Hi guys,
Long time no post :)
Just wanted to run something past the clever folk here (an up front thank you) :)
I received a DCA75 recently for review from Jez from Peak. I have been messing around with its curve tracing ability for hFE / Ic for measuring germanium transistors. The test conditions for the DCA are different from the Small Bear transistor tester and the R G Keen transistor tester and I'm wondering how I could simulate those circuits on the DCA. The Small bear circuit is easy enough, 9v supply and 9ua on the base and there is no load resistor. Edit: actually im thinking a 9v supply is incorrect because of the DCA's 700R load resistor (small bear circuit doesn't have a load resistor).
But the R G keen circuit is a little tricker. The RG Keen circuit has a 2.472k (roughly) resistor for load on the collector of the DUT. The DCA uses a 700R load resistor on the collector for the DUT. Base current is easy enough to work out, 4ua for the current source on the base. But im wondering if the supply voltage should be changed to compensate for the lower load resistor on the DCA? I have calculated the supply voltage on the DCA should be dropped to 2.5v approximately and not 9v like the R G Keen transistor tester. Is this correct?
R G Keens Well know test circuit: http://www.geofex.com/article_folders/ffselect.htm
Hope that makes sense, thank you for any help / confirmation.
Cheers
Paul
I did not know Peak had curve tracing.
Keen's tester uses R values made for easy math. If the Peak will do math, no reason to duplicate the values.
Classic curve tracing usually puts a constant voltage C-E. In Silicon and late Ge devices, the exact voltage is not critical, it is OK to let it rise/fall a bit with current (so we can read I with a voltmeter). In the oldest GE the curves do vary with Vce (and almost everything else).
Don't know how the Peak works, and I would think it is proprietary. Without the secret plans, or extended reverse-engineering, not sure what we can do.
Thanks Paul, was hoping for your reply and appreciate your knowledge. The DCA75 can cruve trace yes.
I guess im trying to simulate R G Keens circuit because he sort of set down the guidelines as far as hFE requirements for fuzz faces etc. Everyone seems to follow this (some blindly) without questioning further. I suppose for me, the underlying question is, why do we put (x microamps) of base current into the germanium transistor and hope for (x hFE)? For Keens its 4ua and the usual hFE's we know and love (Q1 ~70 Q2 ~120 etc). But the question is why those values? Is this the best representation of the conditions of a fuzz face? Some transistor hFE will vary significantly in response to increased base current / voltages.
EDIT: Although its probably over analysing I'm really mainly just curious to know.
Quote from: chromesphere on June 15, 2016, 09:28:51 PM
The test conditions for the DCA are different from the Small Bear transistor tester and the R G Keen transistor tester and I'm wondering how I could simulate those circuits on the DCA.
The Small bear circuit is easy enough, 9v supply and 9ua on the base and there is no load resistor.
Edit: actually im thinking a 9v supply is incorrect because of the DCA's 700R load resistor (small bear circuit doesn't have a load resistor).
The 700R is for Ic measurement / current limiting.
Collector Voltage will be measured above collector not above the resistor.
For Small Bear like measurement:
Tab BJT Hfe/Vce -> setting Ib = 9µA -> start measurement -> lookup graph at Vce =9V
The value is Hfe corrected for leakage.
Quote
But the R G keen circuit is a little tricker. The RG Keen circuit has a 2.472k (roughly) resistor for load on the collector of the DUT.
The DCA uses a 700R load resistor on the collector for the DUT.
Base current is easy enough to work out, 4ua for the current source on the base.
But im wondering if the supply voltage should be changed to compensate for the lower load resistor on the DCA?
I have calculated the supply voltage on the DCA should be dropped to 2.5v approximately and not 9v like the R G Keen transistor tester. Is this correct?
The R.G.Keen-Method uses a constant Ib for measuring HFE
resulting in HFE-dependent Ic and Uce (R.G.: Ub= 9V Rc = 2.472k Rb= 2.2M)
So this is a slightly moving target.
A Hfe=100 transistor will drop 1V above the 2.5k resistor with R.G.s method so Uce is 8V.
Tab BJT Hfe/Vce -> setting Ib = 4µA -> start measurement -> lookup graph at Vce =8V.
The value is Hfe corrected for leakage.
(measument point for THIS specific transistor)
Higher Hfe will drop more Voltage above the resistor and will decrease Uce and viceversa.
Everything is documented in the manual (http://www.peakelec.co.uk/downloads/dca-pro-user-guide-en.pdf).
