Jacco, officially my apartment has three rooms with a total area of 62.4 m2. I don't now what you'd call it, I call it small.
On last century ladies - you take the chikie babies of today, I'l take Halee Berry and be happy.
Fortunately, my wife is allergic to animal skin, and especially mink. She is naturally mink proof.
On last century ladies - you take the chikie babies of today, I'l take Halee Berry and be happy.
Fortunately, my wife is allergic to animal skin, and especially mink. She is naturally mink proof.
Mr. Curl, since you wish a real audio topics, you could explain in details Blowtorch shunt regulator and even publish schematic of this part of preamp.
I saw an accident... car vs motorbike with a father and daughter.... car door open and no driver in sight. young father and daughter laying in the road with bystanders on the side of the road. I asked if anyone called for an ambulance as I had a cell phone. Driver says... we dont have an ambulance. Well what about a doctor. Or hospital? we dont have those either. Oh yes, there is a doctor two miles up river at a border military camp. That is when I realized there is no help for them but wouldnt be for me either.... no 911. Not even police. Anything could happen to me and nothing could or would be done about it. kidnapped. murdered. Ran over. Then I thought, I am an American who bombed the hell out of Laos. And, I was in the northern area. Best I get out of there. This after I had dinner at the ex-dictators home and spoke with his wife briefly and never knew who she was or that it was her house.... I thought it was a restaurant ! True story... short version.
They died and the car owner/driver ran away... standard procedure as the villagers would kill the driver.
Last edited:
Well Kamis, you have a legitimate question. I did, in fact, design a complementary shunt regulator into the CTC power supply, and I did not publish that part of the regulator circuit. It did not work as well as I had hoped, and it was prone to thermal runaway, so I will not publish it now. I do recommend a shunt regulator to be added to a super power supply, so any suggestions?
Does it mean that 40 people have a problem with burned PCB-s ? What is exactly super power supply? What type of regulation do you use in Constellation phono stages?
No, I think I am the only person who had a big problem with the shunt regulator. It just took a lot of adjusting up front to keep it working properly over time. It was only the Vendetta portion of the power supply that had a potential problem, as the output voltages were necessarily lower to operate the Vendetta phono boards. +/- 24V rather than +/-30V for the line stages.
In Salas' thread on "simplistic" shunt regulators, there have been some designs proposed and realized. Preamp (his forum name) suggested dual 15V ones, which allow current sources with healthier voltages by referring them to the opposite rail. He's built something and is currently trying to get noise measurements.
In between paying work I've been doing some designs as well which have not been realized yet, except for a very-low-noise temp-stable voltage reference. The plan if time permits is to build a tracking regulator and make it temperature-stable within reason. Target is 200mA of current per rail. The error amp front ends will be bipolars from the obsolete low rbb' series from Toshiba, 2SA1316 and 2SC3329, and run at about 7mA per device. Except for 1/f noise the overall output noise should be in the low-hundred nanovolt rms range, in a 22kHz noise bandwidth.
Simulations indicate a transient response with a time constant of about 170ns and very little ringing for a step current change of 20mA. Should be a pretty decent regulator and readily scaled to other voltages and currents.
If I get that far I'll probably start another thread.
more LED noise results, string of 16
For a bit more noise compared to the preamp I strung 16 HLMP-6000 together. Same 50mA current. Results for a 400Hz - 22kHz noise bandwidth apparently include a fair amount of 1/f, and dividing the total noise by 4 indicate about 490pV/sq rt Hz average spectral noise density. If the filter is raised to 80kHz the average density is 260pV/sq rt Hz. That 80kHz is probably only a three-pole so a correction for noise bandwidth should be applied. I believe the 22kHz filter is in series with the 80kHz or maybe the 30kHz (Ap Sys One), so it's probably closer to an actual 22kHz NBW.
For a bit more noise compared to the preamp I strung 16 HLMP-6000 together. Same 50mA current. Results for a 400Hz - 22kHz noise bandwidth apparently include a fair amount of 1/f, and dividing the total noise by 4 indicate about 490pV/sq rt Hz average spectral noise density. If the filter is raised to 80kHz the average density is 260pV/sq rt Hz. That 80kHz is probably only a three-pole so a correction for noise bandwidth should be applied. I believe the 22kHz filter is in series with the 80kHz or maybe the 30kHz (Ap Sys One), so it's probably closer to an actual 22kHz NBW.
What excess noise in the flat spectrum would you expect? The equivalent resistance of the diode should tell all or does the 1/f extend far out in frequency?
A diode and diode connected transistor are not the same thing, a diode connected transistor is active (unless the Vcesat is unusually bad) and almost all the current is collector current so the Voltage noise assuming a noiseless current excitation is from .5re + rbb.
Last edited:
A good question of course. I was a bit surprised that the 80kHz reading wasn't higher. The attribution to 1/f is just conjecture.
I really need to do some good spectral measurements as well as impedance measurements, the latter in part to parse out the bulk resistance. There is a definite increase in noise if the filter bandwidth is extended downwards, but unfortunately that can be seen in the residual to be partly line frequency and related.
At least the series string technique gets me fairly well out of the 2SK364BL preamp noise, and the preamp gets me well out of Ap Sys One noise.
In any event the 400pV number mentioned by SY for some LED(s) seems plausible. I'm inclined to suppose the Avago parts are pretty decent, yet another HP spinoff that seems to have done well.
Another datum: the average forward voltage at 50mA is 2.23V.
Yes. Diodes have a different power law. That was something discussed in Horowitz and Hill and one of the times I learned something from that book that I'd missed appreciating before.
Reminds me of when someone published an amplifier schematic using a diode to linearize the transistor. When I saw it I naively thought that a diode-connected transistor might work better, but it didn't. The difference in I-V curves was the trick that made the guy's stage work.
Attachments
Noise is a little lower with 10mA for the string. Possibly a bit of a bulk resistance thermal noise effect.
Do you recall what the temperature coefficient of the output voltages are? There should be partial compensation of the error amp input voltage by the probable tempco of the 2SK170 current generator.
I intend to measure that really soon now, lost quite a lot of time to biking and
changing this & that on my house. Well, not really lost.
While a lot is written on the noise behaviour of LEDs, I could find nothing
on their TK, so I measured a red and a blue one, green to follow. The red one
seems to have the usual -2mV/K we know from silicon diodes, the blue one was
not so constant. Wierd. I would like to use a LED / cap multiplier to feed my
precision oscillators, and getting rid of noise while spoiling the TK does not
seem attractive.
Attachment: Excerpt of an upcoming article on V references. But with the
new lo noise LT regulators this is all moot, probably.
regards, Gerhard
and Scott, thanks for the half-thermal diode resistance stuff, back home I found
the school book, and they also derived it from first principles.
Your welcome, another way to look at it is that in a diode connected transistor the shot noise and the Vbe are 100% correlated so you loose a degree of freedom (or is that 1/2 a degree of freedom?).
- Status
- Not open for further replies.