Transistor equivalency, 2SK170, BC547C, etc

[Bakula]

I see this has been discussed sort of with lots of disagreement and side bars...but no direct answer.

I’m building several mics based on a circuit that uses 2SK170BL, BC547C, and BC557C, which are obviously obsolete.

I’d like equivalents to all of these that I can add to a Mouser project parts list that I can just order when I need to build another one instead of wait a month for an ebay seller from China to ship.

I’ve looked at all the values and compared equivalent transistors, but obviously not all values exactly match up. I know a decent bit of what I’m doing but I’m by no means an electrical engineer and not sure which values are most important to match and such or which will produced negative results if unmatched.

Can anybody recommend equivalent transistors to the 3 I mentioned above that are in current production and will work as a replacement in a mic circuit? It’s essentially a Schoeps with some modifications.

I know I can get them on ebay or expensive legit remakes from boutique places, but I just want a similar transistor that will adequately do the same job and produce good results in a mic circuit, and that I can order from Mouser. And I just don’t know enough of the technicals to be sure.

Thanks!

 

[rackmonkey]

Are you willing/able to work with surface mount parts? For the JFET at least, that’ll be the only option for close to original noise performance at Mouser.

BC547 and BC557 can still be found from legitimate sources in the US and Europe

the problem with replacing the 2SK170 is that it’s got excellent noise characteristics and unusually high admittance (which contributes to its excellent noise characteristics). Just replacing it with something that works electrically won’t give you the same results. Your best bet today is an LSK170 from Linear. You can get those at Trendsetter Electronics.

I wouldn’t even consider buying JFETs such as 2SK170 from Chinese sources on eBay, unless you like wasting money. I’ve extensively tested many of those sourced from all over the world, and posted the results here. Rarely can you get real ones from Chinese eBay sellers these days.

 

[Bo Deadly]

Mmm, BC547C is no longer on Mouser. Bummer. Fortunately I think I once bought piles of those things. I think the parameters you want to look at is package and voltage of course, gain (Hfe), noise, generally what type of transistor it is (low noise amplifier vs switching for example) and maybe optimal source impedance. You can search Mouser for parts that meet those criteria. BC549C is a high gain, low noise TO-92 package. If I search Mouser for an equivalent, MPSA18 doesn't look obviously wrong. It's low noise, high gain, right package but it doesn't say anything about voltage or current noise figures so can't say anything about OSI. 2SK170BL is long gone. The ones on Ebay from China are likely fake but their performance is good enough to qualify as an equivalent part (the only difference I could detect was that the gain was ~4db lower but that's not a big deal for such a high gain JFET).

 

[rackmonkey]

I’d say you got lucky, Squarewave. Some of those parts I tested had curves/Idss/Vgs etc more like a 2n3819 or even a J109. Neither of which would work well if at all in most mic circuits. The ones from Israel were particularly bad.

 

[Bo Deadly]

I’d say you got lucky, Squarewave. Some of those parts I tested had curves/Idss/Vgs etc more like a 2n3819 or even a J109. Neither of which would work well if at all in most mic circuits. The ones from Israel were particularly bad.

I did see some of the China ones tested poorly compared to most. They definitely need to be tested and selected. But I had one in a mic circuit and I thought it worked perfectly. Then I swapped it out for what I decided was a legit one and it worked pretty much the same near as I could tell. I still use that mic and it works great.

Were the Israel ones from seller alweit? After some testing I decided those were actually legit. Of course that was a long time ago and when you buy something on Ebay you're rolling the dice every time.

We have talked about this before extensively and strangely I thought we were mostly in agreement. Maybe not.

