Germanium Transistors with Shunt Resistor HFE measuring

[Kesh]

So...the question...why are the HFE's so low, when the data sheet specifically says they should be 80-150hfe.

In Soviet Russia, transistor shunts you.

[chromesphere]

So...the question...why are the HFE's so low, when the data sheet specifically says they should be 80-150hfe.

In Soviet Russia, transistor shunts you.

I think this is the most accurate description of whats giong on here

Found this thread.  I had read it, but i didnt realise they actually came to conclusion that these transistors dont have a shunt resistor.  Which means, and goes back to what prr and iviark have said.  Different test configuration.  Maybe they get those gains at 15 / 20 volts are something?  Who knows.  I'm over it

http://www.freestompboxes.org/viewtopic.php?f=17&t=6123

Paul

[jrod]

So...the question...why are the HFE's so low, when the data sheet specifically says they should be 80-150hfe.

In Soviet Russia, transistor shunts you.

[jrod]

Yeah, like IVIark said it should be the 313's that have the shunt resistor. All the ones I have show so on the Peak.

I haven't read through this thread completely but all the gt308s i have are much lower hfe than the datasheet says they should be. Most on the ones i have are in the high 50's to  mid 60's. Out of about 100 there may be a few in the low to mid 70's.

Sorry if im way off topic here since i really did read all the thread.

[PRR]

> Ic 437ua

Russian 308 hFE is specified at 10mA. >20 times higher.

With a looks-like-B suffix, hFE at 10mA should be 80-150.



There is a "typical" (not guaranteed) hFE versus I curve which shows *steep* fall-off at lower current. At half-mA, a lot closer to 50 than the ~~100 gotten at 10mA current. (Yet awful close to your observed 46.8 hFE @ 0.4mA.)



> how do you determine if a transistor has a resistor shunt?

The diode junction is "dead" for very low voltages, yet perfect resistors are linear to infinitesimal voltage (real resistors are pretty close).

Apply a very small voltage/current and measure the current/voltage.

If you own a good-old VTVM with an ohms scale, it is easy. If you Ohm a naked junction and turn the range switch, it reads "different Ohms" on each range yet the needle hardly moves. If you look at the Volts scale it hangs with 0.6V (for Si)..... the "ohm meter" is actually finding the diode drop voltage for different currents! But if there is a shunt resistor, as you go to high ohms ranges, the "Ohms" becomes constant at some point well below 0.6V on the Volts scale. Yes, this is less useful with the 0.2V of Germanium because you get jammed in the bottom of a 1.5V range.

[azrael]

Where are you guys buying your 308s? I'd like to get my mitts on some consistently low leakge Russkies.

[Lurco]

New under the sun? http://www.nxp.com/products/automotive/discretes/resistor_equipped_transistors/  the rest of the clicking is up to you (unless you want to stick to through-hole germaniums only).

[IvIark]

Where are you guys buying your 308s? I'd like to get my mitts on some consistently low leakge Russkies.

http://stores.ebay.co.uk/Orpheus-2005
http://stores.ebay.co.uk/Next-Lux
http://stores.ebay.co.uk/BgEbus

FWIW, I've bought about 15 different Russian germanium PNP transistors and with the exception of the GT402B, they all had very low or zero leakage.  The GT402B did have some but out of 50 only 2 were above 0.6mA and they were higher gain than the others, ranging from 78 to 193 hfe.

[chromesphere]

Thanks for the detailed explaination PRR!  Seems you were on the money the first time around!  I'm getting low gain due to different test conditions, my rig's only supplying approx half a milliamp, the test conditions: 10 milliamps.  And looking at that graph you found, you can see where 1ma is.  Not going to get much more gain them 80!  Wow it all makes sense

Ok so these would be more suited to higher current applications i suppose.  Well still I'm getting a few around the 80's mark, so half will probably be usable in some way.  I have found better results / better suited germaniums with other russians though and probably wouldnt buy these again.

Paul

[chromesphere]

Of course, another question that springs to this persons mind is, how do we increase the current to get more gain  

Would it be possible to bias say a fuzz face or tone bender to be supplying the tranny with this sort of current?  Or do you have to increase the voltage to the circuit?  Might be getting a bit over my head with this question but thought i would throw it out there.

Paul

[IvIark]

It's all relative.  A pair of AC128s that read 80 and 120 hfe and sound superb in a Fuzz Face have exactly the same issue.  The reading of 80 and 120 is nothing like the actual gain they're going to be providing in the circuit, it's just an easy way for us to measure them to allow us to compare for gain levels in our test, and select transistors appropriately.

