Beta Matching!

[Joster]

Hi All,

I am running a 4 transistor ss sg. Two of the devices get warm and the other two are cold. Is this a case of unmatched devices? I'm thinking so as I have not done any matching on any of my circuits yet but will be on my next build. Anyway, I am assuming that the beta of the two devices that are getting warm would be close and same for the two that are cold. Also then, the ones with the higher beta will be doing most of the work?

This gets me thinking though....the beta matching has to do with getting all the devices switching at the same time correct? So how could two of them be doing most of the work when all four of them are switching. I suppose this could be that they are fighting one another. Meaning some are switching on slightly before the rest creating a crazy situation in the coil where not all the windings are in-phase.

I have been having luck using mje13009 at $1 each. Roughly how many would I need to purchase to get 4 matched mje13009's??

I am just having trouble understanding how the transistors could be switching at slightly different times even though they are all being driven by the same trigger signal. But I suppose that transistors with higher beta will turn on quicker than those with a lower beta.

Just looking for confirmation here. Thanks.

Thanks,

Joster

[Tom C]

it is not about when they switch but how far they open. if you look at the data sheet on a transistor there is a curve as to how much current on the base opens the junction, it varies a little bit to a lot. does nt matter on a single transistor but with multiple ones it does. on class A amps that is one of the main things that makes them class A. imagine 4 WATER VALVES all of them have 2 lbs of pressure applied to open the valve, but 2 valves open a quarter of a turn and 2 valves open a full turn, the full turn valves have more water flowing thru them even though they all have the same pressure applied to the valve handle.

not a perfect picture.... you are trying to get all transistors to open their bases the same amount for the same milliamps coming off the trigger.

Tom C

[Joster]

it is not about when they switch but how far they open. if you look at the data sheet on a transistor there is a curve as to how much current on the base opens the junction, it varies a little bit to a lot. does nt matter on a single transistor but with multiple ones it does. on class A amps that is one of the main things that makes them class A. imagine 4 WATER VALVES all of them have 2 lbs of pressure applied to open the valve, but 2 valves open a quarter of a turn and 2 valves open a full turn, the full turn valves have more water flowing thru them even though they all have the same pressure applied to the valve handle.

not a perfect picture.... you are trying to get all transistors to open their bases the same amount for the same milliamps coming off the trigger.

Tom C

thanks now i get it. I just pulled out the 4 470 ohm base resistors and tested them all and they are all over the place too. I figure what I'll do is get a bunch of 1% tolerance resisters and match them from there.

Any idea how many transistors I will need to get 4 matched devices...I understand it's sort of luck of the draw but from your experience how many do i need to buy?

[Tom C]

we buy them in lots of 500 and match them... never bought less. if you want them already matched you can get them from teslagenx. yes they are more expensive but they are power matched and you don't have to buy 500.....

Tom C

[Joster]

we buy them in lots of 500 and match them... never bought less. if you want them already matched you can get them from teslagenx. yes they are more expensive but they are power matched and you don't have to buy 500.....

Tom C

Ok thanks...thats good to know. In your experience how many of the 500 do you have to test on average to get 8 matches?

I did some poking around on the "all about circuits" forum and got this reply which I thought was interesting and wondered what you thought about it...

"You don't want to go with matched transistors as that would be difficult to do and expensive. You should be able to design the circuit with some emitter negative feedback so the value of the beta has only a minor effect on circuit operation. That's the normal way to design a circuit. If you post more details about what you need (frequency response, current values, coil inductance, voltage values, etc.) someone on this forum can likely help you with the design."

here is the link to the thread

http://forum.allaboutcircuits.com/th...tching.102176/

[ErikN]

Ok thanks...thats good to know. In your experience how many of the 500 do you have to test on average to get 8 matches?

Joster, Buying more transistors a time reduces the number of transistors we can't match. Start with a tube of 25 and you may get lucky and find 1 or more matched sets of 4, or you could end up with no matched sets. There is a reason we buy 500 or more at a time...the price of the transistors should give you a good idea of the loss we find but it is also a time consuming process.

