mmbfj201 compare to the original fairchild

[HeavyFog]

So how close do they compare to the originals, say in the context of a common source amplifier overdrive like the many ROG designs? Obviously the datasheet would point to them being identical but the parameters are pretty wide. I bought a handfull of interFET J201s and even though the datasheet said it was exactly the same as the fairchild, measuring them showed they consistently spec'd different than the older part, but still within datasheet parameters. Even the 3 different makes of J201 i got from smallbear (Siliconix, National semi, Fairchild) all measured and sounded different from one another.

Waiting on some SMD adapters now so in a week or 2 il be able to actually test it out properly, but i'm curious what everyone else's experiences have been with the mmbfj201. If they sound exactly the same as the fairchild part that would solve a huge issue of supply for me, and i have no issue soldering SMD adapters if it means i dont have to worry about running out of JFETs.

Not here to argue about the merits of the ROG style of pedals, just looking for some advice before i make an investment on a bunch of these.

[stonerbox]

I have spent too many days comparing TO-92 J201s to the SMD counterpart and in my humble opinion there is no difference in characteristics. I tried the SMD and TO-92 in different setups, amp configurations, pedals, voltages etc.. when my ears/head was rested and tired, when I was happy, indifferent and angry and I never picked up any texture, dynamic or frequency response difference. That said each J201 can and will differ significantly from one another but there is absolutely no audible difference between a SMD and a TO-92. Others may say differently but that is my personal conclusion.

[aion]

They're exactly the same, except the SMD ones are tighter tolerance with far less variance in Idss and Vgs(off). But the average spread is well inside the average spread of the through-hole version, so in practice they'll not only be identical but even more consistent unit to unit.

This is only true comparing Fairchild/ON Semi J201 to ON Semi MMBFJ201 though - the same manufacturer using the same recipe. J201's from another manufacturer like InterFET will technically meet the J201 datasheet specs but will likely be nowhere near ON Semi's in practice.

[bowanderror]

I've had similar experiences with SMD JFETs; more consistent Vp & Idss than through hole. Comparing one manufacturer to another is not really feasible as they vary more than enough already from batch-to-batch within a manufacturer.

I don't have much personal data on through hole J201s, but here is my JFET testing spreadsheet with 2x TH J201s for reference, and SMD J201s from 2 different sources:



I've found a pretty similar standard deviation between sources of SMD J201s, but there is ~15-20% difference in their averages. Fortunately, most circuits can handle such variation.

[Rob Strand]

A useful parameter to check JFETs are similar is to plot or tabulate rds_on = VP/(2*IDSS).Similar JFETs will show quite tight rds_on agreement.  The part tolerance shows up mainly in VP and rds_on categorizes the part.  Of course you can get similar rds_on and *widely* different VP on different part models.

You can see the constant slope in the data plotted on this page, (thanks to member Eb7#9),
https://viva-analog.com/characterizing-and-matching-2n5457-jfet-transistors/


Re-jigging stats

Stat Summary         
               VP [V]   IDSS [mA]   rds_on [ohm]
av (TH)        0.925   0.795     584.7
sd%           8.4      15.1        6.8
av (SMD AU)   0.728    0.562     648.5
sd%            3.9       6.9        3.2
av (SMD amazon) 0.634    0.465    688.3
sd%            7.9      14.8     6.9
         
avg (pooled SMD) 0.663    0.495    676.0
sd%            9.4         15.3    6.6


So rds_on is about 16% higher on the SMD.
Probably a bit higher than usual for rds_on but not bad.
There's only a few samples on the through hole.

EDIT:
I don't have all my JFET data on my current computer but I found some info saying
the J201's have some dependency between rds_on (actually 1/Gfs) and VP.
From that info the fact the larger VP for the through-hole parts leads to smaller rds_on.

EDIT2:
I chased down where that info came from.  It's from the datasheets.
The datasheet graphs often show Gfs with a VP dependency but measurements show different.

I got this info from the ROG site for 10 x J201's.


Out of ten (10) J201
average Vp: -0.83V (spanning from -0.57 to -0.97)
average Idss: 0.64mA (spanning 0.30mA to 0.83mA)

then,
Gfs = 1542uS
Rds_on = 648 ohm

The 10 samples are more representative than a sample of 2. 
You can see the Rds_on value isn't far off the SMD devices.

[HeavyFog]

Thats interesting, the fairchild j201s ive been using are on the older side (pre 2016 ON acquiring fairchild atleast) given through hole jfet manufacturing has been pretty slow in the last 15 years. Whether or not the mmbfj201 is a reproduction of the fairchild j201 specifically or the pre fairchild aquisition ON j201 will have to come down to a good test. I've never used or even really seen ON branded J201s before but i can only assume they are very similar all thing considered.

