4 Pin Mosfet or am I confused?

[Stonent]

Q1 says RF 511, I know there's an IRF511 but it has 3 pins or am I misunderstanding the schematic? Or is this an IGBT?

[sleemanj]

Mosfets are 4 pin devices.  But only 3 are usually broken out.  The 4'th is the substrate.  Which is generally internally connected to the source.

[Sigmoid]

There's a thing I've been wondering about MOSFETS... Are they symmetrical by default? I mean if there's a 4 pin packaging, could I connect the substrate to any of the two channel pins, and thus "nominate" it source, and the other drain?
From what I've read about their structure and manufacture, they look symmetrical, but often things just aren't that simple in real life.

[w2aew]

There's a thing I've been wondering about MOSFETS... Are they symmetrical by default? I mean if there's a 4 pin packaging, could I connect the substrate to any of the two channel pins, and thus "nominate" it source, and the other drain?
From what I've read about their structure and manufacture, they look symmetrical, but often things just aren't that simple in real life.

It depends on the device. Often they have the substrate connection tied to the source which will prevent you from using them 'the other way around', because the substrate has to be biased at the lower of the drain and source contact voltages. If the body or sub contact is brought out separately, you 'might' be able to run it with the drain and source reversed, but it is not guaranteed. Often the packaging will prefer one orientation, offering less capacitance between the gate and drain, thus reducing the miller effect.  So, I'd say in general - no - but there are exceptions.

[tree]

Some (most) discrete mosfets have a reverse-biased diode from drain to source, so if you try to use them in opposite polarity (e.g. N-channel mosfet; current from source to drain) then you will just forward bias the diode and your mosfet will look like a diode.

[Stonent]

So I should just assume it is an IRF511?  I know how Geiger counters work in general but this circuit as simple as it is has me a little confused. It looks like one of the 555 timers is being inverted through the NOT gate. The tip120 is working as the audio amp and the MOSFET is driving the HV transformer. But what I wasn't sure of was the purpose of the second 555. And what functionally is the NOT gate doing to the circuit? Why do they need to invert the 555? I haven't looked at all the pinouts yet so my understanding at the moment is just by seeing the parts and where they go more so than the exact function.

[free_electron]

Mosfets are indeed symmetrical. The point where you tie the 'bulk' becomes the source.
in an integrated circuit you can make floating mosfets.

you can actually make a 3 pin mos reverse as well.

[SeanB]

4 pin MOSFET's were available a while ago and were used to make analogue switches, but as integrated circuit analogue switches improved they kind of all but disappeared. They had a varying resistance with voltage above the substrate, but would provide better isolation than the original 4016 CMOS analogue switch. With the development of the 4066 and others they became the non chosen option, as they needed a driver and assorted bias components whereas the 4066 just would work with a single decoupling capacitor and give better performance.

[uwezi]

So I should just assume it is an IRF511?  I know how Geiger counters work in general but this circuit as simple as it is has me a little confused. It looks like one of the 555 timers is being inverted through the NOT gate. The tip120 is working as the audio amp and the MOSFET is driving the HV transformer. But what I wasn't sure of was the purpose of the second 555. And what functionally is the NOT gate doing to the circuit? Why do they need to invert the 555? I haven't looked at all the pinouts yet so my understanding at the moment is just by seeing the parts and where they go more so than the exact function.

The 4049 inverters are used as poor-man's gate drivers for the MOSFET. While it is true that a MOSFET is a voltage controlled device, which does not require a current into the gate at steady-state, a powerful driver is needed for fast switching of the device, especially to get it quickly from OFF to ON and back. Using a couple of standard CMOS inverters is not uncommon when driving power MOSFETs and can replace more difficult to obtain, dedicated driver chips. What I don't understand here is, why only 3 of  5 available inverters inside the 4049 are used. At the very least the 2 unused inverter inputs need to be grounded - but why not use them as well?

The second NE555 is connected as monostable and makes the count pulses longer, it also makes all pulses the same length. This is necessary if you e.g. would want to drive an analog meter with the signal and get a display which is proportional to the count rate. Also it makes sure that the pulses really are audible.

This circuit appears to be quite old, but the principle will of course still work today. For the MOSFET you can almost use any decent power MOSFET of today and yesterday, not only the (I)RF511 - a higher Vds breakdown voltage will make it stable agianst the back-emf from the transformer.

While the transformer does not seem to be specified I have successfully built similar circuits with a small 240V/6V mains transformer run backwards.

What this circuit lacks is any kind of feedback and the only means of voltage setting is the 3 zener diodes on the secondary side. Make sure to adapt those to the "needs" of your GM tube, since these tubes can be damaged (or at least suffer from a reduced lifetime) easily by too high voltages.

