diyAudio

diyAudio (http://www.diyaudio.com/forums/)
-   Power Supplies (http://www.diyaudio.com/forums/power-supplies/)
-   -   New (to me) regulated power supply design (http://www.diyaudio.com/forums/power-supplies/63298-new-me-regulated-power-supply-design.html)

wes-ninja250 28th August 2005 03:57 AM

New (to me) regulated power supply design
 
Hey, All!

I bumped into a new-to-me power supply design today while fixing my ~1976 JVC JT-V310 Tuner.

The transformer was a little unusual, with a fancy jumper matrix, allowing the selection of 100, 117, ???, and 230VAC input voltages. the only low-voltage winding was a 7VAC secondary.

So far, so good, right?

The 7VAC goes into a two-diode bridge and a bunch of capacitors. Somewhere down the line, 20VDC comes out (I'm assuming 7VAC->10VDC + voltage doubler). Then here's where it gets really interesting.

The 20VDC goes through a 470 ohm resistor, which functions as current limiter to a 12VDC zener diode. The zener diode then feeds the base of a PNP transistor, D235D (xref: NTE152, http://www.nteinc.com/specs/100to199/pdf/nte152.pdf).

The transistor's collector is connected to the 20VDC line, although it was actually about a volt low; there might have been a diode in the middle or something. Anyhow, then the emitter shoots out 11.5VDC.

So, it would appear that the transistor is using the zener-voltage as a reference, for regulating the power supply. Unfortunately, the transistor was flaky -- it only worked some of the time. I suppose it's acceptable for 30-year-old transistors to fail.

Question One: I am the only one who's never seen this design before? (Yes, I am most definately an amateur).

Question Two: I used a C4161 (xfer: NTE2337, http://www.nteinc.com/specs/2300to2399/pdf/nte2337.pdf) to replace this transistor. Any idea if this will provide any kind of regulation, or did I get lucky? (It was the closest-thing-on-paper I had at hand -- no parts stores in these parts. Except Radio Shack. Ha haha).

Question Three: Are there any benefits to this type of supply, over the designs I'm used to? (LM317, 78xx, etc). Drawbacks? Seems to me that if you can get the right transistors and zeners, you could built a high-current regulated PS, cheaper than with the dedicated chips (the transistor I used was from an old PC SMPS.. and it had two of 'em). And just changing the zener might be a way to switch voltages for said PS..

At any rate, the tuner works -- with all-new PS caps (several were leaking), it sounds better than it has in years.

And I finally replaced the incandescent bulbs which blew during a move in 1980... with LEDs run in series to limit the current, off a Q&D four-diode bridge running from the original 7VAC for the panel bulbs.

Cheers,
Wes

DcibeL 28th August 2005 04:16 AM

If I am reading this correctly, it sounds like your standard old school voltage regulator circuit. This is what they did before regulator IC's were available. The zener is used to maintain the 12VDC (or 11.5 as it may be), and the transistor is used for overcurrent protection. For large supplies this gets expensive and the transistor passes all the current used in the circuit. This type of regulation is less effective and less efficient than a chip regulator (78xx, 79xx type).

rinox 2nd September 2005 09:51 PM

YES, DcibeL say right.... why loose you time?
Use a lm 78xx or if you need a little more power (1,5 A instead of 1A) use al LM317, the last is variable so use a trimmer to adjust the output voltage.... internet is plenty of this kind of application.

wes-ninja250 2nd September 2005 10:15 PM

Hey, guys!

Thanks for the feedback.

Using the LM317-style parts definately seems to make the most sense, I was *mostly* just curious to see a design I didn't recognize. Upon further reflection, it looks a bit like the inside of a 78xx regulator! :)

That said, it's definately a neat way to get zener-matched output if all you have on your bench is some power transistors and zener diodes, and don't want to wait for your next parts house order. :)

(I have a lot of zeners, and power transistors are easy to find inside broken PC SMPS etc).

Wes

anatech 2nd September 2005 11:00 PM

Hi Wes,
Your 2SD235 NPN transistor is very common in those power supplies, the replacement part 2SC4161 will not perform as well due to low DC gain. The D235 is rated 40V,3A 25W, hFE=40~240, the C4161 is rated at 400V, 7A,30W, hFE=15~50. It's not even close, but will sort of work.

NTE do not supply quality parts, you don't even know what the real base part number is. You can only get a rough idea of the specs from these replacement manuals. I mean rough (=vague) idea. Don't use them if you can get around it. You can mail order parts easily from a number of vendors.

I would not use an IC regulator in this case without redesigning the supply a bit. Just replace the old electrolytics and the pass transistor. The advantage to this supply is less gain and simplicity (+ price). If you want to improve it you can use a constant current source for the zener, and a darlington for the pass transistor. Use a diode in series with the zener to maintain the same voltage, you can now reduce the zener current. Or go the IC regulator route. More tendancy to have HF noise if you aren't a bit careful.

