Measuring Rbulk of transistors

From circuitbending to homebrew stompboxes & synths, keep the DIY spirit alive!

Moderators: Joe., lisa, luketeaford, Kent

User avatar
devinw1
Super Deluxe Wiggler
Posts: 1190
Joined: Tue Aug 07, 2018 11:20 am
Location: Portland, OR
Contact:

Re: Measuring Rbulk of transistors

Post by devinw1 » Wed Jan 29, 2020 4:02 pm

Crap, yeah OK I had an error in my spreadsheet. Makes sense, thanks!

User avatar
devinw1
Super Deluxe Wiggler
Posts: 1190
Joined: Tue Aug 07, 2018 11:20 am
Location: Portland, OR
Contact:

Re: Measuring Rbulk of transistors

Post by devinw1 » Wed Jan 29, 2020 4:09 pm

Funny how weak that squared term even is that they even include it... If you treat it as linear you get 3548 vs 3535... Peanuts. :despair:

User avatar
guest
Super Deluxe Wiggler
Posts: 3954
Joined: Mon Aug 19, 2013 11:49 am

Re: Measuring Rbulk of transistors

Post by guest » Wed Jan 29, 2020 5:27 pm

yeah, that squared term is what gives the graph its slight curvature, which is 0.1% error.
openmusiclabs.com

User avatar
devinw1
Super Deluxe Wiggler
Posts: 1190
Joined: Tue Aug 07, 2018 11:20 am
Location: Portland, OR
Contact:

Re: Measuring Rbulk of transistors

Post by devinw1 » Fri Jan 31, 2020 12:39 am

So, I got some more matched pairs in to test and was trying to do apples and oranges and was getting much different values on the same units that I had measured before, so I re-did the setup and am getting numbers that seem more believable. I base lined with a THAT320 since the datasheet at least gives a value of 2ohms (max), so I have that at least. My measurement of 1.54Ohms seems right in line with what one might expect.

The setup is the same Renee Schmitz setup used before but I use a -8V voltage regulator going into a TL052 buffer for the base current. I use resistors 1k, 3k, 5.1k, 10k, and 15k and record the current and open collector voltage with a precision DMM. Results are below.
MatchedPairs2.png
I tested the top contenders in the expo of the TZ0 prototype and the only surprise was the BCM53DSF was not as good as the winner (NST4501MW6TIG), which performed a little better than my old favorite PMBT4403YSX and has the advantage of being an actual Vbe/HFE matched pair. Unfortunately it is again the small 0603 package, but what can ya do? I'm pretty sure I will use this for the final release of the TZ0.

The NSS40300MDR26 is funny in that in this range you actually get a realllly low rBE, but that is am I guessing becasue this is the very lowest end of what this pair is designed for, which means also that it would probably be useless trying to use it under 1mA for an expo. I didn't even try it.
You do not have the required permissions to view the files attached to this post.

User avatar
guest
Super Deluxe Wiggler
Posts: 3954
Joined: Mon Aug 19, 2013 11:49 am

Re: Measuring Rbulk of transistors

Post by guest » Fri Jan 31, 2020 7:39 am

thanks for compiling all of this. i would be curious to see how they perform below 1mA, as that is the intended range, and youre seeing some nonlinearity on at least one of them down there. whats also nice, is that most of these have a complimentary NPN, so one might be able to extrapolate how those work. im glad to see that the PMP matches the BMC, which was my expectation from looking at the datasheets. it would be interesting to see if the NSS was actually bad at the lower range.
openmusiclabs.com

User avatar
devinw1
Super Deluxe Wiggler
Posts: 1190
Joined: Tue Aug 07, 2018 11:20 am
Location: Portland, OR
Contact:

Re: Measuring Rbulk of transistors

Post by devinw1 » Fri Jan 31, 2020 11:08 am

Yeah I was thinking I might re run the test with 8.2k to 100k or something (about 75uA - 900uA). The Vcs will be very low though, although I should be able to measure them with my benchtop DMM.

