Component Testing Is Breaking My Brain

[PedalBuilder]

Can I ask how one might reduce the voltage coming from a 9v battery leading to a Keen tester? I'm using a rechargeable lithium 9v and it's usually at around 9.3v... Trying for as close to 9v as possible. I've seen it mentioned to insert a 10k pot (where exactly?), but I've also seen warnings that it could burn out the wiper... Would love some tips!

@PedalBuilder I swear I've taken your advice to chill out about this stuff to heart! Ha... Still, this Keen tester is the best chance I have to get ballpark figures, and feeding it more than 9v may skew the results too much.

The TC1 that you have should be fine for ballpark figures. Really. But if you really want to make the Keen tester and are dead set on having exactly 9 volts, you can set up a voltage divider to do that. If you're getting 9.3 volts from your battery, then a 68k/2M2 or 75k/2M2 voltage divider will get you 9 volts.

 

[Ginsly]

The TC1 that you have should be fine for ballpark figures. Really. But if you really want to make the Keen tester and are dead set on having exactly 9 volts, you can set up a voltage divider to do that. If you're getting 9.3 volts from your battery, then a 68k/2M2 or 75k/2M2 voltage divider will get you 9 volts.

Gotcha, thanks so much PB!

Let’s say I actually did install a pot like this:

Would I need the pot on the black lead instead of the red if I was testing a PNP?

 

[PedalBuilder]

Gotcha, thanks so much PB!

Let’s say I actually did install a pot like this:
View attachment 72082
Would I need the pot on the black lead instead of the red if I was testing a PNP?

Well, to start with, you'd want to have a resistor in series with the pot so that you don't burn it out when it's set to a minimum resistance. And you are fine using a TC1. I feel like I can't emphasize that enough—it's just as valid a tool for ballpark hFE/leakage figures as the Keen rig, and it's much easier and faster to use. And yes, you would want to swap the battery terminals if you're using PNP transistors. But I think you still really need to just chill out about transistor testers. Lots of people, including some people who do this for a living, use the TC1 to test transistors. It works. Instead of undertaking a quest to build the perfect RG Keen transistor tester circuit, just use the TC1. Or better yet, breadboard the circuit and see how the circuit sounds with different transistors—that's what really matters.

 

[Ginsly]

Well, to start with, you'd want to have a resistor in series with the pot so that you don't burn it out when it's set to a minimum resistance. And you are fine using a TC1. I feel like I can't emphasize that enough—it's just as valid a tool for ballpark hFE/leakage figures as the Keen rig, and it's much easier and faster to use. And yes, you would want to swap the battery terminals if you're using PNP transistors. But I think you still really need to just chill out about transistor testers. Lots of people, including some people who do this for a living, use the TC1 to test transistors. It works. Instead of undertaking a quest to build the perfect RG Keen transistor tester circuit, just use the TC1. Or better yet, breadboard the circuit and see how the circuit sounds with different transistors—that's what really matters.

Oh, I’d MUCH rather not deal with this tangled breadboard tester! Ha…

Unfortunately the TC1 just doesn’t seem to give very useful Ge readings, as far as I can tell.. Could be wrong. All of this is really in service of getting “good enough” hfe and leakage for Ge.

But even with Si, when the Keen test or multimeters say 190 and the TC1 says 285, that’s a pretty big gap. I think your point is that even if there IS that big a discrepancy is just really doesn’t matter until you pop em in and LISTEN.

 

[PedalBuilder]

Oh, I’d MUCH rather not deal with this tangled breadboard tester! Ha…

Unfortunately the TC1 just doesn’t seem to give very useful Ge readings, as far as I can tell.. Could be wrong. All of this is really in service of getting “good enough” hfe and leakage for Ge.

Why don't you think that the hFE/leakage readings aren't useful?

But even with Si, when the Keen test or multimeters say 190 and the TC1 says 285, that’s a pretty big gap. I think your point is that even if there IS that big a discrepancy is just really doesn’t matter until you pop em in and LISTEN.

Why do you think that the Keen hFE reading is a more appropriate metric than the TC1's hFE reading?

 

[Ginsly]

Why don't you think that the hFE/leakage readings aren't useful?

Why do you think that the Keen hFE reading is a more appropriate metric than the TC1's hFE reading?

That’s a good question. It seems like the Keen tester’s readings lean closer to the datasheets, at least with Si.

With Ge, the Keen results have been much closer to the hfe/leak numbers provided by the kind folks who have sent me a few. The TC1’s hfe readings don’t seem to take leakage into consideration, either.

Prowling around here, it seemed like most people don’t consider the TC1 reliable for Ge. That may be completely wrong, though! I’d much rather use it for quick checks/pinout on absolutely everything than dealing with separate testing setups.

Still very, very new to all of this and trying to learn best practices. I really, truly take your advice to heart!

 

[PedalBuilder]

That’s a good question. It seems like the Keen tester’s readings lean closer to the datasheets, at least with Si.

That doesn't mean that it's a better tester; it just means that it is testing the transistors at a current that is closer to the current that is described in the data sheet. As far as building a pedal is concerned, that's not a reason to trust the Keen tester over the TC1.

With Ge, the Keen results have been much closer to the hfe/leak numbers provided by the kind folks who have sent me a few. The TC1’s hfe readings don’t seem to take leakage into consideration, either.

Does the TC1 list an ICEO or ICLEAK value? If so, that's the leakage. If it does not include that, then you are probably correct that it doesn't account for leakage.

Prowling around here, it seemed like most people don’t consider the TC1 reliable for Ge. That may be completely wrong, though! I’d much rather use it for quick checks/pinout on absolutely everything than dealing with separate testing setups.

