Just like this one! Although tbh, if was going to that much trouble I'd build something a little better!!
Attachments
Yeah, but that's half the point: for less than 1euro you can approach 100dB ripple rejection if you want.
Replace R2 with a red LED and C2 with 1uF+2R7 in series (or 100nF +10ohms if you like) and it'll perform even better noise-wise. Not the ultimate, no: but dirty-cheap, and very effective.
Replace R2 with a red LED and C2 with 1uF+2R7 in series (or 100nF +10ohms if you like) and it'll perform even better noise-wise. Not the ultimate, no: but dirty-cheap, and very effective.
cough, like this:
It's the tracking pre-arrangement, not the minor changes that makes for good performance. Look at how it's arranged: the first 317 ensures there is *no* ripple voltage appearing across the second, in fact the second reg sees a constant voltage difference across it regardles of whayt the load does. So the second reg applies all its gain to load regulation, while the first deals only with the ripple/noise on the input. This is a subtley, but it audibly improves on just using two regs in series.
So - my suggestions are using an LED voltage drop to set c.2.8-3v across reg2, which absolutely minimises noise from the first reg (because the led at 5-10ma has a dynamic impedance of only c.20-30ohms). The added resistor(in blue) critically-damps the input to the second reg, stopping any spurious oscillation due to high loop gain / layout strays. Use 10R with C2 = 100nF down to a min. 2R7 for 1uF, or 1ohm with 10uF.
A circuit well worth playing with for the very low cost of entry.
It's the tracking pre-arrangement, not the minor changes that makes for good performance. Look at how it's arranged: the first 317 ensures there is *no* ripple voltage appearing across the second, in fact the second reg sees a constant voltage difference across it regardles of whayt the load does. So the second reg applies all its gain to load regulation, while the first deals only with the ripple/noise on the input. This is a subtley, but it audibly improves on just using two regs in series.
So - my suggestions are using an LED voltage drop to set c.2.8-3v across reg2, which absolutely minimises noise from the first reg (because the led at 5-10ma has a dynamic impedance of only c.20-30ohms). The added resistor(in blue) critically-damps the input to the second reg, stopping any spurious oscillation due to high loop gain / layout strays. Use 10R with C2 = 100nF down to a min. 2R7 for 1uF, or 1ohm with 10uF.
A circuit well worth playing with for the very low cost of entry.
Hi Ian,
Yes been a member for a while but not posted much. Enjoyed reading about many mods hear though.
The diagram posted is a basic tracking reg and I am actually using Avondale's pretrackers for my 5v supplies. Old Les has added a few bits to the basic layout to reduce ripple so I am not certain that Martin's tidy mod applies and that's why I asked a couple of questions.
Regards
Yes been a member for a while but not posted much. Enjoyed reading about many mods hear though.
The diagram posted is a basic tracking reg and I am actually using Avondale's pretrackers for my 5v supplies. Old Les has added a few bits to the basic layout to reduce ripple so I am not certain that Martin's tidy mod applies and that's why I asked a couple of questions.
Regards
Ah, thought you were interested in a 'from scratch' ideas. I'd leave LesW's TPRs alone, not much to be gained by chopping them about 
RCruz - not necessarily. A tracking reg arrangement requires about 5-6v minimum headroom or 'dropout' voltage from input to output, whihc you may not have. A good shunt reg uses a current source to feed the load anyway, so what it upstream should be fairly irrelevant. If you want to add anything, a little RC decoupling before will help at high frequencies when the active current source starts to become less effective (but a little RC ahead of most regs is rarely a bad idea IMO!)
RCruz - not necessarily. A tracking reg arrangement requires about 5-6v minimum headroom or 'dropout' voltage from input to output, whihc you may not have. A good shunt reg uses a current source to feed the load anyway, so what it upstream should be fairly irrelevant. If you want to add anything, a little RC decoupling before will help at high frequencies when the active current source starts to become less effective (but a little RC ahead of most regs is rarely a bad idea IMO!)
Ray, did your daughter get her Christmas present ?
Simply Swank Pink Solder Iron 60W Jewelry Crafts Tool on eBay (end time 23-Jan-11 13:15:29 GMT)
.
Simply Swank Pink Solder Iron 60W Jewelry Crafts Tool on eBay (end time 23-Jan-11 13:15:29 GMT)
.
Not so early; 1985. I have looked everywhere but there is no existing reference to this.
Bonne année Frenchie! Il vas faire beau parait-il.
à propos... of nothing and everything. I turned off the current, opened up the wall sockets, tightened the wiring connections and treated everythig with Craig Gold cleaner and enhancer. Nice cheap and effective
tweak. Sounds cleaner too!
Bonne année Frenchie! Il vas faire beau parait-il.
à propos... of nothing and everything. I turned off the current, opened up the wall sockets, tightened the wiring connections and treated everythig with Craig Gold cleaner and enhancer. Nice cheap and effective
tweak. Sounds cleaner too!
Last edited: