BS170 issues

[M.Spencer]

Hi folks, New builder first time poster. I built three SHOs from 3 tagboardeffects layouts, 2 on breadboard one on eyelet, but out of the 50 our so BS170s only two sound good, the others make the circuit sound garbled and gross. I got them all from an estate sale lot. Of the two that work well my peak atlas dca55 says the VSG is.90 and 1.23; all the others have measured 2.0 or above (all of which seems within the datasheet parameters). I've read this type of transistor is highly susceptible to static damage.
My question is would I have destroyed the lot by keeping them loose in a plastic storage compartment? More likely that they were toast before I bought them? Or what would cause only 2 out of 50 to work in the SHO? Any input appreciated.

Thanks

[R.G.]

If the storage bag was not semiconductor type conductive plastic, yes, the storage and resulting static electricity when they were moved around could have killed them. Maybe.

Try measuring the drain voltage of the BS170 in circuit and powered for both good and bad examples. How does that compare?

[M.Spencer]

Hi R.G., thanks for replying. I tried doing as you asked, not sure if I did it correctly (imagine the opening scene in 2001 Space Odyssey...that's how I feel when working with electronics so far).

On my eyelet circuit I tested the two "good" ones with the DMM hooked up to the ground lead to ground and the positive lead to D, G and S...but with the circuit not engaged:

1st good bs170 (off)
S .21
G 1.18
D 2.9

2nd good bs170 (off)
S .25
G .98
D 2.21

Now some of the "bad" ones, first with the circuit disengaged...and then I figured it would be helpful to try with the circuit on::
off/on
S .11 / .06
G 2.2 / 1.4
D 4.9 / 4.26

S .1 / .04
G 2.15 / 1.4
D 4.9 / 4.28

S.59 / .06
G 2.19 /1.4
D 5.92  / 4.25

S .24 / .02
G 2.16 / 1.4
D 5.19 / 4.26

Then I went back and tested the good ones with the circuit off AND on:

off/on
S .09 / .05
G 1.32 / 1.15
D 2.9 / 2.64

S .12 / .07
G .91 / .78
D 2.0 / 1.75

So some of my calculation might be off due to my inexperience (again, ape, monolith, smashing bones, a rousing Strauss score  ), but I am not sure why I got different measurements the second time I measured the good bs170s. Does running current through them change their "resting" values?

Either way can I take it that it's a good Rule of thumb to not store transistors loose in a storage bin like this?:

[bluebunny]

Either way can I take it that it's a good Rule of thumb to not store transistors loose in a storage bin like this?:

It's only the MOSFETs that are particularly delicate.  All your BJTs and JFETs will be quite happy.  If you have any of those pink anti-static bags, put your (good) MOSFETs in one of those, then put the bag in the drawers.  If not, I'm sure some kind soul would send you a couple.  Where are you located?

[Brisance]

You can always also store them in aluminium foil and then a baggy

[samhay]

Try breadboarding one of your 'bad' MOS-FETs with a trimmer in place of the 5k1 drain resistor. Looks like you like them biased a little differently than most - one would usually aim for roughly 5V on the drain, so your 'bad' voltages look reasonable if you have a 9V supply.

Note that the position of the 'crackle' pot will effect the voltages - perhaps best to set it to minimum - i.e. with 5k resistance to ground.

As a side note, I am a little puzzled about what happens when you 'engage' the circuit. You obviously have current flowing through the FET when not engaged (or else you would not be measuring any voltage), so what happens when you 'engage' that is loading down the gate?

[anotherjim]

Tag board? Are you soldering them in? Your soldering iron could be killing them if it's not got a well grounded tip.

[M.Spencer]

Where are you located?

Portland, OR

[M.Spencer]

Try breadboarding one of your 'bad' MOS-FETs with a trimmer in place of the 5k1 drain resistor. Looks like you like them biased a little differently than most - one would usually aim for roughly 5V on the drain, so your 'bad' voltages look reasonable if you have a 9V supply.

