Forums :: NEW! Web Resources :: Features :: Photo Gallery :: Vintage Radio Shows :: Archives
Support This Site: Contributors :: Advertise


It is currently Jul Sat 31, 2021 5:27 pm


All times are UTC [ DST ]





Post New Topic Post Reply  [ 14 posts ] 
Author Message
 Post subject: LM317 regulator design
PostPosted: Jun Tue 08, 2021 8:12 pm 
Member

Joined: Jan Thu 01, 1970 1:00 am
Posts: 343
Location: Raleigh, NC
I'm trying to use an LM317 linear regulator (and also an LM337 for the complementary negative supply) to provide a regulated 175vdc from an approx 185vdc power supply. Below is the circuit I've built. The circuit does work, but so far I've smoked 2 each of the ICs--once due to a wiring error (adjust resistor not connected to Vout), and once due to overload (output transistors s/c). I thought these were somewhat robust with the ability to protect themselves from overload and short circuit and not die in the process.

I do know the LMs are only rated for 37v Vin-Vout, which I'm staying within.

Is the problem that the adjust pin references ground and has a potential of ~200v? Would it help to add zener diodes between ground and this circuit so it didn't see a high potential to the adjust pin? I realize this regulator will pass 1.5A without care, so my load is not protected from that point of view, but if the load shorts and pulls the main p/s voltage down am I killing the IC with reverse voltage?

Or am I required (best practice) to use a pass transistor for any HV applications?

For reference, the main power supply is full wave rectified off a CT txf with 330uF filters.

Any feedback appreciated,
Michael

Circuit as-built:
Image

LR8N3 design using TIP50:
Image


Top
 Profile  
 
 Post subject: Re: LM317 regulator design
PostPosted: Jun Tue 08, 2021 9:38 pm 
Member

Joined: Jan Thu 01, 1970 1:00 am
Posts: 6285
Location: Minneapolis, MN USA
Yes, shorting the output with high input voltage will destroy the LM317 regulator due to the high "reverse" voltage.
It can regulate high voltage subject to max in-out difference but short circuit protection is good only up ~37-40 volts input.

I have seen some LM317 circuits with pass transistor and other components that supposedly provide good regulation and protection but I prefer the LR8 with pass transistor.


Attachments:
LM317 two.PNG
LM317 two.PNG [ 9.18 KiB | Viewed 1014 times ]
Top
 Profile  
 
 Post subject: Re: LM317 regulator design
PostPosted: Jun Tue 08, 2021 10:19 pm 
Member

Joined: Jan Thu 01, 1970 1:00 am
Posts: 5204
Location: Rochester NY USA
You could protect the LM317 with a zener from input to output, but there would be nothing to limit current in a short circuit so the zener would likely fail shorted with an output short... a series resister sized to take most of the voltage drop at full load MIGHT help, if it limits short-circuit current to less than the IC's internal limit. And of course it must have an adequate heat sink - if not, it goes into thermal shutdown and then fails from overvoltage.

The LR8 has a much higher input-output rating, but is a lower power device, needing a pass transistor in most cases.

_________________
My web page: https://bit.ly/2rxq4qx


Top
 Profile  
 
 Post subject: Re: LM317 regulator design
PostPosted: Jun Tue 08, 2021 11:37 pm 
Member
User avatar

Joined: Jan Thu 01, 1970 1:00 am
Posts: 1715
Location: Wayside, NJ Monmouth
The LM317 was not designed handle High Voltage. I would use the TL783 High-voltage Adjustable Regulator. Outputs 1.2 to 125 Volts at 700 ma.


