hey guys,
will the regulator be able to power a squeezebox touch which has a 5v 3amp requirement? what do i need to replace on the scheme to output 3amp?thanks
V1.0 has done 3A at 13V for a T-Amp. The dissipation at 5V is almost 1/3rd than that, so I guess its possible. Its only Rset that needs go low value and you need substantial sinking. Since that post many used it at 5V for digital. DACEND, DVBPROJECT's 1541 REF, Merlin El Mago on Buffallo remote sensed, etc.
I don't think GP is necessary in analog pcbs, and, as pointed out by Salas, a solid GP is like adding a big cap. If you want it, use a grid or net, but better without IMO.
Verify hole size: if you vary hole size for each components, the pcb price goes up. Consider that through hole plaiting reduces the actual size by 0.1 -0.15 mm. Small bjt/fet pads are still too small for me. Easy to destroy them if you have to desolder a component.
Move 33 pf cap to the left (on top of 1k resistor) in order to reduce track lenght. Its pitch can be 5 mm or smaller if you use a good ceramic component.
Thanks for all the feedback.
I designed a new PCB for the LED-reference BJT-cascode and tried to incorporate all the feedback.
The small transistor pads have been increased (they now only have 0.3mm space between them - I don't think that it'll be a good idea to further reduce it - any comments, maximo?)
Due to popular request I also added TO220 holes to the PCB layout.
The new design looks cleaner IMHO.
Since the new design does not require IDSS sorting, can go for higher input voltages and performance is the same as the older version, this new design will become the standard design I will be working on.
@Salas:
1.) Somewhere in the thread it was mentioned that LED-reference could be used in the remote sensing and adjustment part, too. Does that improve overall quality? (There seems to be enough space left in the new design so that I could even add in that option with a "daughterboard".)
2.) I understood correctly that the performance of the LED-reference version is at least the same as the one with 2SK170/2N2559, did I?
3.) At the moment the circiut contains BC560B and BC560C. Can't one model be used for all without compromising performance?
I designed a new PCB for the LED-reference BJT-cascode and tried to incorporate all the feedback.
The small transistor pads have been increased (they now only have 0.3mm space between them - I don't think that it'll be a good idea to further reduce it - any comments, maximo?)
Due to popular request I also added TO220 holes to the PCB layout.
The new design looks cleaner IMHO.
Since the new design does not require IDSS sorting, can go for higher input voltages and performance is the same as the older version, this new design will become the standard design I will be working on.
@Salas:
1.) Somewhere in the thread it was mentioned that LED-reference could be used in the remote sensing and adjustment part, too. Does that improve overall quality? (There seems to be enough space left in the new design so that I could even add in that option with a "daughterboard".)
2.) I understood correctly that the performance of the LED-reference version is at least the same as the one with 2SK170/2N2559, did I?
3.) At the moment the circiut contains BC560B and BC560C. Can't one model be used for all without compromising performance?
Attachments
If I remeber right, Salas advised on using a Silver Mica 33p and I took the footprint from the datasheet. I don't know about performance of different capacitors there, but I'm eager to learn and changing the pitch is not a big deal...
3.Yes, use 550B-560B for those CCS TO-92s, so some will not stick out for hfe too much. Had no problems with mixing B or C, or all being C, but the B grade is more conservative around 300 hfe.
2. Yes, equivalent. BJT or FET cascode same PSRR.
1. LEDs are for fixed reference on a daughter board if you fancy. In that case you should have to jumper the source to ground of the trimmer controlled JFET, because the LEDS are going to need stronger current to be happy. They are stiff, will not come up thermally anymore than +/-0.4V, the resistor will slide with the small Jfet current coming to thermal final -but stay then-. Also less impedance. A bit different subjectively than a resistor Norton ref. Semis are active things. Much of a hassle for 24V, not recommended there since you will also lose noise performance gradually with number, but nice in the 5V.
P.S. Don't know if TO-220 will work equally well given the extra length to gate stoppers in your last iteration. But they can always get soldered directly to legs and just leave their other leg lengthier to is pad.
2. Yes, equivalent. BJT or FET cascode same PSRR.
1. LEDs are for fixed reference on a daughter board if you fancy. In that case you should have to jumper the source to ground of the trimmer controlled JFET, because the LEDS are going to need stronger current to be happy. They are stiff, will not come up thermally anymore than +/-0.4V, the resistor will slide with the small Jfet current coming to thermal final -but stay then-. Also less impedance. A bit different subjectively than a resistor Norton ref. Semis are active things. Much of a hassle for 24V, not recommended there since you will also lose noise performance gradually with number, but nice in the 5V.
P.S. Don't know if TO-220 will work equally well given the extra length to gate stoppers in your last iteration. But they can always get soldered directly to legs and just leave their other leg lengthier to is pad.
Another reworked version:
- made some cosmetic changes
- made the TO-92-pads oval
Keep the good comments coming, I feel that version 1.0 is coming pretty close...
About the TO-220 pads:
Personally I don't see why someone would be using the TO-220 if the performance of both types is identical and the TO-247 have way better heat dissipation - am I missing something there?
