I have recently built a phaser which my son like so much he wanted me to build him one too
the only problem is, that the jfets i bought i think turned out to be fakes. After buying some more the new batch dont work and measuring completely different
so i dont know what they really are.
I was wondering if anyone could assist how to identify or use the info i have to find something similar?
using my chinese $7 component yellow box
(https://i.ytimg.com/vi/7Br3L1B80ow/maxresdefault.jpg)
they come up as a NPN transistor with hfe 96.8K! vf 469mv
so i used my RG jfet tester and got a reading range of -0.13 to -0.2 on my working jfets
on page 13 and 14 is a good little comparison chart of some jets
can anyone assist / advise what alternatives that may be suitable?
is it the VGS(off) mins and maxs i should be looking at?
http://www.mouser.com/catalog/supplier/library/pdf/fairchildsmallsignal.pdf
more info please - what parts did you put in the working phasor phaser, what parts have you sourced that aren't as you think? most importantly, what do your "bad" fets say when R.G.'d?
this just in: which phaser circuit?
Seems to me they're testing as very leaky Ge transistors!
What were they supposed to be?
roland phase 5
2sk30
i originally got the working ones about 3-5 years ago from futurlec, recently got some more from them and completely different to new offerings
will R.G.'d the new batches results shortly
K30A - check your pinout. again. I'll just leave this here, y'no, just in case .....
yeah pinout is fine :)
anyone??
Here you go..
(https://i.postimg.cc/ZvyY3SfQ/2-SK30-Pinout.jpg) (https://postimg.cc/ZvyY3SfQ)
is that what this symbol means?
as stated earlier i have jfets which aren't what they are labelled , they come up as npn transistors when tested, all the others ive bought which measure as 2sk30 measure as intended.
im trying to use the little information i have using the geofx testing reults etc to work out their specs so i can possible look at ordering something with similar specs
5457
j201
j112
5452 etc i have on hand all dont work, is there a way of identifying jfet specs another way?
(https://i.postimg.cc/WpwxXtyp/IMG-2862.jpg)
here is its readings
and here is a pic of the jfet
(https://i.postimg.cc/V6CmJ1sr/IMG-2864.png)
Ah...Idiot here thought that you were after the 2SK30 pinout :-[ Sorry.
I have the exact model component tester that you have and another one (but different ) that I built.
Both can get rather confused with JFETs and think that they are testing a standard transistor.
But when testing my current stocks of 2SK30 JFETs both testers do show that the device being tested is a N channel JFET. But I am testing 2SK30A GR devices which usually have a higher Idss than the Y type. This difference seems to be the thing that causes these little testers to work or not work. Also the data that they display when they do work with a JFET is not very reliable nor believable.
The geofx tester is a great way to test a JFET and I use it often but only to test Vgs. Drain current is another reading that can help with matching.
the range using the jfet matcher ive been getting is...
-0.140 to -0.150
where 5458's are giving me -1.69
real 2sk30a are -0.825
j112 -3.05
5452 -1.724
so as you can see the fake ones that funnily enogh work great are getting very low
readings , and its trying to use this info to select an alternative?
My test results using the geofx tester when testing a 2SK30A GR is;
Vgs = -1.320V (which is about what I would expect )
I'm not set up at the moment to test Idss but give me an hour I can do so.
But the LCD tester (same as yours ) shows;
Vgs = 0.82V
I = 1.2mA
So I can't really believe that what the tester is indicating is correct.
From my limited experience with measuring jfets - forget about using this component tester (I have a similar unit and readings are not reliable). Use battery and multimeter (https://www.diystompboxes.com/smfforum/index.php?topic=121610).
T.
QuoteFrom my limited experience with measuring jfets - forget about using this component tester (I have a similar unit and readings are not reliable). Use battery and multimeter.
I'd go back to basic pin identification measurements and use a multimeter to identify the pins.
Use the diode function to find the common pin which gives a voltage reading of around 0.6V (normally within 0.55 to 0.75V) to the other two. For NPN BJTs and N-channel JFET the diode's anode should be the common pin. That would be the base or gate. Now switch to 2k ohms range and measure the resistance between the *other* two pins. If you get a reading in both directions say less than 1k then it's a JFET, if you don't get any reading at all then it's a BJT.
For the MOSFET case there will be a common pin which gives no reading to the other two (for both polarities).
Once you do that see what matches up with the pin-outs you have and what your gizmo says.
Oh, there might be a software update to the gizmo that comes-up with "Fake" :icon_mrgreen: :icon_mrgreen:.
(https://i.postimg.cc/02mszFFh/2sk30-pin.jpg)
Diode test:
Red - Green : nil
Green - Blue : 0.675
Red - Blue : 0.25
resistance test:
Red - Green : nil
Green - Blue : 475K
Red - Blue : nil
$7 yellow box says the following:
NPN
RED - Collector
GREEN - Base
Blue - Emitter
hFE = 85.9K
Uf = 464mV
I'd think any bjt indicating a gain of 86,000 would be iffy, one way or another. and if I thought I was testing a jfet that read as a bjt, with an hFE of 86k .....
I've never had one of those boxes to test with. an R.G-o-Meter has always done the good thing.
Ducks arse, which rgmeter have you had most success with? I've built a few of the online fet testers and have had very inconsistent results. I've measured them once then gone back to them and obtained different results. I've also got one of those yellow boxes and haven't had that problem but I'd like a bit more granularity.
sorry, got my -o-Meters confused. this is the ROG, the one I've used most often:
(http://runoffgroove.com/fetzertech10.png)
I think there is another, possibly an RG, that uses an opamp, but I couldn't see it when I searched my local files.
QuoteDiode test:
Red - Green : nil
Green - Blue : 0.675
Red - Blue : 0.25
resistance test:
Red - Green : nil
Green - Blue : 475K
Red - Blue : nil
The 0.25V is an odd value. Like it could be germanium but that's inconsistent with the 0.675, unless it's a Darlington germanium which is extremely unlikely. Then there's the possibility there an internal resistance across Red-Blue but that's inconsistent with the resistance test.
My best guess at this stage is it's an NPN transistor.
blue = base
green = emitter } these could be flipped but that's my best guess.
red = collector }
Quote$7 yellow box says the following:
NPN
RED - Collector
GREEN - Base
Blue - Emitter
I suspect it's confused since GREEN-base is not consistent with your manual test. That's backed up by the very high gain value which I suspect comes from a heap of current going down whatever lead the box uses to measure current.
I'm also confused but not as much as the box.
It would be very useful to wire up something like this tester:
https://cdn.hackaday.io/images/4589791484797920897.png
You don't need the diode part and the 9k, and a 1M is fine for the 900k. Replace the 1mA current meter with a 220 ohm resistor and measure the voltage across that resistor; *roughly* 100mV is hFe=100. As an additional sanity check to make sure the base current has an effect on the collector current I'd do a second check with the 1M replaced by 470k. You should see double the current.
If you get low gains try flipping red and green.
i began breadboarding those 2 circuits but got quickly frustrated!
I went back and read over the post.
it just occured to me that pending the diode tests probes orientation can give different resuilts!
Doh!
heres what i got second time round
Diode test:
(https://i.postimg.cc/02mszFFh/2sk30-pin.jpg)
Red - Green : 0.675
Green - Blue : 0.675
Red - Blue : 0.25
i have paused on pursuing the transistor testing for the moment hoping that this new info may refine what to do.
what is strange is this transistor works great in this phase pedal!
just the hassle to identify what it is or what it may be similar too!
here is the jfet geofx tester results
(https://i.postimg.cc/J4xwrSpR/IMG-2872.jpg)
(https://i.postimg.cc/sXY6nZ83/IMG-2873.jpg)
QuoteRed - Green : 0.675
Green - Blue : 0.675
Red - Blue : 0.25
OK that's looking much more sensible - well maybe.
N-channel Jfet
Green = Gate
Red, Blue = Source and Drain (use pin-out as it's often not possible to distinguish the two.)
I believe the Green pin does map to the gate on the K30A.
The hanging "issue" is the fact that the resistance measurement didn't show anything between Red and Blue.
So maybe the next step is the circuit duck_arse posted.
ok cheers ! i will get onto it:)
ok, using this tester heres what i got
(http://runoffgroove.com/fetzertech10.png)
(https://i.postimg.cc/02mszFFh/2sk30-pin.jpg)
Drain- Blue
Gate - Green IDSS 0.053v (not multipled) VP 0.022
Source - red
Drain- Red
Gate - Green IDSS 2.626v (not multipled) VP 0.43
Source - Blue
QuoteDrain- Blue
Gate - Green IDSS 0.053v (not multipled) VP 0.022
Source - red
Drain- Red
Gate - Green IDSS 2.626v (not multipled) VP 0.43
Source - Blue
The results aren't "normal" but at least now they are starting to look like some sort of JFET.
The things that still look odd are the Drain=Blue results and the lack of symmetry.
Also if I calculate Rd = Vp / (2 Idss) = Vp / (2 * VIDSS * 10 *1e-3) = 50 Vp / VIDSS
Blue=drain: Rd = 50 * 0.022 / 0.053 = 20.8 ohms
Red = drain: Rd = 50 * 0.43 / 2.626 = 8.2 ohms
I would have expected those to agree a bit more. If they were signal JFETs you would expect something around say 200 ohms to 400 ohms.
