I generallly try to sort devices by looking at Vgs at the intended drain current. I use a home made jig to test this that runs off a pair of 9V batteries and uses a pair of DVMs for voltage and current sensing. I misplaced my sorting jig during a move, and was forced to sort using IDSS. This didn't work too badly, either. A note of caution - this (IDSS) technique works best for JFETs where the IDSS isn't too high, like the 2SJ74, 2SK170, 2N5457-59, etc. Chopper FETs like the PN4391 series will have a high IDSS, and you may run the risk of burning out devices during a sorting session.
With the caveat that you want enough voltage so you're sure you're limiting due to pinch-off, not just insufficient voltage. This is definitely an issue for chopper-type FETs with high off voltages.
Thanks for the replies.
I will have to measure Idss of some jfets that I matched by pinch off voltage. I am curious as to what the correlation will be.
Does a jfets "Rds on" follow from its Idss? I can't quite get my hands around the idea of why we want to match for Idss -- as long as we are operating the jfets well below Idss.
PS--sorry for the tiresome questions, but surely everyone must be used to it from me!😀
JJ
I will have to measure Idss of some jfets that I matched by pinch off voltage. I am curious as to what the correlation will be.
Does a jfets "Rds on" follow from its Idss? I can't quite get my hands around the idea of why we want to match for Idss -- as long as we are operating the jfets well below Idss.
PS--sorry for the tiresome questions, but surely everyone must be used to it from me!😀
JJ
jupiterjune said:
The Rds is not directly related to the Idss. Generally speaking,
we match for Idss because it's easy. You ground the Gate and
the Source and apply a consistent voltage on the Drain and
measure the current. You can measure the current at Vgs
values other than 0V if you want to.
Nelson's comment that it's easy is right on the mark. Idss pops up when you hook up the JFET as a full-on current source. Pinch off, you have to search for.
If you want to measure something else, wrenchone's suggestion to measure at the intended operating current is good. Better still, do a preliminary sort by Idss, then sort that subset at your intended operating current. This isn't as big a drag as it is with power devices since JFETs usually stabilize their temperature pretty quickly, which gives you an accurate read on things.
Grey
If you want to measure something else, wrenchone's suggestion to measure at the intended operating current is good. Better still, do a preliminary sort by Idss, then sort that subset at your intended operating current. This isn't as big a drag as it is with power devices since JFETs usually stabilize their temperature pretty quickly, which gives you an accurate read on things.
Grey
Someone correct me if I'm wrong, but, isn't operating near Idss the most temp-stable point, the most linear area and even more so in a high Idss devices?
Unfortunately, the temperature coefficient isn't quite that accommodating. There's a formula for it, but it's not one of the ones I carry in my head.
There are any number of arguments that can be advanced regarding the amount of current to run in a given device in a given topology. I'm kinda fond of the one that says that it gives you more current with which to drive the next stage.
Grey
There are any number of arguments that can be advanced regarding the amount of current to run in a given device in a given topology. I'm kinda fond of the one that says that it gives you more current with which to drive the next stage.
Grey
The problem with running at IDSS is there's no place to go but down...I have seen designs for moving coil amps with the JFETs running at IDSS, meaning there must be a small amount of forward gate bias.
Yeah, a moving coil signal is about the limit I could justify running at Idss, although other folks do more. I've never been comfortable about running anything--tubes or solid state--past the "10" setting on the volume knob.
No, I've never even tried a moving coil design that way. Maybe once I get this amp design put to bed I'll try that, just for fun. I've been wanting to do another phono for ages.
Grey
No, I've never even tried a moving coil design that way. Maybe once I get this amp design put to bed I'll try that, just for fun. I've been wanting to do another phono for ages.
Grey
jupiterjune said:
My reason is that the same Idss means that the probability of the same transconductance curve is higher than I believe.
Apparrently not, it really depends on the type of device. This and more is competently explained in the Borbely JFET articles: http://www.borbelyaudio.com/special_articles.asp
There are more sources of information off course.
P.S. And I "think" there is no reason to compromise other parameters of a circuit just for the sake of temparature stability
Temperature stability is good, but given that it's not necessarily at a convenient point for folks building audio circuits, I regard it as a nifty bonus, not a requirement.
Grey
Grey
Well, to put information where it is germane if not timely, there are two TO92 packages for J174. Looking at the flat side with the leads down, left to right, National Semi & Fairchild J174 is DSG. Same perspective, Vishay Siliconix J174 is DGS. My dead transistor has a little square looking S on it, so I think it is siliconix. It is sure the pin in the middle is the gate, by the board schematic diagram (Peavey PV1.3k amp). My replacement transistor is a Fairchild. The minimum IDss matches up, the on resistance matches up. The maximum IDSS is a little wimpy onthe fairchild, 100 ma vs 135 on the Siliconix, I hope that is okay.
Edit, the fairchild datasheet dated (no date I downloaded it fall 2011 from datasheetcatalog.com) shows the same gate on the end that National Transistor databook 1975 has. The actual F part from Newark has 600 ohms from both ends to the middle pin. I guess they have changed.
Edit, the fairchild datasheet dated (no date I downloaded it fall 2011 from datasheetcatalog.com) shows the same gate on the end that National Transistor databook 1975 has. The actual F part from Newark has 600 ohms from both ends to the middle pin. I guess they have changed.
Last edited:
- Status
- Not open for further replies.