I finished testing the circuit from post 506. Attached is the circuit diagram that I used in my layout software. I used 3 LEDs as voltage reference for maximum flexibility, they can be bridged individually or replaced by a zener, depending on voltage requirements.
Also, I have miller caps for T3 (C6/C16) = 470pF and for T6 (C7/C17) = 1000pF. Only one cap is required in each position, the redundant caps in the layout software allowed me to have two different packages, one SDM cap (1206) and one leaded cap with 2,54 mm spacing. I finally decided on using the SMD caps (NP0 ceramics) as they need less space and can be placed closer to the transistor (less inductance).
The miller caps are required for stability as the circuit oscillates without them at some 20 - 30 MHz (setting R7 to 65 Ohms and R8 to some arbitrary 30 Ohms reduced those oscillations but did not eliminate them completely). With the caps I can only see a couple of millivolts of noise on my oscilloscope but my testsetup does not allow me to analyse this further.
Also, I have miller caps for T3 (C6/C16) = 470pF and for T6 (C7/C17) = 1000pF. Only one cap is required in each position, the redundant caps in the layout software allowed me to have two different packages, one SDM cap (1206) and one leaded cap with 2,54 mm spacing. I finally decided on using the SMD caps (NP0 ceramics) as they need less space and can be placed closer to the transistor (less inductance).
The miller caps are required for stability as the circuit oscillates without them at some 20 - 30 MHz (setting R7 to 65 Ohms and R8 to some arbitrary 30 Ohms reduced those oscillations but did not eliminate them completely). With the caps I can only see a couple of millivolts of noise on my oscilloscope but my testsetup does not allow me to analyse this further.
Attachments
This is the final layout with components. The size is 61 mm x 44 mm. Originally it was 60 mm but I added C7/C17 to my already "completed" layout after I found that the CCS oscillates without them. My prototype has this cap still soldered accross the board, btw..
Attachments
The circuit is in the zip file in post 567, which is basically the same circuit as in post 506, except that it now has an additional miller cap (C7) as described before.
Actually no improvement can be seen - in simulation, performance just looks identical. I compared various configurations: my test version (C5 on C3), the "corrected" version (C5 to +), a CCS for R12, and finally no bootstrap at all, i.e. just one resistor instead of R2/R12 and C5. It appears that this is not a critical element and the bootstrap can be omitted.
Here is the PCB of the circuit from 567 (seen from the top, i.e. component side):
MRupp said:Well, it certainly looks niceBut I now need to complete my alephono clone to put the regulator to some good use. The time I spent on the regulator would just have been enough to completete the phono stage, so kind of a catch 22 situation now.
Very nice, I'm glad it works now for you, I know how frustrating a new design can be. I've been designing/developing yet another shunt regulator and it's been a huge time sink.
JimT said:
Jim, this improvement was over which version? The compensation caps if anything, would lower the performance slightly in the higher frequency.
Martin, of course, any change on some existing design, or a new design, is a learning experience. I do find it frustrating when the new thingy behaves too different than the simulation. But it's a good reminder that one should not read too much in a simulation result. I speak of course, only of my experience.
I just received some boards that I drew up based on the latest iteration of Colin's shuntreg design, and a simple CLC pre circuit. I'm looking for a few people to try these out and make comments on them and depending on the results, perhaps this GB might get resurrected If you're interested, please contact me and we can work the details out. The more fancy the measurement equipment you have and willingness to get cracking on this, the better. I'm willing to ship worldwide and would like to recoup a tiny bit of my investment so far.
Bigger pic here:
http://picasaweb.google.com/lh/photo/JapdboKbRhH8U6q5Kr66Gw?feat=directlink
If you're interested, please contact me and we can work the details out. The more fancy the measurement equipment you have and willingness to get cracking on this, the better. I'm willing to ship worldwide and would like to recoup a tiny bit of my investment so far.
Bigger pic here:
http://picasaweb.google.com/lh/photo/JapdboKbRhH8U6q5Kr66Gw?feat=directlink
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.