MUFF WIGGLER Forum Index
 FAQ & Terms of UseFAQ & Terms Of Use   Wiggler RadioMW Radio   Muff Wiggler TwitterTwitter   Support the site @ PatreonPatreon 
 SearchSearch   RegisterSign up   Log inLog in 
WIGGLING 'LITE' IN GUEST MODE

Matching Transistors for Vbe
MUFF WIGGLER Forum Index -> Music Tech DIY Goto page 1, 2  Next [all]
Author Matching Transistors for Vbe
Jackdamery
I was looking at this circuit for matching.

http://www.oldcrows.net/~patchell/matcher/matcher.html

As far as I understand it you connect the two leads of your multimeter to the "4 1/2 Digital multimeter" nodes. There is a pnp already in there. So then there is a node labeled "D.U.T." Which I presume means device under test. So which leg of the transistor you're testing would this go to? And what would you connect the other two to?

Thanks,
J
decaying.sine
Check this out.
http://home.comcast.net/~ijfritz/projects/transmat001.pdf
bananeurysm
Sorry if this is a dumb question, I've read a bit about transistor matching, looked at circuits etc, but haven't found a clear answer.

Are these circuits essentially the same as what is inside fancy multi meters that have holes for plugging in and testing transistors? Or is that a different kind of test?

Maybe this belies my confusion over HFE, vbe, etc?
Bergfotron
bananeurysm wrote:
Sorry if this is a dumb question, I've read a bit about transistor matching, looked at circuits etc, but haven't found a clear answer.

Are these circuits essentially the same as what is inside fancy multi meters that have holes for plugging in and testing transistors? Or is that a different kind of test?

Maybe this belies my confusion over HFE, vbe, etc?


Multimeters usually only measure Hfe. This is useless for Vbe matching.
I have made a board for a transistor matcher according to Ian Fritz' method. You can find it here:
Bergfotron transistor matcher
daverj
Vbe (base-emitter voltage) matching is the most common type of matching done around here. It is what is used for matching transistors in an expo converter. It can be done with a circuit like Ian Fritz's, or can be roughly done using the diode setting on a volt meter (Ian's circuit gives a more accurate result).

Hfe (current gain) is a different type of matching that requires a different type of circuit. Some multimeters include an Hfe tester. It is the type of matching needed on some older discrete VCAs and VCFs that have differential transistor amplifiers and differential signal paths. (a linear differential amplifier with a single ended signal path typically doesn't need matching Hfe, though matching Vbe in those cases can help with temperature tracking)
davemoog
Music from Outer Space also has a useful page on different techniques and circuits.

http://www.musicfromouterspace.com/analogsynth_new/TRANSISTORMATCHER/T RANSISTORMATCHER.html
bananeurysm
Ahh Ok, that clarifies it. Thanks!
Jackdamery
decaying.sine wrote:
Check this out.
http://home.comcast.net/~ijfritz/projects/transmat001.pdf


Thanks! This is probably the best thought out and explained scientific document I've ever read!
EATyourGUITAR
I do not have a good meter so I just pay the hefty price to buy transistor packages that have vbe matched and all transistors thermally coupled by default. makes a nice flat surface to glue your smd tempco to.
theabsent
EATyourGUITAR wrote:
I do not have a good meter so I just pay the hefty price to buy transistor packages that have vbe matched and all transistors thermally coupled by default. makes a nice flat surface to glue your smd tempco to.


What's your source?
nikmis
If a VCO was built with the typical dual transistor thing in the exponential converter, but they were unmatched, what would happen? Would the vco scale drifts out of tune over time, or would the overall pitch drift but the scale remain the same?
EATyourGUITAR
theabsent wrote:
EATyourGUITAR wrote:
I do not have a good meter so I just pay the hefty price to buy transistor packages that have vbe matched and all transistors thermally coupled by default. makes a nice flat surface to glue your smd tempco to.


What's your source?


https://www.muffwiggler.com/forum/viewtopic.php?t=86749
Jackdamery
How can you apply Ian Fritzs matching method if you need to match an NPN to a PNP?
EATyourGUITAR
not with that circuit
rpocc
I've just tried to breadboard Ian Fritz solution and got smoke from the Rref and diode. That was predictable with 7912 source shorten to ground thru 3.9k and 1n4148. But what's the trick? seriously, i just don't get it
daverj
with -12v and a 3.9K there should have been 2.9ma flowing through the resistor before transistors are connected, and 3.1ma once transistors are connected. That's 35mw. A 1/4 watt resistor should not smoke. Are you sure you used 3.9K? Or wired it correctly?

Even directly between the +12 and -12v, a 3.9k is using less than 1/4 watt and shouldn't smoke.
rpocc
daverj wrote:
with -12v and a 3.9K there should have been 2.9ma flowing through the resistor before transistors are connected, and 3.1ma once transistors are connected. That's 35mw. A 1/4 watt resistor should not smoke. Are you sure you used 3.9K? Or wired it correctly?

Even directly between the +12 and -12v, a 3.9k is using less than 1/4 watt and shouldn't smoke.