I've worked on this circuit many times. It is reliable.
-Chris

AndrewT 3rd September 2005 08:43 AM

Hi,
the regulator chips are usually based on lowish speed amplifiers using feedback to maintain regulation. They tend to produce noise and hi frequency hash.
The zener although noisy is reputed to produce good sound for the down stream stages and has no feedback and NO risk of oscillation. With appropriate caps on the output it can produce a clean sound.
Have you put current control ( either resistor or ccs ) in the LED string?

anatech 3rd September 2005 02:22 PM

I get excellent performance from Red LED's in CCS circuit. Simple and clean.

-Chris

wes-ninja250 3rd September 2005 03:16 PM

Anatech:

> NTE do not supply quality parts

That's too bad! They have a good cross-reference application. I couldn't find that transistor anywhere else.. I miss the days of school when I had access to an IC Master.

But -- you provided an important clue! I did not realize that a transistor in a Japanese tuner marked "D235D" would be a 2SD235. I should have guessed, but I didn't.

I'm admittedly kind of clueless when it comes to replacement transistors, doubly limited by what I had on hand. Both parts had "very linear gain" (according to the spec sheet), were NPN and silicon, so I figured, "hey, I'll try it!". The replacement got REALLY hot, so I added a heatsink -- but the output voltage measured okay. The heatsink doesn't even get warm (it's pretty big).

The tuner's been working for about a week now, I'm fairly happy -- but I'll think I'll order a 2SD235 from Digikey on my next order run. I don't think I want to bother redesigning the PS (even though it looks straightforward), because all I have to do is replace a TO-220 transistor [which is now mounted in an easy to access location] -- the radio worked fine for roughly 29 years, why mess with it. ;)

What does the wide difference in Hfe mean? Is that the amount of gain the transistor provides? So, if the line voltage alters, the replacement transistor's relative voltage swing will be much wider than the original? That would yield.. Poor regulation, right? :)

And yes, I replaced almost all of the electrolytics already (some were leaking badly), with the exception for a few tiny ones right near the tuner electronics. I used commodity caps (pulled from my parts drawer and older power supplies), and managed pretty good matches for the originals. It doesn't seem to have affected the sound -- to my ear, anyhow. It actually sounds better than I can remember it ever sounding.

AndrewT: Hey, that's cool, your name is Andrew and your flag is St. Andrew's Cross!

Thanks for the comment re. that power supply design. I think I get how it works now. I wish I had more formal education in this area.

As for the LEDs -- I didn't provide any current limiting, I figured the string could do that for itself. I made a quick bridge off the VAC tap with whatever diodes I had on hand (I think they were 1N4001s or similar), stuffed a 220uF cap on it, ran five in series, decided they were too dim, and pulled one out. Then it was just right. Goldilocks school of circuit design!

Actually, I did it that way because I have no clue how to calculate the current-limiting resistors for a series string, and didn't want to run them in parallel because the P2P layout would have been annoying that way. I also remember that the green LEDs I chose sink a little more current than my red ones, but I couldn't remember how much.

Incidentally, I had originally thought about using the LEDs to "be the bridge", thinking that would be a very zen-like design. Unfortunately, I couldn't quite get head around how to do it and keep them from flickering at 60Hz (and hence be too dim or over-current).

OBTW -- Define CCS?

Cheers,
Wes

anatech 3rd September 2005 05:01 PM

Hi Wes,
My feeling was to just repair it. Buy new caps for it though. They are not expensive and you are ordering anyhow. 2SD613, TIP41B(or C) are fine. Your original has not been available for many years. The new part will run hot as well, should have had a heatsink from the factory.

The difference in hFE means poorer regulation as you guessed. Perhaps a slightly lower voltage and the possibility the zener may "drop out". This would allow more noise to come through as the zener would be a high impedance when not conducting.

CCS is short for Constant Current Source. A good way to run your LEDs You must limit the current through these. The current determines the brightness and lifetime (opposite - of course). It would make the brightness constant with AC supply voltage variations. Right now the limiting factor would be the resistance of your transformer. Not good. Current would be approx. 1.2V / (emitter resistance) for a red LED as a reference. With a resistor you calculate the voltage difference between your LED stack and available supply voltage / (dropping resistor). You will find the voltage rises as the current demand drops, so each current value has a different differential voltage.

LEDs will not make a good bridge. The reverse breakdown voltage is too low. Never mind they cost too much as well over diodes.

-Chris

N-Channel 3rd October 2005 04:49 PM

Voltage Regulator
 
I'd say the 78XX-series or LM317T solution, too. I'd lean toward the LM317, due to its adjustability, and good noise rejection.

Steve


All times are GMT. The time now is 01:44 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