"You don't want to go with matched transistors as that would be difficult to do and expensive. You should be able to design the circuit with some emitter negative feedback so the value of the beta has only a minor effect on circuit operation. That's the normal way to design a circuit. If you post more details about what you need (frequency response, current values, coil inductance, voltage values, etc.) someone on this forum can likely help you with the design."

You are free to design the circuit however you wish. Yes, it will run without matched transistors, but the SG isn't just a coil banger. It's designed as an audio amplifier, without matching the transistors, you reduce the efficiency of your SG. Our suggestion is build it like John Bedini says; after understanding how it works, then start making modifications.

Erik

[Joster]

Joster, Buying more transistors a time reduces the number of transistors we can't match. Start with a tube of 25 and you may get lucky and find 1 or more matched sets of 4, or you could end up with no matched sets. There is a reason we buy 500 or more at a time...the price of the transistors should give you a good idea of the loss we find but it is also a time consuming process.



You are free to design the circuit however you wish. Yes, it will run without matched transistors, but the SG isn't just a coil banger. It's designed as an audio amplifier, without matching the transistors, you reduce the efficiency of your SG. Our suggestion is build it like John Bedini says; after understanding how it works, then start making modifications.

Erik

ok thanks...that was what I was looking for...on a good day 4 matches out of 25

Are your matches exact or is there a small tolerance window?

[ErikN]

Are your matches exact or is there a small tolerance window?

Ours are typically matched to 1 percent or less. It's very unlikely you will achieve that low tolerance with a batch of 25 transistors, but again, you could get get lucky.

Erik

[Joster]

ok! Ill try my luck with 25 thanks alot for the info...I'm sure having matched base resistors and output diodes would be awesome too. That should be easier at least with the resistors. Do you guys match the diodes too? If so what is your luck with that on a good day?

[Joster]

found 100 tip41c's for $16 bucks

I know its just 6A Collector Current and 100V collector/emitter but I think it will do for now anyways. I do not have alot of dough right now to play with.

Just curious what people think of using the tip41c

there are some 2n3055's on ebay for a good deal but they are only 60 Volts collector emitter. They do have a 12A collector max current which is good.....

What should I go with?

[wild weasel]

hey joster

i use these bd243C
not to expensive and they work just fine
but watch your trigger seting on the pot
bad triggering burns them so when timing watch the temerature

i run them on a single coil at 19 volt and 22 degrees celcius
on the multi coil they sit at 19 degrees with 12 volt input

regards
weasel

[Lman]

Joster,

I will start with some history that I know of. Back in time when transistors were invented the technology had been such that they were point-contact. Highly non-linear, going negative as well and as far as I know mostly made of germanium. A great breakthrough and alternative to the vacuum tubes in many areas but their non linear behavior had been considered not good in some electronics. Later the silicon had begun to replace the germanium and with the technology improving the junctions were layered. This had led to more linear results, improving electronics and making them more reliable, higher power transistors had appeared as well. There are fights on some forums between audio guys about the "germanium sound" versus "silicon sound". Some people prefer the qualities the germanium gives to audio in different devices.
John Bedini witnessing all the development from the birth of the transistors and all the changes during the years till now, working in the semiconductor industry and doing transistors knows very well what to look for in a device. I remember having seen a post of him recommending 2N3055H transistor from a certain manufacturer for use in audio amplifiers, meaning if you get the same model from another manufacturer it is not the same already.
He has been testing many transistors and if he chose MJL21194 for the SG that is for a very good reason. If there was a cheaper alternative to it he would have told us.
MJLs are high power and in some of Johns videos you can see they go negative as well when switched off meaning they show non-linearity. With the radiant energy we want the transistors to do that and the more it does the better the results. If I am not wrong it helps the transistor to stay colder. Not all the transistors go negative.

