Building A VHF Amplifier For Valve Sets

[OldTechFan96]

Hello,

Today I built a VHF amplifier based around a BF199. I constructed it using panel pins and an off cut of wood. It looks like something that would have been put together at Bell Labs!

Unfortunately, it did not work first time and needs some troubleshooting.

The injected signal is not amplified at all. The first thing I checked was the voltage between the base and emitter. It was too high at 0.85V. This could be the cause of the issue.

I'll start with increasing the resistors in the potential divider R1 and R2 to get 0.6V.

[paulsherwin]

Check the leadout. The BF199 is c-e-b.

That looks like an oscillator in disguise, not a very good disguise either. Better on a sheet of tin from a can, I find mustard tins a good source of tin plated steel.

[Nymrod121]

Bread-boarding at its finest. In fact it bears a passing resemblance to a 'Philips EE8 kit' layout!
Guy

[paulsherwin]

It should actually work despite the agricultural approach to construction. The bias being a bit out shouldn't be a showstopper. Have you checked all the circuit voltages and the current?

[OldTechFan96]

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

B: 180mV

C: 1V

E: 148mV

Ic: 9mA

The construction booklet does not give any voltages or current information.

[Andrew2]

That isn't going to work very well even if you got the bias right. It needs an earth plane, a more compact layout and the supply line should be decoupled at several points along its length. It also needs connecting into circuit with coax.
I suspect that once you've got the bias right and it shows some useful gain it will show a tendency to oscillate merrily!

[G6ONEDave]

Are all the wire wraps tight to their pins and were the pins cleaned before assembly? They don't appear to be soldered. I would check the 22 Kohm from base to earth since the transistor appears to be switched fully on.

[OldTechFan96]

The circuit has been soldered. The photo was take before I soldered everything. I'll check the 22Kohm resistor.

[Skywave]

That build seems to have too much stray C and L for the intended freq. of operation. Do you have access to a signal generator, etc., to test its performance at lower frequencies?

Al.

[Radio Wrangler]

Originally Posted by OldTechFan96

B: 180m
VC: 1V
E: 148mV
Ic: 9mA

9mA doesn't seem bad as a quiescent current for such a thing.
You haven't said where those voltages were measured with respect to. It's important, but everyone takes it for granted. I'll guess 'ground'

There seems to be only 0.032 volts on the base with respect to the emitter. Now that isn't enough to even begin turning a silicon transistor on. So why is 9mA flowing? the emitter voltage is raised with respect to (assumed) ground, so the current seems real and the emitter voltage seems real.

I assume you measured the current in the lead from the power source.... if so that's decoupled so it should be a safe measurement. The emitter resistor is decoupled too, so I'll trust the emitter voltage as well.

But the base to emitter voltage is impossible for a good transistor and that current, unless something else is going on.

That circuit is probably oscillating its little head off. The base-emitter junction is self-rectifying the oscillation and producing additional bias for itself in a negative direction. This is how oscillators stabiliae at some amplitude.

This isn't a field many people have bothered with. Booka and books have been written on frequency control in oscillators, and on getting the frequency stable. Damn all has been written on their amplitude control mechanisms. This is a bit unfair, you need amplitude control in an oscillator else you wind up with circuits that don't start and circuits that hammer themselves into the end stops, distorting horribly.

The silly bias voltages are impossible in an amplifier, but quite routine in an oscillator. So they are a walloping great clue.

Builders of amplifiers of all sorts read up on amplifiers, naturally. Why would they spend time reading up on oscillators? Well, there are clues there on how to build things that don't oscillate.

This is a well-trampled pitfall. Absolutely everyone falls into it unless they do nothing constructive at all.

Seasoned circuit developers don't go into a non-working circuit with a DVM, they reach for the oscilloscope as their first measure. I very rarely use a meter, it's the scope unless there is a special need for accuracy. The scope shows oscillations, instabilities, wobbles, hum and all that stuff. It gives you either more confidence that nothing is wrong, or it gives you good clues if something is wrong. Don't worry about 'scopes are too complicated for me' They aren't. They are manageable and they make the really difficult things, the things that stump you a lot easier.

DVMs have other problems, too. Their readings can go completely loopy if there is RF around or any frequency beyond their bandwidth. There are protection diodes all overr the place to stop people blowing them up, either by high voltage lines or from static zaps. These diodes act as rectifiers and the resulting readings can cause great confusion if you don't suspect what's going on.

