2N4401/2N4403 substitutes

[Martin G7MRV]

Ive got this circuit breadboarded at the moment. Its a flyback type HV generator.

Ignore the 1N914 diode - thats replaced by a string of zeners and neons!

Its the 2N4401 and 2N4403 transistors im interested in here. As I dont have any, to quickly mock-up the circuit, I used what was immediately to hand - namely a BC212L and a 2N3904. The circuit works, but the current drain at 3V is much higher than im told it should be (5mA rather than about 1mA).

The MPSA42 transistor is what I have used as shown, and it was this which I considered to be responsible for the actual HV generation, the other two being the oscillator and, for want of the right term at present, the "quench" control.

Im interested in how the parameters of the 2N4401 and 2N4403 affect the operation of the circuit and hence the current drawn.

I do have some BC327 and BC337s (but only a couple of each) which have very similar collector current and power ratings, which I intend to try later.

How important in this circuit is the hfe of these two transistors?

Any suggestions would be great - before I order the specified devices!

Cheers
Martin

[Radio Wrangler]

2N4401/2N4403 maintain their Hfe quite well at higher collector currents than most small TO-92 transistors. They are also quite good for low noise amplifiers operating from low impedance microphones etc (low base spreading resistance)

It's likely that your boost converter runs short, high current pulses.

2N2222A might be a possibility.

You may be better looking through the types made by Zetex (once Ferranti, now part of Diodes inc.)

David

[paulsherwin]

I don't understand why your transistor substitutions are causing excess current. They might go pop as they're low current devices, but if they survive they should work.

Is there anything unusual about the MPSA42? I haven't looked it up (I'm on a phone.)

[Martin G7MRV]

Originally Posted by paulsherwin

I don't understand why your transistor substitutions are causing excess current. They might go pop as they're low current devices, but if they survive they should work.

Is there anything unusual about the MPSA42? I haven't looked it up (I'm on a phone.)

Only unusual thing with the MPSA42 is that its a higher voltage device.

Im not 100% sure exactly how in 'excess' the current is - the write up isnt particularly clear on that point! The circuit comes from this site http://www.techlib.com/science/geiger.html and it looks like English may be the authors 2nd language.

The BC327/BC337 look to also have similar hfe specs to the 2N4401/2N4403, so i'll try them out. Im trying to use whatever I have to hand, but I generally dont use a lot of PNPs!

Im wondering if its the gain of the 'feedback' transistor thats the issue - maybe not cutting the oscillator off quick enough and so allowing more current to flow?

[paulsherwin]

I've got loads of BC327/337s. Send me a PM if you need some more.

[Argus25]

Some thoughts on this supply:

Your circuit is merely a version of the classic two wire lamp flasher:

https://mastercircuits.blogspot.com/...torcycles.html

(This circuit was commonly used in my neighborhood back in the 1960's, to feed the secondary of a small valve audio output transformer to make a rough and ready high voltage generator, without a voltage multiplier, for prankster applications)

Your circuit has been slightly modified with a diode to allow the collector voltage to go much more positive than the supply without affecting the timing.

There are three issues with the circuit as it stands.

Firstly there is no energy recovery diode which also prevents possible destruction of the output transistor when the second half cycle of oscillation occurs at the collector after the output transistor is turned off, so you need to add a diode, say UF4007, across the collector-emitter of the output transistor (cathode to collector)

Secondly, the magnitude of the positive peak voltage on the collector will depend very much on the properties of the transformer, its inductance and self capacitance, so it becomes unclear as to the collector voltage rating you will need for the output transistor. Having additional lumped capacitance added is always favorable in a flyback supply. It pays in this case to measure it on the scope and select a suitable output transistor where its voltage rating is not exceeded, or is reduced by the lumped capacitance to about 80% of the transistor's Vce max.

Also, its not a wonderful idea to use the 1N914 as an avalanche or breakdown device for the feedback, though they can zener around 80 to 100V it is very variable. There should be some resistance added in series with that diode to limit the base current of the transistor it feeds.

Also, the circuit might not be efficient, in that likely the duration of time the output transistor is on for is too long and the inductor is has reached a stable high current for an unnecessarily long period, before the current is interrupted, and during that part of the cycle its just wasting energy. So that is also easily checked on the scope to adjust the operating frequency so its high enough that this does not happen.

PS, the original version of this circuit used an AC127/128, had an LDR associated with it, and was used on those yellow body/orange lens roadside lamp flashers at roadworks sites, but the load was merely a small lamp and there are other considerations noted above when its an inductive load.

[Martin G7MRV]

Originally Posted by Martin G7MRV

Ignore the 1N914 diode - thats replaced by a string of zeners and neons!

Its the 2N4401 and 2N4403 transistors im interested in here.

As I stated quite plainly - it is the transistor equivalents im interested in, and the 1N914 IS NOT being used.

[Martin G7MRV]

Originally Posted by paulsherwin

I've got loads of BC327/337s. Send me a PM if you need some more.

Thanks for the offer Paul. Ive a pair of each at the moment, so long as I dont blow them up!

[Argus25]

I could advise for a circuit like this, as replacement transistor options, a good npn pnp pair to start with is the BC639/BC640. They have very robust junctions and are much more difficult to destroy in the experimental process.

[Argus25]

Originally Posted by Martin G7MRV

As I stated quite plainly - it is the transistor equivalents im interested in, and the 1N914 IS NOT being used.

Yes I know, it was only one of three things wrong with the arrangement that I could see.

