• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

-50V negative supply

mark_titano

Member
2004-10-19 10:10 pm
Hi all,
I want to use Gary Pimm's CCS serf bias on the cathode of a long tailed pair and I need a negative supply that can provide something between -40V and -50V. Ripple is not a problem, I just need something quite stable. The long tailed pair is polarized with 4mA. The reference chain absorb only 0.35mA. Maybe a simple zener shunt regulator can works good but, if there is on the market something similar to Lm337 that can support -50V on its output, maybe a three pin regulator could be a better idea. Have you an advice?

Mark
 
Mr. Triatic said:
like EC8010 says, use a zener on the adjustment pin to lift it from groundpotential.. there are several schematics on the web, good luck!

What EC8010 means is that the voltage rating of the device is what it sees between the "input" and "output" pins. The device doesn't know if it is providing 50V or 500V. In fact, devices such as this can be used to regulate HT for valve amps (albeit with supporting circuitry).

The zener would be superfluous.
 

mark_titano

Member
2004-10-19 10:10 pm
Here I am...thankyou for the replies.
Please look at the schematic. What do you think of it?

Mark
 

Attachments

  • reg--50v.gif
    reg--50v.gif
    10.6 KB · Views: 635

mark_titano

Member
2004-10-19 10:10 pm
Thankyou EC8010 ;)

What kind of diodes should I use? MUR120 (200V/1A 50ns) maybe? I think that I will use the chassis as an heatsink (with mica insulator).

Regarding the ripple...maybe I could use a simple crc filter. Something like 330uF/500Ohm/330uF. The voltage in front of the LM would be 64V and the ripple shouldn't be a problem, I think.

I'll try to check this issue...;)

Mark
 

EC8010

diyAudio Moderator Emeritus
2003-01-18 7:57 am
Near London. UK
I expect MUR120 will do fine if they're soft recovery (don't have a data sheet to hand). I can't imagine they need to be on a heatsink - you're not passing much current! Try using Q = CV = It for calculating the ripple on the back of an envelope, or PSUD if you like graphical answers and like to play.
 
Thankyou ec8010.
For the crc idea I've used PSU Designer II. The ripple is not a problem. Mur120 are very fast and soft recovery indeed ;)
I'll try to build the circuit and evaluate if it works properly.

If the load is higher ( same tension but more current) the lm337 would be happier, right?
So, if I'd try a different tube in the driver stage (as 6h30 or 5687 instead e88cc), I could use the same circuit, maybe only with an adjustment of the R in the C-R-C filter. Do you agree?

Mark
 
Hold on a moment, it's not quite as simple as it may seem. Yes it is true that the differential voltage across the regulator is the key issue when operating, but notice the electrolytic on the reference terminal and also on the output - there is a small but significant delay due to the presence of that electrolytic on the reference pin, and depending on how fast the raw supply comes up you may be surprised to see more than the rated differential voltage across the input and reference pins of the LM337 even if for just a few milliseconds. I would add a moderately high voltage pnp transistor in series with the input pin of the regulator and reference its base to the output of the regulator through a 9V zener, note that you need to provide the base current as well so a resistor is needed from the base to the raw supply. (Make sure that you provide enough current to take care of both the zener and the base current. Note that it is not a good idea to ever allow the voltage on any terminal of a LM337 to exceed 40V, particularly relative to any other terminal. I would use the LM337HV which has a 60V rating in this application for slightly more margin. Note that different brands and vintages of LM337 can have different voltage ratings. Many are rated for no more than 35V/40V.

Here is the link to the national part.

http://www.national.com/pf/LM/LM337.html

Kevin

Edited to correct a typo and added another thought..
 
Thankyou kevinkr.
Well, maybe the Lm337 is not the better solution for that work. I'll try to change something...

