How to Measure and Transistor Leakage and Gains for Germanium, Silicon, and Sziklai Pairs

How to determine leakages and gains with Germanium Transistors, and how to improve upon them by using Sziklai Pairs.

[@paulepruss]

Electronics theory, practical application knowledge AND history. What more could an aspiring pedal punk want? Great job & thanks!

[@freeelectron8261]

Really interesting and useful. May the fuzz be with you!

[@Uraim]

im about to test my old transistors but i dont have that willy nilly multitester, this will do the job!

[@EDGARDOUX1701]

Great video, learned. Hope to see more! Be safe

[@flaconsius]

Really great Video. Super easy to understand and the subject is very usefull in designing/ making Pedals. Thanks a lot.

[@the_nondrive_side]

Yes this the video I've been waiting for. I've got a 1k resister between the E and B from 3904 3906 which would be the "complimentary pairs" but in the Sziklai Germanium Silicon hybrid you kinda want to tame that HFE like the last video you had but with Variable Resistor and the only other controls would be Bias and Volume.

[@jighead81]

Thanks Paul and co!

[@mathiasohlson6265]

Great, thanks! Finally understood the R G Keen paper...

[@tommyfoley]

Great video.

[@jeremywinnett6352]

Love the vid. One of the best yet but... Please opt out of the additional adverts. It's a change in KZbin policy and it suuuucks. Preroll and post roll only pleeeeassseee.

[@jeremywinnett6352]

Also, if you didn't know, they have applied this to all previous videos EEVBlog did a good vid about it.

[@fer_fdi]

excellent! thank you so much

[@erikvincent5846]

Your welcome

[@aldadanger802]

Hi mate, when you explain a germanium for a gain.. i still dont inderstand where you can get 4uA to get 0.9888v.. can you explain it please?

[@ivolol]

You could put an ammeter in series with the 2.2M to measure the microamps, or put a voltmeter across it to measure through ohms law. Although be aware most DMM volt measurements have a 10M resistance themselves, which lowers the resistance at time of measuring. Lastly you could also work backwards after measuring the main channel current and the effective gain properly

[@untro7287]

Great video! But I may need to come back after year two of wizard school...

[@theczarofguitar2109]

Do you find that you have to wait sometimes for voltage readings to "settle" when measuring germanium transistors? I'm measuring transistors and it seems to take a few minutes for voltage readings to drop to a steady voltage.

[@soulrobotics]

I wonder what would be the frequency response...

[@ChemXFX]

Excellent video! Question: if you’re using two germanium transistors in a Sziklai pair do you just add the leakage of both together to get the total leakage or is there some fancy math you need to do to figure it out? Also I would love to hear your thoughts about using a Darlington pair to increase gain in two germanium NPN transistors?

[@erikvincent5846]

So, in the case of either, it's not that they add, but rather, they multiply leakages. However, it isn't a "straight multiplication" either. But this is why a silicon germanium pair works great, because a leakage of near 0 from the silicon can greatly reduce the leakage of its germanium in the set. The reason we go more with PNP Germanium is because they were more common, and they lasted longer than their NPN counterparts. Most NPN Germaniums that you find now are either leaking too much or their gains are abnormally low, which is why they were B stock transistors in the first place to become NOS.

[@soulrobotics]

Since the currents measured are minimal, you should take care on the design. You can play a lot with Spice, but in the real life, breadboards has a lot of stray capacitance and if you left floating in the breeze the base of the transistor, like an antenna, you are making a very good electrostatic receiver... it'll capture whatever glich or noise is on your bench,... even the static electricity of your body. I suggest to make the circuit in a proper board. also, put a 47nf (the less ESR as possible) between the "ground" and the base. In that way you can get rid off a lot of noise in your bench. The last thig is that you let all the transistors on the bench for a few hours and the use ceramic tweezers to plug it into the socket. your body temparature will modify the reading for sure! Take note of the ambient temperature when fill up your spread sheet with the data. this are my two cents.

[@frederickshelborn6911]

Thanks for the informative video. I'm an 'absolute beginner' (I used a breadboard 30 years ago for a week in college). I want to take on the task of building a Tonebender MKII Pro. Possibly with OC81D's, maybe OC75's, or 77's. I have read that the actual Ge transistor model doesn't matter as much as matching the three. Supposedly, matching meaning there is a magical combination of gain & leakage ranges for each of the three Ge positions in the design. I'm "guessing" leakage doesn't matter, but it's the TOTAL GAIN (as you demonstrated by subtracting out leakage) that needs to be magically matched. I will experiment and use my ears, but have you heard of this? Do you know where I can source the info on what gain each of the three should be? I believe mismatched gain on the three is how the tone bender gets it's great sound. This could be a fun rabbit hole to go down. THANKS for any tips!

[@erikvincent5846]

For a Solasound TB MKII Pro, there is a little bit of fiddling required, but nothing too ridiculous. The first thing to note is the first transistor (input stage). It will likely have a resistor that connects to the base of the transistor and goes to ground. It is a resistor to divert excessive base-current which is caused by leakage from the base to the collector a Germanium transistor. On PNP Germanium Transistors, like the AC125, AC126, AC128 or OC81D, the value of the resistor should be around 100K. However, if using an OC75, it should be reduced to around 10K. The first transistor, usually you are targeting a 50-70 hfe transistor that is slightly leaky. Something like between 100μA and 400μA. The second transistor needs to be at around the same gain, so 50-70 hfe, but you want it to be the one with the least amount of leak. Try to have it under 100uA if possible. The third and last transistor is the one that needs the higher gain, somewhere between 100-125 hfe, and it should be slightly leaky as well, between 100μA and 400μA. When the ToneBender was being developed, they took leak into consideration into the design.