Thank you! Your post made me double check the value and I found that resistor was 3R9 instead of 3K9. Seems that I've mixed contents of two packets into one drawer. Of course it should smoke smile
daverj
Dr. Sketch-n-Etch
So, I recently tried to build a (fairly popular) transistor ladder clone, and I just hFE-matched the transistors. Bottom line, it really didn't work at all -- it isn't doing what I would call "filtering" -- more waveshaping, and it barely resonates at all. The hFEs were matched precisely. Is Vbe matching that important in a ladder filter?
rpocc
Dr. Sketch-n-Etch wrote:
So, I recently tried to build a (fairly popular) transistor ladder clone, and I just hFE-matched the transistors. Bottom line, it really didn't work at all -- it isn't doing what I would call "filtering" -- more waveshaping, and it barely resonates at all. The hFEs were matched precisely. Is Vbe matching that important in a ladder filter?


In Prodigy ladder filter clone used in Ritm-2, was used large amount of CA3046 equivalents and matched transistors. According to datasheets, that 3046 clones have 15% hfe precision and +/-4mv vbe matching. (By the way, original CA3046 have +/-5mv matching.
Usual hfe for ladder filters should be around 100
Dr. Sketch-n-Etch
I was trying to build the Yusynth Minimoog VCF clone, and I used CA3146 (as that was all I could find locally -- Yves said that was one of the options), and I used 2N3904 (instead of BC547) for the NPNs and 2N3906 (instead of BC557) for the PNP. As I said, all the 2N3904 pairs were perfectly matched for hFE (but completely unmatched for Vbe). And, yes, I am aware that the pinouts for those transistors are reversed, so I turned them all around relative to Yves' layout. (I etched my own PCB from the graphic on his website.)

The filter basically takes any waveshape and turns it into this strange-looking humpy waveshape. Also, I could only get any resonance after short-circuiting the leads of the 10uF AC coupling capacitor in the resonance loop, and even then it was very weak. Mostly the resonance just made the "filter" sound dirty -- it created little wiggles on the waveform around 0V. The circuit also had some other very strange quirks -- I didn't have the right values for most of the control pots, but the audio input would suddenly turn on when the pot was at about 60% of its range, and then there was very little volume change after that.

In short, what I ended up with is completely unusable as a filter, and the only thing I can think of to do is to try again with Vbe-matched transistors of the proper type -- I guess what I'm wondering (and asking) is, will this make a fundamental difference?

Has anyone else here tried to build this filter? Did it work for you? Here's the link:

http://www.yusynth.net/Modular/EN/MOOGVCF/index.html
rpocc
I think, that hfe matching can be limited to +/- 10 units, but Vbe matching should be precise.

Also, BC547C have triple hfe over 2n3904. (My meter says 200-300 vs 400-800) Unlikely that it is very important, but character may change.
frijitz
nikmis wrote:
If a VCO was built with the typical dual transistor thing in the exponential converter, but they were unmatched, what would happen? Would the vco scale drifts out of tune over time, or would the overall pitch drift but the scale remain the same?

The tuning would drift, but the scale factor would be OK. This is because the residual temperature dependence is in the exponential prefactor, not in the exponent.

Ian
daverj
Dr. Sketch-n-Etch wrote:
So, I recently tried to build a (fairly popular) transistor ladder clone, and I just hFE-matched the transistors. Bottom line, it really didn't work at all -- it isn't doing what I would call "filtering" -- more waveshaping, and it barely resonates at all. The hFEs were matched precisely. Is Vbe matching that important in a ladder filter?


Based on the typical ladder schematic I would think Vbe matters a lot on the bottom pair since they have their emitters tied and their bases are biased the same. Any Vbe difference on the top pair will cause a voltage offset at the op amp, but as long as that doesn't cause clipping I wouldn't think that's critical. For the rest of them I wouldn't think Vbe matching is as critical since the emitters are tied capacitively.

EDIT: I just looked at the schematic in your link above. The one I had looked at before had the op amp directly off of the top pair, which is what my comments above were based on. With this schematic I would think not only the top pair, but the two pairs used in the darlington diff amp would need to be Vbe matched since that would cause a serious amount of voltage offset if they aren't.
gdavis
Dr. Sketch-n-Etch wrote:
I was trying to build the Yusynth Minimoog VCF clone, and I used CA3146 (as that was all I could find locally -- Yves said that was one of the options), and I used 2N3904 (instead of BC547) for the NPNs and 2N3906 (instead of BC557) for the PNP. As I said, all the 2N3904 pairs were perfectly matched for hFE (but completely unmatched for Vbe). And, yes, I am aware that the pinouts for those transistors are reversed, so I turned them all around relative to Yves' layout. (I etched my own PCB from the graphic on his website.)

The filter basically takes any waveshape and turns it into this strange-looking humpy waveshape. Also, I could only get any resonance after short-circuiting the leads of the 10uF AC coupling capacitor in the resonance loop, and even then it was very weak. Mostly the resonance just made the "filter" sound dirty -- it created little wiggles on the waveform around 0V. The circuit also had some other very strange quirks -- I didn't have the right values for most of the control pots, but the audio input would suddenly turn on when the pot was at about 60% of its range, and then there was very little volume change after that.

In short, what I ended up with is completely unusable as a filter, and the only thing I can think of to do is to try again with Vbe-matched transistors of the proper type -- I guess what I'm wondering (and asking) is, will this make a fundamental difference?

Has anyone else here tried to build this filter? Did it work for you? Here's the link:

http://www.yusynth.net/Modular/EN/MOOGVCF/index.html


I've built one from Yves' layout and probably 9 or 10 from my own layout of Yves' circuit. I didn't do any matching (just pulled them from the same reel) and they seem to work fine for me seriously, i just don't get it
MUFF WIGGLER Forum Index -> Music Tech DIY Goto page 1, 2  Next [all]
Page 1 of 2
Powered by phpBB © phpBB Group