Here's the sneak preview negative version circuit for those who want to build it on their own. Pretty much all the comments from the positive version apply. I decided some time ago not to go with complementary devices (i.e. instead of n-channel mosfets to use p-channel mosfets, etc.). Instead the topology is turned upside down
but the benefit is that you can expect a much closer match in terms of performance for the positive and negative rails, plus, using the same devices in both circuits makes it easier when shopping around for components.
This will be revision 5d. The only changes from 5c is using only one LED in biasing M2, and the addition of capacitor C4, which is intended to be a film type, or whatever. This position seems to be important and several people have mentioned that the type and value makes an audible difference. Look back in the thread for hifinutnut's comments, it'll be helpful (thanks hifinutnut!).
REMINDER: all capacitor and resistor value are not intended to be exact values. 1% tolerance is not what this circuit is about and it would make no difference in reality. Some resistor values shown in the circuit MUST be changed according to the parts that you end up with. It is possible to narrow the values down, but well, do you want a few pages of math? It's much easier if I give you a recipe, and I'm going to write one.
Example: the CCS mosfet M2 limits the main current output. Now, you can use a lot of different mosfets for M2. Some people may not like this, but I like it. I've always liked to have options AND a default that is well chosen. In this case I chose the default to be irfbc40 or irfp240, both will work like a charm. But hey, maybe you have an irf630 on hand. It can be used just fine. But each may be biased slightly different. So we use a mini-CCS (constant current source) to pass a relatively constant current through R2 and the LED. Both the LED and R2 cause a voltage drop. It could be about 2V across the LED, and another 1.5V across R2, depending on its value and the current passing through it. The total voltage across LED+R2 is the same as the voltage that M2 has at its G and S pins. But this works well if we pass a certain relatively constant current through D1+R2. That is achieved via the mini-CCS made up of J2, J3, and R7. I won't go into details, but there is theory that says you get good regulation (and tempco) in a CCS by using a value of R7 such that the current limit of this mini-CCS be about 15% of the Idss of J2. When I say "about" it means approximately. No need to go crazy here with exact measurements because it won't make a difference. So hook up J2 in series with a 3V battery and your DMM in series, set on 20mA. J2 G and S pins touching. J2 D pin on +3V. G+S on V+ probe of DMM. -3V on V- probe of DMM (in the how to I'll draw a picture
). This is how you measure Idss of J2. Let's say it is 10mA. Good, 15% of 10mA is 1.5mA. Now hook up J2 + J3 + 500R trimmer as shown in the circuit, to a 3V battery in series with your DMM set on 20mA. Adjust the trimmer until your DMM shows 1.5mA. Take the trimmer out and see how many ohms it measures. It could be 200 ohms. Find the closest fix value resistor you have. Don't obsess about a few tens microamperes difference. I know it's hard not to obsess for an audiophile, but you have to try. That's how R7 value is decided. This is the recipe. Then, in the real circuit you'll use this R7, and you will use a trimmer for R2 to get the current limit that you want. Then replace the trimmer with a fix value for R2.
Now, back to the drawing board for me get that negative rail pcb done.
6th December 2009, 02:44 PM
Didn't think there would be so much interest in it. OK, more work needs to be done then. The circuit design I have done some time ago, and even built a negative prototype. Will report progress as it happens. The wheels on the bus go round and round ... 
). This is how you measure Idss of J2. Let's say it is 10mA. Good, 15% of 10mA is 1.5mA. Now hook up J2 + J3 + 500R trimmer as shown in the circuit, to a 3V battery in series with your DMM set on 20mA. Adjust the trimmer until your DMM shows 1.5mA. Take the trimmer out and see how many ohms it measures. It could be 200 ohms. Find the closest fix value resistor you have. Don't obsess about a few tens microamperes difference. I know it's hard not to obsess for an audiophile, but you have to try. That's how R7 value is decided. This is the recipe. Then, in the real circuit you'll use this R7, and you will use a trimmer for R2 to get the current limit that you want. Then replace the trimmer with a fix value for R2.
Similar Threads
