Need some advice. Accomodating different voltage rails in console.

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

JW

Well-known member
Joined
Jun 8, 2005
Messages
1,100
Location
Portland USA
My console runs on +/-24VDC rails. (It's a Yamaha PM2000) I'm in the process of building Jeff Steiger's API summing section clone.

My desire is to use the existing power rails, but I thought I'd pick folks' brains here first.

In the API summing buss, there are 4 DOA's.  I want to be able to swap DOA's in and out and compare. Right now I have a bunch of GAR2520 kits as well as some of the GAR Melcor style kits, a couple Jensen style opamp kits from Whistlerock,  and some original Hardy 990C's that are perfectly happy running at +/-24V. I want to be able to listen and pick and choose from all of these, but there is obviously a discrepancy concerning what some of these opamps want to run at and what is available.

When I bought the 2 Hardy 990c's (which are awesome BTW) I made sure to get the one's that are built to be +/-24V capable. I think they'll probably be happy running at less, but I'll have to check on that.

Otherwise, most of these opamps want +/-18V at the most.

There's obviously a few options, but I wonder what you guys would recommend?

I could use the console's existing voltage rails and drop the voltage going into the opamps individually with 1 approximately 7V zener diode in series on both rails.

Or I could build a separate power supply just for the API summing buss.

There are probably other options I'm not thinking of.

What would you folks recommend?
 
Dropping from +/- 24 down to +/- 18 via series Zeners is a BAD idea.  I suggest adding a LM317 (for the positive rail) and a LM337 (negative rail).

Best,

Bri

 
It seems like the zeners in series could work as long as you use adequate decoupling capacitance to maintain a low PS impedances, And perhaps a modest resistor between them to insure the zeners drop enough at light load current. Of course real regulators are always better, and this doesn't seem like a project to cut corners on.

While I am not familiar with idiosyncrasies between sundry DOAs, it seems their internal operating current may be dialed in for specific rail voltages, so they probably won't be very happy running at other than target design rail voltages.  Higher than design rail voltages could draw excessive current and overheat. Too low rails could starve some internal sections.  IC opamps are much better behaved at different rail voltages.

JR
 
Okay,
So, I've never done the math using a an LM317 regulator before.

We're talking about say, the schematic on this page? http://diyaudioprojects.com/Technical/Voltage-Regulator/

For the equation on that page, I'd plug in 18V for Vout. 
The right side of the equation is what confuses me a bit.  For Iadj, is that "I" which is input voltage * the voltage adjust, which would be 6V?
So, 24V*6V(R2)
Or is it only voltage adjust * R2, which would be 6V*R2?
 
http://www.reuk.co.uk/LM317-Voltage-Calculator.htm

Use this one... Just pick your r1, which should be between 120-240. I use 220 cause it's what I stock. Then dial in your output voltage. It will give you a resistor for R2 for that voltage.

I usually calculate the +-4 volts range so I can dial it in with a trimmer.

The317 is basically a 1.25V regulator, and you're lifting the "ground" (adjust) pin up until you get the output you want.
 
Ah, okay,
Much easier.

I don't see how input voltage is accounted for though.
Or is V out actually the voltage difference, so I'd be looking at the values around 6V?

Or the ones around 18V?
 
probably one of the best calcs for this is built into the electrodroid app for android. you can enter In/out voltages, it will tell you what real world resistors will get you close. It will even calculate wattage dissipated.
 
In this instance the regulator may not care but generally it does.

The input voltage is selected by taking what is called the "drop out" voltage into account, which is temperature and output current dependent. This is the voltage difference between the input and output. You have to factor that in. See page 6 on the following data sheet;

http://www.ti.com/lit/ds/symlink/lm117.pdf
 
JW said:
Ah, okay,
Much easier.

I don't see how input voltage is accounted for though.
Or is V out actually the voltage difference, so I'd be looking at the values around 6V?

Or the ones around 18V?