Maybe the technical implementation and software is propietary but not the methods and underlying mathematics.
http://www.diystompboxes.com/smfforum/index.php?topic=108294.msg986737#msg986737 (http://www.diystompboxes.com/smfforum/index.php?topic=108294.msg986737#msg986737)
electrip
Thanks Electrip, thats very useful information.
Wish i had of seen the thread you linked to about a week ago, I have gone down a similar path. Like you mentioned in that post, the first thing i noticed was the variable ib on the quick test mode of the dca75 @ 5ma Ic isn't very useful for fuzz transistors and will obviously report back high hFE measurements. Which is why i ended up going down the curve trace path. The difficulty is finding a method of testing with the dca75 and the curve trace that will give us a hFE measurement that is "accurate". I know with germanium transistors there isnt such a thing because its a moving target, the target moves more so with more gain / leakage.
I've always used small bears simple tester and the measurements i have taken have always sounded good in fuzz face circuits so im thinking at this point a curve trace equivalent of this is the way i'm going to go. Its also obviously the simplest.
Do you have a preferred method that you would be willing to share?
> why do we put (x microamps) of base current .... For Keens its 4ua
I hurt too much to think.
What IS the current in a FuzzFace? You can easily derive it from published resistor and voltage values.
Is it similar to Keene's condition?
(Obviously not: Q2 runs higher current than Q1. But are they close?)
Side-note: for most design, collector current is a Target and we contrive the circuit to deliver whatever base current is needed. Keene's fixed base current is "wrong". He does that to make the math simple (and even then, many people are tripped-up). And if we are hunting for some narrow range of hFE, then setting the base current fixed and sorting for a small range of collector current finds the right parts. For *this* application, not necessarily others.
If you actually understand circuits, something as simple as a FuzzFace, you just build it with sockets. Measure the voltages. hFE of Q1 is easy to read (assuming Q2 hFe is not way-low).
I'll look at that Peak theory another day.
Welcome back. Long time no hear.
Thanks Paul I will digest the information you have given. Appreciate your time!
Thanks Kipper, long time no see! Hope everyone here is doing well.
i have proven the methods are indeed the same..all you really need is a good dmm
Most of us are fine Paul. Some have had bumps in the road but mostly fine.
Thanks and you?
Thanks LSG ill check out the video when i get a sec.
EDIT: watched the video. The test you performed is fine for most germanium transistors. Unfortunately there are some that wont react the same with the peaks higher test current, which is "whatever base current it takes to see 5ma on the collector" and unpredictable. Try an IT308B (mil spec GT308V) with the same test and you will find the readings are very different. There are a few others that i have found that react the same. This is the reason i would like to test with a curve trace or simulated small bear / r g keen test on the computer so that i can specify "more fuzz face like conditions" then the peaks unpredictable auto test. Its probably essentially overkill and also overly complicated but that sums up my life in general. Side note: From my understanding, Leakage didn't change in your test because it will only change with temperature and not increase voltage / current. I have yet to test this but this is what i have read.
Cheers Kipper, busy as hell...youtube, webstore, day job, 2 young children and god knows what else....Somedays they let me have some sleep :-\
Quote from: chromesphere on June 18, 2016, 07:23:44 AM
Unfortunately there are some that wont react the same with the peaks higher test current,
which is "whatever base current it takes to see 5ma on the collector" and unpredictable.
Try an IT308B (mil spec GT308V) with the same test and you will find the readings are very different.
There are a few others that i have found that react the same.
This is the reason i would like to test with a curve trace or simulated small bear / r g keen test on the computer
so that i can specify "more fuzz face like conditions" then the peaks unpredictable auto test.
Did the test with an GT308b.
With Ib=4µA Hfe is around 27,
with Ic=5mA Hfe is 59.
That's twice as much.
One could say the R.G.Test ist unpredictable.
But it is the 'standard' for comparison at least for fuzz faces etc.
And it is a fast test.
I use the PEAK always with a PC for its curve tracer function,
the clips connected to a socket for easier handling.
Testing at Fuzz Face operating points:
Q1 should be tested Ic=100-200µa @1.5-2V Uce
Q2 should be tested Ic=500µA @5V Uce.
But feel free to tweak collector resistors to make a lot more of transistors usefull.
electrip
The IT308B is the extreme, as you found out with the peak big difference. But from my brief experience with the peak i have found that all transistors measure slightly higher (10-20%).
Thanks for the info electrip ill try your method!
Paul
Quote from: chromesphere on June 19, 2016, 08:17:52 AM
The IT308B is the extreme, as you found out with the peak big difference. But from my brief experience with the peak i have found that all transistors measure slightly higher (10-20%).
Thanks for the info electrip ill try your method!