These flyback supplies have some interesting features:

Where you can with a supply like this you are always better , if you are able, to avoid using the voltage multiplier on its output as it substantially raises the supply's output impedance and requires more parts. If you can it works better simply to find a transformer with a suitable turn's ratio. As I mentioned a classic easy to get transformer is a valve type audio out used in reverse.

Also, assuming the transformer is not being switched on for too long a period of time that the rate of rise of current tapers off, the rate of rise will be V/L amps per second, where V is the supply voltage. And since the stored energy is proportional to the square of the current and since the voltage peak you get is proportional to the square root of the stored energy, it turns out that the voltage peak you get is directly proportional to the power supply voltage.

What this means of course if you double the supply voltage V, you double your peak output voltage.

So when making a flyback supply, a good way to regulate it, if you want it regulated, is to control the length of time the transistor is switched on for per cycle, so at the end of each cycle you have controlled the peak primary current I. This is where a 555 can come in handy or many types of pwm IC too.

Or, alternatively you can also manipulate the supply voltage V with a series pass transistor, which controls V, which is ok and not too lossy, if its a low power output supply.

So there are a number of configurations that work & are fun. An all discrete type is good of you are starting with a low level supply voltage like 3V.

[Radio Wrangler]

These circuits rely on the output rectifier/reservoirs to limit the transient when the transistor turns off. A simplistic viewpoint says that transformer action will limit the transient at the transistor to a scale model of the output voltage.

Reality, in the shape of leakage inductance spoils this idyllic view and the transient from dI/dt of the leakage inductance can destroy or degrade transistors.

Cautious design is to have a clamp on the collector but set so that it only operates on transients bigger than the scale model of output level which is essential to the circuit's operation. Care in winding the transformer helps.

David

[Argus25]

Originally Posted by Radio Wrangler

These circuits rely on the output rectifier/reservoirs to limit the transient when the transistor turns off.

David

That assumes though the output is under some reasonable load. Otherwise if not the capacitor/s on the output side simply climb to the peak voltage and the diode stops drawing any significant current on peaks.

In the Push-Pull Royer converters though, one trick to snub the voltage transients off the collector circuits (caused by the transformer's leakage inductance) was to peak rectify the collector voltage and feed it to a capacitor with a low value bleeder resistor to discharge it significantly between peaks.

One thing about the flyback supply though, due to the fact that you are not actually transferring power from one winding to the other (instead it is generated by a collapsing field) the whole issue of leakage inductance between the primary and secondary becomes academic, unlike the push-pull converter.

This was one thing that some automotive electronics engineers failed to get to grips with when CDI was invented. The standard Kettering coil, with its high leakage inductance, was fine, for a flyback type of Kettering supply. However, when used in a CDI mode (where its acts as a transformer) its performance was badly degraded by leakage inductance. This is why CDI's work much better with a transformer style ign coil. But of course when "add on CDI boxes" came out, the manufacturers of those didn't want to tell people they should buy a new IGN coil as well.

[Martin G7MRV]

Originally Posted by paulsherwin

I don't understand why your transistor substitutions are causing excess current.

After some playing, Ive come to understand that ive misread the original article, or it is perhaps badly worded. It says the current should be very low, maybe 1mA or so, under no load. Ive taken that to mean no load whatsoever on the output of the multiplier, but it seems this doesnt include the series chain of zeners/neons etc acting as the regulator. After adding a few wire ended neons to the regulator side of the circuit, the current with a 9V supply dropped to 2mA.

I do now have it working with the BC327 and BC337 at 3V, which has been a bit of a learning curve!

[Martin G7MRV]

Originally Posted by Argus25

Where you can with a supply like this you are always better , if you are able, to avoid using the voltage multiplier on its output as it substantially raises the supply's output impedance and requires more parts. If you can it works better simply to find a transformer with a suitable turn's ratio.

this is likely to be my only option for a 1.2V supply. Getting the circuit working at 3V with a high enough output has been tricky. I eventually found that the miniature audio transformers I was trying could cope at 9V supply but wouldnt get above 230V from 3V. Swapping to what I now know to be a common mode choke from a SMPSU has allowed me to get a little over 400V from it with a 3V supply

What this means of course if you double the supply voltage V, you double your peak output voltage.

Indeed, as the above experiments have shown!

So when making a flyback supply, a good way to regulate it, if you want it regulated, is to control the length of time the transistor is switched on for per cycle, so at the end of each cycle you have controlled the peak primary current I. This is where a 555 can come in handy or many types of pwm IC too.

PWM is one of the techniques ive tried, using a 555 and the common mode choke, from 9V. This works but the regulation is a little loose, varying around +/- 10V at 430V out. This isnt actually too bad for a G-M tube, so long as the nominal voltage is set for the middle of the G-M plateau

So there are a number of configurations that work & are fun. An all discrete type is good of you are starting with a low level supply voltage like 3V.

This is my goal, ultimately a single 1.2V cell powering a cute little 2cm long alpha tube, and all in a box small enough for a shirt pocket! I think a true transformer might be the only way to go here. I believe the transformers from disposable flash cameras can be used, if I can find one!

The circuit as shown, with the transformer replaced by the common mode choke and the BC327/BC337, didnt like running from 3V with a 100uF cap for the pulse timing. With a neon bulb as part of the regulator chain, it could be seen visibly to flash! A bit of experimentation showed that the highest output voltage was had with a 4u7 capacitor. As you say, timing the pulses so as to reach the peak current in the inductor but not go over.

When I started this project, the aim was simply to generate 400V to drive the G-M tube, and all my attention was on the indication/counting/metering circuits. Im actually now getting quite interested in the power supply side! Its proven to be quite a learning curve.