Fuzz Face build - help swapping out silicone for germanium

[LucifersTrip]

The PCB is getting a bit past best now (I'll have to etch another).

remember, alligator clips with Q's hanging off board

and have gone back to using the silicone transistors. Again, if I adjust to get Q2C at 4.5v Q2B won't go below 1.3v.

there's nothing wrong with that...remember the si voltages above

si

q1e: 0
q1b: 0.6
q1c: 1.4

q2e: 0.8
q2b: 1.4
q2c: 4.5

[pinkjimiphoton]

also... GE transistors sometimes are better BACKWARDS...  reverse beta'd. if you find you're not getting the sound you expect, try flipping them 180 degrees.

Very Interesting, not read this anywhere - I will experiment! So e becomes c, and vice versa?

yes, exactly. some FETS work that way too, where you can flip 'em 180 degrees. i've found ge's that i thought were useless sounded great backwards. peeps will say it can damage it i'm sure, but 300 fuzzes later, i still haven't had a problem with it.

sometimes the gain will be almost the same each way. sometimes, one transistor right and one backwards will make the difference. try it on your breadboard and see.

i know one of my old friend devi ever's effects uses the trick as well... the torn's peaker is the same as another of her boxes with one transistor reverse beta'd.

some of the GE's i have, forget finding a data sheet... house numbers and NOS JAN stuff can be a real pain to figure out. i gotta little 20$ gizmo on ebay (search "this looks totally useful and cheap) which works great for figuring out stuff.

it DOES drop the ball sometimes... a reverse beta'd ge may show up as a jfet... but it's been handy to determine pinout and gain.

[lmorse]

remember, alligator clips with Q's hanging off board

I have just ordered a bunch from ebay 

there's nothing wrong with that...remember the si voltages above

si

q1e: 0
q1b: 0.6
q1c: 1.4

q2e: 0.8
q2b: 1.4
q2c: 4.5

With the silicones transistors I now have:
Q2B 4.51v
Q2C 1.378v
Q2E .7

Q1B .629v

Resistors = 4k6 and 47k

GE:
Q2B = 1.70v
Q2C = 4.56v
Q2E = 1.57v

Q1B = 0.043v

Resistors = 8k6 39k2

Again, I cannot bring down the voltage on Q2B any lower without Q2C climbing to near 9v, and I can't find a way to bring up Q1B. I am still stuck!  

It seems that matching the resistors is a very fine art, do all pedal builds require this much fine work in adjustment, or is it just the variability of ge?

i gotta little 20$ gizmo on ebay (search "this looks totally useful and cheap)

I failed to find anything on ebay, sounds intriguing.

[pinkjimiphoton]

i gotta little 20$ gizmo on ebay (search "this looks totally useful and cheap)

I failed to find anything on ebay, sounds intriguing.


[/quote]

http://www.ebay.com/itm/201087634115?ssPageName=STRK:MEWNX:IT&_trksid=p3984.m1497.l2649



    Features :

    1,2013 latest M328 version of the software ,more functions.Chip: Atmega328


    2.128*64 big Backlight LCD display,only 2mA when stand by.


    3,Using 9V battery (Not included)

    Test ranges:

    Inductors, capacitors , diodes, dual diode , mos, transistor, SCR , the regulator, LED tube , ESR,
    Resistance,Adjustable potentiometer

    Resistance :0.1 ohm resolution, maximum 50M ohm
    Capacitor :25pf -100,000 uf
    Inductors : 0.01mh-20H


