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: 1376
Joined: Tue Aug 07, 2018 11:20 am
Location: Portland, OR
Contact:

Measuring Rbulk of transistors

Post by devinw1 » Wed Jan 15, 2020 12:38 am

I'm trying to measure the Rbulk on some PNP transistors and would appreciate if anyone can chime in if I did this right. I used the description that Renee Schmitz gives:

"...measure Vce (collector is open) while variing the base current from a few hundered micro amps to about 5 miliamps. I used a set of a few fixed resistors in the range from 1k to 15k for the base current. The Rbe is the high current slope of the Vce vs Ib graph"

I did this with a BC557A (PNP) and used my benchtop MM to measure the base current and my DMM to measure the Vce. I used resistors 15k, 10k, 8.2k, 5.6k, 3.9k, 2k, and 1k and -12V for the base current. The data is shown below and had a linear fit of Rsquared = .996 so pretty damn good, but the value for Rbulk I get is 0.74 Ohms, which seems low but maybe that is right? I just say it seems low since Renee's page shows all the PNPs he measured being over 1.1Ohms or so.

The data:

Resistor Used Ib(mA) Vce(mV)
15k 0.760 5.9
10k 1.14 6.2
8.2k 1.40 6.4
5.6k 2.03 7.0
3.9k 2.92 7.7
2k 5.7 10
1k 11.4 13.8
Last edited by devinw1 on Wed Jan 15, 2020 10:49 am, edited 1 time in total.

User avatar
guest
Super Deluxe Wiggler
Posts: 4186
Joined: Mon Aug 19, 2013 11:49 am

Re: Measuring Rbulk of transistors

Post by guest » Wed Jan 15, 2020 1:19 am

did you mean BC548C? thats an NPN transistor.
openmusiclabs.com

User avatar
guest
Super Deluxe Wiggler
Posts: 4186
Joined: Mon Aug 19, 2013 11:49 am

Re: Measuring Rbulk of transistors

Post by guest » Wed Jan 15, 2020 1:32 am

also, if you have a TL072 handy, set that up to buffer the Vce voltage. your readings might be effected by the resistance in the meter. im also not sure what effect temperature would have on it, as higher currents would warm it up a bit. i measured Rbe for a bunch of NPN transistors, but did it with a thermally compensated exponential converter circuit, and then calculated a fit using an Rbe based error model. i got very similar results to what schmitz has posted (for 3046 and MAT04), but i didnt test any PNP transistors.
openmusiclabs.com

User avatar
devinw1
Super Deluxe Wiggler
Posts: 1376
Joined: Tue Aug 07, 2018 11:20 am
Location: Portland, OR
Contact:

Re: Measuring Rbulk of transistors

Post by devinw1 » Wed Jan 15, 2020 10:48 am

guest wrote:
Wed Jan 15, 2020 1:19 am
did you mean BC548C? thats an NPN transistor.
Doh! I meant BC557A... Original post corrected. Got my wires crossed there. It was sort of a random PNP I pulled out to verify my setup before measuring some SMT matched parts I have.

I'll try it again buffering the Vce as well. :tu:

User avatar
devinw1
Super Deluxe Wiggler
Posts: 1376
Joined: Tue Aug 07, 2018 11:20 am
Location: Portland, OR
Contact:

Re: Measuring Rbulk of transistors

Post by devinw1 » Wed Jan 15, 2020 8:43 pm

OK, I think I'm doing it right. I buffered the Vce with a TL072 and retook the data. I don't have any of the exact transistors to what Renee has, but I did have something old and NPN to compare somewhat to (CA3086). I measured this one as well as a few other PNPs. It seems perhaps modern PNPs are not so bad in regard to Rbulk compared to NPN? I also have a DMMT3906 that I'm going to test (as well as some other NPNs), but here's what I have so far:


BC557A(PNP TO-92): 0.72 Ohms
2N3906(PNP TO-92): 0.83 Ohms
BCM857DS,135 (PNP SMT Pair): 0.71 Ohms <-what i'm using on my new TZ0
CA3086(NPN Quad, DIP): 3.78 Ohms

User avatar
metasonix
Tube Pioneer
Posts: 4400
Joined: Fri Aug 03, 2007 2:35 pm

Re: Measuring Rbulk of transistors

Post by metasonix » Wed Jan 15, 2020 8:48 pm

why.jpeg
You do not have the required permissions to view the files attached to this post.

User avatar
devinw1
Super Deluxe Wiggler
Posts: 1376
Joined: Tue Aug 07, 2018 11:20 am
Location: Portland, OR
Contact:

Re: Measuring Rbulk of transistors

Post by devinw1 » Wed Jan 15, 2020 9:16 pm

Lol!