The Rd value is the same as 1/gm0 which you can see from,
www.kennethkuhn.com/students/ee351/jfet_basics.pdf
The resistance between D and S measured with a multimeter should be in the ball park of Rd.
So at this point:
- it's looking like an N-channel JFET.
- The low Rds and lowish Vp look more like a switching JFETs.
- The parameters are way off signal JFETs so you will probably struggle getting
them to work in most effect pedal circuits.
- The asymmetry could be an indication of a factory reject.
You could check a few others to see how consistent they are.
If it were me I probably try to see if they actually obeyed the JFET square law for Id vs Vgs. That would be very convincing the devices are JFETs.
BTW, nice multimeter!
i bought these from futurlec , i was so happy with the results, that few years later i bought some more from them,
similar markings . the first batch worked great in a roland jet phaser clone too.
the new ones identify in the yellow box as jfets
I=0.084a
vgs=0.59v
i think your right regarding switching jfets.
i can make a great phase by bridging the jfet socket source and drain with a pair of tweezers on and off.
i measured the resistance between the source and drain and they vary between 1.6k to 2.1k
do you suggest any specs or ballpark varieties of jfets i may order to test?
i was having a look at the small signal transistors and jfets guide which i linked at beginning of thread
to see what examples may be closer to spec required
i think what im looking for is jfets with low vgs (off)? im not sure still trying to comprehend your
advanced knowledge!
yes buying that dmm second hand on ebay was one of the best things i did, the continuity tester is fantastic.
Quotei bought these from futurlec
To be honest, you should tell futurelec you have been shipped dodgy parts.
If they are reputable at all they should exchange the parts and carry out their own investigation
to see which of their suppliers is shipping dodgy parts. In this day and age they have to admit there's
a lot of fakes out there and they should be on top of it. I don't see why the customer should suffer.
Quotei measured the resistance between the source and drain and they vary between 1.6k to 2.1k
So that seems inconsistent with the low Rds values. The bottom line is there's something weird going on
with those parts.
Quotedo you suggest any specs or ballpark varieties of jfets i may order to test?
If you get parts with Vp values of 1 to 2V (and Yfs specs in the 2000uS to 8000uS) zone
you should be safe in most circuits. That's not to say parts with *specs* for Vp of 0.5V to 8.0V won't work,
For one you rarely get the extremes in practice. Parts with Vp=0.5 can be more finicky to adjust and high Vp's with like 3.5V will certainly work but may benefit from increasing the LFO output. When you get to Vp's of 5V the bias point has to be set to undesirable values.
The 2SK30A-Y's are fine but the problem is your parts aren't really 2SK30A-Y's!
Some other common ones are 2N5457, 2N5458, MPF102, BF244 (A or B), BF245 (A or B). People have got J201's to work but they probably aren't the best choice for a phaser. I don't know what is available these days. The lists keep shrinking and (fortunately) I'm not buying a lot of parts these days for DIY.
QuoteTo be honest, you should tell futurelec you have been shipped dodgy parts.
the bodgy ones that worked i bought 3 years ago so i just assumed they were good!
I have most of the other normal jfets for phasing and none work , it appears the volatges are too low range for their ideal workings. the only other option i think i may have to do is use more common working jfets and adjust the circuit so it lfo cycles better to their values?
Quotethe bodgy ones that worked i bought 3 years ago so i just assumed they were good!
Interesting indeed!
QuoteI have most of the other normal jfets for phasing and none work , it appears the volatges are too low range for their ideal workings. the only other option i think i may have to do is use more common working jfets and adjust the circuit so it lfo cycles better to their values?
That's a bit weird. What circuit are you using? Maybe the circuit needs a tweak to help it work with other JFETs. The voltage at the gate needs to swing down to Vp below the source and up close to the source; hw close determines the range of the sweep. Normally the bias adjustment sets the Vp condition. If you don't get the bias right it's pretty much doomed from the start. The upper limit of the gate swing is determined by the swing of the LFO and the resistor values that combine the LFO and the bias voltage. Often the resistors affect both the bias and the LFO sweep range so you need to tweak those so both aspects are just right.
There is a JFET linearization article here as applied to an ST-1000A FM signal generator:
http://ham-radio.com/k6sti/st-1000a.htm
This quote is important:
"A trick to linearize JFET channel resistance may not have been known when the ST-1000A was designed. Applying one-half of the drain signal to the gate makes any variation in channel resistance symmetrical, cancelling even-order distortion. This kills the dominant second-order JFET distortion product, greatly reducing THD."
This is for a JFET used in an oscillator but the principle may apply.