Super Regulator, collecting the facts - Page 22 - diyAudio
Go Back   Home > Forums > Amplifiers > Power Supplies

Please consider donating to help us continue to serve you.

Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving
Reply
 
Thread Tools Search this Thread
Old 28th October 2003, 06:56 PM   #211
diyAudio Member
 
jan.didden's Avatar
 
Join Date: May 2002
Location: Great City of Turnhout, Belgium
Blog Entries: 7
Walt,

Thanks for the inputs (verbal/literary I mean...). I remember measuring the JFET source as aroung 150kOhms dynamic impedance, with the "normal" LED/xsistor source below 20k. I found that improvement worthwhile. But I'm sure the compound you describe would better that by some margin.

Jan Didden
  Reply With Quote
Old 28th October 2003, 09:02 PM   #212
diyAudio Member
 
john curl's Avatar
 
Join Date: Jul 2003
Location: berkeley ca
I think that you could change the green led to a red one and get lower effective voltage drop. Heck, you could use 1 diode instead of an led. Then a FET current source would be really more important. You should use current source fets if you can, and are not expert. They are already selected and optimized at the rated current. The Vishay j202 seems to be a pretty good deal, if it is like the j502 in performance.
Don't expect VERY high effective impedances with fets, unless they are long gate devices. This would be unusual with cheaper fets. Cascode is recommended in many cases.
  Reply With Quote
Old 28th October 2003, 10:10 PM   #213
diyAudio Retiree
 
Join Date: Oct 2002
Location: Spain or the pueblo of Los Angeles
Default Once again..........

The J202 is a regular jfet and not a two terminal jfet CCS. fairchild makes it as well.

http://www.datasheetcatalog.com/datasheet/J/J202.shtml

I suggested it for biasing the LED and using the bipolar transistor for the actual current source to the regulator. It will be very difficult to make a fet current source for ten mA and low drop out voltage.

Jfet current sources:

http://www.vishay.com/document/70596/70596.pdf
  Reply With Quote
Old 29th October 2003, 04:34 AM   #214
diyAudio Member
 
john curl's Avatar
 
Join Date: Jul 2003
Location: berkeley ca
If you NEED a current source, the j202 on reflection is actually a poor choice for a current source, UNLESS you know how to sort for current or calibrate it with a series resistor. This is beyond most people here, and it takes a certain amount of time and a large number of parts. There are better parts with a lower Vgs, such as the 2SK170 Gr that would work OK, once selected out to the right current value of Idss.
  Reply With Quote
Old 29th October 2003, 04:43 AM   #215
Banned
 
Join Date: Feb 2002
Location: As far from the NOSsers as possible
Default Don't use a J202?

I thought that you said that you need a long gate device. According to Siliconix AN103, a J202 is such a device.

Yeah, I would use the '170 too, but only because I have a lot of them around.

Depending on the current range you are designing for, you can make a FET CCS with a dynamic Z > 1 Megohm.

I think we are generally in agreement, +/- a few tenths of a dB, in a manner of speaking.

Obviously not in terms of actual PSRR. (Hey, it is only a joke, guys......)

Jocko
  Reply With Quote
Old 29th October 2003, 06:00 AM   #216
diyAudio Member
 
jan.didden's Avatar
 
Join Date: May 2002
Location: Great City of Turnhout, Belgium
Blog Entries: 7
Default Re: Don't use a J202?

Quote:
Originally posted by Jocko Homo
[snip]Depending on the current range you are designing for, you can make a FET CCS with a dynamic Z > 1 Megohm.
[snip]Jocko

Jocko, I don't have the type numbers I used in my head right now (it was on of the Jxxx series) but the ones that get close to 1MOhms are the ones with low current (1mA). They would be usefull for the LED current source. I just wanted one to replace the whole shebang, 5-6mA and that was much lower in impedance (150K IIRC).

Still, I would be interested to know if anyone has done some measurement on the actual impedance of the LED cum CS compound CS. Anyone has any figures?

jan Didden
  Reply With Quote
Old 29th October 2003, 07:16 AM   #217
diyAudio Member
 
john curl's Avatar
 
Join Date: Jul 2003
Location: berkeley ca
Ok folks, we are talking in circles.
First of all: Do we just want to use a fet as a current source from the supply input? If we do, then we have to use a low Vp device. That means, high Gm. It will have marginal output impedance and it will amplify its own noise, but it can be made to work.
Or do we want to use a fet as a current source to bias a LED that is connected to the supply, while the fet current source is connected to ground? Then, a relatively high voltage 40-50V part, usually a long gate type, like the J202 is recommended, as it will allow relatively high voltage across it, and because it has low Gm, will not amplify its own noise as much. A current source fet will work as well, but they are fairly noisy, and relatively expensive. Their greatest advantage is that you know the current they will pass, because they are selected at the factory to have a specific current.
Effective impedances above 20K are practical with the fets discussed here, but only a cascode will get you there, for sure. Watch out for short gate devices (high Gm) They sometimes look like a triode, rather than a pentode.
I just checked a j232 and it has a very high output impedance, much higher that I suspected that it would have.
  Reply With Quote
Old 29th October 2003, 09:31 AM   #218
diyAudio Retiree
 
Join Date: Oct 2002
Location: Spain or the pueblo of Los Angeles
Wink Conspiracy of complexity?

