szekeres power supply
Nov 2, 2004 at 4:23 PM Thread Starter Post #1 of 24

ECM

100+ Head-Fier
Joined
Jan 9, 2004
Posts
175
Likes
23
i'm building a szekeres amp and i have some questions concerning the power supply.

i plan on using an 18V transformer and regulating the voltage down to 12VDC. is the drop in voltage too much for a 7812 1.5A regulator?

i have two 2700uf 35VDC caps before the regulator and a 2200uf 35VDC + 0.1uf 250V cap after the regulator. will this blow up on me?
 
Nov 2, 2004 at 5:41 PM Post #2 of 24
why only 12V ? It's not gonna blow, you'll just need more cooling. A LM7815 would be perfect.

35V caps are perfectly ok and 2200uf after the regulator will work like a charm, don't worry.
 
Nov 2, 2004 at 5:57 PM Post #3 of 24
It's all about cooling. Many linear regulator datasheets (particularly the ones from Linear Technology) have good info on how to calculate the size of the heat sink you need.
 
Nov 2, 2004 at 6:05 PM Post #4 of 24
i'm using 12v because that is the regulator i had on hand. i went to radioshack and couldn't find a 15 volt regulator (can you believe that?).

if i mail order the 15 volt regulator i'll end up buying more stuff than i need!!

the heatsinks i'm using are these from Allelectronics.com:

http://www.allelectronics.com/cgi-bi...04&type=store#

are they big enough? these are the same heatsinks i'll be using on the mosfets too. should i go bigger?
 
Nov 2, 2004 at 6:15 PM Post #5 of 24
well, i just looked up the temperature rating on the regulator and it requires 3C/W. the heatsink i was planning on using is more like 2.5C/W. guess i'll grab a bigger heat sink.
 
Nov 2, 2004 at 7:41 PM Post #6 of 24
Quote:

are they big enough?


Aavid-Thermalloy doesn't list that part in their catalog, and I can't see anything like it from browsing. Just eyeballing it, I'd say it's 5C/W or smaller.

Quote:

it requires 3C/W


?? That sounds like an oversimplification. Regulators don't "require" a particular thermal resistance. You have to plug in the voltage drop, the output current, and the desired temperature rise to come up with a thermal resistance. And, of course, you need to account for the thermal resistance of the regulator package and the heat sink compound.
 
Nov 2, 2004 at 8:40 PM Post #7 of 24
i just went to the thermalloy website and cross referenced the 6098 heatsink. the replacement is this one:

http://www.aavidthermalloy.com/bin/s...e=v2&SortBy=V1

the datasheet states it has a thermal resistance of 13.4.

so how would i compute how many watts are going to be dissipated by the regulator?
1.5A * 12VDC = 18 watts.

or do i compute the voltage drop accross the regulator?
(18V-12V)*1.5A = 9 watts.

how do i use this data to compute the heatsink size?
 
Nov 2, 2004 at 9:00 PM Post #8 of 24
Quote:

Originally Posted by ECM
i'm using 12v because that is the regulator i had on hand. i went to radioshack and couldn't find a 15 volt regulator (can you believe that?).

if i mail order the 15 volt regulator i'll end up buying more stuff than i need!!

the heatsinks i'm using are these from Allelectronics.com:

http://www.allelectronics.com/cgi-bi...04&type=store#

are they big enough? these are the same heatsinks i'll be using on the mosfets too. should i go bigger?



you can get an LM317 variable voltage regulator at ratshack and have it put out 15v
 
Nov 3, 2004 at 5:51 AM Post #9 of 24
good call flecom. i think i might have a couple of them lying around. i'll check tomorrow.

should i just bolt the regulator to the case of the amp for cooling? i have the same case as the one in the original article.
 
Nov 3, 2004 at 8:22 AM Post #10 of 24
Quote:

or do i compute the voltage drop accross the regulator?
(18V-12V)*1.5A = 9 watts.