Top
 Profile  
 
 Post subject: Re: LM317 regulator design
PostPosted: Jun Wed 09, 2021 12:42 pm 
Member

Joined: Jan Fri 27, 2017 8:41 pm
Posts: 340
Location: Springwater, NY
Alternatively, if you need a bipolar regulated high voltage power supply, you could use two of these:

https://www.antiquewireless.org/homepage/pcb-item-1/


Top
 Profile  
 
 Post subject: Re: LM317 regulator design
PostPosted: Jun Wed 09, 2021 2:09 pm 
Member

Joined: Jan Thu 01, 1970 1:00 am
Posts: 343
Location: Raleigh, NC
That makes sense--I didn't realize I'm using the wrong regulator for the job. I found a TI document (https://www.ti.com/lit/an/snva583/snva583.pdf) for a HV adjustable regulator circuit using the LM317, provides 1.6-160v output (170v input) using the LM317, but it requires 2 external transistors + heat sink and I don't have the board space for that. Plus I don't need full range output adjustment, I only want 170-175v regulation from 185v to eliminate the last of the ps ripple. And I'm also trying to go beyond the 170v spec... so that has a few drawbacks.

The LR8 with a TIP50 won't take more board real estate than I have now--that may work out best. Plus I'm using TIP50s for other parts of this design and have plenty on hand.

The TL783 looks like a drop-in replacement for the LM317... but it has a 125v input-output differential. Won't that have the same issues as the LM317 if I'm trying to get 175v out? Or will it be less problematic on output s/c due to the higher working voltage?


Top
 Profile  
 
 Post subject: Re: LM317 regulator design
PostPosted: Jun Wed 09, 2021 2:55 pm 
Member

Joined: Dec Tue 29, 2015 8:23 am
Posts: 284
Location: Australia
An unmentioned issue in your circuit is C101 - at the instant of power-up it's effectively a short-circuit, and the LM317 sees 185V between IN and ADJ (or, perhaps more relevantly, 175V between OUT and ADJ). Rarely is it mentioned in the datasheets/appnotes, but excessive voltage between IN/OUT and ADJ is just as bad as between IN and OUT.

A quick sim suggests your circuit will have > 40V between OUT and ADJ for ~3ms, and between IN and ADJ for about 60ms, at startup.

The canonical appnote for HV LM317 supplies is NatSemi Linear Brief 47 / TI SNVA583 (on preview: I see you've found that). The circuits given there have a couple of features that limit the problem:

  • C1 is only 1uF, and there's a 100R resistor between C1 & ADJ. This alone limits the time the voltage is above 40V between OUT and ADJ to less than 0.5ms, and between IN and ADJ to a little over 1ms.
  • The addition of the series regulator setup before the LM317 means the voltage differential between IN, ADJ, and OUT is never more than a few volts.


Top
 Profile  
 
 Post subject: Re: LM317 regulator design
PostPosted: Jun Sat 12, 2021 5:54 pm 
Member

Joined: Jan Thu 01, 1970 1:00 am
Posts: 343
Location: Raleigh, NC
Is there a negative voltage equivalent to the LR8? A quick google didn't reveal anything...


Top
 Profile  
 
 Post subject: Re: LM317 regulator design
PostPosted: Jun Sat 12, 2021 6:54 pm 
Member

Joined: May Tue 30, 2006 4:46 pm
Posts: 10764
Location: Santa Rosa, CA
I never saw a negative version.

These little devices were developed in the 1980s as start-up circuits for highly efficient switch-mode (flyback) power supplies. Once the main control IC starts up, it can be powered from an auxiliary winding on the output inductor.

Image

The version we built at Siliconix used a depletion mode HV MOSFET. Once the main control circuit started up, the "house keeping" circuit turned off, in order to make a very efficient 1 Watt supply.
https://www.vishay.com/docs/70000/si9100.pdf

Rich


Top
 Profile  
 
 Post subject: Re: LM317 regulator design
PostPosted: Jun Sat 12, 2021 8:44 pm 
Member

Joined: Jan Tue 16, 2007 7:02 am
Posts: 4589
Location: Lexington, KY USA
The NS App Note circuit for the LM317 should work with the LM317, with polarized components turned-round. If you want to try this, someone will sit down and look for gotchas for you.

The heatsinking required will depend on factors external to the regulator; so will remain about the same for any circuit you use, barring SiC components.

A power MOSFET or IGBT can be a good choice for the high voltage transistor, in place of the bipolar TIP.

If you want useful help on this, a rundown on the volts in and out, and amps requirements, would help things along.