    NEW Function:
    1:Automatic detection of NPN and PNP transistors, n-channel and p-channel MOSFET, diode (including double diode), thyristor, transistor, resistor and capacitor and other components
    2: Automatic test the pin of a component, and display on the LCD
    3:Can detect the transistor, MOSFET protection diode amplification coefficient and the base to determine the emitter transistor forward biased voltage
    4: Measure the gate and gate capacitance of the MOSFET threshold voltage
    5:Use 12864 liquid crystal display with green backlight
    Specifications: For you reference
    1,One -button operation, automatic shutdown .
    2,Only 20nA shutdown current.
    3,Automatically detect NPN, PNP bipolar transistors , N -channel and P -channel MOS FET, JFET , diodes , two diodes, thyristors small power unidirectional and bidirectional thyristor.
    4,Automatic identification components pin arrangement .
    5,Measuring bipolar transistor current amplification factor and base - emitter threshold voltage.
    6,Via the base - emitter threshold voltage and high current amplification factor to identify Darlington transistors.
    7,Can detect bipolar transistors and MOS transistors protection diodes.
    8,Measuring the gate MOS FET threshold voltage and the gate capacitance.
    9,Can simultaneously measure two resistors and resistor symbol is displayed. Displayed on the right with a decimal value of 4 . Resistance symbol on both sides shows the pin number. So you can measure the potentiometer. If the potentiometer wiper is not transferred to an extreme position , we can distinguish the middle and both ends of the pin.
    10,Resistance measurement resolution is 0.1 ohms , 50M ohms can be measured .
    11,Can measure capacitanceCan measure capacitance of 30pF-100mF , resolution 1pF.
    12.2uF more capacitors can simultaneously measure the equivalent series resistance ESR values. The two can be displayed with a decimal value , resolution 0.01 ohms.
    13,Can be in the correct order and the diode symbol display two diodes , and gives the diode forward voltage.
    14.LED is detected as a diode forward voltage higher . Combo of the LED is identified as two diodes.
    15,Eeverse breakdown voltage is less than 4.5V Zener diode can be identified.
    16,Can measure a single diode reverse capacitance. If the bipolar transistor connected to the base and collector or emitter of a pin , it can measure the collector or emitter junction reverse capacitance .
    18 can be obtained with a single measurement rectifier bridge connection.
    Notice: Before measuring capacitance , the capacitor must be discharged , otherwise very likely damage the meter .

    Package
    100% Brand New
    1 pcs New Transistor Tester Diode Triode Capacitance ESR Meter MOS PNP NPN R/C/L M328（As the picture ，Not included battery )
     

[lmorse]

http://www.ebay.com/itm/201087634115?ssPageName=STRK:MEWNX:IT&_trksid=p3984.m1497.l2649

Cool, thank you

[electrip]

http://www.ebay.com/itm/201087634115?ssPageName=STRK:MEWNX:IT&_trksid=p3984.m1497.l2649

Cool, thank you

Some of the testers with the big "fancy" diplay do not measure/show leakage.

We should start a new thread were we can post images/info/seller link for that testers
and wether they show leakage or not.

electrip

[LucifersTrip]

With the silicones transistors I now have:
Q2B 4.51v
Q2C 1.378v
Q2E .7

Q1B .629v

Resistors = 4k6 and 47k

Again, I cannot bring down the voltage on Q2B any lower without Q2C climbing to near 9v, and I can't find a way to bring up Q1B. I am still stuck!  

I really have to think there's something wrong on the board. It's not that probable with two good silicon transistors with correct orientation that you'd have to adjust the resistors that much to hit 4.5V.  What are the hfe's of the 2907's? Some of the old ones can be under 50 hfe. 

I built a FF test box years ago and put together many, many sets of FF pairs and rarely had to alter anything more than the 8.2K by 2-3K either way. I wouldn't worry about the ge's yet until you find out why the big adjustments are needed with the silicon...and remember, your ge Q2 is a bit high leakage for optimum FF.

A I wrote before, you can always post a clear closeup pic for others to check out.

It seems that matching the resistors is a very fine art, do all pedal builds require this much fine work in adjustment, or is it just the variability of ge?

All fuzzes are different...and everything is made easier (less adjustment) with the right choice of transistors. 

It's kind of funny that so many choose the FF as one of the first ge fuzzes to build because of popularity and small parts count, but there are far less potential problems with a Fuzzrite or believe it or not even a Tonebender MKIII (where you only need a single decent transistor, Q3, that shouldn't even be super low leakage, the fuzz sweep is good through the entire rotation, and there's no mushiness). With a FF, you need 2 good low leak transistors, the fuzz sweep is mainly useable in the final 10-20% (which is why so many set it at full and use a pre-gain instead) and it can be mushy (which is why so many lower the size of the vol pot and/or decrease the size of the input cap.

