| Onvinyl |
Hi,
attached is a simple but hopefully good (in relation to parts count) PSU that I plan to build. Only positve rail shown.
The quality of the attachment is bad (is there a fast&easy freeware circuit painter out there?), I hope you can read it.
My goal is to have a reasonable quiet and sonically well performing PSU, or the best what can be made out of that few components.
Is there something missing, any mistakes? What do you think of the ccs feeding the zeners (it's in fact cheaper than a current diode).
Rüdiger |
|
|
| Onvinyl |
| uh, oh: the Collector and the emitter of the BC560 are swapped... |
|
|
| AndrewT |
Hi,
why 4*6.8V instead of 27V?
Why the 1k0 before 220uF elna?
Why use a plastic cap at the start of the RCRCRC string.
I think you would get better audio performance if the plastic cap was either at the end of the string or replaced the 47uF output cap. That 1600V 100uF must be as big as a car battery and cost the same as the car it fits in!
Can you modify your 1 (one) script, it is very ambiguous? |
|
|
| Tweeker |
| I think thats 100nF 1600V. |
|
|
| Onvinyl |
| quote: | Originally posted by Tweeker
I think thats 100nF 1600V. |
Yes, indeed.
I have a few hundred 6.8V Zeners, thats why...
I made a quick sketch of this but used only 3x6.8Z, because I found only a 18 VAC Transformer in my bin.
The 1k/220uF String is for filtering zener noise.
It works so far, now I need the negative rail.
How big can the output Cap (now 47uF) get, is there a upper limit?
| quote: | Originally posted by AndrewT
Can you modify your 1 (one) script, it is very ambiguous? |
Not sure what you mean :scratch:
Rüdiger |
|
|
| Tube_Dude |
| quote: | Originally posted by Onvinyl
The 1k/220uF String is for filtering zener noise. |
Hi Rüdiger
For that purpose , the 150 Ohms resistor that come from the gate of the Mos-fet must be connected to the junction point of the 1K Resistor with the 200uF capacitor. This is what Technics call "Virtual Battery Operation".
The resistor that bias the LED must be connected between the base of the BC 560 and ground , that way the transistor could behave as a CCS.
Regards. |
|
|
| Onvinyl |
| quote: | Originally posted by Tube_Dude
Hi Rüdiger
For that purpose , the 150 Ohms resistor that come from the gate of the Mos-fet must be connected to the junction point of the 1K Resistor with the 200uF capacitor. This is what Technics call "Virtual Battery Operation".
The resistor that bias the LED must be connected between the base of the BC 560 and ground , that way the transistor could behave as a CCS.
Regards. |
Ok, I'll fix it. But, since the thing works as it is, how does the transistor behave now?
Regs,
Rüdiger |
|
|
| Tube_Dude |
| quote: | Originally posted by Onvinyl
Ok, I'll fix it. But, since the thing works as it is, how does the transistor behave now?
|
As a CCS with a resistor in parallel ...;)
Putting another way...a very bad CCS. |
|
|
| Onvinyl |
| quote: | Originally posted by Tube_Dude
Putting another way...a very bad CCS. |
Silly me :headbash:
Thanks,
Rüdiger |
|
|
| Onra |
Hi,
i think you will waste a lot of power with this PSU.
For linear operation up to 1A the IRF610 needs a VGS of about 6V. The CCS needs at least 2V to stay in regular operation.
At higher currents the 10R-resistor will consume some voltage too.
Added you need an input voltage of about 30V to achieve a 23V output. That seems to much for me. Do you have to use this FET? A BJT wold do a better job here, because the staple of zeners can only be adjusted to a certain operation point and the output voltage will vary with the current.
For better drawings (and simulation) you may download switchercad III from:
http://www.linear.com/company/software.jsp
It´s a free spice without restrictions and very common.
Onra |
|
|
| Eva |
| Put the N-channel MOSFET in the negative side and use a PNP transistor with its base connected to the negative (regulated) output, the emitter connected to the positive (fixed) output through a suitable zener and resistor, and its collector driving the gate with a bleeder resistor to the negative unregulated input. I've employed that kind of regulator in the past and it works fine, not to talk about its awesome simplicity and the convenient grounding of the tab of the power MOSFET. |
|
|
| Onvinyl |
Hi Eva,
you're giving me a hard time...
Are you talking about something like
borbely link see Page 5 with Q2 and Q4 omitted? Seems to be very cool indeed!