Additionally you could always bend the pins on the TO-220 to fit into the TO-247 holes.
- made some cosmetic changes
- made the TO-92-pads oval
Keep the good comments coming, I feel that version 1.0 is coming pretty close...
About the TO-220 pads:
Personally I don't see why someone would be using the TO-220 if the performance of both types is identical and the TO-247 have way better heat dissipation - am I missing something there?
Additionally you could always bend the pins on the TO-220 to fit into the TO-247 holes.
Attachments
I agree. Delete TO220 pads. Bending legs apart isn't that easy
but can be done sometime.....
Last week a friend of mine left me some speaker cables to try : Super Flatline MkII Speaker Cable from Nordost
I have listening to the v12 on the benz riaa and with these cables I can really evaluate it´s trebble performance.
Even with a cheap 10u polyprop on vref, detail is so much better than v1.
I missed some of the v1 attack but with these cables I got the sweeter, even, detailed, full on highs ever.
A pitty these cables do not match well with the meridian amp so I loose bass dramatically.
Hi THKL
I can not identify the sense wire connections on your board.
Also I can not see what is the square part with 12345 lettering.
Why the 47u on input ?
Splendid pcb work BTW.... I will use some of your ideas in my next build
I have listening to the v12 on the benz riaa and with these cables I can really evaluate it´s trebble performance.
Even with a cheap 10u polyprop on vref, detail is so much better than v1.
I missed some of the v1 attack but with these cables I got the sweeter, even, detailed, full on highs ever.
A pitty these cables do not match well with the meridian amp so I loose bass dramatically.
Hi THKL
I can not identify the sense wire connections on your board.
Also I can not see what is the square part with 12345 lettering.
Why the 47u on input ?
Splendid pcb work BTW.... I will use some of your ideas in my next build
That square thingie is the sense-wire
Actually it's an SMA-socket.
Because the sense-wire should ideally be a coaxial rg59, an SMA-connector does the trick nicely. You don't have to fiddle with soldering the shield and always get good looking and securely shielded connections...
The 47uF bypassed with 0.01uF is for decoupling if longer cables from the transformer/rectifier/cap-reservoir are used. It has just been added, because there was unused space on the board and it might come handy under some circumstances.
I will sleeve the rg59 and the power connection together and have another small PCB on the "receiver" side, where rg59 and power are joined together and fed into a 2,54 or 5,08mm two pole connector for the power connection to the consumer.
A good power-supply does not only have to be highest quality, but look nicely, too.
You talked about a cheap PolyPropylene on the Vref - what would you recommend there? I mean the PCB was designed for (upto) 10uF MKP10 PP
Thanks for the compliment on the layout, but I couldn't have done it without the feedback here. I would have had many tracks running too long and wouldn't have thought how close the stopper resistors really have to go. I wouldn't have thought about pad sizes either - I would have used what the pcb cad software had as standard sizes...
Out of curiosity what's your new project?
Actually it's an SMA-socket.
Because the sense-wire should ideally be a coaxial rg59, an SMA-connector does the trick nicely. You don't have to fiddle with soldering the shield and always get good looking and securely shielded connections...
The 47uF bypassed with 0.01uF is for decoupling if longer cables from the transformer/rectifier/cap-reservoir are used. It has just been added, because there was unused space on the board and it might come handy under some circumstances.
I will sleeve the rg59 and the power connection together and have another small PCB on the "receiver" side, where rg59 and power are joined together and fed into a 2,54 or 5,08mm two pole connector for the power connection to the consumer.
A good power-supply does not only have to be highest quality, but look nicely, too.
You talked about a cheap PolyPropylene on the Vref - what would you recommend there? I mean the PCB was designed for (upto) 10uF MKP10 PP
Thanks for the compliment on the layout, but I couldn't have done it without the feedback here. I would have had many tracks running too long and wouldn't have thought how close the stopper resistors really have to go. I wouldn't have thought about pad sizes either - I would have used what the pcb cad software had as standard sizes...
Out of curiosity what's your new project?
Last edited:
SMA REV BUE-180 SMA-Steckverbinder - reichelt elektronik - Der Techniksortimenter - OnlineShop für Elektronik, Netbooks, PC-Komponenten, Kabel, Bauteile, Software & Bücher - ISO 9001:2000 Zertifiziert
That is the sma-connector I was thnking about - it is available in 90degrees turned also.
I have heard of people using Mundorf Supreme Silver/Gold or Jensen Silver Oil in some dedicated equipment, but is the difference to Wima MKP10 that extreme?
I mean the price difference is tremendous and not even the Wimas are cheap...
We are not talking about cross-overs here, but about a power-supply...
That is the sma-connector I was thnking about - it is available in 90degrees turned also.
I have heard of people using Mundorf Supreme Silver/Gold or Jensen Silver Oil in some dedicated equipment, but is the difference to Wima MKP10 that extreme?
I mean the price difference is tremendous and not even the Wimas are cheap...
We are not talking about cross-overs here, but about a power-supply...
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.