What do you think about the use of an IRF9510? With a reference chain made using some zeners this mosfet should work quite well. Look at the datasheet
http://chdist35.distrelec.com/distr...4EF2D2E1F468C1256FF6004C54F2/$FILE/Data E.pdf

The Vds is -100V and power dissipation at 25° is 43W. Maybe it can support the heat dissipation without an heatsink for the load that it should drive.

Or maybe an IRF9610 (20W, Id 1.8A, Vgs -200V).

http://chdist35.distrelec.com/distr...44C96372D459C1256FF6004CE059/$FILE/Data E.pdf

Mark
 
kevinkr said:
Hold on a moment, it's not quite as simple as it may seem. Yes it is true that the differential voltage across the regulator is the key issue when operating, but notice the electrolytic on the reference terminal and also on the output - there is a small but significant delay due to the presence of that electrolytic on the reference pin, and depending on how fast the raw supply comes up you may be surprised to see more than the rated differential voltage across the input and reference pins of the LM337 even if for just a few milliseconds. I would add a moderately high voltage pnp transistor in series with the input pin of the regulator and reference its base to the output of the regulator through a 9V zener, note that you need to provide the base current as well so a resistor is needed from the base to the raw supply. (Make sure that you provide enough current to take care of both the zener and the base current. Note that it is not a good idea to ever allow the voltage on any terminal of a LM337 to exceed 40V, particularly relative to any other terminal. I would use the LM337HV which has a 60V rating in this application for slightly more margin. Note that different brands and vintages of LM337 can have different voltage ratings. Many are rated for no more than 35V/40V.

Here is the link to the national part.

http://www.national.com/pf/LM/LM337.html

Kevin

Edited to correct a typo and added another thought..


Kevin,

You are right of course about the max voltage at start-up, but this can be avoided by putting a zener of, say, 33V across the LM337 so that the output cap can charge through this zener at start up. When the circuit is stable, the zener becomes invisible. Similarly with the cap at the ref point.

The alternative mentioned, a FET and a ref string, is a totally different circuit, it's not a regulator but just a stabilizer. It may be enough, but you cannot really compare the two.

Jan Didden
 
Thankyou Jan. Yes of course a FET plus a ref string is just a stabilizer, but I think that it should work well because, in case of minimum difference in main voltage , the ccs should adjust its working point by itself.

However your advice to use a zener limiter sounds good!
Please look at schematic, post six. Should I change the 1n4002 with a 33V zener? That's all? I'll use a crc in front of the lm337, not only the single can shown in the schematic.

Mark
 
mark_titano said:
Thankyou Jan. Yes of course a FET plus a ref string is just a stabilizer, but I think that it should work well because, in case of minimum difference in main voltage , the ccs should adjust its working point by itself.

However your advice to use a zener limiter sounds good!
Please look at schematic, post six. Should I change the 1n4002 with a 33V zener? That's all? I'll use a crc in front of the lm337, not only the single can shown in the schematic.

Mark

Yes, replace D2 with a 33V zener, same polarity. Use a zener of 1 or 2 watts, although with a CRC the 337 input voltage will not rise that fast, and with the small cap at the output (22uF) the one-time charge current will be modest.

Jan Didden
 
Hi Mark,

Some good ideas for setting up and protecting the negative regualtor here.

My input is why regulate at all? The negative regulator is adding complexity with very little if any performance gain.

With the high performance CCS all you would need is a nice simple CRC filter that puts out more than -40 volts and less than -500 volts and your set. At the low current you specified (4ma) the CCS mosfet would not need a heatsink even if you ran the CCS at it's voltage limit of -500 volts... :mischiev:

Gary
 
OK Gary, thankyou for the advice. I liked idea to use a negative regulator, simply to have a stable ref voltage even if I should change the driver stage and the idle current. But if you say that with something between -40V and -500V the ccs works happy I'll try to use a simple CRC filter. Maybe I'll use a power transformer with a secondary that can provide at least 60Vac/1A, for this purpose.

Mark