I'll give that a shot. As far as liking them biased differently, not sure I'm doing that on purpose...I'm just plugging components in based on the schematic/tagboardeffects layout. Yes, I am using a 9v battery.

As a side note, I am a little puzzled about what happens when you 'engage' the circuit. You obviously have current flowing through the FET when not engaged (or else you would not be measuring any voltage), so what happens when you 'engage' that is loading down the gate?

I am wondering this myself. I figure when the 3pdt switch is engaged it sends the signal to the circuit, but regardless the battery/9v is always going through the circuit whether it's on or off (if my guitar cable is plugged in of course). Am I right in this thinking?

[M.Spencer]

Tag board? Are you soldering them in? Your soldering iron could be killing them if it's not got a well grounded tip.

Sorry, not tagboard, I'm not sure what it's called. I built the circuit on something that's like vero/stripboard, but none of the copper traces are linked together. Guess it's not called eyelet.

But either way I am using sockets for the transistors so it's not likely my soldering iron.

[davent]

Tag board? Are you soldering them in? Your soldering iron could be killing them if it's not got a well grounded tip.

Sorry, not tagboard, I'm not sure what it's called. I built the circuit on something that's like vero/stripboard, but none of the copper traces are linked together. Guess it's not called eyelet.

Perfboard?

[bluebunny]

I am wondering this myself. I figure when the 3pdt switch is engaged it sends the signal to the circuit, but regardless the battery/9v is always going through the circuit whether it's on or off (if my guitar cable is plugged in of course). Am I right in this thinking?

Yep, the 3PDT switches the guitar signal, not the power.  The power is usually switched on by inserting the input jack.

[highwater]

If your MOSFETs work (meaning the circuit passes audio through, whether it sounds good or not), then they probably haven't been zapped. You are right to be careful with them, as they are very sensitive, but my (amateur) understanding is that if that was your problem they probably wouldn't work at all.

However, there is a second, arguably more important, difficulty in dealing with MOSFETs: large variation in threshold voltage - that threshold voltage being what your meter is measuring when it tells you the "VSG". Those readings indicate that your two transistors which "sound good" simply have a lower threshold voltage than the ones that "sound garbled and gross", and your in-circuit voltages are telling the same story. This article on R.G.'s site explains quite-well why this is a problem; the fifth diagram (the one labeled "MOSFET Gain Stage - Bias Network") is closest to the SHO, and the next three diagrams with their associated text explain what you can do to have a good chance of making the "bad" MOSFETs sound good. Once biased properly, you might still hear subtle differences in sound versus the , and once you make that effort, you could even find that you prefer the sound of the MOSFETs that didn't bias right with the circuit as it currently stands.

[M.Spencer]

Perfboard?

Bingo!

[M.Spencer]

If your MOSFETs work (meaning the circuit passes audio through, whether it sounds good or not), then they probably haven't been zapped. You are right to be careful with them, as they are very sensitive, but my (amateur) understanding is that if that was your problem they probably wouldn't work at all.

However, there is a second, arguably more important, difficulty in dealing with MOSFETs: large variation in threshold voltage - that threshold voltage being what your meter is measuring when it tells you the "VSG". Those readings indicate that your two transistors which "sound good" simply have a lower threshold voltage than the ones that "sound garbled and gross", and your in-circuit voltages are telling the same story. This article on R.G.'s site explains quite-well why this is a problem; the fifth diagram (the one labeled "MOSFET Gain Stage - Bias Network") is closest to the SHO, and the next three diagrams with their associated text explain what you can do to have a good chance of making the "bad" MOSFETs sound good. Once biased properly, you might still hear subtle differences in sound versus the , and once you make that effort, you could even find that you prefer the sound of the MOSFETs that didn't bias right with the circuit as it currently stands.

Fantastic, thanks for the explanation and pointing me to R.G.'s article. I'll try biasing them differently and see how that affects their sound. I had been under the impression that while Ge transistors were temperamental and prone to varying hFe levels, Si transistors were more stable and thereafter would have a lower tolerance. So MOSFETS are fairly stable but canvary greatly from component to component?

