LM329 as voltage reference in a CCS.

[Ryssen]

Any one tryed it,does it have the same drawbacks as a Zener (noice e.t.c)?

[SY]

It's quieter than most zeners (take a look at the datasheet for noise density curves). Very stable. Lowish current.

[Ryssen]

Ok,I thought of trying it in a CCS instead of a Led,is there any gain in bypassing it?

[SY]

It will probably be noisier than a LED. As far as bypassing, that depends on the details of the circuit you're using it in; you have to see what the impedance at that point is and if any reasonably sized bypass will get you a significant improvement.

[Ryssen]

Well the circuit is:

I read in the pdf file for the CCS that I could make it better by using the LM329 for the first Led,but if you say it is noisier maybe it´s not an improvement?

[SY]

A 329 will give you 6.9V, so the emitter resistors would have to be adjusted accordingly. What the extra voltage gets you is those larger resistors and hence higher output Z from the CCS.

Noise shouldn't be an issue- it's divided down by the ratio of the CCS source Z to the plate resistance of the tube.

[Ryssen]

Noise shouldn't be an issue- it's divided down by the ratio of the CCS source Z to the plate resistance of the tube.

Ok,It´s worth a try then,will it hurt to bypass with a 0,1 filmcap?

[jackinnj]

After reading the WJ article in AudioXpress I switched to a LM329DZ + 2N5639 + 2N5089 Reference -- (WJ used LM329 + J113 + PN2222A.)

I haven't done a noise test on this yet but the graph in the article was pretty impressive -- here's the circuit and some data:

[Fanuc]

Originally posted by jackinnj
After reading the WJ article in AudioXpress I switched to a LM329DZ + 2N5639 + 2N5089 Reference -- (WJ used LM329 + J113 + PN2222A.)

I haven't done a noise test on this yet but the graph in the article was pretty impressive -- here's the circuit and some data:

Hello,

We would need to see the Zout versus frequency plot. (50khz plots would at least cover the audioband! Ho Hum!!) . Also any capacitances from PCB's & wiring etc will reduce this. The Zout for higher current looks attrocious - even compared to ring of two at high currents. A good one I seen was two LM329 in series biased by a J50* and then two bjt's cascoded. This is the best I have seen in terms of Line Rejection.

Max Zout doesn't neccesarily improve Line Rejection (or there is some competing mechanism) - Witness cascoded jfets etc. I have seen some Jfet CCS's with >100M Zout at LF but PSRR was about 85db and diminished greatly at HF. See what I mean.

I think from what I have read that very low gm FETs are best for current sources. Plus could be RC filtered for lower noise.

Best Regards

Kevin

[jackinnj]

Did you see the article? This is the actual circuit which has an equivalent impedance of 1.8MegOhms...ymmv