Some rangemaster transistor testing

[mozz]

Finally starting to sort out my stash of old transistors. Hooked up some test equipment and was running some numbers. I do not know if the excel file will format properly. I have a few hundred but this is a sample i felt like running just to see the results. Not scientific but if you have any suggestions or test i can do please post.

Test stuff
Rangemaster clone /half breadboard. PNP version. 10k vol pot. Stock except 22uf used instead of 47uf, 5k pot in place of 3.9k resistor. Carbon zinc cheap battery running at 9.85 volts.

HP 54510A SCOPE
HP 339A DISTORTION ANALYZER
FLUKE 27/FM METERS
ATLAS PEAK DCA55

What i did was measure each transistor gain and leakage. Then it was placed in the unit and emitter resistor pot was adjusted to -7.0vdc.

Input was set to 1khz and 100mv, distortion was measured. Then 200mv, then 300mv. It was just a test made out of thin air. You can see how the transistor gain effects the output. There are some really low and really high gains i threw in there just to show the numbers for them. I have to go through my amps yet and see what is what, collector voltages, leakages, gains and how each has an effect on the sound. Always open to suggestions. There is a point of lowest distortion but it was at about -4.5vdc for most all transistors.

What i was looking for was the approx gain to give a clean sound with the guitar volume down, then when you bump it up, it goes into drive. I can do some screen shots, the distortion analyzer has a output that removes the original signal and shows you the distortion component.

I have other test equipment i should be able to measure the hiss. Once i get a amp set up and find some hissy ones, i want to see if i can find the noise with a freq analyzer or scope.

As i said open to suggestions, i can test at other frequencies, other input levels. I have a Syscomp usb scope unit i can also do freq sweeps.

[blackcorvo]

You could try a silicon MPSA92. I've used it's NPN counterpart (MPSA42) on a Brian May treble booster with good results. They're high-voltage devices, so they usually don't have much gain (avg. 130 hfe from what I've measured in my batch).

There's probably other high-voltage transistors with mid-to-low avg. hfe that you could try as well.

[mozz]

Just found a old circuit card (from which a lot of these came from) with Motorola SA106, pnp ge, gain of about 82-85, leakage is fairly high for the stash, about 200ua. Found 18 of them, they are pretty consistent so i like that aspect. Few 2N398 on the same circuit card, which are slightly lower gain at 55. These are all date coded 1962.

Another odd thing. 85 gain transistor, biased at -7.00vdc, 90mv input at 1khz, i have the vol so it puts out about 90mv (rms)on the scope. I have a .022uf on a switch across the input capacitor (.0047).  When i hit the switch, gain goes up to 140mv.

Now if i stick a 2n398 in there, gain of 55, i get about 74mv out. That in itself seems normal. But when i hit the switch, gain is now about 150mv. So lower initial gain but higher gain when i increase the input cap.

[blackcorvo]

hFE is the ratio between base current and collector current, where:

Ic / Ib = hFE

But just because one transistor has more hFE than another, doesn't exactly mean it'll have more gain if placed in the same circuit, with the same resistor values. Transistors, just like any other semiconductor, have a curve of operation, usually included in their datasheet. Depending on the transistor's curve, it might be biased for "better" or "worse" operation in the circuit intended for another transitor model. There's some wiggle room for similar models, but for OPTIMAL performance, it's always good to take that information into consideration.

For more information on common-emitter circuits:

https://www.electronics-tutorials.ws/amplifier/amp_2.html

[PRR]

> 85 gain transistor, biased at -7.00vdc, 90mv input at 1khz, i have the vol so it puts out about 90mv (rms)on the scope. I have a .022uf on a switch across the input capacitor (.0047).  When i hit the switch, gain goes up to 140mv.

90mV input signal "should" clip the heck out of the transistor.

BUT-- 0.0047uFd is awful small going right into a transistor. Your "bias" description is way incomplete. But assuming 1mA, hIE is like 30 Ohms, time hFE of 85 is 2.5K input impedance. A 0.0047u gives a low-cut from 13KHz, so WAY reduced. A 0.022u low-cuts from 2.8KHz, less reduced. Different hFE will change these relationships.

[antonis]

@Paul:
(maybe I'm wrong, but..)

Considering a "standard" Rangemaster circuit  with 470k/68k voltage divider bias resistors, 10k Collector pot (Boost) and 3k9 Emitter resistor (w 47μF bypass cap) we have a working current of about 110μA resulting in a h_ie of about 230R (negleting 47μF capacitive reactance at 1kHz) times h_FE of 85 -> input impedance about 20k, further dominated to about 14k5 due to voltage divider bias resistors shunt..

That said (I'm notorious for algebraic faults made by heart..  ) we've a High pass filter of 2.3kHz cut-off point exibiting about -10dB (again by heart..) attenuation at 1kHz..

For Boost setting all the way up, stage gain should be 10k/230R -> 43.5 (33dB), hence for 1kHz signal actual gain should be 23dB (Voltage gain about 14)
(ignoring Collector leakage current and output loading..)

So, mozz: An input signal of 90mV at 1kHz should measure 1.26V on output with 4nF input cap where should measure full clipping headroom (depended on your specific bias point) with 22nF 'cause HPF cut-off point is now set at about 318Hz..

That said (again), any transistor with h_FE greater than stage's particular gain (depending on input HPF & signal frequency) should behave ideally.. 
(not driven into "germinal" saturation..)

[PRR]

...maybe I'm wrong, but.. ... working current of about 110μA ....

I picked "1mA" from thin air. Taking your 110uA as probably right:

hIE is like 240 Ohms. hFE of 85 makes base impedance like 20K. Let's be lazy and ignore the resistors.
0.022uFd makes 317Hz
0.0047uFd makes 1,700Hz

Relative to a 1,000Hz test tone, the big cap passes most of the signal, the small cap passes like half the signal.

If you want to test the *transistor*, use a small (10mV) signal and a plenty-big cap (10uFd).

If you want to test the "fuzz", forget the numbers, play it.

[mozz]

So, with all these numbers, each transistor is not going to have a flat frequency response? Also, some transistors are going to be able to take a larger input signal before clipping. Maybe i should have bought the DCA75 with the curve tracer function.

[antonis]

We can't even have flat response with zero frequency (DC) and pure resistive components and you want it with reactive components and frequency domain involved..?? 

[PRR]

> with all these numbers, each transistor is not going to have a flat frequency response?

Not with these caps connecting ins and outs.

> some transistors are going to be able to take a larger input signal before clipping.

No. The input characteristics hew essentially to Shockley's Law, not to the device dimensions. 20mV significant distortion, 60mV gross distortion. To shift this you add resistors, several ways.