I recently built the PASS 6-24 Diy crossover. I got two boards to do a three way but figured it's also an opportunity to build each one with different jFets in order to see if I can actually hear any difference between the kit's J113 and the suggested substitute -SK170.
The J113 with datasheet ldss of 2mA is biased at about 8mA with a 100ohm resistor and roughly 22VDC across two of these in a vertical series pair, the lower as a CCS the upper as a follower. (So roughly 11VDC across each jFet/Rbias.)
J113 DataSheet. ------ LSK170 DataSheet. ------ PASS 6-24 DIY Manual linked from Post #1 of thread.
My question is which grade of LSK170 would work here ? I made up a jig for B versions but at 12V across it, the most current I can get is ≈7mA with Rs=0, Vgs=0, Vds=12V.
I'm having trouble connecting the dots with the two different devices' Idss and their expected operating points . I have read a couple of web pages on jFets , one being Erno Borbely's page but it just aint clicking and need a little help. Anyone got a little light to shine on it?
Thanks !
The J113 with datasheet ldss of 2mA is biased at about 8mA with a 100ohm resistor and roughly 22VDC across two of these in a vertical series pair, the lower as a CCS the upper as a follower. (So roughly 11VDC across each jFet/Rbias.)
J113 DataSheet. ------ LSK170 DataSheet. ------ PASS 6-24 DIY Manual linked from Post #1 of thread.
My question is which grade of LSK170 would work here ? I made up a jig for B versions but at 12V across it, the most current I can get is ≈7mA with Rs=0, Vgs=0, Vds=12V.
I'm having trouble connecting the dots with the two different devices' Idss and their expected operating points . I have read a couple of web pages on jFets , one being Erno Borbely's page but it just aint clicking and need a little help. Anyone got a little light to shine on it?
Thanks !
The Idss of J113 is 2mA MINIMUM and if yours are operating at 8mA then obviously their Idss is more than 8mA. It looks like the SK170 parts have a lover pinch off voltage so you may have to reduce the BIAS resistors and you will have to use SK170C (or D) parts to get 8mA. But the circuit should work OK with less than 8mA? Maybe B parts would work with zero Ohm BIAS resistors? In any case, I don't know why Nelson bothered with the upper BIAS resistors except maybe to center the voltage but this has a single supply and it will not improve the peak AC voltage. I would just short the upper "BIAS" resistors.
Edit: OK for the low-pass, the upper bias resistors effect the bias voltage of the following stage so best keep them.
Edit: OK for the low-pass, the upper bias resistors effect the bias voltage of the following stage so best keep them.
Last edited:
Thanks - I didn't know that figure represents the j113 minimum. So then does its upper range end at the j112's min. ?
(With the provided bias resistor the j113 is conducting about 10mA.)
I didn't understand saturation. I thought it was independent of Vgs . . . a sort of max current if the device is operated as a diode and to exceed that the gate would have to be positively biased.
I'm really just trying to get an understanding of how jFets work , how to relate their data sheets to what they actually do. (For example, the j113 data sheet gives Idss as 2mA yet it will pass 10mA with 100Ω Rs while the SK170B has Idss of 6-10mA and can only pass 7mA with Rs=0 . . . .It doesn't make any sense to me.)
I found another schematic for B1 buffer preamp that shows vertical jFet pairs with no bias resistors. I'm wondering if I can try that in the crossover with the 170B's I have.
I want to be sure the ones I have are workable in one way or other before soldering them in. I'll sell them or trade for something that will work if not.
Anyway, Thanks again!
(With the provided bias resistor the j113 is conducting about 10mA.)
I didn't understand saturation. I thought it was independent of Vgs . . . a sort of max current if the device is operated as a diode and to exceed that the gate would have to be positively biased.
I'm really just trying to get an understanding of how jFets work , how to relate their data sheets to what they actually do. (For example, the j113 data sheet gives Idss as 2mA yet it will pass 10mA with 100Ω Rs while the SK170B has Idss of 6-10mA and can only pass 7mA with Rs=0 . . . .It doesn't make any sense to me.)
I found another schematic for B1 buffer preamp that shows vertical jFet pairs with no bias resistors. I'm wondering if I can try that in the crossover with the 170B's I have.
