YABSa: Yet another bench Supply (version A)

[Jebnor]

I've been working on my new bench supply. I've successfully created a positive supply which has <5mVpp Noise at 12V, over a current range from 0 to 1.2A.

I use no ICs.  Thought anyone? Is there anything I'm missing?

[BravoV]

Not an expert, just a thought, since that D6 1N4148 diode is series with the zener forming the voltage reference, that D6 tempco is about -2.1 mV per celcius, isn't that will affect the output stability ? cmiiw

[Jebnor]

Not an expert, just a thought, since that D6 1N4148 diode is series with the zener forming the voltage reference, that D6 tempco is about -2.1 mV per celcius, isn't that will affect the output stability ? cmiiw

Oh, Probably... Everything else is temperature compensated, as far as I can tell, so why skimp on the most important part.

I have a bunch of the 1N4742a's around, I could select one and adjust the current source so that it is above 12V and then use RV1 to fine tune.

[codeboy2k]

Not an expert, just a thought, since that D6 1N4148 diode is series with the zener forming the voltage reference, that D6 tempco is about -2.1 mV per celcius, isn't that will affect the output stability ? cmiiw

Oh, Probably... Everything else is temperature compensated, as far as I can tell, so why skimp on the most important part.

I have a bunch of the 1N4742a's around, I could select one and adjust the current source so that it is above 12V and then use RV1 to fine tune.

I thought it was obvious... the zener's reverse biased and thus it's tempco is approx +2mv/C... the diode D6 being forward biased is approx. -2mV/C... the two tempcos cancel each other out and this stabilizes the voltage ref over temperature.

[BravoV]

Cool  , thanks guys.

[quantumvolt]

Very nice. I like it all discrete. Matching transistors for current mirrors and diff amps is for the dedicated 

Not an expert, just a thought, since that D6 1N4148 diode is series with the zener forming the voltage reference, that D6 tempco is about -2.1 mV per celcius, isn't that will affect the output stability ? cmiiw

Oh, Probably... Everything else is temperature compensated, as far as I can tell, so why skimp on the most important part.

I have a bunch of the 1N4742a's around, I could select one and adjust the current source so that it is above 12V and then use RV1 to fine tune.

I thought it was obvious... the zener's reverse biased and thus it's tempco is approx +2mv/C... the diode D6 being forward biased is approx. -2mV/C... the two tempcos cancel each other out and this stabilizes the voltage ref over temperature.

There is more to it. The tempco varies with Zener voltage. So for example to make a decent 9 V temp compensated voltage it is common to use a 7.5 V Zener and 2 small signal PN diodes in series. Here is a decent short read: http://www.techlib.com/reference/zenertempco.htm

[Jebnor]

I thought it was obvious... the zener's reverse biased and thus it's tempco is approx +2mv/C... the diode D6 being forward biased is approx. -2mV/C... the two tempcos cancel each other out and this stabilizes the voltage ref over temperature.

Turns out I may have accidentally did something better.

Very nice. I like it all discrete. Matching transistors for current mirrors and diff amps is for the dedicated 

[cut]

There is more to it. The tempco varies with Zener voltage. So for example to make a decent 9 V temp compensated voltage it is common to use a 7.5 V Zener and 2 small signal PN diodes in series. Here is a decent short read: http://www.techlib.com/reference/zenertempco.htm

And now, to run down the zener Diode temperature compensation rabbit hole.

Thank you all.

[Jebnor]

Probably more than you wanted to know about Zener Diodes (127pg): http://www.onsemi.com/pub_link/Collateral/HBD854-D.PDF