Ted


Top
 Profile  
 
 Post subject: Re: LM317 regulator design
PostPosted: Jun Sun 13, 2021 11:50 am 
Member

Joined: Jan Thu 16, 2020 12:29 am
Posts: 1401
There are a number of special challenges creating high voltage regulated power supplies, far too many to address on a single post.

We could all try to build a space rocket in our backyards. But it does pay to remember that others have done it and succeeded. You might go through some number of iterations, attempts and failures, not dissimilar to the trials in the space program.

With that in mind, the best thing you can do for a project like this, is to take an academic approach initially and learn about regulated high voltage power supply design. A very good book for this topic is The Art of Electronics, by Horowitz & Hill, Cambridge University press.

I have attached only a part of their section on High Voltage power supplies and there are a number of solutions. (obviously I cannot scan the entire book, you need to buy the real book, it is readily available).

Once you have studied the appropriate sections, you will be in a much better position to design and build a reliable regulated high voltage power supply.

Shortcuts, quick quips, brief suggestions and words of advice on forums have a very limited utility for the complex nature of this task. Your best move is to study the material in H & H's book, before proceeding with your design. This will significantly improve the probability of it working out for you.


Attachments:
HVreg.jpg
HVreg.jpg [ 586.56 KiB | Viewed 781 times ]
Top
 Profile  
 
 Post subject: Re: LM317 regulator design
PostPosted: Jun Mon 14, 2021 1:35 pm 
Member

Joined: Jan Thu 01, 1970 1:00 am
Posts: 343
Location: Raleigh, NC
Quote:
If you want useful help on this, a rundown on the volts in and out, and amps requirements, would help things along.


Sorry, I thought I had given these specifics in my original post. Regulating approx +/- 187vdc down to +/- 175v with a 60mA load. This is for a Class A amplifier project I'm working on. Main objective for the regulation is to eliminate the ripple and also to keep DC offset of the amp from drifting with supply variations.

Here is the complete schematic of the power supply. Note D11-D14 (30v zeners) are not in the current implementation, they were suggested to be added for protection. The regulator circuits do work, just too unreliable (maybe too fragile for testing/development purposes).

Image

Quote:
With that in mind, the best thing you can do for a project like this, is to take an academic approach initially and learn about regulated high voltage power supply design. A very good book for this topic is The Art of Electronics, by Horowitz & Hill, Cambridge University press.

I have attached only a part of their section on High Voltage power supplies and there are a number of solutions. (obviously I cannot scan the entire book, you need to buy the real book, it is readily available).


Thank you! I will read through this and look at getting a copy of the complete book.


Top
 Profile  
 
 Post subject: Re: LM317 regulator design
PostPosted: Jun Tue 15, 2021 7:28 pm 
Member
User avatar

Joined: Jan Thu 01, 1970 1:00 am
Posts: 4649
Location: Perrysburg, OH, 43551 U.S.A.
NOTE: I developed the following before I knew you needed a positive and negative supplies. The circuits below can be adapted to a negative supply without much effort (mainly finding a 400v P-Channel MOSFET for M!).

It is possible to use an LM317 in a high voltage regulator circuit as shown in the following circuit taken from the Analog Devices (formerly Linear Technology) application note AN2 by the late Jim Williams. (See
https://www.analog.com/media/en/technical-documentation/application-notes/an2f.pdf).
Attachment:
HV Reg AN2 Fig11.jpg
HV Reg AN2 Fig11.jpg [ 84.85 KiB | Viewed 655 times ]

The circuit description begins on the right-hand column of page AN2-6 and continues on to page AN2-7. Ther are a couple of drawbacks to this circuit. First is the use of a high voltage power bipolar Darlington transistor for Q1. The required base current, even for a Darlington device requires a low resistance and high power value for the collector-base resistance of Q1 (2K/5W). This in turn requires a high wattage Zener (the 1N3031 is a 1-watt device).
The second problem is the fact that the current limit is set by the saturation current of the power transformer. This will be covered later.
The first problem is handled by replacing the power bipolar Darlington transistor with an N-Channel power MOSFET as shown in in the 90v supply below.
Attachment:
HV Reg AN2 Revd Version.jpg
HV Reg AN2 Revd Version.jpg [ 55.42 KiB | Viewed 655 times ]