Paul
Yes, I got the same results with Peak DC75 not only HFE is 10-20% higher but also Leakage measurements are much lower than when taken with RG Keen circuit.
Leakage with the RG Keen circuit always measures much higher than with any other transistor tester I've used, sometimes it measures 2x more leakage than the other testers.
I'm confused
Why is it a problem? Which number you assign to the hfe or amount of leakage is not important as long as you know how your transistors compare with each other. Once you see how the transistor biases in the circuit you're building you'll knowif you need a higher or lower leakage one.
QuoteDid the test with an GT308b.
With Ib=4µA Hfe is around 27,
with Ic=5mA Hfe is 59.
That's twice as much.
One could say the R.G.Test ist unpredictable.
But it is the 'standard' for comparison at least for fuzz faces etc.
And it is a fast test.
hfe = hfe(ib), or , hfe is a function of base current.
For ACxxx this effect can help to clean the sound as you roll back the guitar volume.
mac
Quote from: Electric Warrior on May 25, 2017, 09:45:59 AM
Why is it a problem? Which number you assign to the hfe or amount of leakage is not important as long as you know how your transistors compare with each other. Once you see how the transistor biases in the circuit you're building you'll knowif you need a higher or lower leakage one.
I'm inclined to agree. Who cares what the number is? Use it as a guide to evaluate a pile of transistors and then decide if you need "higher" or "lower". While the Peak measurements won't match up exactly with R.G.'s method, it'll do the same job.
Tom
Quote from: Electric Warrior on May 25, 2017, 09:45:59 AM
Why is it a problem? Which number you assign to the hfe or amount of leakage is not important as long as you know how your transistors compare with each other. Once you see how the transistor biases in the circuit you're building you'll knowif you need a higher or lower leakage one.
Well for anyone that follows RG Keen advise it's a problem for sure. You would discard transistors with an higher leakage, higher than 300uA or 500uA.
From RG Keen in http://www.geofex.com/article_folders/ffselect.htm:
"How much leakage is too much? 100uA is common, 200 happens pretty often. More than 300uA means the device is suspicious, and more than 500uA I would say is bad."
Small Bear also only selects transistors with leakage under 300uA.
It gets a bit frustrating then when you have some transistors that measure 250uA of Leakage in the Peak DC75 and 620uA when measured with the Geofx circuit.
So what I was saying is that with the RG keen circuit measuring Leakage much higher than any other tester, and sometimes even 2x more, a lot of people that built that circuit will be inclined to trash a lot of transistors that might be perfectly fine.
In the same way the oposite can happen, you measure some transistors with the Peak DC75 and it gives a small leakage value, when in fact the leakage in much higher and the transistor is pure crap.
And this is a stompbox building forum, not the NASA Headquarters, so newbies are welcome and people are learning, there's nothing wrong with that.
Some people that already built fuzz pedals and that tried to select germanium transistors know that listening in the circuit is also a part of the selection process and that you may find that transistors out the guideline ranges suggested by RG Keen can work well for a nice sounding fuzz also, even not following the suggested HFE values for Q1 and Q2. Transistors with an HFE vlaue of 40 can be used as transistors with an HFE value of 220 also. The same for leakage, I would even suggest to never ditch transistors with an high Leakage value without listening to them first in the circuit. But most people will just follow the guidelines and thats totally fine and I complete respect and understand the reasons why.
In my case I would like to understand the reasons for such a difference when measuring leakage with the 2 meters.
For the HFE 10/20% higher reading on the DC75 it's easy to understand has the DC75 measures Gain using an higher base current than the RG Keen method, so to match the 2 measurements you have to use the curve tracing feature of the DC75 and look for the base current you want.
For the leakage big difference I still don't understand the reasons.
So Peak DC75 makes the transistors look good, RG Keen Circuit makes the transistors look bad,
what's the more appropriate of the 2 to help on this Job (selecting transistors for Fuzz pedals)?
Thats what I'm trying to understand and learn
Thanks
Quote from: ElectricDruid on May 25, 2017, 06:41:55 PM
Who cares what the number is?
RG Keen cares, and a lot of people seem to follow his advice. Also it's the biggest reference in selecting transistors for Fuzz pedals since the year 2000.
"How much leakage is too much? 100uA is common, 200 happens pretty often. More than 300uA means the device is suspicious, and more than 500uA I would say is bad."
from Smal Bear Electonics website, Steve Daniels seem to follow RG Keen also:
"I sort and test to the following spec: With a 9-volt supply, collector-emitter leakage under 300 microamps at room temperature with the Base open."