Yes, that's it. Although, you're probably not actually drawing 1.5A from the regulator. Calculate for what you're drawing, not the theoretical max. If you are indeed pulling 1.5A, I'd say you need a bigger regulator, since that's the limit of the LM317 and 78xx series.

Quote:

should i just bolt the regulator to the case of the amp for cooling?


The tab of most power regulators is 'live'. Even if none of your panel components ties a voltage to the case, you'll have a negative regulator, too, won't you? If so, the tab of the negative regulator is usually at a different potential than the positive.
 
Nov 3, 2004 at 4:06 PM Post #11 of 24
the 12 volt regulator has the tab connected to the "common" leg, however, if i use the lm317, i will have to insulate it since the tab is connected to the "output" leg. i need some of those nylon shoulder washers for to-220 parts. i went to the mouser website and they have two different sizes. what size should i order?

i won't be using a negative regulator. i built the dual supply amplifier circuit last month, but i scrapped it after i could not control the thermal drift. i allowed the amp to run for 20 minutes before i adjusted the bias, but it would not stay close to zero.

so i'm building the single supply version this time. i feel safer with the cap blocking the DC.

if i recall correctly, the circuit will draw about 1 amp max. the regulator will have to dissipate about 6 watts. i have a few different heatsinks i can try if one gets too hot.
 
Nov 3, 2004 at 11:31 PM Post #12 of 24
go to your local radio shack and pick up the "TO-220 Mounting Kit" which has both the shoulder washers and mica insulator.A bit pricey but instant gratification has a price.

The current draw on the original is a total of 500ma if i remember right though you can vary that by changing the value of the source resistor.I had mine as high as 1A per ch and that was not only overkill but sound wise had no benefit so it amounted to wasted heat (great to sit by in the winter though
tongue.gif
).
I think my present version is running at 300ma per ch though it really is not a szekeres any longer but more of my own design .highly recommended design though very power supply and layout dependant.
Quiet supplies and low input capacitance layout are a must and in my case I isolated the mosfet with a dual jfet white/broskie follower front end so the input to the mosfet is a constant
 
Nov 4, 2004 at 6:07 AM Post #13 of 24
i'm using the standard values for the circuit so it might come in under 1 amp. i thought about implementing tomo's regulator as a current source mod, but i really just want to make it simple the first time.

i hope my layout is ok. the amplifier "modules" are built on the small protoboards from radio shack. the ones that you snap in half. at this point i just want to listen to the damn thing!

i have a couple of radio shack nylon washers, but i bought some silicone insulators a while back and i need just the washers. not the mica and the bolt.

i'm still wondering how to calculate heat sink area for a given regualtor.
 
Nov 4, 2004 at 10:17 AM Post #14 of 24
Quote:

i'm still wondering how to calculate heat sink area for a given regualtor.


The teory is rather simple (slightly simplified):
Voltage drop * current draw = wasted heat (in Watts)

accepted temperature gain / wasted heat (in Watts) = rating on heatsink

Note that the rating of the heatsink depends on how much air is traveling around the heatsink, so in a closed case it will be much worse and with a fan against the heatsink it will be much better. The specified value usually works fine to use in a typical case. You should preferably have some hole under and above the heatsink so you get some airflow (compare with the look of the case of any commercial amplifier).

So, an exampel with made up numbers:
Voltage drop is from 18V to 12V. 18-12=6V.
Current draw is 0.8A.
Wasted heat=6*0.8=4.8W
You can accept that heatsink becomes about 20C above room temperature. This is a bit warm, but not so hot that you will get burned.
20 C / 4.8 W = 4.167 C/W. So any heatsink with a value lower than 4.167C/W should work fine.

Edited to reduce confusion and slight errors.
 
Nov 4, 2004 at 2:47 PM Post #15 of 24
thanks silvervarg.

i have a few different heatsinks i can try.

rick42, does your amp use an output cap?
 

Users who are viewing this thread

Back
Top