Well, in a nutshell, the lower the Rbe, the better it should be in an exponential converter (better linearity at higher currents)

User avatar
guest
Super Deluxe Wiggler
Posts: 4186
Joined: Mon Aug 19, 2013 11:49 am

Re: Measuring Rbulk of transistors

Post by guest » Wed Jan 15, 2020 9:25 pm

those numbers seem reasonable. that BCM857 did a lot better than i thought it would. when i get a chance i wil rerun some of my transistors using this method.
openmusiclabs.com

User avatar
guest
Super Deluxe Wiggler
Posts: 4186
Joined: Mon Aug 19, 2013 11:49 am

Re: Measuring Rbulk of transistors

Post by guest » Wed Jan 15, 2020 11:45 pm

i just measured a 3906 and 3904 using this method and ~0.6ohms, which seems reasonable. the reason why im suprised by the BCM857, is that it an identical transistor the PMP4201 (couldnt tell any differences in the datasheet, all graphs look the same), and the PMP4201 performed the worst on my exponential tracking tests, with an estimated Rbe of 1.2ohms or so. the 3904/6 are the only transistors i have on hand right now, or else i would get a bit more data.

for my test rig, i built a non-inverting opamp with gain of 100, and drove the base with a function generator through a 500ohm resistor. i then set my scope up to X/Y mode and measured the slope of the line. with the function generator at 100Hz i figured this would negate any temperature effects.
openmusiclabs.com

User avatar
devinw1
Super Deluxe Wiggler
Posts: 1376
Joined: Tue Aug 07, 2018 11:20 am
Location: Portland, OR
Contact:

Re: Measuring Rbulk of transistors

Post by devinw1 » Thu Jan 16, 2020 12:58 am

Nice. Your test setup sounds more accurate potentially than mine but at least the numbers are agreeing pretty close. I was counting on thermal effects being minimal by not touching the transistor, my shop is pretty cold solid at about 60F right now and I only turn the power supply on for a fixed amount of time for the current to settle (about 5 seconds) and take the reading, shut it off, switch resistors and wait about 30 seconds and repeat. Making sure to start with the big resistors and work my way down.

That is very surprising about the BCM857 but perhaps it could just be this sample of it...or there's a spec that's being made much better than the datahsheet. So many possibilities!

Thanks for sharing your data. I'll have some more tomorrow to add to this thread

ixtern
Common Wiggler
Posts: 203
Joined: Mon Jan 22, 2018 4:34 am

Re: Measuring Rbulk of transistors

Post by ixtern » Thu Jan 16, 2020 2:13 am

guest wrote:
Wed Jan 15, 2020 11:45 pm
for my test rig, i built a non-inverting opamp with gain of 100, and drove the base with a function generator through a 500ohm resistor. i then set my scope up to X/Y mode and measured the slope of the line. with the function generator at 100Hz i figured this would negate any temperature effects.
It remainds me of the Rbe test circuit from the page 6 of the MAT02 datasheet (https://www.analog.com/media/en/technic ... /MAT02.pdf).

I have a lot of different transistor array types and I am curious about their Rbe too. Will also try to assemble test rig (perhaps simpler, like devinw1 - Rene's type rather than scope and function generator based).
Last edited by ixtern on Thu Jan 16, 2020 2:41 am, edited 1 time in total.

User avatar
guest
Super Deluxe Wiggler
Posts: 4186
Joined: Mon Aug 19, 2013 11:49 am

Re: Measuring Rbulk of transistors

Post by guest » Thu Jan 16, 2020 2:32 am

thanks for posting that link, i was just thinking about the best way to do log conformance testing, as i was wondering if maybe there is something other than Rbe that is causing a difference in log conformance (which was my initial test) between the PMP and the Rbe of the BMC. the linked method is good, as it allows for temperature to be compensated for as long as the current sweep is done faster than the thermal transients.
openmusiclabs.com

User avatar
devinw1
Super Deluxe Wiggler
Posts: 1376
Joined: Tue Aug 07, 2018 11:20 am
Location: Portland, OR
Contact:

Re: Measuring Rbulk of transistors

Post by devinw1 » Thu Jan 16, 2020 1:34 pm

So, guest, on your setup you are measuring Vin and Vce with it open still? Would you mind sketching the circuit up? I have a function generator and would like to rig this up myself.