Notice that R1 is now 33K and the Zener is a 1N5256B, a 500mW part. The only additions to the circuit are D1, a 15v/500mW Zener for MOSFET gate protection, and R11, a Q-killer to prevent high frequency oscillation of the MOSFET (a good idea to include with any MOSFET circuit because they're very fast devices). The second problem still remains - there's no current limit.

The next circuit below , a 175 volt regulator, adds current limiting with the addition of Q1, Q2, R2, R7 and R10. R7 senses the load current by providing a voltage to the base-emitter junction of Q1. When the voltage drop across R7 rises above 500mv or so Q1 turns on, turning Q2 on, pulling the gate voltage of M1 low, turning it off.
Attachment:
175v Reg CL PS.jpg
175v Reg CL PS.jpg [ 89.91 KiB | Viewed 655 times ]

The following plot shows the output voltage (top trace) and the voltage across the regulator versus load current (second trace). The maximum output current is about 84mA. At that current the maximum power dissipation in M1 is 22.2W (third trace) and 1.9W in the LM317 (fourth trace). Note that the power dissipation in M1 is roughly equal to the current limit times the input voltage so it's best to keep the input voltage as low as possible while maintaining the maximum output current at the desired output voltage. One other addition to this circuit is D3 a 1N4936, 400v/1A fast recovery rectifier to protect the regulator from negative inductive voltage spikes from the load.
Attachment:
175v Reg CL PS Voltage and Pwr Diss Plots.jpg
175v Reg CL PS Voltage and Pwr Diss Plots.jpg [ 213.78 KiB | Viewed 655 times ]

The high power dissipation in M1 can be reduced significantly by introducing foldback current limiting to the last circuit. This change is shown below.
Attachment:
175VREG SCR foldback CL Schem.jpg
175VREG SCR foldback CL Schem.jpg [ 85.5 KiB | Viewed 655 times ]

The additional parts are Q3, R9, R12, R13 and C1. Q3 with Q2 form an SCR that latches on if the current limit is reached, dropping the output voltage to near zero. The additional resistors and C1 serve to stabilize the circuit. The value of the current sense resistor, R7, had to be increased so that the maximum output current is close to 84mA as in the previous circuit. Below is a plot of VOUT versus load current.
Attachment:
175VREG SCR foldback CL Plot.jpg
175VREG SCR foldback CL Plot.jpg [ 138.37 KiB | Viewed 655 times ]

The last plot is load current (first plot), the power dissipation in M1 (second trace) and he power dissipation in the LM317 (third trace).
Attachment:
175VREG SCR foldback CL Pwr Diss Plot.jpg
175VREG SCR foldback CL Pwr Diss Plot.jpg [ 268.67 KiB | Viewed 655 times ]

Notice that the maximum dissipation in M1 is about 5W and 1.9W in U1. Both drop to zero in current limit. This is a vast improvement in the MOSFET power dissipation. In fact, a lower power part such as the IRF720 could be used where it would be inadequate in the previous circuit.

Data sheets:
IRF720: https://www.vishay.com/docs/91043/sihf720.pdf
IRF740: https://www.vishay.com/docs/91054/91054.pdf
ZTX458: https://www.diodes.com/assets/Datasheets/ZTX458.pdf
ZTX558: https://www.diodes.com/assets/Datasheets/ZTX558.pdf
1N4936: https://www.onsemi.com/pdf/datasheet/1n4933-d.pdf

John

_________________
“Never attribute to malice that which can be adequately explained by stupidity.”
― R. A. Heinlein


Top
 Profile  
 
 Post subject: Re: LM317 regulator design
PostPosted: Jun Sat 19, 2021 3:25 am 
Member
User avatar

Joined: Nov Mon 02, 2009 7:01 am
Posts: 5313
Location: Lincoln City, OR 97367
Greetings to the Forum:

I recently had an occasion to attempt design of a regulated power supply designed to deliver 270 volts at up to 200 mA using an LR8N regulator chip and a TIP50 pass transistor. I started out with the design from this article in Nuts and Volts magazine:

https://www.nutsvolts.com/magazine/arti ... wer-supply

Unfortunately, when attempting to tailor it for my needs, I discovered that there was an error in the design. The minimum voltage from the supply was considerably lower than the 65 volts stated in the article, and further, the useful voltage range for tubes occurred over the last 3% or so of the pot's rotation.... making the supply very difficult to use. I discovered this error (confirmed by comments appended to the article by people who actually built the thing) by creating a spread sheet to calculate the expected output voltage using the LR8N regulator. The spread sheet also allows for the insertion of a pot to adjust the voltage and will calculate the voltage output by the chip at any percent of the pot's rotation from zero to one hundred percent.

The voltage developed by the LR8N is given by the formula (from the chip maker's data sheet) as: Vout = 1.20V*(1+(R2/R1))+Iadj*R2. This is the formula employed by the spread sheet.

Unfortunately, the formula gives us the Vout for two known resistors. Usually, we have a desired Vout and we want to know the resistor values needed to get there. Here's how we can do that: For a ball park figure for an adjustable power supply, one should pick a value for R1. This typically seems to be around 2K but you can choose anything you want and see what pops up. First, we can simplify the above equation quite a bit by simply throwing away the Iadj *R2 part of the formula. Although it should be taken into account, especially for non-adjustable supplies, even with a 1 meg R2, the result only adds 10 volts... which is not terribly significant when dealing with B+ voltages of 200 volts or more. Since we are trying to find a ball-park figure for R2, we can simply ignore this part of the formula. Then, we can distribute the 1.2 volts through the parentheses like this: Vout = [1.2+(1.2*R2)/R1]. I picked a value for R1 of 2.2K. Since we have a desired Vout (in my case, 270 volts, our equation then becomes: 267.8= [(1.2*R2)/2200]. Solving for R2, we get R2=491K.

In order to facilitate calculation by the spread sheet, I have renamed the resistors according to the following schematic:

Attachment:
LRN8 Generic Regulator.jpg
LRN8 Generic Regulator.jpg [ 35.52 KiB | Viewed 546 times ]


In order to find the values for a fixed resistor (non-adjustable) regulator, one need only set the value of Rpot to zero and the pot goes away. If you plug 2.2K and 491K into the spread sheet (in ohms, not Kohms), you will see that we get pretty close to our desired 270 output volts.

The problem comes when we add a pot. I have some 1K trim pots in stock, so I inserted that value in the spread sheet. I then dialed the pot to zero percent, which places the wiper at the bottom of R upper limit, the 2.2K resistor. I then changed the 491K resistor to 490K and sure enough, we get the same output. However, when we start to move the pot away from zero percent, we depart quickly and radically from our desired output voltage. Here's where the spread sheet comes in handy. We can drop in experimental values for R lower limit and see what happens when we turn the pot. As you will observe with the attached spread sheet, the final value of R lower limit turns out to be 620K. If you look at the percent entry for line 9 of the spread sheet, you will see that we hit our desired value of Vout at 63% of the pot's rotation, which is close enough to the center to give us some range. The range is also given on the spread sheet; lines 5 and 7 give the minimum and maximum voltages available with the specified combination of resistors and pot.

I hope this will be of some use to anyone contemplating using the LR8N chip in an adjustable regulator circuit.


Attachment:
LR8N3 Resistor Work-up V2.xls [25 KiB]
Downloaded 16 times


Edit: Misplaced decimal point in Iadj column. Corrected spread sheet and text to concur. JKT


Regards,

_________________
Jim T.
KB6GM


Top
 Profile  
 
Post New Topic Post Reply  [ 14 posts ] 

All times are UTC [ DST ]


Who is online

Users browsing this forum: No registered users and 5 guests



Search for:
Jump to:  


































Privacy Policy :: Powered by phpBB