Where to buy genuine Toshiba 2SK170 BL?
2SK170 on Ebay

 

[abbey road d enfer]

I think this discussion has two sides; one, the legalistic aspect, and the other, do these parts work?
In terms of bipolars, most of the transistors used in low-level audio are easily interchangeable. I would not hesitate one second replacing BC547C and BC557C with BC550 and 560 that are commonly available, thanks to ON-Semi.
A well-designed circuit should be capable of operating with transistors ranging from one end to the other of the tolerance range, which pretty much covers a 2:1 ratio if using grade versions.
FET's are a different case, because they are quite often used in circuits that rely on trustable specs, and the noise figure of two FET's that seem to work equally well otherwise (same Idss, Vp) may vary considerably.
As squarewave demonstrated in his comparisons, even FET's that have obvious discrepancies are capable of providing suitable performance, on the condition that the circuit is tolerant enough.
I use LSK170's in two products. One is an instrument input stage where the FET is embedded with an opamp; I could put almost any N-channel FET in there and it would "work".
The other product is an in-line phantom powered 20dB mic booster, designed to use LSK389 grade A. Grade B doesn't work, period.
I've tried the same circuit with LSK170's, that are the single version.
Again, the "A" version works, but not the "B". I suspected that from the start. But there is no practical solution for making this circuit work with the "B".
So, finding alternatives boils down to how tolerant the circuit has been made.
Now, I wouldn't replace 2SB737's with any old PNP of vaguely similar parameters...

 

[JohnRoberts]

No easy answer, but the hard (correct) answer is to compare data sheets wrt the specific parameters that matter to your design. This requires some rudimentary understanding of the design.

Swapping out JFETs can be a little more complicated due to variance between devices of actual pinch off voltages (Vgs). Data sheets only spec ranges of Vgs.

JR

PS: wrt (obsolete) 2sb737 there are modern devices that perform as well or better for mic preamps, since the 737/786 series was developed for MC head amps expecting around 10 ohm source impedance, vs low Z mics more like 150-200 ohm source impedance.

 

[Winston OBoogie]

The one thing about the 2SK170 and Linear's LSK170 is that you do, at least have options in terms of which grade of IDSS you use. Lots of other J-Fets that are available might indeed work in lieu of a 170, but the spread on them would require some sorting.

In a circuit where a J-Fet is used with active current sink/source etc, you would have leeway in being able to zero in to your target.

But, I'm assuming that your Schoeps type mic amp is using the J-Fet as a zero gain polarity splitter feeding your output transistors.
In which case, noise might not be an issue with a substitute part, but you'd have to be willing to experiment with changing the resistive loads etc. as there is a finite amount of current available from your 48V phantom in which to dial things in for adequate headroom/operating point.

Buying a few Linear LSK170 would be my choice if you just want more of a plug-and-play device.

 

[rackmonkey]

We have talked about this before extensively and strangely I thought we were mostly in agreement. Maybe not.

Not clear how quoting works in this new format.

I don't think we disagree, Squarewave. I just didn't post my complete tests in the thread where we talked about this. I agree with you that when they work in circuit, and you're satisified with the performance, it's all good regardless of what the actual nature of the part is.

BUT - I've gotten even more cynical since then. A kid in the 'hood who plays guitar asked me to build him a Rat pedal. You could argue that one could use any old 741-pinout, slew limited opamp (741, OP-07, LM308 etc) and get the same results with that circuit. But for reasons I won't get into, I decided I'd make him a pedal with the original LM308. I had 2 of those sitting around that I bought in the 90s, but I went ahead and tried the ebay China route just to see what I'd get (and because I didn't want to use up one of my last 2 known-originals). First batch - drop in circuit - nothing. Second batch; worked, but didn't quite sound Rat-ish to me. Third batch...nothing.

I tested batch 2 against my original LM308s by configuring them as voltage followers and running a square wave signal through them. Increase frequency until it's no longer a square wave, but rather a sawtooth/triangle wave (slew rate limiting). Roughly 120k difference. So it's a 741-compatible opamp, but it's not an LM308. Okay. What about the others? They didn't even work in circuit. Quick DCR measurements between VCC-Gnd, +/- inputs etc - not even close (no reading at all for each pin combination). So those aren't even 741 pinouts, whatever they are.

All that to say, the more I go back to that well, the more I'm convinced that the crap shoot for the average DIYer (who may not have the ability to test such parts) just isn't worth the money. If I'm giving advice to someone on the forum here, and I have no idea what level of skill they have, I'm just going to steer them away from the temptingly cheap but highly risky ebay China purchase.

 

[Winston OBoogie]

... the crap shoot for the average DIYer (who may not have the ability to test such parts) just isn't worth the money. If I'm giving advice to someone on the forum here, and I have no idea what level of skill they have, I'm just going to steer them away from the temptingly cheap but highly risky ebay China purchase.

