2N3773 Frequency transition

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
AX tech editor
Joined 2002
Paid Member
The transition frequency spec can be specified at different conditions. For instance, it can be the 'Ft' at the frequency where the Hfe has dropped to 1.
Or you can give the frequency where the gain has dropped to say 3dB.

So you need to look at the datasheet to find what the spec really is.

Jon is right it is obsolete for many years and veeeery slow. That means that the gain is a) low to begin with and b) seriously decreased by the time you get to 20kHz. Not good for an audio amp.

Jan
 
Ft wasn't a Jedec registered specification. What designers really needed to know, Jedec decided, was Vcesat. which is useless for audio, obviously.
When 2n3773 came out, the standard Ft was about 200 khz. However, my RCA datasheets from those days didn't publish Ft. Motorola wasn't giving away databooks in the seventies so I had to wait until I had $40 extra to buy a set from them in 1980.
These days a 2n3773 will be a modern parts like a MJ15003 or something which will have 3-4 mhz. So your part Ft will depend on how old the part is. As long as the part meets the 6 or so specifications logged with jedec, the vendor can call it 2n3773. Even if the modern part will cause oscillations on old boards if a 10 ohm resistor is not inserted in the base drive.
I have one surviving 1977 part, rumored to be 2n3773 on dynakit fan sites like greg dunn. The part number is rusted over. But I don't know how to measure Ft anyway.
To quibble with jonsnell, the sound of the old slow parts was notoriously bad on high frequency sources like top octave piano & tinkly bells. Somebody on here did a sim of ST120 with original parts and TIP3055, which confirmed the complaints of the reviewers of the ST120. Vacuum tube amps in 1971 really did sound better than solid state amps like the ST120 with 40406 . TIP3055 was a way faster part. I bought the ST120 with burned output transistors and my speakers weren't great at reproducing highs in 1983 anyway, so I can't speak from personal audition. Home Entertainment was using Simon & Garfunkel to demonstrate AR3, KLH5's etc, so who knew what the amp & speaker sounded like? Voice & guitar has no high frequencies to mess up. Lows, either, they sold a lot of bookshelf speakers that sounded "just as good" with that demo source.
 
Last edited:
If it’s newer than about 1980-85 it’s 4 MHz. Pre-1980 RCA versions (and some second sources) were about 200KHz. Motorola of any vintage was always 4MHz. Between 80 and around 85 RCA did continue to make hometaxials, but mostly for house numbered parts and a few types they had an exclusive on. By then they switched the 3773 to be like everybody else’s, even listing the 3773 on the same data sheet as their version of MJ15003.
 
No, but they are approximate anyway. Hfe falls to 1 somewhere around 4 MHz on a run of the mill epitaxial base power transistor. Could be anywhere between 2 and maybe 10 MHz at the peak, falling off at very high or low current of course. Hometaxial (or single diffused) struggle to get to 1 MHz, and can be spec’ed as low as 50kHz (ie, RCA 2N3442). On some old data sheets, f(hfe) was often spec’ed and defined as the 3dB down point. More common back in the Ge days.
 
No no no. The spec sheet just lists the JEDEC specs for a generic 2N3773 which call for a 200kHz minimum fT. Both the modern epitaxial version and old types will meet this spec. 4MHz > 200kHz, and no maximum is specified. Anything made now is around 4MHz - most of the time. It could however, be much lower and still be “good”, but it won’t be typical. If you did make an epi-base device and it’s fT came in at 200kHz, there is probably something else wrong with it and it would never make it out the door from ON.

4 MHz is not considered a high speed part anymore, and some 20 and 30 MHz parts approach the robustness of even the old slow RCA 3773’s.
 
In old design amplifier, original SD424 had 5MHz Ft and fix a limit for instability.
MJ15003 have a 2MHz Ft.
Recent 2N3773G offer 4MHz Ft, according to this thread and seems to be the best choice.
My beloved BDY58 with 10MHz present a risk and 2N6341 (40MHz) will lead to malfonction.
 
MJ15003 have 2 MHz minimum, no maximum it could be higher but not usually by a lot. They all use similar processing. There is not a lot of difference between them and 3773G. In fact I have been using 15004’s as 2N6609 substitutes for a while now - until the Mospec clones have become widely available again.

2SD424’s were what I used as 3773’s back in the 80’s. Very few distributors carried the full Moto line, and getting 6609’s was like pulling teeth. Consolidated and MCM which serviced the audio/video repair business carried a lot of Jap parts, so getting D424/B554 was easy. The only thing they wouldn’t work in was the PL700, which had to have the expensive and elusive MJ15024.