It depends on the firmware. Some versions of the TC1 firmware (i.e. ones that have an ICEO or ICLEAK measurement) are perfectly good tools for measuring germanium transistors. Others are not.

 

[chris.knudson]

I have a TC-1, a DC-55, a DMM -- none of them are perfect, but they all serve a purpose. One thing I never do though is compare apples to oranges. If I'm using one meter to measure hfe, for example, I stick to that meter. Sure there are some meters that can measure a lot of different measurements on components, but a lot of them do it in different ways, so I just pick a meter and use that to baseline my measurements.

 

[Ginsly]

Does the TC1 list an ICEO or ICLEAK value? If so, that's the leakage. If it does not include that, then you are probably correct that it doesn't account for leakage.

It does list Iceo, but I don't think it takes that into account when measuring hfe, resulting in an inaccurate gain reading for Ge (according to some). Plus, reading through threads here gives me the impression that many people here don't trust a TC1's Iceo as a decent leakage figure. Maybe they're mistaken.

It depends on the firmware.

I wonder if mine uses the latest firmware... I looked up ways to update the firmware on a TC1, but it all seemed to be third-party weirdness...

I have a TC-1, a DC-55, a DMM -- none of them are perfect, but they all serve a purpose.

Do you use your TC1 to measure transistors? How about Ge?

Either way, I'd love to see the results of a test between the TC1 and DCA55 - think you could test a standard silicon and a Ge and share the gain/leakage results? I'm just wondering how different they'll be...

I'm certainly comfortable with ballpark, but there seem to be situations where a fairly precise measurement is needed. I have a Dizzy Tone ready to build, and it sounds like the Ge NPN transistors need to be pretty darn close to 50, 60, and 110 hfe. I have to figure out which testing method will give me figures that align closest with this, and I have to kind of guess as to which testing method was used in the build doc to come up with those hfe numbers. It's tricky to know what I'm dealing with when one tester says 56 and one says 96... It's this kind of situation which lead me to the rabbit hole I'm in!

 

[PedalBuilder]

It does list Iceo, but I don't think it takes that into account when measuring hfe, resulting in an inaccurate gain reading for Ge (according to some). Plus, reading through threads here gives me the impression that many people here don't trust a TC1's Iceo as a decent leakage figure. Maybe they're mistaken.

If it's measuring ICEO, then it's measuring leakage. It would make no sense make a device that measures leakage but does not factor it into the hFE calculations, so I think you're safe to assume that it's measuring the gain accurately under the testing parameters

I have a Dizzy Tone ready to build, and it sounds like the Ge NPN transistors need to be pretty darn close to 50, 60, and 110 hfe.

That's simply wrong. You can build a Dizzy Tone with a wide range of gains and it'll sound great. For example, I've used transistors with gains ranging from 30/40/80 to 80/80/170 (as measured on my DCA55), and a bunch of values in between. All of them sounded great. I wouldn't be surprised if values outside of those ranges also sounded good. It might be the germanium fuzz circuit that is least picky about transistors. You just want to put the lowest leakage transistor in Q1, as Q1 and Q2 are set up as a Darlington pair and any leakage in Q1 gets multiplied by the gain of Q2. Any measurement tool will be fine for this, just don't do an apples-to-oranges comparison. Q3 needs to be leaky, but I've used transistors with anywhere from 200 µA to 700 µA of leakage. The most important component in the circuit IMHO is the diode. The diode's leakage helps set the bias for Q3 by shunting some of the leakage current to ground, so you'll want to audition diodes with Q3 before building for optimal results. I don't have numbers a ratio between diode leakage and transistor leakage, or anything like that. The target Q3 collector voltage is a really wide range, too, anywhere between 2 and 4.5 volts can sound good. This is prime "use your ears" territory.

It's tricky to know what I'm dealing with when one tester says 56 and one says 96... It's this kind of situation which lead me to the rabbit hole I'm in!

Again, you're getting too caught up in "what is the right hFE number." A Dizzy Tone has less than 25 components. It would take 5 minutes to set it up on a breadboard and then you can test transistors with the one tool that really matters—your ears.

So to repeat the thing that we've been telling you for the last 30 posts—it's really ok to stop obsessing about hFE and just pick the transistors that sound best to your ears.

 

[Ginsly]

@PedalBuilder First off, I now have a much better understanding of what's going on with the Dizzy Tone circuit - thanks! It's great to know there is some wiggle room - lots of it, apparently. There is a lot of precision in building, but also a lot of guessing and auditioning, clearly.

I wasn't really too concerned with hfe until I recently went back and "fixed" an Si Fuzz Face. It was my second build, and everything worked but it was just unbearably noisy - even for fuzz. I realized that both transistors were pretty high gain, so I popped one of my only super low-gain Si's in Q1 and it was like night and day. Still had plenty of fuzz, but the noise floor dropped considerably. It also cleans up much better.

That (paired with wading into Ge waters) got me thinking about how important hfe is, and how drastically it can affect not just the sound of the pedal but the overall usability. Your "shut up and listen" advice is certainly what makes the most sense, but while learning the fundamentals of all of this I need to understand what's what - especially when it comes to choosing components wisely.

It's really confusing for a person in my postition when there are several accepted methods of testing which come up with such different results. I now understand why that is, but I still kinda have to choose one or the other if I want to get some baseline figures once in a while. The advice in this area is all over the map, and that tells me that your overarching theme of "don't worry about it, just try a bunch and listen" is really the only way to know what components are going to sound like in-circuit.