By the sounds of it though they should be close, and a more even spread actually in the average range of vintage specs (unlike the interFET) would be a perfection. The adapters i ordered are the wrong pinout apparently so il have to order in a new set, then i can test it properly. I will definitely buy or build a good JFET tester, given how often i use them it would be useful to be able to properly sort them out. 

[Chillums]

Not sure if your adapters are the same as the ones shown here but this is what I did to correct the pinout problem.  Doesn't look pretty but for testing it should suffice.

[Ice-9]

Not sure if your adapters are the same as the ones shown here but this is what I did to correct the pinout problem.  Doesn't look pretty but for testing it should suffice.

Are those adapters not double side ? Many i have seen are double side and are routed differently so that you can alter the pin outs depending on what side you solder the jfet to.

[Rob Strand]

By the sounds of it though they should be close, and a more even spread actually in the average range of vintage specs (unlike the interFET) would be a perfection.

To me the SMDs are looking pretty close to the through-holes.   You can't have high expectations with JFETs matching.

For example,
- there is about the same disagreement of VP between the two through-hole batches (ROG and bowanderror) and the two SMD batches.
- the rds_on of ROG's through-hole vs the SMD's agrees better than ROG through-hole vs the bowand error through-hole batch.

All the parameters seem to be bundled around the same region.

The only pattern I can see with this small sample is the VP's of the SMDs are lower.    One pattern which seem be emerging for all modern JFETs is lower VP's.

If you want to split hairs down below the sub 10% region the differences in the test jig used to get IDSS and VP needs to be factored in.

For example the ROG test method, which is fairly accurate, will produce VPs upto 3% low when using a 1M ohm input impedance multimeter and about 1% low when using a 10M ohm input impedance multimeter.    The Vgs measurement on RG's test jig (and the kits which copy it) is lower than the true VP value by a factor of 1.4 to 2.0 (some of the VP posts on the forum have this error factored into the results).   

IDSS measurements can be affected by self heating and the Vds test voltage (due the gate modulation effect which is the Lambda parameter in spice).

[HeavyFog]

Truthfully i never really had many problems with variation between JFETs by the same make, i usually dont even match them and it always works out for me (at least in an overdrive). I'm more concerned with variation between different brands of JFETs, like each of the 4 brands of J201 being slightly different from one another. As long as the through hole ON/fairchild mmbfj201 specifically is close in sound to the to92 fairchilds i got then i'd be happy. I have lots of Siliconix and National through hole J201, but only around 75 Fairchild, which i prefer the sound of.

Probably should have clarified earlier that when i meant the ROG method i was referring to that style of JFET overdrive, rather than the testing method. I have used the rog testing circuit before though, and had some luck with it, i would probably be better off getting myself a good digital tester though given how often i use JFETs. Il take some new measurement of my own sometime of what i have and the mmbfj201. Just have to wait for my adapter boards to show up.

[HeavyFog]

Well interesting timing, I was just doing some looking around, and i noticed that Digikey has a new listing for the Linear systems through hole J201s, and for a fairly reasonable price! https://www.linearsystems.com/product-search-result.html?type=products&partnumber=J201
They describe it as a direct replacement for the Vishay/siliconix and ON/Fairchild part which is exiting, and it looks like they'll be selling for $2 (usd) a piece. Still a far cry from the 0.50c you could get them from through smallbear, but still much better than the $5 for the off-spec interFET.

Digikey has them listed but it doesn't look like they're for sale just yet, but when they do il buy a handfull and pit them against the classics. With any luck they'll actually spec close to the vintage fairchilds unlike the interFET. But either way its very exiting to see new current production of a classic through hole JFET! If they do fit well into the vintage specs, this could be a big deal for the DIY community! Someone ought to tell smallbear about this.

EDIT: looks like these might have been available for some time now, i just haven't ran into them before. Not sure if production stopped, and was recently picked back up again or if they've been going this whole time, but either way it is nice to see 10k of these on Digikey!

[Rob Strand]

I cobbled together some data I had on my laptop to increase
the sample size for the through-hole.

J201 pooled data from a number of sources
sample size  29

Indicated 'avg' variables are averaged from data and other
variables are calculated so set of parameters is consistent

                      avg        avg
                      VP & IDSS   VP & Gfs    min        max
        VP [V]        0.83        0.83        0.54        1.03
        IDSS [mA]     0.61        0.60        0.23        0.88
        Gfs [uS]      1479        1457        850        1796
        Rds_on [ohm]   676        686         557        1176

Not overly different to the ROG data (which I've massaged into these figures anyway).

[POTL]

Years ago I had a stack of mmbfj201 and regular j201 they easily replaced each other. SMDs were within the specifications of 10 pieces, only 2 (or so) differed in parameters from others. of course I didn't compare different parties, but within the same party everything was fine.