[@frederickshelborn6911]

@@erikvincent5846 GREAT info! Thank you. This will be a big task for a first build.

[@erikvincent5846]

@@frederickshelborn6911 Let me know if you need help. I have an unreleased PCB called the Clone Bender. It builds a Tone Bender MK2, plus several other PNP Germanium based fuzzes and has an inverter onboard. If you are in the US, just hit me up and I can send you one, along with the build doc.

[@J.C...]

I assume I can use resistors in series to make up 2.472k and 2.2M, right? It makes no difference electrically.

[@xplorer3475]

What if my battery is 9,5 volts instead of 9 volts ?

[@JM-ym8mm]

Do multimeters that have an hFE setting account for leakage in a germanium transistor? Just got my first germanium transistor and plopped it into my cheapo multimeter, being careful of the pinout ( they didn't make it easy for you back then :D) and it came out at 80. It's a GT402A. If it isn't leaky, that should be pretty good I think.

[@erikvincent5846]

multimeters don't have anything for hFE, and for Germanium it would be tricky as leak calculations are needed to subtract from the measured hFE. But, as shown in the video, you can measure leak and gain with a multimeter and breadboard and subtract.

[@JM-ym8mm]

@@erikvincent5846 My multimeter has 2 sets of 4 holes labelled, ECBE for NPN and ECBE for PNP. I guess they make them that way to accommodate all sorts of pin outs. Anyway, I only ask because I have a germanium transistor that I wish to use and the multimeter says it has a gain of 90, which would be pretty awesome if leakage wasn't a factor in germaniums. ;) I'll have to dig out my old breadboard and see

[@erikvincent5846]

@@JM-ym8mm ah, I think I might have seen chip/component identifiers that have that. I haven't seen any of those that take leak into consideration, though. Not saying they don't exist, but I haven't seen them before. I did make an Arduino circuit that does do that, so it is possible. We made a video about that a bit ago.

[@JM-ym8mm]

@@erikvincent5846 indeed I think you are correct. I would assume that they build multimeters with today's silicon chips and transistors in mind, which don't have the same leakage issues I believe. It would make sense for them to ignore it seeing how the germanium transistors we buy nowadays are new, old stock. Thanks for replying, was good picking someone's brain about it!

[@xplorer3475]

Please, what does it tell when the voltage value keeps decreasing and doesn't really stop ? i don't have a fixed value to endly make the calculation and get the total gain ..

[@paulepruss]

Could it be that you have used you fingers to insert the transistor and it hasn't temperature stabilised yet? Or maybe your power source is sagging?

[@xplorer3475]

@@paulepruss thank you . it's a basic 9v battery.

[@paulepruss]

@@xplorer3475 I'm no expert, just learning like you but - It's a simple set-up so check all of your components for the correct value. Re-seat all of your components in the breadboard and/or try a different transistor to see if the voltage stays steady. try soldering or using jumper clips to connect the circuit to rule out the bread board. Try to find a 120/240vac to 9vdc power pack to give you a constant 9vdc source. Good luck!

[@xplorer3475]

@@paulepruss i observed that the difference between with and without 2M2 remains the same, whatever the oscilation goes to. so it means : checking each voltage quickly thanks to the switch, then do a substraction, and then you have the total gain.

[@sebastianmattsson9816]

so... If I measure a leakage of just under 9 volts it's pretty safe to say that transistor is pretty dead, right?

[@erikvincent5846]

Yeah, that's not going to work very well.

[@buzzcrumhunger7114]

Are you running a de-esser on your mic?

[@erikvincent5846]

Nope. On screen capturing, just an SM58 with a pop filter straight into my Scarlet 2i2 interface. From there, into Bandicam. Otherwise, if using the Sony Cam, it's just that.

[@buzzcrumhunger7114]

Maybe try a de-esser. You’ve got quite the whistling ‘s’ going, there.

[@erikvincent5846]

@@buzzcrumhunger7114 on the Bandicam screen capturing or on the regular garage cam scenes?

[@buzzcrumhunger7114]

Erik Vincent I dunno about Bandicam, my daughter insists OBS is better but can’t tell me about plugins. I don’t know what your post editing process is like, but wherever in your chain you would use plugins. It’s loudest in the 1st section with the DMM, but noticeable in the graphics presentation section as well. That’s about as far as I got.

[@erikvincent5846]

@@buzzcrumhunger7114 I can look into it. Might be a plugin within Sony Vegas.

[@ivandel7788]

I think I'm going to build a sziklai Big muff 🧐

[@bgcurbani]

Wait a minute, that's how the Fuzz Factory works with PNP and NPN at the same time?

[@ErebosGR]

I think so. Although the Fuzz Factory has 2 stages of Germanium transistors after the Silicon IIRC.

[@jimbucket2996]

Can you explain that in English please?😉

[@erikvincent5846]

If a I could a speak a da English!

[@user-fp8eh6mq8p]

Video is long and slow Here's the PNP instructions R1 2.2M at B R2 2.472K at C 9v: - at R1/2 ends and + at E DMM: DC V across R2 Note V w/ R1 (X Volts) Note mV w/o R1 (0.XX Volts) subtract small # from big and x100 = HFE