When you are using very fast transistors for VHF and upwards amplifiers, the instability can result in oscillations much higher in frequency than affordable oscilloscopes can handle, but analogue scope designers are aware and arrange things so that too high frequencies don't cause problems they even provide bandwidth limiting switches. This may not be the case with digitising scopes.

So with VHF/UHF oscillations we have to fall back on indirect indications. Watching what happens during some finger-poking is the usual method on low voltage low power circuits.

The style of construction obviously has to be suited to the frequency range you want the thing to work over. Less obviously, the style also has to be appropriate to the ranges of frequencies you want it to NOT oscillate at. At some point, the transistor gain has rolled off and it's not going to be able to get into trouble. You don't have to worry about this and higher, but you do have to worry about the full frequency range up to here.

Look at the Ft spec for your transistor, and note that it's a MINIMUM figure you get. So you have to make a guess how much higher the part you've bought can go.

So for any transistor worth using in a VHF preamp, you need a groundplane and some precautions.

Build a breadboard on a scrap of unused PCB material. Solder grounds straight onto the copper foil. Build your circuit as small as you can in the air above the plane. Use decoupling capacitors resistors going to ground as support pillars. If needed add megohm resistors as pillars.

You may also need to add 20 or so ohms in series with the base pin as a stopper.

After a while, you'll have come across most of the issues and developed and armoury of tricks to combat them.

David

[GMB]

I ran a simulation of this circuit, and the seemingly low base voltage is what you might expect if it was wildly oscillating, which with that layout is one of the expected outcomes.

Working correctly, my simulation gives a base voltage of about 0.82V, collector voltage about 4V and emitter voltage of about 0.12V, assuming a +12V supply.
At 100MHz is gives a voltage gain of 160.

[Terry_VK5TM]

The transistor is wired wrong?

Motorola BF199's are c,b,e, Philips are c,e,b.

[paulsherwin]

C-E-B is correct for this part. Did Motorola really break the spec?

I agree it's always worth a quick check though.

[Terry_VK5TM]

Originally Posted by Terry_VK5TM

The transistor is wired wrong?

Motorola BF199's are c,b,e, Philips are c,e,b.

Cancel that, looking at the wrong page again.

[G0HZU_JMR]

Yes, the layout isn't really suitable for a VHF amplifier like this and it could be unstable.

However, there are a few other things to be wary of.

The 18R emitter resistor is a very low value and it looks like the dc operating point depends on the dc current gain of the transistor. A typical BF199 might have a dc current gain of 90 meaning the Vce could be 5V on your circuit. If the gain is much higher than this (gain = 150?) the dc operating point will shift to an undesirable point where the collector voltage will dip down towards 1V. Even with the 'correct' current gain the operating point will vary with temperature because the emitter voltage is only about 0.1V assuming a typical BF199 is fitted.

Even when built with a tight layout on a groundplane this amplifier won't be close to being unconditionally stable unless the output is terminated in a lowish impedance. You could try fitting a 390R resistor across the output terminals as a temporary measure. Otherwise, with no output load, the act of probing it with a set of DMM leads could be enough to make it go unstable. You could try adding a series 1k resistor or a ferrite bead at the tip of the +ve DMM lead when making voltage measurements and this might help a bit.

Assuming you have a genuine BF199 the datasheet suggests that the noise figure could be less than 3dB at Vce = 10V and Ic =4mA. I'd expect the noise figure to be similar at Vce = 6V but that is just a guess.

[OldTechFan96]

Thanks for the detailed responses! Here are a few answers:

I do have an oscilloscope and a Heathkit RF signal generator to test the circuit.

Everything was measured with respect to the negative line. 12V supply.

To measure Ic, I measured the voltage across R4 and divided it by the resistance of R4.

The BF199 is a Philips part.

I would like to get this working for the learning experience, but at the same time I understand that the construction of the circuit leaves much to be desired.

Is it worth continuing or should I go back to the drawing board?

Is it worth continuing or should I go back to the drawing board?

Not quite, I would rebuild on a ground pane (bit of tin) making it as small as reasonably practical. Luckily going smaller doesn't require longer component leads!

[paulsherwin]

Before doing that, connect the output to an actual radio input coil or 75 ohm resistor to 0V, and measure the voltages again. That may be enough to stop it oscillating. While the construction is far from ideal, lashups like this can usually be made to work after a fashion.

[Restoration73]

I do not think that a common emitter amplifier will have the gain or impedance matching suited to the intended application.
A common base amplifier is more appropriate e.g. page 68 of https://www.americanradiohistory.com...01981%2006.pdf which will work fine with most npn bipolar junction transistors e.g. the types already
described above. I have built this and confirms it works well.