User avatar
guest
Super Deluxe Wiggler
Posts: 3954
Joined: Mon Aug 19, 2013 11:49 am

Re: Measuring Rbulk of transistors

Post by guest » Fri Jan 31, 2020 11:56 am

yeah, im not sure this technique will work down that low. there are two effects, Vbe increasing and Ie*Rbe increasing. the latter is really small in comparison at low currents, so it will be hard to disambiguate the two. but, your PMP values matched well with mine (albeit different polarity devices), so maybe its fine. i was suprised to see the 3906 do so poorly, though.
openmusiclabs.com

User avatar
devinw1
Super Deluxe Wiggler
Posts: 1190
Joined: Tue Aug 07, 2018 11:20 am
Location: Portland, OR
Contact:

Re: Measuring Rbulk of transistors

Post by devinw1 » Sat Feb 01, 2020 8:45 pm

Hmm, so I measured 4 of the previous devices with lower currents and I'm a little confused as I get 2 different types of curves. 1 is a downward facing thing with some power, and the other is upward sloping and very linear. I did the setup identiically on all devices and made sure to note the sign of voltages and currents. Here's the graphs:
DMMT5401.png
NST4501.png
PMBT4430.png
THAT320P.png
You do not have the required permissions to view the files attached to this post.

User avatar
guest
Super Deluxe Wiggler
Posts: 3954
Joined: Mon Aug 19, 2013 11:49 am

Re: Measuring Rbulk of transistors

Post by guest » Sun Feb 02, 2020 12:36 am

whats your setup exactly? i was a little confused as to how you were using the opamp, and where you were taking your measurements in the circuit.
openmusiclabs.com

User avatar
devinw1
Super Deluxe Wiggler
Posts: 1190
Joined: Tue Aug 07, 2018 11:20 am
Location: Portland, OR
Contact:

Re: Measuring Rbulk of transistors

Post by devinw1 » Sun Feb 02, 2020 1:26 pm

It's like so (Just the Schmitz setup with a buffered voltage):
rBE Test2.png
You do not have the required permissions to view the files attached to this post.

User avatar
guest
Super Deluxe Wiggler
Posts: 3954
Joined: Mon Aug 19, 2013 11:49 am

Re: Measuring Rbulk of transistors

Post by guest » Sun Feb 02, 2020 2:15 pm

so, whats the opamp doing in this case? shouldnt the 7908 be able to produce a steady voltage? also, your output is not buffered, so you will be drawing a small amount of current and effecting the readings. are you putting your metere in-line to measure the base current, or are you measuring the drop across the resistor and dividing by the resistance?
openmusiclabs.com

ixtern
Common Wiggler
Posts: 187
Joined: Mon Jan 22, 2018 4:34 am

Re: Measuring Rbulk of transistors

Post by ixtern » Sun Feb 02, 2020 2:50 pm

I have only small knowledge of measuring bulk emitter resistance by this method (Schmitz or Kulke-Miller or Giacoletto) but I had an impression that this method is valid only for large base currents (e.g. 1-10 mA) as with lower base currents nonlinearities of BJT transistor model are larger and larger - so Vce voltage is not proportional any longer to emitter's Rbulk in small base currents area.

User avatar
guest
Super Deluxe Wiggler
Posts: 3954
Joined: Mon Aug 19, 2013 11:49 am

Re: Measuring Rbulk of transistors

Post by guest » Sun Feb 02, 2020 2:58 pm

yes, thats my impression as well. when i did my scanned version, and looked at the graph on the scope, there was a knee in the curve around 1mA or so from where it went from flat to linearly increasing.
openmusiclabs.com

User avatar
devinw1
Super Deluxe Wiggler
Posts: 1190
Joined: Tue Aug 07, 2018 11:20 am
Location: Portland, OR
Contact:

Re: Measuring Rbulk of transistors

Post by devinw1 » Sun Feb 02, 2020 5:38 pm

ixtern wrote:
Sun Feb 02, 2020 2:50 pm
I have only small knowledge of measuring bulk emitter resistance by this method (Schmitz or Kulke-Miller or Giacoletto) but I had an impression that this method is valid only for large base currents (e.g. 1-10 mA) as with lower base currents nonlinearities of BJT transistor model are larger and larger - so Vce voltage is not proportional any longer to emitter's Rbulk in small base currents area.
That's kind of what I thought, and it seems I have proven this. I just thought it would be interesting to try.