I always thought the GT308's sounded great in a Fuzz Face and you'd certainly have Q1 covered, plus Q1 and Q2 in Tone Bender circuits.  Maybe it's worth looking for those Q2 transistors.  Like I said the GT402Bs are a good choice for higher gain, and I also got some 2N404s which were also low leakage but were again higher gain on average than the other Russian transistors I've got.  Plenty over 100 hfe.  Neither of these were very expensive either.

[PRR]

hFE=50 doesn't suck. Just use them.

[chromesphere]

hFE=50 doesn't suck. Just use them.

Yeah i agree.  From my experimenting, ive found only the slightest difference in gain from 50 hfe upwards.  Lower though, and they lose a notificable amount of gain (in a tone bender, or i from memory, even a soul bender). I think its been drummed into people how important gains are though.  I sell a few GE's every now and then, to help fund my DIY addiction, and people tend to be very specific.  "Must be 70-80!!!". 

Im probably just going to start a debate with these comments though

Paul

[PRR]

> slightest difference in gain from 50 hfe upwards

When I was younger the prof's quizes always stated "assume hFE >50".

By that time, improved Si processes were yielding much higher hFE at jellybean prices; I guess prof didn't want to re-type the quiz mimeo masters.

Early Ge could be sorted to hFE >50 with small yield, or you could buy the weed-outs with hFE<50 for a dollar a bucket. (Even before that, CK722 "experimenter grade" was hFE of 10 or even 5.) There might be a borderline grade 40-80 at an inbetween price.

So many designers of that era targeted hFE near 50. Less made design awkward, higher cost more.

When higher hFE became common/cheap, there was more emphasis on circuit techniques which worked with "any" hFE. When my prof said "hFE>50", he meant he might check my design with hFE=5,000! Indeed it isn't that hard to keep things within reasonable limits with any hFE to infinity, though you do need to target a minimum hFE below which you don't expect the design to meet all specs.

[chromesphere]

So i have measured about half of these and they all sit, min 55, max 80.  So they are useable.  if your fussy about q2 / q3 being 90+ then i'd skip these.

But just a word of warning, i know some GE's are pretty tolerant to inserting the legs into a circuit incorrectly.  These are not one of them!  I've killed 6 of 40 so far, and im pretty sure now that its from the base or collector, brushing across the wrong socket as i put it into the leakage tester.  Even a half sec touch -> transistor is now a diode.  So if you buy some, yeah, be EXTRA RIDICULOUSLY careful.  If you insert the legs incorrectly theres a good chance you'll kill it.

Paul

[bancika]

I got two lots of 50 x GT308V's and measured them all with my Peak and this is the spread I got:

I got these a couple of years ago and have heard a lot of people getting much lower gain batches since.

these sound too low, I bought 50 recently and majority was around 90.

[Kesh]

Does anyone know, when people say something like "for a fuzz face Q1 is best about 70-90, Q2 100-130" what collector current they are measuring it at? Or base current will do, as it's easier to fix.

[chromesphere]

Does anyone know, when people say something like "for a fuzz face Q1 is best about 70-90, Q2 100-130" what collector current they are measuring it at? Or base current will do, as it's easier to fix.

Yeah thats the question isnt it.  As PRR described above, the HFE can change with collector current.  (I was going to start a new topic about this actually but it seems approperiate here):

Which makes me steer even more towards the conclusion that HFE can mean jack squat with fuzz's, (within reason of course).  70-90, should be redefined as 50 - 200.  Seriously, anything in that range sounds good to my ear.  Sure you get a bit of extra sustain over 100, but they all still sound 'good'.  Some low gainers i've found to sound better in the 50-100 mark (the ones im currently selling in the trade section) then others ive had that sit perfectly in the 'recommeded' ranges.

Last night i was curious to see what the HFE's were of the transistor set i bought off small bear some time ago.  To my surpirse they measured around 140 and 170 hfe. Leakage was around 300ua and 200ua.  These dont sit within the 70-90 / 100-130 HFE range.  I'm not saying he's transistors are no good.  On the contrary, they sound awesome!   But i think thats more to do with the actually brand (2n's) then the hfe's?

Prehaps im going to start a war with these comments, but you know.  Thats what i do.  I'm a practical learner.  And from what i've actually heard thus far,  anything over 50, sounds good, with only subtle differences, all the way up to 200hfe.  Even over 300ua leakage sounds good, again within reason.  over 1000ua i can hear it starts to sound squishy.  But i wouldnt be suprised if in a blind test some guitarists would actually perfer this sound.

So, either, im going deaf from sitting infront of my cranked up amp for too many hours (highly likely) or theres not quite as much importance on HFE as what has been recommended.

If this information is incorrect please educate me.  I mean that seriously.

Paul