Salas: For Reflektor D Mini, I would even suggest going smaller using SMD parts where possible. Good SMD resistors exist, the transistors seem like they are easy to do in SMD, and I have found SMD film caps (Panasonic PPS) to perform very well also.
Heck, maybe an SMD layout could even perform a bit better due to shorter trace lengths and more tightly controlled inductance?
Heck, maybe an SMD layout could even perform a bit better due to shorter trace lengths and more tightly controlled inductance?
The MOSFETS need be the same as in Ref-D regular to can sink mount and they set the general dimensions frame. Surely there is tighter layout with SMD but it will also put many people off who are unfamiliar & unequipped for surface assembly when its not going to give it three pin self standing form factor anyway. Because it will not change the high dissipation big parts so to make it self contained on a super mini PCB. There are Vref and CCS trials to set it also that will become a pain on SMD for some. Just to confirm hFe it takes a special measurement adapter for instance. So I will skip that path.
I like the idea of the Reflektor-D Mini. I wonder if it might be nice to have a separate little PCB to hold the AC input stuff, connectors, rectifiers, big filter cap (maybe even provisions for those 4pole Mundorf Rolls-Royce guys?).
An advantage of SMD resistors is the cost and availability of high precision 0.1% 25ppm parts... likewise JFets getting harder and harder ... I have some 2SK117 cheap from China (thought I was buying USA on Amazon ...), chance they are real? 2SK880s are easy to get.
Worthwhile?
Worthwhile?
The naturally closest to board surface and less capable in mW SMD mirror inside or around the heatsink area on a small board will be more susceptible to drift and you can't thermally bond its transistor pairs like with the face to face TO-92s we are using now, but I guess it will basically work well enough.
P.S. Slim chance they are real, post a close up picture. We can usually tell from a macro photo.
Both Ref-D and Ref-D Mini have provisions for K880 pads also already. In the Mini they are underside.
P.S. Slim chance they are real, post a close up picture. We can usually tell from a macro photo.
Both Ref-D and Ref-D Mini have provisions for K880 pads also already. In the Mini they are underside.
OK, I was thinking that the transistors could be thermally bonded with a small stick on heatsink like used for memory chips etc (e.g. 12x12mm very inexpensive on ebay etc). Those transistors are easy to source so if TO-92 is better, then no problem.
I'm thinking that if I need to do the 2SK880 SMD anyways, then the resistors aren't too much of a problem. It seems easier/cheaper to get tight spec SMD resistors than through the hole. Any problem with these?
Yeah, when I received the package, I tossed it somewhere in my workshop in disgust. Not sure what I was thinking but thought that purchasing on Amazon.com would be reliable... I will track it down and get a macro photo...
Buying JFets is like buying drugs...
If you use -for example- 9VAC-graetz-22000uF-r-C at 0.5A load, the diode peek spike would be even 5A! By the way C1 peek charge current is about 4-4.5A!!!
I use there 16A diodes (MBR1635) and lower (4.7-10mF) first capacitor.
I use there 16A diodes (MBR1635) and lower (4.7-10mF) first capacitor.
Hi all, would there be any benefit to bypassing CM cap with a 0.1uF polyprop or polyester cap? I can see that that a poly cap would be able to fit on the underside of the board. Would this provide better noise decoupling?
No it won't do anything good to noise. It may be detrimental even if it will give a peak by resonating with the main reservoir's parasitic inductance at some very high frequency. Use soft recovery diodes instead.
Thanks Salas!
A couple more questions if you would:
1) Would these be even better diodes?C3D04060A Cree, Inc. | Mouser
2) For the value of CM, there is a range of 10000-22000uF given. Is higher in that range better (e.g. 22000 vs. 10000). I have a 9VAC 3.33 amp transformer supply.
A couple more questions if you would:
1) Would these be even better diodes?C3D04060A Cree, Inc. | Mouser
2) For the value of CM, there is a range of 10000-22000uF given. Is higher in that range better (e.g. 22000 vs. 10000). I have a 9VAC 3.33 amp transformer supply.
Thanks again Salas!
For #2, what if I am pulling upwards of ~500mA? Would 22,000uF be better? Can my transformer handle the in-rush current?
For output, I'm looking to run somewhere between 6-7V and at ~500mA current draw (R1 = 1ohm). I'll have the mosfets insulated and sinked to large aluminum case (no risk of over-heating).
I was under the impression 500ma was above the capability of the Reflektor-D so i used a BiB for my DAC, also running 7v at approx 500ma.
Question is, which is the better board for the job, the BiB or R-D?
Question is, which is the better board for the job, the BiB or R-D?