[lmorse]

It's not that probable with two good silicon transistors with correct orientation that you'd have to adjust the resistors that much to hit 4.5V.  What are the hfe's of the 2907's? Some of the old ones can be under 50 hfe.

I am measuring around 171hfe for the silicon transistors (does this sound correct?), using the technique outlined by R.G. Keen on his site http://www.geofex.com/Article_Folders/fuzzface/fffram.htm.

A I wrote before, you can always post a clear closeup pic for others to check out.

I'll get onto this as soon as I can.

It's kind of funny that so many choose the FF as one of the first ge fuzzes to build

I remember reading on a website somewhere that it is an ideal starting project, that's why I chose it. I have to say I am enjoying the learning progress, it far outweighs the minor frustrations. Maybe a mild baptism of fire?

[LucifersTrip]

It's not that probable with two good silicon transistors with correct orientation that you'd have to adjust the resistors that much to hit 4.5V.  What are the hfe's of the 2907's? Some of the old ones can be under 50 hfe.

I am measuring around 171hfe for the silicon transistors (does this sound correct?)

yes... Hfe 35/300

http://transistor-spravochnik.ru/description/2n2907/2625

[lmorse]

I've had to leave home for work. I work away - 8 days at work, then 6 days off. I'll continue my efforts when I get home middle of next week,
Thanks
Lee

[Electric Warrior]

With a FF, you need 2 good low leak transistors, the fuzz sweep is mainly useable in the final 10-20% (which is why so many set it at full and use a pre-gain instead) and it can be mushy (which is why so many lower the size of the vol pot and/or decrease the size of the input cap.

Finding low leakage transistors is not hard. Just get a couple of 2SB175s or 2SB172s. They can be really cheap. I'm having much more trouble finding leaky transistors for a MKII Tone Bender.

I don't think you must have low leakage transistors for a fuzz face anyway. The NKT275s were often low gain (around 50hfe is pretty common and they're usually closely matched) and medium leakage (100 to 200µA). I guess with gains that low they need a bit of leakage to sound good..

But that sure isn't the only way to set up the circuit. It works with a wide range of different transistors. After all, they used everything from low gain germaniums to high gain silicons back in the day. I do find it much easier to make a nice sounding germanium one, though.

edit:
I also suspect 242uA leak for Q2 is a bit too high for an optimum ff. (the higher leak, the lower Q2C voltage)

I found that Q2C is lower with low leakage transistors, but that is with both transistors being low leakge as compared to two higher leakage trannies.
I wouldn't expect a vintage NKT fuzz face to measure as low as 4.5V on Q2's collector either. The one vintage unit I used to have had a crazy high voltage there, but that one was biased a little extreme...

[lmorse]

I guess with gains that low they need a bit of leakage to sound good..

Could someone explain why this might be the case? I don't really understand how leakage might effect the sound.

I wouldn't expect a vintage NKT fuzz face to measure as low as 4.5V on Q2's collector either. The one vintage unit I used to have had a crazy high voltage there, but that one was biased a little extreme...

What do you think would be reasonable\unreasonable to read here?

[LucifersTrip]

With a FF, you need 2 good low leak transistors, the fuzz sweep is mainly useable in the final 10-20% (which is why so many set it at full and use a pre-gain instead) and it can be mushy (which is why so many lower the size of the vol pot and/or decrease the size of the input cap.

Finding low leakage transistors is not hard.
Just get a couple of 2SB175s or 2SB172s. They can be really cheap. I'm having much more trouble finding leaky transistors for a MKII Tone Bender.

That's correct. It's not difficult to find low leak ones especially if you know what you're doing, but still more difficult than finding leaky ones, which are a dime a dozen...unless you are looking for a leaky version of a specific model # that's usually not leaky.  On the other hand, out of the 1000's of transistors I have, the least common would be higher (300ua+) leakage with low (50 hfe or less) gain.

