RHS scale) and this shows just under 2dB noise figure.

The input match is really poor and would benefit from a classic L match and this could be designed to give a reasonable match across 88-108MHz without much change in the noise figure.

[G0HZU_JMR]

I had a rummage to try and find a transistor that has similar specs to the BF199 and found a 2SC1674L. See the attached datasheet.

These transistors are typically used in VHF RF amplifiers and they sometimes get used in CB receive circuits as well. The Ft, dc current gain and noise figure specs are similar to the BF199.

The datasheet states that the 2SC1674L typically has a 3dB noise figure at 100MHz when driven from a 50R source. However, the datasheet graphs show that for 5-6mA collector current and a 50R source impedance the noise figure is typically just over 3dB.

I quickly tacked together a 2SC1674L amplifier based on the circuit described in this thread and tested it for gain and noise figure in a 50R test system. It managed just over 17dB gain at 88MHz and just under 17dB gain at 108MHz. The noise figure was flat at about 3.7dB across the band. This is a fairly good result but I suspect that the BF199 will manage to deliver a slightly better noise figure especially in a 75R system.

I'm measuring the noise figure and gain with a decent lab spectrum analyser using an Agilent noise source. This system works very well as the analyser has the required internal LNA option fitted to allow direct measurement of noise figure at these frequencies.

So hopefully this will give a warm feeling that the amplifier circuit should work OK but I'd recommend using a tight layout on a groundplane. It only takes a few minutes to tack the circuit together and I would recommend keeping the component legs quite short. Be prepared for overload problems though. I don't think this amplifier will perform well where there are lots of local/large VHF signals on the band.

[paulsherwin]

The OP is in Newcastle though, not central London, and is using a very modest aerial.

The BF199 was the standard Mullard/Philips VHF transistor of the era. It succeeded the lockfit BF194/195 and is often chosen as a a replacement for them. I don't use it, but only on cost grounds - I use MPSH10s or 2SC2999s. I don't think the precise transistor is going to make much difference in a simple circuit like this though - a BC548 or 2N3904 would probably work just as well, and might even increase stability.

[G0HZU_JMR]

To get the best from the circuit I think it's worth cherry picking the best candidates for the transistor based on the datasheet. The ideal part would maintain a low noise figure and good performance across collector currents of 2-7mA and it would have to have quite low internal capacitance to minimise Miller effect.

Probably the best 'sensible' choice would be the 2SC2026 although these are quite exotic and expensive these days (and probably obsolete and faked a lot too).

If SMD is allowed them the cheapest (and easiest to buy reliably) would be the BFS17 in SOT-23. This should perform quite well here. This is the basic BFS17 not the BFS17A.

The 2SC1674L isn't really an option because these parts went obsolete ages ago apart from a few modern/dodgy clones.

The MPSH10 will probably work quite well. I have quite a few of these in SMD here somewhere. I'm not sure about the noise figure performance but it will probably be in the 2.5-3.5dB range in this circuit.

Looking at the datasheet, the 2SC2999 would make me nervous because this part is really optimised for operation below about 3mA. Above this current the gain bandwidth product and noise figure can nosedive as Vce dips below 6V. Below 3mA it is a nice low noise device and will perform really well I think.

I think the classic jellybean BJTs with a typical Ft of 250MHz at 5mA will struggle in this circuit. I could try a few but I'd expect to be disappointed...

[G0HZU_JMR]

I'm not sure if OldTechFan96 has given up on the circuit or not...

As I mentioned in my first post, one of the limitations of this circuit is that the bias network is design centred for a typical BF199 with a dc current gain of about 90. If the current gain of the chosen transistor is quite different then the dc operating point can change quite a bit. If a transistor with a current gain of 200 was fitted then the dc operating point would shift a lot and the collector voltage would dip down to below 1V dc.

If the circuit is modified as below, the increased feedback will stabilise the dc operating point at around 4 to 5.8mA collector current for transistors with a dc current gain range of maybe 75 -200. This will keep the collector voltage within a sensible range for this circuit.

[paulsherwin]

You could simplify it further - R3, C4 and R5 don't really do very much and could be deleted. C5 is probably redundant too, but does no harm.

[G0HZU_JMR]

Yes, I think it could be simplified as you say.

I managed to find some KSC1674YBU transistors in my stash and I'd forgotten I'd bought these from Mouser. Sadly they don't seem to have any more stock but I tried one in the circuit and it gave about 17dB gain and a noise figure of 2.8dB. This transistor has a different pinout compared to the 2SC1674L I have here but otherwise it seems to be a very similar part.

I'll see what I can find tomorrow because I think I still have some old stock of 2SC2062 transistors. I should also have some even lower noise parts but these are overkill for a simple VHF preamp. I think anything that can deliver >14dB gain and less than 4dB noise figure would help to perk up a valve receiver but I'm really just guessing. I don't know how sensitive a typical VHF valve receiver will be.

[OldTechFan96]

I'm still here! I've been trying to get my head around all of these posts and I am wondering where I should start with regards to getting the circuit working.

I've also had a Dynatron Atlantis on the bench which I've made loads of progress with.

[paulsherwin]

Have ýou tried damping the output, as I suggested back in #19?

[G0HZU_JMR]

I'd recommend you try the test in post #27 where R2 (100k) is temporarily removed and the voltage at the collector is measured. This is a basic test of the health of the transistor.

If the BF199 tests out OK then you could try simplifying the circuit as per the circuit below. Ideally this should be built on a groundplane with short connections but you can try your original layout first to see if it works without oscillating. This circuit below includes the 75R damping resistor suggested by Paul.

[Radio Wrangler]

Originally Posted by G0HZU_JMR

I think the classic jellybean BJTs with a typical Ft of 250MHz at 5mA will struggle in this circuit. I could try a few but I'd expect to be disappointed...

I rather think so. At 100MHz it wouldn't give enough gain to significantly depress the noise figure of the following receiver. At best it would make a preamplifier that wasn't worth using.

I'd think of something better than 750MHz Ft and a noise figure in the 2dB or better region. I wouldn't bust a gut trying to get much under 2dB.

There's also the overload level to consider, if there are any strong local signals.

I'd consider one of the transistors intended for CATV distribution amplifiers... 2N5109 or 2N3866

I have some 2N3866 if anyone wants to have a go. These are TO-5 devices designed to run a moderate bit of power for linearity reasons, say 25 or so mA the Ft peaks at 50mA. They would definitely need building over a groundplane. Quite frolicsome if not tamed.

David

[OldTechFan96]

I have just tried the test suggested in post #19 and nothing changed. Not voltage change and no output The closest resistor I had to 75 ohms was 82 ohms.

I will try the suggestions in post #27 and #38