I'm just looking at your website to know how you built yours.
If I understand, there are two ways to control the volue : change the value of the feedback resistor or short the input to ground, but I suppose in fact it is the same as both solutions are connected to line in and signal ground.
Is it this ?
If I understand, there are two ways to control the volue : change the value of the feedback resistor or short the input to ground, but I suppose in fact it is the same as both solutions are connected to line in and signal ground.
Is it this ?
Setting the gain of the amplifier is not the same as having a method of controlling the volume while playing.
Think of it like water. If you allow all the water to flow through a pipe (to the amp) you have maximum volume. Because this is too loud, you need to allow some of the water to branch off from the main pipe, through another pipe and back to the source. How much water flows to the amp depends on how much water flows back to the source and that in turn depends on the size of the pipe branching off the main one.
This is what an attenuator does except that instead of different diameter pipes, we use resistors of different resistance.
The attenuator shown on DD is a single switch shunt attenuator that works well with a GC. If you want to use two switches to provide finer control, I can probably work out the values for you. I used a two-switch attenuator for a while but didn't really like it and am happier to put up with a single 12-step attenuator. Having said that, I now use a 23-step attenautor that I bought through Ebay!
Think of it like water. If you allow all the water to flow through a pipe (to the amp) you have maximum volume. Because this is too loud, you need to allow some of the water to branch off from the main pipe, through another pipe and back to the source. How much water flows to the amp depends on how much water flows back to the source and that in turn depends on the size of the pipe branching off the main one.
This is what an attenuator does except that instead of different diameter pipes, we use resistors of different resistance.
The attenuator shown on DD is a single switch shunt attenuator that works well with a GC. If you want to use two switches to provide finer control, I can probably work out the values for you. I used a two-switch attenuator for a while but didn't really like it and am happier to put up with a single 12-step attenuator. Having said that, I now use a 23-step attenautor that I bought through Ebay!
In fact it means that the attenuator or the pot shorts the input to its ground, simply.
So the pot shorts progressively the input to the ground, and sends the rest in the middle pin, hence the 3 pins, whereas the attenuator shorts to the ground and what is not sent to the ground is connected directly to the chip input.
It also means that the attenuation depends on the input impedance isn't it ?
So the pot shorts progressively the input to the ground, and sends the rest in the middle pin, hence the 3 pins, whereas the attenuator shorts to the ground and what is not sent to the ground is connected directly to the chip input.
It also means that the attenuation depends on the input impedance isn't it ?
Well, I'm about to order the components for my NIGC, using snubberized 2200mfd on the PSU pins.
Looking at the LM3875 datasheet, the chip seems to dissipate very much power !
According to the graphs, it dissipates 30W for a 15W output at +-35V supply !!
It means that I need theorically a 1.5K/W heatsink !
Looking at gainclones realizations, it seems that some people use very large heatsinks, like some you can see on briangt website, and others that use ridiculously small ones like http://idefixes.phpnet.org/sections.php?op=viewarticle&artid=2
As heatsinks are really expensive here in France (10 euros for a 1.5 K/W) , do you really think I need such a big heatsink ?
Looking at the LM3875 datasheet, the chip seems to dissipate very much power !
According to the graphs, it dissipates 30W for a 15W output at +-35V supply !!
It means that I need theorically a 1.5K/W heatsink !
Looking at gainclones realizations, it seems that some people use very large heatsinks, like some you can see on briangt website, and others that use ridiculously small ones like http://idefixes.phpnet.org/sections.php?op=viewarticle&artid=2
As heatsinks are really expensive here in France (10 euros for a 1.5 K/W) , do you really think I need such a big heatsink ?
A well-built LM3875 will not get very warm at all! So you can get away with a smallish heatsink. As a guide, a CPU heatsink is OK to cool one channel. The smallest that I have used is a piece of copper bar about 10 mm thick and 75 mm by 50 mm.
And have a look at my original GC monoblocks. Half of the rear is heatsink - 10 mm thick aluminium but no fins and it runs at around body temperature.
An externally hosted image should be here but it was not working when we last tested it.
I saw very cute heatsinks in an electronics shop which cost only 1€ each. They have a 3.1K/W rating. Will it be enough if I use one per channel ?
Though you seem to use small heatsinks, they are really thick and therefore can take much heat away from the chip, even if they have no fins. Do you approximately know what is their rating ?
Considering the CPU heatsinks, I've always wondered how were they rated w/o any fan. I suppose they aren't so efficient aren't they ?
Though you seem to use small heatsinks, they are really thick and therefore can take much heat away from the chip, even if they have no fins. Do you approximately know what is their rating ?
Considering the CPU heatsinks, I've always wondered how were they rated w/o any fan. I suppose they aren't so efficient aren't they ?
No I don't. I just tried them and they are adequate in use.
The fins are closer together and ideally need a fan to force the air over the fins. But again, in practise, I have found that they work adequately for the LM3875. The ones shown below, were even cut down to fit inside the case. There is an aperure in the base and lid to encourage air-flow over the fins.
An externally hosted image should be here but it was not working when we last tested it.
youyoung21147 said:
The National Semi design tool allows you to determine the appropriate heat sink to a given supply voltage -- http://www.national.com/appinfo/audio/files/Overture_Design_Guide15.xls
with 3.1 C/W you are good for about +/- 34V with the LM3875, or the LM3886T (but NOT the LM3886TF).
a copper bar will store some energy as heat, but won't do a particularly good job of dissipating heat -- for this you need surface area and air movement. CPU heat sinks are pretty darned efficient and have the benefit of continuous airflow -- aluminum is a decent conductor of heat (that's why baseboard heating has alum fins).
Thick heatsinks can take much heat from the chip at first but they can't get rid of it to the air, so they are not very effective. Even thought it isn't very effective, it works just fine for the chip amp. Fins help a lot-especially if they are widely spaced.
Don't worry
The small heatsinks you linked to are fine. Use something that size or a little bigger and there will be no problem
Mine are less than twice as big, and they barely get warm.
Look up some of Peter Daniels creations- they appear to have only approx. 50mm x 150mm x 12 mm solid blocks as heatsinks for both channels combined-no fins.
If you calculate the surface area of this lock you will see that it is the same as a very small finned heatsink.
Maybe you can find a flat piece of aluminum about the size of Peters but at least 5mm thick - that would work fine. I have often been tempted to use an old thick alu cooking pot- you can put the circuit inside, chips bolted through the bottom!
Don't worry
The small heatsinks you linked to are fine. Use something that size or a little bigger and there will be no problem
Mine are less than twice as big, and they barely get warm.
Look up some of Peter Daniels creations- they appear to have only approx. 50mm x 150mm x 12 mm solid blocks as heatsinks for both channels combined-no fins.
If you calculate the surface area of this lock you will see that it is the same as a very small finned heatsink.
Maybe you can find a flat piece of aluminum about the size of Peters but at least 5mm thick - that would work fine. I have often been tempted to use an old thick alu cooking pot- you can put the circuit inside, chips bolted through the bottom!
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.