I also thought it would be interesting to put a triangle wave going from -200mV to -8.2V through the 8.2k base resistor and look at the open Vc on the scope (after a 1000x gain stage). Here are the 3 units I measured before. Interesting how 2 of them are pretty linear and one is highly non linear!:
THAT320
THAT320_Scope.jpg
PMBT4403
PMBT4403_Scope.jpg
NST4501
NST4501_Scope.jpg
You do not have the required permissions to view the files attached to this post.

User avatar
guest
Super Deluxe Wiggler
Posts: 3954
Joined: Mon Aug 19, 2013 11:49 am

Re: Measuring Rbulk of transistors

Post by guest » Sun Feb 02, 2020 6:39 pm

do those waveforms look any different if you run it down at 10Hz? its odd that they are not symmetric on the funky looking one.
openmusiclabs.com

User avatar
devinw1
Super Deluxe Wiggler
Posts: 1190
Joined: Tue Aug 07, 2018 11:20 am
Location: Portland, OR
Contact:

Re: Measuring Rbulk of transistors

Post by devinw1 » Mon Feb 03, 2020 11:10 am

Crap, I just realized I screwed that up. I was AC coupling since the small signal is offset and the gain is so high, but I was using way too small a cap (1.5u). Looks like I need more like 22u. :doh:

User avatar
devinw1
Super Deluxe Wiggler
Posts: 1190
Joined: Tue Aug 07, 2018 11:20 am
Location: Portland, OR
Contact:

Re: Measuring Rbulk of transistors

Post by devinw1 » Wed Feb 05, 2020 12:04 am

OK, so I got it hooked up with a 100uF coupling cap and turned down to 100Hz. I think I basically have just found (and this was probably already known) that Vc,open increases rapidly as Ib increases, but when Ib gets to a certain threshold (depends on the transistor), it turns around and goes the other way. The rates of Vc change seem to be fairly linear on some transistors, and not at all on others, but it probably depends on the window you're looking at and what the transistors max current is.

I don't know if any of this has anything to do with how good a given transistor would work as an exponential converter, becuase I've tested the crazy-looking NST4501 below and it works quite well. It also had a high-current slope method rBE of a respectable 1.60Ohms.

Rbase=8.2k
PMBT4403 with voltage to Rbase~ -200mV to -8V triangle
PMBT4403_Moreoffset.jpg
NST4501 with voltage to Rbase~ -200mV to -8V triangle
NST4501_Moreoffset.jpg
PMBT4403 with voltage to Rbase~ -10mV to -8V triangle
PMBT4403_Lessoffset.jpg
NST4501 with voltage to Rbase~ -10mV to -8V triangle
NST4501_Lessoffset.jpg

You can get some interesting wave shaping by varying the offset of the input signal, like this "sine-notch" wave, with variable notch width. It sounded kind of cool anyway. :lol:
Notchwave.jpg
You do not have the required permissions to view the files attached to this post.

User avatar
guest
Super Deluxe Wiggler
Posts: 3954
Joined: Mon Aug 19, 2013 11:49 am

Re: Measuring Rbulk of transistors

Post by guest » Wed Feb 05, 2020 10:47 am

the reversing is when Vbe turns off.
openmusiclabs.com

User avatar
devinw1
Super Deluxe Wiggler
Posts: 1190
Joined: Tue Aug 07, 2018 11:20 am
Location: Portland, OR
Contact:

Re: Measuring Rbulk of transistors

Post by devinw1 » Wed Feb 05, 2020 11:36 am

Makes sense, at saturation.