+1

 

[abbey road d enfer]

But, I'm assuming that your Schoeps type mic amp is using the J-Fet as a zero gain polarity splitter feeding your output transistors.

No, the FET's are used in the basic common-source mode. Larger value of Idss puts them in triode mode with very little gain and very little Vds.
I've published the schemo somewhere in one of the threads.

Buying a few Linear LSK170 would be my choice if you just want more of a plug-and-play device.

I've sold all my PCB's to persons who shed the money for genuine LSK389A's.
I have built two with same. and kept only one.
I've done tests with genuine LSK170A's and EBay 2SK170's (very likely fakes). The LSK170A's work as expected, the "2SK170's" not. They can't be paralleled. They may be genuine but BL instead of GR.

 

[Winston OBoogie]

No, the FET's are used in the basic common-source mode. Larger value of Idss puts them in triode mode with very little gain and very little Vds.
I've published the schemo somewhere in one of the threads.

Ah, mea culpa. Thanks
In that case, even more jiggery-pokery would possibly be needed to substitute another J-Fet.

Edit: just to clarify - minus the cascode, I assumed the OP was wanting to use the 170 inside a mic a la Schoeps as shown below:

 

[abbey road d enfer]

In that case, even more jiggery-pokery would possibly be needed to substitute another J-Fet.

That's right. Some of the people who got my PCB asked me what FET they could substitute for the LSK389A and I just mentioned the LSK170A. I don't want to spend hours rolling FET's for finally giving an unsatisfactory answer.

 

[Bakula]

I appreciate everyone’s input.

I think I’ve got replacements figured out. I’ve ordered a few different kinds and will swap/test them and post the results when they come in.

I will post the schematic here also when I get a minute. Maybe some discussion/dissection/explanation would be helpful, not just to me, but others who are wading through the same learning curves. The information on this site is wonderful, but sometimes it gets lost in the disagreements and other segues.

The circuit is actually from a Blue Bluebird. Not the best mic I’ve ever used at first. I bought it 10 years ago, used it once, hated it, and never used it again until recently. I found when I changed a couple of caps and swapped out the capsule, it’s actually a very nice, flexible mic that works well for many sources. I needed a lot of mics fast and had plenty of capsules and bodies, so this seemed the quickest route.

 

[abbey road d enfer]

The circuit is actually from a Blue Bluebird. Not the best mic I’ve ever used at first. I bought it 10 years ago, used it once, hated it, and never used it again until recently. I found when I changed a couple of caps and swapped out the capsule, it’s actually a very nice, flexible mic that works well for many sources.

The capsule being the most significant element in a mic, changing the capsule is pretty close to changing the whole mic.
Makes me think of the old guy who proudly says: this knife I had for all my life, and it's still perfect, I only replaced the handle twice and the blade thrice.

 

[Rusan]

Has anyone tried buying LSK389s (or other Linear Systems JFETs) from NAC Semiconductor in Florida, U.S.A.?

https://store.nacsemi.com/products/detail?part=LSK389A-TO-71&stock=LIS000000001340

 

[Ricardus]

Has anyone tried buying LSK389s (or other Linear Systems JFETs) from NAC Semiconductor in Florida, U.S.A.?

https://store.nacsemi.com/products/detail?part=LSK389A-TO-71&stock=LIS000000001340

Yes. I built a few inline preamps designed by Abbey, and bought some 389s from NAC and they work great. A little pricey though.

 

[Bakula]

This is the schematic with original values. And yes, I know the layouts garbage. I had a PCB fab order going in that day and just needed to get the schematic down quick so I could make sure the PCB connections were right.

Like I said, this is a Blue Bluebird from about 2010-ish. I've actually seen a few people asking about the schematic and it's nowhere to be found, so here it is. I think I mostly understand how it works but any discussion/exposition of it would definitely be helpful to me and anyone else trying to learn this stuff without the extensive background some of you have.

My one other weird thing on this schematic, I couldn't figure out what a C205T diode was. My guess was a 20V Zener. Could find no datasheet with those labels though, and that's all the diode says on it.