User avatar
guest
Super Deluxe Wiggler
Posts: 4186
Joined: Mon Aug 19, 2013 11:49 am

Re: Measuring Rbulk of transistors

Post by guest » Thu Jan 16, 2020 5:37 pm

heres the circuit:

Image

i offset the triangle wave a bit (200mV?) so that the scope image would fit the screen better and be less jumpy. the current through the resistor is not exactly linear with the applied voltage, but is pretty close up in the measurement range (4-8V), as Vbe changes become a pretty small percentage of the total voltage at that point. i was thinking that if i was going to be doing a lot of this, id print up a clear sheet with various angled lines on it, each labeled with an Rbe value. that way i could do the 'math' a lot faster. just see which line it matches to.

also, you could probably increase the gain to 1000x and get a bit better resolution, depending upon your scope (ive got a digital scope that isnt so good down at the low end).
openmusiclabs.com

User avatar
devinw1
Super Deluxe Wiggler
Posts: 1376
Joined: Tue Aug 07, 2018 11:20 am
Location: Portland, OR
Contact:

Re: Measuring Rbulk of transistors

Post by devinw1 » Thu Jan 16, 2020 6:16 pm

Cool, thanks. So pretty much what I was doing but instead of varying R and directly measuring the current, you are fixing R at a low value and varying Vin and using this to indicate the base current, and then letting the scope do the slope math by plotting XY. :tu:

I have a Rigol DS1102E, and actually havent ever even tried XY mode on it. This gives me an excuse to play with it.

User avatar
guest
Super Deluxe Wiggler
Posts: 4186
Joined: Mon Aug 19, 2013 11:49 am

Re: Measuring Rbulk of transistors

Post by guest » Thu Jan 16, 2020 9:25 pm

i also reran some numbers from data i had on some transistors, which is slightly different from my other post on this topic:

THAT300 1.2ohm
MAT14 0.66ohm
2N3904 0.7ohm
PMP4201 2.1ohm
LM13700 6.2ohm
openmusiclabs.com

User avatar
devinw1
Super Deluxe Wiggler
Posts: 1376
Joined: Tue Aug 07, 2018 11:20 am
Location: Portland, OR
Contact:

Re: Measuring Rbulk of transistors

Post by devinw1 » Fri Jan 17, 2020 11:18 am

I tried the XY setup and it did work but yielded something like .61Ohms for the same 3906 that I got .83Ohms with the other method. It ends up being a little disapponting because my scope can't do any math for slope, so far as I can tell the only way to do it is to use manual cursors by eyeballing your 2 endpoints and then dividing the delta Y by delta X. Oh well. Glad I know how to use some more features on my scope now!

I was going to measure the DMMT3906 that I bought, but realized that for whatever reason, the one that Mouser and Digikey sell in the larger (SOT-26) package is common base! So, it would be useless for a standard PNP expo converter. :doh: So far as I can tell, all the other PNs of DMMT3906 available (All SOT-363 tiny package) have 3 discrete pins per transistor EXCEPT the SOT-26 one. :despair:

User avatar
guest
Super Deluxe Wiggler
Posts: 4186
Joined: Mon Aug 19, 2013 11:49 am

Re: Measuring Rbulk of transistors

Post by guest » Fri Jan 17, 2020 7:20 pm

sucks about the small 3906 package. i suppose you can still test the one you have, as it will probably be the same as the one in the smaller package.

i did the setup from the MAT02 datasheet, and tested a 3904 and got 0.5ohms, so pretty similar. the result was pretty noisey, and i think i can get better data if i build a cleaner amp for it. the signal is a few hundred microvolts superimposed on a few hundred millivolts of DC.

as for measuring the slope, you really only need one datapoint, so its pretty fast to get. if you use a constant deltaX (pick 2 lines on your scope), then adjust the Y offset so the curve aligns at 0V with one of them, then you can you can just read off the value where it crosses the other. its not horribly accurate (scopes are 8bit typically), but then again, im not expecting any of these measurements to be better than 5% accurate, and will probably vary by more than that over a batch anyways. i think the fast scanning of the transistor gives more accurate readings due to eliminating temperature effects.
openmusiclabs.com

User avatar
guest
Super Deluxe Wiggler
Posts: 4186
Joined: Mon Aug 19, 2013 11:49 am

Re: Measuring Rbulk of transistors

Post by guest » Sat Jan 18, 2020 6:23 pm

i was also just thinking that if you used a sine wave as the input, and just went from 3V to 8V (for example), then you should be able to use your multimeter on AC mode to measure the output amplitude, and then its just a matter of scaling for the input drive level.