Using a 6341 in an audio amp (over about 40 watts) is a recipe for disaster. Their second breakdown capability is horrible. If you NEED that speed you’re better off re-tooling an amp to take TO-3P Sanken LAPTs. Or use ON 3281/4281, at a bit of speed penalty.
 
I never regarded the 2N3773 as an audio device, though RCA would have you believe it was. The only designs I've used a 2N3773 for are in D.C. power supplies.
But I agree, all recent (post1988) devices will be epitaxial base.
It's not too difficult to measure fT. You need a signal generator (not necessarily low distortion, <1% may suffice) which will work to 1MHz but deliver a reasonable output current, and an oscilloscope. Or possibly a sensitive AC voltmeter with high bandwidth. Possibly even a ferrite current transformer (but one capable of working at around 1MHz) to provide a low impedance load in the collector.
Bias the transistor in class A using current feed into the emitter (and lots of decoupling to ground), then feed a modulated base current (I built a modulatable bias current supply for this purpose, but you can feed a current into the base from a signal generator and moderate series resistor (higher value than the input impedance of the transistor) (as in an ordeinary capacitively coupled amplifier) then measure the gain at several frequencies and/or currents depending on how detailed you want your characterisation.
A measurement I made some time ago on a perhaps one of the last RCA 2N3771's gave a reading of 40kHz fhfe, compared with the spec. of ... well, no spec. but a much older, hometaxial base, measured 8kHz.
If you need to measure at 100kHz to get a reading it is a slow device, otherwise you should be able to measure at 1MHz and still have gain.
 
Last edited:
Thanks for the 6341 warning.

I play with a 405 clone pcb and it's a way to fill my knowledge gap between firsts jobs in broadcast studio and today.
As John said, 3773 was used in power supply and BDY58 as output transistors of amplifiers.

what references do you use in TO3 ?
 
Over the years I have used the following, for audio, either as replacement for original or some unobtainium device, or in something DIY.:

D555 / B600
D424 / B554
3773 / 6609
5630(31) / 6030(31)
5879 / 5881
5886 / 5885
3055 / 2955
5303 / 4399
15003 / 15004
15024 / 15025
15015 / 15916
21193 / 21194
802 / 4502
11015 /11016
11032 / 11033 ( yes, the 50 amp thing, as class G commutators)

NPN only:

D427
D873
3772 (both varieties)
3773 (hometaxial)
6254
6258
6259
6547 (driver service)
6308 (driver service)
RCA 1B05
RCA 40411

B407 (PNP Ge)
B449 (PNP Ge)

Of course I’ve used some types from the complementary list in same sex designs, and I’ve been known to mix and match (ie, 3773 with 6030 or 15004 or even 15016). This list doesn’t include house numbers, which I’ve often pulled out of stuff not even knowing its specs, and put it into service in something that I know won’t stress it based on its original duties. If it quacks like a duck and acts like a 3055, you use it as a 3055 or where one would work and don’t spend your 3773’s.
 
I guess the argument is that even the epitaxial 2N3773 isn’t “hi-fi enough”. By that argument neither is the MJ15024. I do t know, I’ve always been able to get decent sound out of most epi-base types, IF they are driven by good sustained-beta drivers or run in EF3, preferably with good high fT drivers. Some still will insist on Jap triple diffused or LAPT types, which will work well even in EF2. But I’ve always had a devil of a time finding those in TO-3 even back in the day. Flatpacks all over the place with fT’s as high as 80 MHz, but very scarce TO-3’s. True the D424/B554 are triple diffused but they are still around 5 or 6 MHz, and are closer to a MJ15024 than to a C3281. The D424 would routinely pass the 2N6259 100V/2.5A second breakdown test. Which is why I really liked the damn thing. PL400’s (or even CS800’s) fitted with D424’s just wouldn’t blow again.
 
I incidentally tested some 2N377x for Ft, and here are the results (Vce=5V, Ic=300mA, Ft non-extrapolated (Ft=Fm)):
2N3771 Motorola 1993: 3MHz
2N3772 Motorola 1993: 1.15MHz
2N3772 Tesla 1989: 12MHz

The lower Ft for the 2N3772 is logical, and the 12MHz of the Tesla is quite surprising, but do not rush to find them on Ebay: the rest of their characteristics is awful.....
 
I incidentally tested some 2N377x for Ft, and here are the results (Vce=5V, Ic=300mA, Ft non-extrapolated (Ft=Fm)):
2N3771 Motorola 1993: 3MHz
2N3772 Motorola 1993: 1.15MHz
2N3772 Tesla 1989: 12MHz

The lower Ft for the 2N3772 is logical, and the 12MHz of the Tesla is quite surprising, but do not rush to find them on Ebay: the rest of their characteristics is awful.....


The saying of that era was "the faster the transistor acts the faster it will die". There was some truth in that - until the lateral FET entered the scene;)
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.