"If you NEED a current source, the j202 on reflection is actually a poor choice for a current source, UNLESS you know how to sort for current or calibrate it with a series resistor. This is beyond most people here, and it takes a certain amount of time and a large number of parts. There are better parts with a lower Vgs, such as the 2SK170 Gr that would work OK, once selected out to the right current value of Idss."

"Or do we want to use a fet as a current source to bias a LED that is connected to the supply, while the fet current source is connected to ground?"

YES, THAT'S WHAT I HAVE BEEN TALKING ABOUT ALL ALONG.......


Your just having fun with us now John...... The J202 was recommended as replacement for the 2N4340 mentioned in the Siliconix "The FET Constant Current Source" ap note. It has very a low output conductance of a few micro-mhos typical of those jfets that make good current sources and a drain to source capacitance of about one pF. It seems that jfets with large foward transcondutance like the 2SK170 don't make the best jfet current sources.

I agree that a two terminal current source would be better still, but I created a near panic by recommending one THAT COST A DOLLAR AND SIXTY CENTS! Putting a trim pot together with a jfet to dial in the current to the value desired is easy. It is not beyond any one here with a soldering iron, a VOM, and the initiative to read the article referred to above. In the context of biasing the LED in the BJT current source, the actual value of the current biasing the the LED is not that important. I would shoot for a value in the 0.5 to one mA range and not get anal about it in this particular circuit.

Once again, all of this was about biasing the LED for the bipolar constant current source and making the BJT current source more immune to voltage variations in the supply voltage. I believe it is worth doing whether one is using a preregulator or not, but obviously it's more important when not using a preregulator. Assuming an LED impedance of about 30 ohms (based on measurements) the change in LED voltage with input voltage is about 30/10000 or 3 mV per volt. One hundred ohms of emitter degeneration gives about 30 microamps per volt change for an output impedance of about 33.3 K ohms.

It appears to me that the bias current change the in the LED voltage reference is the dominant factor controlling the effective output impedance of the BJT current source under discussion. Replacing the 10 K resistor with a jfet current source with a 1 megohm impedance should give an improvement in output impedance of about 2 orders of magnitude, a pretty good return for the investment of one jfet.

As I recall the regulator article series was called "Regulators for High Performance Audio." By way of analogy, it seems to me that we are talking about building race cars and not about the most practical and economical car to drive to work and the grocery store.


Fred Dieckmann

PS I forgive you because I know you'll drop us another subtle but valuable circuit hint. I pay very close attention to both you and Mr. Pass when you throw those casual BTW pearls of wisdom out,
just to see if anyone notices. It seems that often, the more casual and low key it is, the more useful the information. I think you and Nelson have fun whispering the good stuff during all the shouting over the endless debates on usless topics. Don't worry the sensible members are watching very closely when you guys post the good stuff that we are here to learn in the first place.
Attached Images
File Type: jpg modccs.jpg (28.4 KB, 1064 views)
  Reply With Quote
Old 29th October 2003, 01:37 PM   #219
diyAudio Retiree
 
Join Date: Oct 2002
Location: Spain or the pueblo of Los Angeles
Wink No one expects the jfet current source

My last post was before I saw Mr. Curl's last post and I hope he will forgive me for the "comfy chair inquisition".

J232......was it? One of those BTW goodies I was talking about.


Thanks John,
Fred
  Reply With Quote
Old 29th October 2003, 01:52 PM   #220
Electrons are yellow and more is better!
diyAudio Member
 
peranders's Avatar
 
Join Date: Apr 2002
Location: Göteborg, Sweden
Blog Entries: 4
Default Re: Jan didden's post #387

Quote:
Originally posted by WaltJ
In other words, those posters asking about what the reg output does for various Isource topologies are on the right track. The answer is that it will change dynamically due to Isource imperfections, if you allow them to happen.

Now, after all that, maybe the prereg looks good?
The prereg seem to be a good idea and very cost-effective also but how much does this circuit improve the PSRR of the whole regulator? Anyone who knows?

My wondering was about the fact the you have 1k+1k across the pass the transistor (for the LM317) and therefore make the inpedance of this transistors much less (=less PSSR?).

I think it's wonderful that Walt has turned this long thread back into the main track. It's wonderful to just talk about electronics and ventilate different ideas, crazy or not.
__________________
/Per-Anders (my first name) or P-A as my friends call me
Super Regulator SSR03 Group buy
  Reply With Quote

Reply


Hide this!Advertise here!
Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off



New To Site? Need Help?

All times are GMT. The time now is 07:15 PM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright ©1999-2014 diyAudio

Content Relevant URLs by vBSEO 3.3.2