When designing with LM317-type regulators, you concern yourself with:

a) input-to-output differential. This is the voltage difference between the input and the output. I think on the 317 it's something like 36 V. So if you're trying to do a phantom power thing, your output is 48V. That doesn't mean the 317 won't work!

b) drop-out voltage. The regulator input has to be at least this voltage above the output. On the 317 it's something like 3V and is somewhat dependent on current. So if you're trying to do a 15V supply, the input has to be at least 18V.

c) power handling/heat dissipation. Power dissipated in the regulator is simply the input-to-output difference times the current. The greater the current and/or the greater the input-to-output differential, the more heat that gets dissipated. You'll probably need a heatsink. In most cases, when overheated, the 317 simply shuts down, and when it cools down, it turns back on. This can lead to interesting on/off/on/off cycles.

So there it is.

-a
 
Thanks guys. Great info. I was planning on using heatsinks as well.
So, in that last linked schematic, if I used 220ohms for R1 and a 5K trimmer for R2 I should be good to go right?

Thanks!
Jon
 
Hmmm.

Well, it looks like 4 2520 opamps is too much current draw for 1 LM317 and 1 LM337 pair? They dropped voltage fine with no load, but as soon as they're loaded with the 4 DOA summing buss, they drop down to about 3V

I do have heatsinks. Should I look for a higher current version of these regulators, or use an additional pair somehow?
 
LM317 and LM337 will run them with ease. You have an implementation error.

some common possibilities, roughly in the order of probability:

1. Your heat sinks are inadequate size.
2. You have inadequate thermal coupling from the regulators to the heat sinks.
3. You're trying to drop too much voltage with the regulators. Too much heat dissipated for your sinks, go to 1.
 
JW said:
Hmmm.

Well, it looks like 4 2520 opamps is too much current draw for 1 LM317 and 1 LM337 pair? They dropped voltage fine with no load, but as soon as they're loaded with the 4 DOA summing buss, they drop down to about 3V

I do have heatsinks. Should I look for a higher current version of these regulators, or use an additional pair somehow?

What kind of heatsinks? How hot do they get?  If you can't touch then without removing your finger prints you may still need better heatsinking.

If the heatsink seems adequate, is the input voltage to the regulator holding up, or sagging?

It's always something...

JR
 
Yeah, the heatsinks get blisteringly hot, but only one of them at a time, strangely. By that I mean, first the LM337, who, after the load was removed, still wasn't able to adjust voltage at all, so I replaced it. Next the LM317 was the culprit, blisteringly hot. I stopped screwing around at that point.

I'm using the schematic that gemini86 posted earlier in this thread.

I used these heatsinks: http://search.digikey.com/scripts/DkSearch/dksus.dll?vendor=0&keywords=HS106-ND
What kind of dissipation characteristics should I be looking for?

Edit: Okay, I'm realizing that the transistors shouldn't be conductive with the heatsinks? Is this an idiot rookie mistake? Do I need to procure some thermal grease?
 
JW said:
Yeah, the heatsinks get blisteringly hot, but only one of them at a time, strangely. By that I mean, first the LM337, who, after the load was removed, still wasn't able to adjust voltage at all, so I replaced it. Next the LM317 was the culprit, blisteringly hot. I stopped screwing around at that point.

I'm using the schematic that gemini86 posted earlier in this thread.

I used these heatsinks: http://search.digikey.com/scripts/DkSearch/dksus.dll?vendor=0&keywords=HS106-ND
What kind of dissipation characteristics should I be looking for?

Edit: Okay, I'm realizing that the transistors shouldn't be conductive with the heatsinks? Is this an idiot rookie mistake? Do I need to procure some thermal grease?

Without seeing how you built the thing, be warned: the tabs of these regulators must be isolated from ground and any other potential.  LM317 tab is connected to the output, LM337 tab is connected to the input.

-a
 
JW said:
Do I need to procure some thermal grease?

grease won't give adequate insulation. You need an isolation pad + plastic washer. The common TO-220 size "kits" are available in just about every electronics store.

You can slightly improve thermal conductivity by applying a very thin layer of that silicon grease to both sides of the pad, but this is not necessary, just an optimisation.

As for heat sink sizes, too big is always good. The sizes used in common ATX computer power supplies work well. Get some broken ones and recycle. The heat sinks from these have never failed even in the most overkill regulation cases for me.
 

Latest posts

Back
Top