Having in mind that compare it with the cheap devices you are thinking of.
MJE13009 needs 9 volts on the base to fully open. MJL21194 needs just 5.
MJE13009 could handle around 100W of power dissipation while MJL21194 around 200W. This means if you drive both transistors with the same power the MJL21194 will be way colder. Add to this that it goes negative and this will give you even better results.
We do not want heat in our devices but in Solid State SG the transistor switches very fast and sometimes it will get hot. Heat means loss of energy, and can degrade the transistor's junctions in time if not cooled properly.
It is not an accident what ErikN said about the SG circuit being designed as an audio amplifier. This is how it senses the output impedance and when it is of a sulphated battery the transistors may start to develop heat at some point. I have had SS SG that would run cold to the touch for hours and than at some point it would get crazy hot. The transistor is not alive anymore and that was a lesson for me. In EFTV #2 you can see one of John's big SS SG circuits (I think that is what it is) and all the transistors are on a big heatsink.


Saying all this I hope that you and everybody who would cross this thread will understand the importance of the use of MJL21194.


I had at one time built a small SS SG done with MJE13007 running Leds and it was ok but still 9 volts on the base ... not good. I tried to charge a battery with this oscillator and in less then a minute the transistor got so hot that I turned it off to save its life. I have some medium power russian transistors that need just 2 volts on the base. I am planning to try them with LEDs or small batteries as I hope to get some efficiency with them.

I have done many times the mistake to think that going for the cheap will save me money and later when I see I am not getting the results I want I go for the right thing and end up spending more money than if I would have bought the right one in the first place. Be it a multimeter, transistor or whatever else. You do not have to do that mistake with the transistors now.

My advice is go for the right transistor - the MJL21194.

Something else to have in mind is that there are lots of fake transistors out there including MJLs. Especially at online shops.
So if you consider all this TeslagenX saves you all the hassle plus matching to 1 percent is pretty impressive and big time saver. This shows you the devotion the guys put in their efforts to help the others get good results. I doubt anybody at any online shop would do it for you to that level if they offer matching at all. Buying from TeslagenX you can be sure also you have originals and not fakes.
At the end it seems to me in your case it would be cheaper to get 4 matched MJLs and have better results than buying higher count low level transistors that would not get close to what MJL21194 would do.


You are right in thinking about matching resistors and diodes also. This is the way to get the best results.



Regards
Lman

[Joster]

Lman,

Thanks so much for all the info...lots of food for thought!

What do you mean by the transistor going negative? I just figure it is either on or off....

[wild weasel]

hello lman

just a snap question for future development
here in europe the MJL21194 is no longer available. so i suppose it will be a problem in america in the future
here they offer now the MJL21194G as other option
i believe the only manufacturer is ON semiconductors

its not exactly the same but do they react the same way?

kind regards
stephan and serge

[Lman]

Lman,

Thanks so much for all the info...lots of food for thought!

What do you mean by the transistor going negative? I just figure it is either on or off....

Joster,

The reason I was telling some history is that some semiconductor materials can exhibit negative resistance. The devices in the past were doing that more easy than those we have today if they do at all.

There is a DVD where John Bedini explains these things and showing some unique things.
http://www.youtube.com/watch?v=oTpqx620Am4
I use to think of it as "the rock" DVD as he is showing what can be done with a piece of a rock. On Energenx channel as well John posted several videos of a small motor that he runs on that rock. Pulling negative energy.

Transistor being on or of is not like a switch on the wall. Do not forget that the junctions inside resists before they begin to open. But this is the usual positive resistance.The MJL in the SG when switched off will go negative meaning it will pull some negative energy for a tiny fraction of the ... micro or nanosecond. And that cools down the transistor as the negative energy is cold plus adds to your efficiency. I think this has been discussed here on the forum or I read about it on the Energetic forum.
This is just a part of the enormous research John Bedini has done and share with us. And it is why he insists and repeats often that we should build the things as he say at least the first time.


Regards
Lman