I want to be sure the ones I have are workable in one way or other before soldering them in. I'll sell them or trade for something that will work if not.
Anyway, Thanks again!
Parts with Idss as low as 2mA can be sold as J113 but odds are you will get parts that could have been sold as J112 or J111. If your parts operate at 8mA with a 100 Ohm bias resistor then they have an Idss of about 10 or 20mA. You can measure the actual Idss simply by using a zero ohm bias resistor. This is the usual method for matching Jfets.
With 12V across the j113 and 100Ω drain current measuring resistor and Rs =0 the current is 25mA. That seems too far different from the data sheet to be reasonable . No sense at all.
I guess that means if an LSK170 is to be used in the crossover in similar conditions then a C or D part is what is needed.
Anyway, much obliged for your help steveu !
I guess that means if an LSK170 is to be used in the crossover in similar conditions then a C or D part is what is needed.
Anyway, much obliged for your help steveu !
The manufacturer only makes one part for all 3 parts J111,J112, and J113. They use the first maybe 50% to fill their orders for J111, some for J112 and the left overs as J113 provided they meet the reduced J113 spec. Odds are that the whole batch could have been sold as J111, but that doesn't matter. It's about marketing. J113 are surplus J111 parts.
I had more or less assumed that everyone who got the kit received the same value bias resistors but perhaps that's not the case.
Still, it must be something of a logistical headache for a manufacturer if going in one direction on the list the difference between part numbers is meaningless. It makes me curious to know how he actually makes it work.
I've had another look at the B1 Schemo and the XO. I think it will work to use the 170B and go with Rs=0 as either with or without resistance the following string in the XO is still at the supply voltage center . . . . . as long as the upper and lower devices are matched.
Still, it must be something of a logistical headache for a manufacturer if going in one direction on the list the difference between part numbers is meaningless. It makes me curious to know how he actually makes it work.
I've had another look at the B1 Schemo and the XO. I think it will work to use the 170B and go with Rs=0 as either with or without resistance the following string in the XO is still at the supply voltage center . . . . . as long as the upper and lower devices are matched.
As long as the Idss of the upper Jfet is more than the programmed current of the lower Jfet then the circuit should work fine. But if the lower Jfet called for a higher current than the upper Idss then the upper JFET would go towards a positive Vgs, ie forward bias, ie not good. The lower bias resistor should reduce the current constant down within the range of the upper Jfet, but you may want to favor using the higher Idss parts in the upper positions. The upper bias resistor just removes the Vgs offset so that following stages get the same Vref as the first stage, but the high-pass is capacitor coupled so that becomes a moot point. The high-pass might be a bit quieter without the upper bias resistors.
So using zero Ohm lower bias resistors is risky, unless you select the upper Jfets with a higher Idss.
So using zero Ohm lower bias resistors is risky, unless you select the upper Jfets with a higher Idss.
Last edited:
I'm leaning toward the zero ohm for both , mostly because the Pass B1 does that and we know the designer knows what he's doing and clearly likes the sound. I've also read on other pages that it's a good way to do it for buffers and low signal level.
How much higher does the upper jFet Idss need to be to be safe? Is 5% enough?
How much higher does the upper jFet Idss need to be to be safe? Is 5% enough?
Hmm, well, I'll say that the Upper Idss should be equal to the lower current constant plus the current required to drive the load to the expected peak audio voltage. So if the load is 10K and the peak audio voltage is 10V then the upper Idss should be 1mA more than the current constant. If the output drives a lower impedance, then you want a greater difference.
Of course, on negative peaks, the current constant has to be at least as much as the difference, but that should be a given. For maximum signal handling, I would pick a lower bias resistor that cuts the Idss in half, or slightly less, maybe 40%.
Of course, on negative peaks, the current constant has to be at least as much as the difference, but that should be a given. For maximum signal handling, I would pick a lower bias resistor that cuts the Idss in half, or slightly less, maybe 40%.
Last edited:
I'll take some time to digest what I've learned so far. Thanks for taking the time with this. Giving me the chance to kick it around with you is a big help. I really do appreciate it. I'll post again once there are some results to report.