Quote from: ElectricDruid on May 25, 2017, 06:41:55 PM
While the Peak measurements won't match up exactly with R.G.'s method, it'll do the same job.
Well if one measures leakage at 250uA it will approve the transistor
if the other measures leakage for the same transistor at 620uA it will say the transistor is Bad
Completely different results, so no, they are not doing the same job
So you're throwing out perfectly good transistors just because someone on the internet said they're bad? ;D
Quote from: ilcaccillo on May 25, 2017, 08:25:56 PM
Quote from: ElectricDruid on May 25, 2017, 06:41:55 PM
While the Peak measurements won't match up exactly with R.G.'s method, it'll do the same job.
Well if one measures leakage at 250uA it will approve the transistor
if the other measures leakage for the same transistor at 620uA it will say the transistor is Bad
Completely different results, so no, they are not doing the same job
Obviously if you're using a different measurement method, you can't use the same threshold values. Leakage isn't going to be fixed point like "the boiling point of water". It's more like "fuel consumption" - you need to know how much weight there is in the car, how fast you're going, etc etc to make the figures even vaguely comparable.
R.G.Keen's "throw away above this" value is his opinion based on his measurements with his tester. If you're doing measurements with a different tool, you'll need to make a different limit, and that can reflect your own tastes.
You seem to have done measurements with both tools on a set of transistors, so you should have a good set of data to find a comparable figure for the Peak. Or isn't that right?
Tom
The hFE of a transistor varies, depending on the collector current. You can look on a datasheet and see a typical hFE of 150 for a transistor, but on closer inspection, that gain is for a current of 150ma and when the current is reduced to 0.1ma (100ua) then the hFE has dropped to 20. (from 2N4401 datasheet)
Testing at different collector currents will give different results.
Also, the gain and leakage of a germanium transistor will change with the temperature. Even a few seconds of handling with your bare hands will cause the numbers to change:
http://www.muzique.com/news/effect-of-heat-on-ge-transistors/
Best regards, Jack
Interesting. When I did the same experiment (using R.G. Keen's circuit) hfes seemed rather stable while the leakage increased considerably with temperature.
Thanks Jack. That was the point I was trying to make, but you've done it much neater!
T.
Who can explain what conditions we need to simulate RG Keen jig or FuzzFace transistors using DCA75?
Ib=4uA, Vce=4.5?
As I've wrote DCA55 and Chinese testers gives measurements similar to Small Bear testing jig, do we really have any "issues" with RG Keen schema? Did anyone compared them?
Quote from: diy-tubes on May 27, 2017, 08:15:51 AM
Who can explain what conditions we need to simulate RG Keen jig or FuzzFace transistors using DCA75?
Ib=4uA, Vce=4.5?
As I've wrote DCA55 and Chinese testers gives measurements similar to Small Bear testing jig, do we really have any "issues" with RG Keen schema?
There might a problem yes, specially in the Leakage measurements.Something is not right.
As for HFE, I have to look at the curve tracing feature in the DC75 so check if the spot where base current match also matches roughly both HFE measurments.
The RG Keen measures at aproximately 4uA base current, assuming some conditions are met
from the website:
"The exact base current is 4.046...uA, assuming that the transistor's base conducts that much with a forward voltage of 0.1V (reasonable with germanium at these currents) and that the battery is *exactly* 9.0000V, and that the resistors are 2.20000M, and... well, you get the picture. 0.5% accuracy is doggone fine for work with such blunt tools, and much better than you actually need to make a fine sounding FF."
Read this to scale RG tester,
http://www.diystompboxes.com/smfforum/index.php?topic=45481.msg332690#msg332690
I use 1k and 1M. I multiply by 119 for Si, 112 for Ge, to get hfe.
mac
Quote from: chromesphere on June 16, 2016, 12:27:30 AM
Thanks Paul, was hoping for your reply and appreciate your knowledge. The DCA75 can cruve trace yes.
I guess im trying to simulate R G Keens circuit because he sort of set down the guidelines as far as hFE requirements for fuzz faces etc. Everyone seems to follow this (some blindly) without questioning further. I suppose for me, the underlying question is, why do we put (x microamps) of base current into the germanium transistor and hope for (x hFE)? For Keens its 4ua and the usual hFE's we know and love (Q1 ~70 Q2 ~120 etc). But the question is why those values? Is this the best representation of the conditions of a fuzz face? Some transistor hFE will vary significantly in response to increased base current / voltages.
EDIT: Although its probably over analysing I'm really mainly just curious to know.
Hello Paul, sorry to bother . I would like to know if the results shown on the curve tracer graph are the "correct gains" or do i still need to substract the leakage the certain point on the graph that im looking at?.