R = dV/dI
dI = (8V-3V)/500ohm = 10mA
dV = Vpp/G (G is opamp gain, 100 for my circuit)
Vpp = 2*2^(0.5)*Vrms (for a sinewave)

R = 2*2^(0.5)*Vrms/(100*10mA) = 2.83*Vrms (Vrms in Volts)

the multimeter should be more accurate than the scope, and faster to read.
openmusiclabs.com

User avatar
soggybag
Wiggling with Experience
Posts: 288
Joined: Thu Aug 10, 2017 2:57 pm
Location: San Francisco

Re: Measuring Rbulk of transistors

Post by soggybag » Sat Jan 18, 2020 9:11 pm

I have one of these



I got it on eBay for $17. It works great. I’m not sure how accurate it is but it might be good for matching parts.

With BJT transistors it shows ube, hfe, Ic, Ices, and Iceo.

Here is a link to the manual. https://www.circuitspecialists.com/cont ... si-tc1.pdf

User avatar
devinw1
Super Deluxe Wiggler
Posts: 1376
Joined: Tue Aug 07, 2018 11:20 am
Location: Portland, OR
Contact:

Re: Measuring Rbulk of transistors

Post by devinw1 » Sun Jan 19, 2020 3:35 pm

I have something similar (Peak Atlast DCA55), which is pretty cool, but it doesn't measure Rbulk. From the looks of the manual, yours does not either?

User avatar
devinw1
Super Deluxe Wiggler
Posts: 1376
Joined: Tue Aug 07, 2018 11:20 am
Location: Portland, OR
Contact:

Re: Measuring Rbulk of transistors

Post by devinw1 » Sun Jan 19, 2020 3:37 pm

guest wrote:
Sat Jan 18, 2020 6:23 pm
i was also just thinking that if you used a sine wave as the input, and just went from 3V to 8V (for example), then you should be able to use your multimeter on AC mode to measure the output amplitude, and then its just a matter of scaling for the input drive level.

R = dV/dI
dI = (8V-3V)/500ohm = 10mA
dV = Vpp/G (G is opamp gain, 100 for my circuit)
Vpp = 2*2^(0.5)*Vrms (for a sinewave)

R = 2*2^(0.5)*Vrms/(100*10mA) = 2.83*Vrms (Vrms in Volts)

the multimeter should be more accurate than the scope, and faster to read.
That's a good idea. I'll have to try that out.

I measured the common base DMMT3906 with the old method, so it would be apples to apples, and I got 0.93Ohms. Not quite as good as the BCM857DS,135 interestingly!

I've got a couple other SMT pairs on the way tomorrow or tuesday to test out as well.

User avatar
guest
Super Deluxe Wiggler
Posts: 4186
Joined: Mon Aug 19, 2013 11:49 am

Re: Measuring Rbulk of transistors

Post by guest » Wed Jan 22, 2020 3:19 pm

using the collector voltage method (2V-10V across a 1k to the base), i ran a few more transistors:

2N3906 - 0.75ohm
2N3904 - 0.61ohm
MAT14 - 0.837ohm
THAT300 - 1.22ohm
LM3046 - 3.14ohm
PMP4201 - 0.92ohm

some of those match my expectations, others do not. the LM3046, THAT300, 3906, 3904 all match really well. the MAT14 should be better (datasheet gives max as 0.6ohm), but i think that transistor is damaged, as its the only remaining working one on the die. i blew the rest of them out. the PMP4201 should be worse than the THAT300, closer to 2ohms. i think this may be a limitation of this measurement technique. i will rerun these same transistors using the MAT02 datasheet method.
openmusiclabs.com

User avatar
devinw1
Super Deluxe Wiggler
Posts: 1376
Joined: Tue Aug 07, 2018 11:20 am
Location: Portland, OR
Contact:

Re: Measuring Rbulk of transistors

Post by devinw1 » Wed Jan 22, 2020 3:56 pm

Sweet. Your 3906 matches up pretty close to what I got with the Renee Schmitz method too.

I've got a bunch more matched pairs I am testing right now.

User avatar
guest
Super Deluxe Wiggler
Posts: 4186
Joined: Mon Aug 19, 2013 11:49 am

Re: Measuring Rbulk of transistors

Post by guest » Wed Jan 22, 2020 4:57 pm

so i built up a variant on the MAT02 datasheet method (the noise and DC offset were really bad using that circuit). i dont have 2 working MAT14 or THAT300 pairs inside of an IC, so i couldnt test those. heres what i got:

3906 - 0.64ohm
3904 - 0.61ohm
LM3046 - 2.1ohm
PMP4201 - 1.8ohm

these are tested at a max Ic of 1mA, which is what you would use in practice, and is therefore a better representation than the 10mA test of the other method. the bulk resistance probably changes with electric field within the device for some of the transistors. i was noticing a not so great fit using a fixed R with my work a few years back. i would have expected the LM3046 to be a bit higher, but ill have to go back and check my data.
openmusiclabs.com

Post Reply

Return to “Music Tech DIY”