[newperson]

post some images of your work

[M.Spencer]

post some images of your work

[bloxstompboxes]

The tips on your jacks look like they are touching right there in the center. Are they?

[davent]

I've never seen BS170's in a can, have i missed something?

The tips on your jacks look like they are touching right there in the center. Are they?

If they're not touching in the photo what about when you insert the plugs?
dave

[Thecomedian]

I've never seen BS170's in a can, have i missed something?

The tips on your jacks look like they are touching right there in the center. Are they?

If they're not touching in the photo what about when you insert the plugs?
dave

I've done that before. What you do is route them so that there's a quarter inch gap or more between input and output plugs, and turn them so that the long, or sleeve, side of the jacks are at a 120 degree or larger angle from each other. you can also work a stereo jack, used as the on/off switch, into the pedal, by making sure that shorter sleeve tab goes between the two lugs on the output jack. I have I think its a 1590A that can fit two jacks, perfboard circuit of a buffer with non-SMD components, and a battery, without any shorts. If the washers and nuts on the jacks come loose, they would easily swivel and short.

If your MOSFETs work (meaning the circuit passes audio through, whether it sounds good or not), then they probably haven't been zapped. You are right to be careful with them, as they are very sensitive, but my (amateur) understanding is that if that was your problem they probably wouldn't work at all.

However, there is a second, arguably more important, difficulty in dealing with MOSFETs: large variation in threshold voltage - that threshold voltage being what your meter is measuring when it tells you the "VSG". Those readings indicate that your two transistors which "sound good" simply have a lower threshold voltage than the ones that "sound garbled and gross", and your in-circuit voltages are telling the same story. This article on R.G.'s site explains quite-well why this is a problem; the fifth diagram (the one labeled "MOSFET Gain Stage - Bias Network") is closest to the SHO, and the next three diagrams with their associated text explain what you can do to have a good chance of making the "bad" MOSFETs sound good. Once biased properly, you might still hear subtle differences in sound versus the , and once you make that effort, you could even find that you prefer the sound of the MOSFETs that didn't bias right with the circuit as it currently stands.

Fantastic, thanks for the explanation and pointing me to R.G.'s article. I'll try biasing them differently and see how that affects their sound. I had been under the impression that while Ge transistors were temperamental and prone to varying hFe levels, Si transistors were more stable and thereafter would have a lower tolerance. So MOSFETS are fairly stable but canvary greatly from component to component?

From what I understand, Ge variations and low gains are based on the designs and fabrication while the industry was young. Si are better because quality control is better, advances in production, and so on. They still make new Ge transistors, you can even buy them from radio shack, so they're not really "bad" because of the material. The old ones have/had "mojo" due to being leaky, the type of circuit, and so on. An old FF that sounded amazing to someone would be produced alongside 100 others that sounded awful because of the biasing compared to the transistor's wide variances and leakage. Samples of how you'd rebias a circuit for specific gains of old Ge transistors can be seen on smallbearelec. I bought one that I believe is AC128s, and the 100 meg feedback resistor is somewhere around 150, with the biases being 22k and so on. Much different, but it was tweaked for the gains of those transistors so it was still nice and fuzzy. Messing around with fixed bias amp circuits for BJTs taught me how important the characteristics of individual transistors are, how they affect the bias due to

http://www.diystompboxes.com/smfforum/index.php?topic=68796.0;prev_next=prev

The big issue here is that Idss is different from JFET to JFET of the same type. So is Vp. So is the transconductance, the ratio of how much a change in the gate-source voltage changes the drain current. These are all interrelated by the device physics, which is nice, but in a way that makes it very complex and variable from device to device as to where a given bias network will make the currents and voltages settle.  

Are mosfets as variable as Jfets? I don't know. I notice you're using a red LED in the circuit, is that in place of the 9.1v zener diode?