Remember that a TO-92 case (unless there is a special lead frame for enhanced power dissipation) has a junction-to-ambient thermal resistance of ~200C/Watt. I wouldn't operate the jfets at much higher than about 60C case temperature - choose your power dissipation accordingly.
If I were you, I would lose those fancy LM317 current sources. I've found LM317s used in that way to be hiss factories (even though the LT versions are probably quieter than 317s from the usual vendors). Depletion-mode mosfets are much quieter, and will be better current sources over a wider frequency range. If you don't go too crazy with the current, you could possibly use a TO-92 device like the Supertex DN2530N3.
If I were you, I would lose those fancy LM317 current sources. I've found LM317s used in that way to be hiss factories (even though the LT versions are probably quieter than 317s from the usual vendors). Depletion-mode mosfets are much quieter, and will be better current sources over a wider frequency range. If you don't go too crazy with the current, you could possibly use a TO-92 device like the Supertex DN2530N3.
good point
...Given ambient here recently has been sneaking its way into the mid 40's, thats not a lot of headroom
My (shallow) understanding was that a j310 was good for upto 350-400mW, so I was targeting around 200mW as an average dissipation, on the grounds that the transconductance of the jfets, according to the datasheet (as I understood it perhaps?
) is greater at a higher current. At 200C/W I can see that i'm biased way too high. Well that's easily sorted at least!
With regards to the current sources, I picked the lm317 as a convenient option (or as I read Juma recently say "to save on shoes") as I have a heap of them here and it was an easy and lazy solution. There's a thread here on diy audio containing an image comparing a variety of different current source/sinks and this solution was amongst the best in terms of accuracy.
I hadn't considered the noisiness of them though.
Depletion mode fet's aren't something I have in the parts box, but I'm sure I can come up with a slightly more complex solution easily enough using whats on hand. I'll be a pretty rough point to point effort initially anyway...
...Given ambient here recently has been sneaking its way into the mid 40's, thats not a lot of headroom
My (shallow) understanding was that a j310 was good for upto 350-400mW, so I was targeting around 200mW as an average dissipation, on the grounds that the transconductance of the jfets, according to the datasheet (as I understood it perhaps?
) is greater at a higher current. At 200C/W I can see that i'm biased way too high. Well that's easily sorted at least!With regards to the current sources, I picked the lm317 as a convenient option (or as I read Juma recently say "to save on shoes") as I have a heap of them here and it was an easy and lazy solution. There's a thread here on diy audio containing an image comparing a variety of different current source/sinks and this solution was amongst the best in terms of accuracy.
I hadn't considered the noisiness of them though.
Depletion mode fet's aren't something I have in the parts box, but I'm sure I can come up with a slightly more complex solution easily enough using whats on hand. I'll be a pretty rough point to point effort initially anyway...
Subscription access only, but it's the only version I could find on the web
Audio Xpress-August 2001
Now you've teased me here ZM - is the learning gained from buying access to this worth it?
edit - added pic of current sources and their effectiveness. I'll probably use (I) now...
Audio Xpress-August 2001
Now you've teased me here ZM - is the learning gained from buying access to this worth it?
edit - added pic of current sources and their effectiveness. I'll probably use (I) now...
Attachments
Last edited:
Somewhere on Diyaudio there a link to a study of current source circuits that include Ixys and Supertex depletion mode parts. They are quite good by themselves, and even better cascoded. The tester looked at capacitance and output impedance vs. frequency. If you can't get the depletion mode parts, a bipolar-mosfet ring-of-two source could be a contender.
as sort of learning drill
, not copy projectthose things are made long before last year's BAF
Here's Gary Pimm's examination of several current sources, including depletion mosfets (they fare pretty well). Stray capacitance considerations for this application are relaxed comapred to Pimm's intended purpose (plate loads for vaccuum tubes), as there is no appreciable voltage swing across the current source when driving a folded cascode. However, the current source must be reasonably quiet, as any noise will be dumped straight down through the cascode transistor into the gain resistor.
Pimm CCS URL: CCS performance measurments
P.S. I'm familiar with NT's RIAA project. I thought it was a nice circuit at the time (still is), but I never felt particularly motivated to build it.
Pimm CCS URL: CCS performance measurments
P.S. I'm familiar with NT's RIAA project. I thought it was a nice circuit at the time (still is), but I never felt particularly motivated to build it.
Last edited:
And here's a follow up from Jung where he specifically apologises for not addressing the hiss in the lm317 current source and specifically states that they're not suitable for user in a line stage. http://www.waltjung.org/PDFs/Sources_101_Letter_Revisit_0409.pdf
I would but I'm very busy reading about current sources here http://www.diyaudio.com/forums/anal...-supplies-phono-line-preamps.html#post3803384
As text occasionally struggles to convey humour I thought I'd clarify that this is a joke - the thread referenced is wrenchone's CCS analysis post
As text occasionally struggles to convey humour I thought I'd clarify that this is a joke - the thread referenced is wrenchone's CCS analysis post
Last edited:
Interesting, I did not search he had revisited the 317 issue. I remember I had built that hybrid CCS and when I tried it as anode load in a line preamp it was hissing. I might actually still have that module somewhere in a parts box.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.