[Stonent]

I've been helping a friend who has a mechanical engineering degree with a minor in nuclear engineering who wants to get a geiger counter for personal use. He isn't opposed to a a kit or DIY (I said I'd help him assemble it) and I was looking at the available kits as well as some of the schematics that I could find online.  This one looks fairly simple to build with no unobtainable parts.  I've considered driving an MCU input from the speaker ticker part with a standard 16x2 LCD.

His cheapest route would be a kit that drives a 450V Russian/Soviet surplus tube.  More expensively would be one that drives an American style 900V pancake probe by LND or Ludlum.  He likes my CDV-700 and was interested in them but I was just giving him some options. He mentioned something about wanting one that could be waterproofed so he could try to use one when diving, but I'm not sure much more than Gamma would get through the water unless extremely close.

I'm also interested in just breadboarding one up for fun and seeing how well it works, so far this seems to be the simplest design that I've found.

My CDV-700 uses Victoreen Cotoron 900V regulator tube to keep the voltage stable, those are bordering on unobtainable now.

[uwezi]

You might consider a ex-Soviet pancake. The best tube I have used so far. They are available on eBay - look for any SBT or SI  types. They run at around 400 V and cost from around USD 50.

My most recent (about 3 years old) circuit doesn't use any transformer. It is based on a MC34063 as boost converter followed by a voltage multiplier, here are some photos. As usual it was a quick fix, I guess I need to do some reverse engineering on my own circuit... 


The 8-pin bug to the left is the MC34063, the one to the right is a 7555 as pulse-shaper (same as in the above circuit).


This is the voltage multiplier - using just plain 1N4148 and 50V ceramic capacitors for the stages. Only the last capacitor is a kV-type (of course mounted the wrong side down for me to read it now...)

Oh, yeah, the thing at the bottom is a Soviet SI8B alpha/beta/gamma tube.


P.S.: Hah, the forum is not unicode  the cyrillic letters came out as smileys and question marks 
Ok, in "code" it works. So here are the transcriptions:

Code: [Select]

SBT    ???
SI     ??
SI8B   ??8?

oh man, this is crazy. It worked in preview, but when submitted the cyrillics is gone again. And this in the year 2014 - I thought we were past character coding issues... 

[T3sl4co1l]

Mosfets are indeed symmetrical. The point where you tie the 'bulk' becomes the source.
in an integrated circuit you can make floating mosfets.

you can actually make a 3 pin mos reverse as well.

Well, any discrete transistor fab'd since probably 1970 or so, or with Vdss > Vgs(max), is not symmetrical.  Not that you can find four-pin FETs anymore (...can you??).

They always work in reverse though, despite the body diode; indeed, assisting it.  Synchronous rectifiers are very handy.

Tim

[Stonent]

You might consider a ex-Soviet pancake. The best tube I have used so far. They are available on eBay - look for any SBT or SI  types. They run at around 400 V and cost from around USD 50.

My most recent (about 3 years old) circuit doesn't use any transformer. It is based on a MC34063 as boost converter followed by a voltage multiplier, here are some photos. As usual it was a quick fix, I guess I need to do some reverse engineering on my own circuit... 


The 8-pin bug to the left is the MC34063, the one to the right is a 7555 as pulse-shaper (same as in the above circuit).


This is the voltage multiplier - using just plain 1N4148 and 50V ceramic capacitors for the stages. Only the last capacitor is a kV-type (of course mounted the wrong side down for me to read it now...)

Oh, yeah, the thing at the bottom is a Soviet SI8B alpha/beta/gamma tube.


P.S.: Hah, the forum is not unicode  the cyrillic letters came out as smileys and question marks 
Ok, in "code" it works. So here are the transcriptions:

Code: [Select]

SBT    ???
SI     ??
SI8B   ??8?

oh man, this is crazy. It worked in preview, but when submitted the cyrillics is gone again. And this in the year 2014 - I thought we were past character coding issues... 

A while back I gathered some sensitivity data from a website on Soviet tubes and put it together in a spreadsheet based on what it was sensitive to, here's what I have:

https://docs.google.com/spreadsheet/ccc?key=0AnnYGMOtVmrVdEVTa2NyeGlMWTRUenNpLVh2NmZKVVE&usp=sharing

[Stonent]

You might consider a ex-Soviet pancake. The best tube I have used so far. They are available on eBay - look for any SBT or SI  types. They run at around 400 V and cost from around USD 50.

Yeah I think you can throw a couple of 180V zener diodes in series with it to make it work on a 900V unit.