To the OP: yes, it's easy to find low leak with hfe 70 and under. When you hit 80-100+, the leakage is generally higher with vintage stuff.
The Japanese stuff EW mentioned and more so, much of the Russian stuff is lower leak.

I don't think you must have low leakage transistors for a fuzz face anyway. The NKT275s were often low gain (around 50hfe is pretty common and they're usually closely matched) and medium leakage (100 to 200µA). I guess with gains that low they need a bit of leakage to sound good..

you don't need low leak for a FF, it just makes it easier, especially for a beginner.

OP has a pair of NKT275 - Q1 is 71 hfe 139uA leakage, Q2 is 117 hfe 242uA

You could trade him a low leak for his Q2 and use it in your MKII 

[Electric Warrior]

Yup, the higher the hfe, the higher the leakage. At least usually. And only when you compare transistors of the same type.

When you use hfes of 70 or higher they sound fine with low leakage. You can get away with a wider spread of gains as well. Lower gains need to be closer matched in my experience.
A lot of stuff is labled NKT275 these days. And there are very convincing fakes out there. There are genuine high gain vintage Newmarket NKT275s out there, but the low gainers are very common in Arbiter Fuzz Faces. They still sound very gainy.

I still have a couple of OC75s left for MKIIs and other vintage Sola Sound circuits and I'm pretty sure I could find more for cheap if I spent some time hunting them down.
I'm only trying to find cheap and available parts that compare well to them, so I can recommend them to beginners. 

[italianguy63]

+1

I generally see that too.. the higher the Hfe, the higher the leakage skews.  I just got some NKT275's and I would say they were just higher gain (re-labled) Tungsram AC128's.  I don't care, it is more important to get the general ratings you seek, and then pass the hearing test.  It does not matter what tranny package or labeling they have.  Basically a small piece of Germanium is a small piece of Germanium, IMHO.

MC

[Electric Warrior]

The Tungsram ones are supposed to be nothing like the original Newmarket devices, though.

Hfe and leakage specs are more important than type numbers for sure, but there are other specs that may matter as well.. Auditioning stuff sure helps. I keep using stock circuits and just swap transistors in and out using sockets. If you have parts with the right specs it should work great without any further modifications.

[lmorse]

Just returned home after working away for 10 days. The first thing I thought I'd do is re-check the hfe and leakage on my ge transistors.... I am measuring very low gain now - could I have blown them somehow?

Using RG Keen's method I am measuring an NKT275 as follows: .019v leakage with the base turned off, and measuring .032v with the base turned on. This seems stoooopidly low! Measuring a 2n2907 in exactly the same fashion I see 0v leakage (as expected) and 1.723v with the base turned on (resistor values are: 2.484k to collector, 2.715M to base). My battery is measuring 8.38v.



http://www.aronnelson.com/gallery/main.php/v/Schematics-etc/IMG_0980.JPG


I read somewhere that to test if a transistor is good you can measure the resistance across B+C and then reversing the clips to measure resistance across C+B... on my NKT275 I get the following: 123K : B+C, 583ohm : C+B.

[lmorse]

I've decided to buy some new ge transistors.

[lmorse]

A brief update in case anyone is interested....

My first lot of transistors were bought as NKT275's from ebay; that is what is stamped on them. I have 10 and couldn't get any combination of them to work. Since then I have bought several other types of germanium transistors, AC128k, ASX12D, GT308v... all of which give me the results that I expect. It seems that the original NKT275's are not really suited to this circuit, and I suspect are not in fact NKT275's - I got very little fuzz, low voltage on the base of Q1.... etc. - all the weird things mentioned in this thread.

Like I say, the other types of ge's work well, I have been able to experiment with combinations, even mixing types of ge for q1 and q2, tweak cap and resistor values to change tone etc. I really like the sound of the ASX12D's and will house these up!

I am going to experiment with the NKT275's, perhaps 3 or 4 stages might produce something useful?

[duck_arse]

try the nkt's in a darlington config, or two.