What's the current thinking on muting transistors? What's still available? (2SC2878, etc)

Jump to Latest
C

cvanc

Member
Joined 2015
#1
So I need to replace 2SC2878 in muting circuits. In the short term I got a couple of (hopefully) NOS at a high price. But I'm curious to know if any current parts have been determined to work well as a drop in replacement?

Thanks...
 
#3
“Muting” transistors have a high Vebo rating so you can apply a relatively high signal voltage before driving them into emitter-base breakdown. Most ordinary transistors have 3 to 5 volt ratings, and can take 2x that before breaking down (and causing distortion). Transistors intended for muting can take 15, 20, or even more volts of signal in the off state. And have low ON resistance (knee voltage). A common ”fix”, however, is to simply put a 1N4148 diode in series with the base, then even one with a 3 volt Vebo rating can be used. I use the KSC1008, which has a 7 volt rating - and put a diode in the base. It’s not a “muting” transistor per se, but it has really low ON resistance and the diode fixes the breakdown problem in case it can’t take the output of a 5532 op amp.
 
  • Like
Reactions: KSC1845F
#4
Um what exactly are we talking about? In my shop, a muting transistor means a shunt to ground of the signal path by a transistor. Such as at the output of a preamp, to avoid a turn-on thump. The signal is not applied to the base Only a timed control is sent to the base.
 
  • Like
Reactions: ejp
#5
Enzo said:
Um what exactly are we talking about? In my shop, a muting transistor means a shunt to ground of the signal path by a transistor. Such as at the output of a preamp, to avoid a turn-on thump. The signal is not applied to the base Only a timed control is sent to the base.
Click to expand...

Why shunt to ground at all? Why not a simple switch that opens signal path? That's what happens when we unplug an interconnect after all.
 
#6
Are we discussing this scheme:

Line output with a series resistor

After the resistor, an NPN transistor with grounded collector and with its emitter connected to the signal path

Mute signal applied to the base, with enough base current during mute to drive the transistor into deep reverse saturation?

If so, I don't understand how a 1N4148 can help much against base-emitter avalanche breakdown. Even with the base open, when the emitter goes too far positive, avalanching in the emitter-base junction will provide base current for the transistor to turn on in reverse (emitter acting as the collector and the collector as the emitter). It could even go into reverse snap back.

The emitter-base junction will go into avalanche sooner when you drive the base negative than when you leave it open, is that the point? (Connecting the base to ground when not in mute would cause the transistor to turn on whenever the emitter goes below -0.6 V ... -0.7 V, so that won't work.)
 
#7
SoaDMTGguy said:
Why shunt to ground at all? Why not a simple switch that opens signal path? That's what happens when we unplug an interconnect after all.
Click to expand...
When you use some kind of transistor as a switch, having one terminal of it grounded usually leads to simpler circuits with less distortion. Besides, a shorting switch can speed up the settling of an output AC coupling cap, if the switch is connected after the capacitor.
 
#8
#10
Why shunt to ground at all? Why not a simple switch that opens signal path? That's what happens when we unplug an interconnect after all.
Click to expand...
At the output of a line signal, a transistor in shunt disappears when it is off, and shorts the signal path to ground when on. And it costs a penny or less in OEM quantities. A series switch needs to be controlled, so that means a relay or at least an analog switch IC. Those cost more than a penny, and work no better. Beyond that, I don't care to defend the engineers' choices in commercial products.
 
  • Like
Reactions: rayma
#11
I had the curiosity of testing Ge transistors:

Mute.png


With Ge transistors, it is not possible to leave the base open when the mute is inactive: the leakages cause distortion; light with the ASY73, heavier with the 2N1304.
With a reverse bias, it works perfectly.
The ASY73 is close to perfection: the ON resistance is lower than 10 ohm in the test conditions, and the offset voltage is <0.5mV.
The 2N1304 is not as good, but it is OK; the ON resistance is slightly higher, and the DC offset is 2.5mV. Perfectly usable

Edit: I have indicated 15Vpp on the schematic, but the voltage handling capability is in fact 15V peak, 30Vpp
 