So now that we know the open collector measurement of rBE by measuring slope of Vc/Ib curve only works at higher currents, I wonder if there is a measurement that would better indicate how well a transistor would be in an exponential converter? Because while this method seems to loosely work, there are instances where I've measured a pretty low rBE of a matched pair but it does worse than one of a slightly higher rBE (in terms of linearity and HF tracking).

User avatar
guest
Super Deluxe Wiggler
Posts: 3954
Joined: Mon Aug 19, 2013 11:49 am

Re: Measuring Rbulk of transistors

Post by guest » Wed Feb 05, 2020 11:40 am

the other method i posted earlier seems to work very well.
openmusiclabs.com

User avatar
devinw1
Super Deluxe Wiggler
Posts: 1190
Joined: Tue Aug 07, 2018 11:20 am
Location: Portland, OR
Contact:

Re: Measuring Rbulk of transistors

Post by devinw1 » Wed Feb 05, 2020 12:18 pm

I'll have to play with that some more. I had trouble getting it to be very stable (using PNPs of course since that's what I'm looking at). Also, then I still have to read the slope off the scope visually with XY which isn't horribly accurate. Perhaps enough though.

User avatar
guest
Super Deluxe Wiggler
Posts: 3954
Joined: Mon Aug 19, 2013 11:49 am

Re: Measuring Rbulk of transistors

Post by guest » Wed Feb 05, 2020 12:54 pm

a clean layout helps with the stability (keep parts low to the board and have short runs), and then adjust the 330ohm/1nF combo till the output is stable. it will always be a bit noisey, but thats different from the sort of oscilltions you see with instability. the adjustment procedure is to short out the capacitor, and then lower the resistance until its stable (you want the largest resistor that gives stability). then put the capacitor back in, and increase its value until its stable (you want the smallest cap that gives stability).

measuring on the scope is noisey, but i figure its within the variation of transistor parameters. for example, you can probably only read the scope to 10% accuracy, but Rbe will vary at least by 10% amongst parts of the same model transistor.
openmusiclabs.com

User avatar
devinw1
Super Deluxe Wiggler
Posts: 1190
Joined: Tue Aug 07, 2018 11:20 am
Location: Portland, OR
Contact:

Re: Measuring Rbulk of transistors

Post by devinw1 » Sat Feb 08, 2020 7:54 pm

Just for kicks I proto'd up the MAT02 datasheet circuit and tested a few of the matched pairs that I had measured before. There wasn't very good correlation and all of the measurements were lower with this circuit than with the previous (Renee Schmitz method). I triple checked everything and had it set up correct and used a nice instrumentation amp just as shown with gain set exactly to 100.

The circuit is quite easy to use as it is very stable and gives a clear output that you just measure the Vpp of and dived by 9mA * 100 (gain). You just switch the 10k and 100k input resistors to measure the rBE of the other transistor in the pair.

Anyway, not quite sure on the descrepancy, but just thought I'd share this:
MatchedPairs2B.png
You do not have the required permissions to view the files attached to this post.

User avatar
guest
Super Deluxe Wiggler
Posts: 3954
Joined: Mon Aug 19, 2013 11:49 am

Re: Measuring Rbulk of transistors

Post by guest » Sat Feb 08, 2020 8:07 pm

thats really interesting. i would be curious to see how the 3906 and THAT transistors do, that way i could compare to the data i got on the NPN variants.
openmusiclabs.com

User avatar
devinw1
Super Deluxe Wiggler
Posts: 1190
Joined: Tue Aug 07, 2018 11:20 am
Location: Portland, OR
Contact:

Re: Measuring Rbulk of transistors

Post by devinw1 » Sun Feb 09, 2020 1:10 pm

Yeah, I will do that next. I only did those 4 because they all have the same pinout on the little SMT to 0.1" adapters I'm using so it was easy to just swap em in and out. :D

Post Reply

Return to “Music Tech DIY”