  • Like
Reactions: bucks bunny
#13
MarcelvdG said:
The emitter-base junction will go into avalanche sooner when you drive the base negative than when you leave it open, is that the point? (Connecting the base to ground when not in mute would cause the transistor to turn on whenever the emitter goes below -0.6 V ... -0.7 V, so that won't work.)
Click to expand...
Yes, that’s exactly it. The transistor keeps itself off with no base current on the positive half cycle. It’s the negative half cycle that’s the problem. Muting transistors have a high enough Vebo to do it on their own. So do JFETs with a large negative bias, but their on resistance is an order of magnitude OR TWO higher - which means it doesn’t turn the volume down all the way. Fine if you just want to go on mute, but not good enough for actual source switching and isolation. A 10 amp mosfet would act like a low pass filter with all that capacitance. Put a forward biased diode in series with the base of an ordinary low on resistance bipolar, and you can drive the negative half cycle harder than you can the positive before having problems. You can also use +/- control voltage that way too, so if you’re driving the mute function from a spare op amp used as a comparator (ie, temp sensor) it works fine. Hi-fi manufacturers claim the diode trick causes problems with distortion but I haven’t seen any. Theres plenty of pro audio equipment out there using 2N3904’s and 1N4148’s, and have no issues with someone getting the gain structure wrong and hitting the input with full clip from the mixer, so it’s robust enough.
 
#14
I don't understand why they don't go in reverse snap back then. When you drive the emitter of a BC547 about 9 V (typical) above its base while the base is open, the avalanche current through the emitter-base junction should forward-bias the base-collector junction and turn on the transistor in reverse. The injected electrons in the emitter-base junction should cause further avalanche multiplication.

I think I have some 2N3904s in the attic, I'll measure what happens when the emitter is driven very positive with respect to the collector.
 
Last edited:
#15
Maybe a BC547 would go into some sort of snap back mode at 9 volts. Apparently 2N3904’s don’t or they wouldn’t be used. I was using MPSA06’s for years and never had issues holding back all an op amp can deliver (through a couple k ohms of resistance). I switched to the Jap-equivalent C1008 in the EBC basing option due to availability and bought 300 of them before they start to disappear too. They still work as expected. Lower noise transistors tend to have low Vebo in general, and perhaps more susceptible to other avalanche modes. Perhaps a blanket statement saying “any” transistor will work isn’t true, but jellybeans have worked and certainly anything I’d select for the purpose has. There doesn‘t seem to be any real need for 25 volt Vebo. It won’t hurt for sure, but when you can’t get them anymore you can’t get them, and something that works is better than nothing.
 
#16
The snap-back behavior is related to the variation of beta vs Ic, as is the Vceo to the Vcb ratio. Older planar transistors had significant surfaces leakages reducing the apparent beta at low Ic, and were prone to a strong snap-back effect.
Modern transistors benefit from cleaner processes, and are almost free from this effect. It can be simulated by shunting the B-E (or B-C) junction with a resistor of a few megohm.

Regular planar transistors have a large forward/reverse beta ratio, meaning they need to be heavily driven to ensure a low rsat under reverse conditions.

Special transistors are almost symmetrical.

Here are some more historical transistors registered for chopper applications:
2N941-6, 2N1640-2, 2N1676,7, 2N2232-37, 2N2356, 2N2432, 2N2474, 2N2968-71, 2N3317-19, 2N3343 -46, etc. (non-exhaustive)
 
#17
I took a 24 V supply, a 10 kohm potmeter, a 12 kohm series resistor and a 2N3904 in reverse with open base. When I slowly turn up the voltage, I see the voltage between the emitter and collector of the 2N3904 under test increase monotonically to 8.2 V. The base settles at 0.65 V. When I ground the base and leave the collector open, the voltage goes to 7.94 V with this 2N3904.

A second 2N3904 shows non-monotonic behaviour, but only just. The voltage between emitter and collector with open base peaks at 7.45 V and decreases to 7.35 V as I turn up the voltage further.

A third sample went monotonically to 7.81 V between emitter and collector with the base open. This one was from Motorola, the other two have a logo that looks like it's from National Semiconductor.
 
Last edited:
#19
A BC547B, probably a quite old one from Philips, peaked at 9.97 V emitter-collector voltage with open base and then dropped to 9.64 V as I turned up the voltage further. The base is at 0.59 V when everything is turned up all the way.

When you put a transistor with 7 V ... 10 V reverse emitter-collector breakdown voltage after an AC coupling capacitor, nothing weird happens when the voltage peaks are below 7 V ... 10 V, and when they are above 7 V ... 10 V, you get a DC shift due to rectification effects. If the next equipment suppresses the DC shift, you might not even notice it.
 
Last edited:
You must log in or register to reply here.
Top Bottom