Rüdiger |
|
|
| Eva |
| Something like that, but with a bipolar transistor driving the MOSFET. It yields very good performance for a 7 part discrete design... |
|
|
| Onvinyl |
| quote: | Originally posted by Eva
Something like that, but with a bipolar transistor driving the MOSFET. It yields very good performance for a 7 part discrete design... |
Well, you just mentioned 5. I guess you need a resistor from Drain to Ground and a feedback res between Drain and Base?
You don't happen to have a schematic handy? I'm eager to try it out!
@Onra:
The Mosfet was in my parts bin. Some people reported a sonically well performance, I'll try it out. It's clear to me that a non 3-pin regulator should sound better than say a LM317, else it makes no sense to try other options. So, if you know a good bjt suitable for this job, just tell me...
Another option to try would be the TL431 as a reference.
thanks,
Rüdiger |
|
|
| Eva |
| It's five components plus input and output capacitors :) |
|
|
| Onra |
Hi,
there is a similar basic circuit i found at:
http://www.mikrocontroller.net/atta...2/Schaltung.png
You can use a P-MOSFET instead of the darlington power stage.
Voltage is controlled by the Resistor divider across the output.
To improve regulation (?) and output noise you can try a LED instead the zener.
The negative rail is then mirrored with respect to the polarity of active devices.
I asked for the need of IRF610 because of its high RDSOn and low current capability.
What is your current range the PSU should supply?
Onra |
|
|
| Onvinyl |
| quote: | Originally posted by Eva
...output through a suitable zener and resistor... |
[voice from off:] What is suitable for around 24V?
I do not really get what happens there.
Rüdiger |
|
|
| Onvinyl |
| quote: | Originally posted by Onvinyl
I do not really get what happens there.
Rüdiger |
I made a quick breadboard. With +/- 12Vin all goes fine, with 8Vout.
I used a 6.8V Zener, BC550/560, and IRF(9)610.
With +/- 24Vin one of the bipolars explodes, even with 13.6V Zener. I have no clue why, since I see no exeeding of voltages.
:confused:
Rüdiger |
|
|
| AndrewT |
Hi Onvinyl,
you are calling up both pnp & npn transistors.
Which schematic did you use?
That one in post1 is wrong. |
|
|
| Onvinyl |
Hi,
I use the one eva mentioned. A corrected version of posting #1 works ok, which is no big surprise...
Rüdiger |
|
|
| AndrewT |
Hi Onvinyl,
I cannot find a schematic in any of Eva's posts.
Are you working from a description Eva posted?
Or fig5 in Borbely?
Please post your breadboarded schematic, so we can try to debug. |
|
|
| Onvinyl |
Hi Andrew,
I'm working from eva's description.
Rüdiger
edit: circuit picture |
|
|
| AndrewT |
Hi On,
are you sure you have posted the correct schematic?
It seems all wrong to my inexperienced eye. |
|
|
| Onvinyl |
something is wrong for sure, else it would work :smash: :smash: :smash:
It's that what I understood from the description
Rüdiger |
|
|
| AndrewT |
Hi Eva,
your description has me beat. The Pchannel in the +ve line has reverse control and similarly Nch in -ve line.
Onvynil,
the + &- halves of your schematic do not match.
+ve half collector to gate, -ve half base to gate.
If you have 24V on output then Vds is only 4V, so I think this may be right on the limit of operation of the FETs.
Consider +ve half control. If emitter is connected to 2*6.8V zeners then emitter voltage is -24+13.6=-10.4V. Base is at +24V, therefore Vbe=34.4V :hot:
A suitable zener would be 2 times 24V- 0.6V =47.4V.
But wait for Eva to reply because his description of the common source reg does not appear to work. If the toplogoy was changed to Source follower (swap N & Pch) then the control can be made to work.
I notice you want to use 6v8 Zeners. To get these into a good operating range it is normally recommended that they carry a minimum current of more than 10% of maximum. Using low voltage Zeners forces you to use very high current to meet this requirement. |
|
|
| Eva |
I'm sorry, I think you didn't understood my description. These are two versions of the simple regulator that I proposed. Note that it's not a source follower at all, and that the drain terminal (the tab of TO-220 devices) is conveniently grounded to allow to use a metal case directly as a heatsink.
The version on the left is the simplest one with only 5 components, but its minimum dropout voltage is determined by Vgs. The version on the right features a dropout voltage only limited by Rds-on at the expense of two additional components. The version on the left may require a zener for gate protection and some form of startup current limiting for the small transistor, while the version on the right has an inherent Vgs limitation to 17.6V (12V*1k/680) and the small transistor is protected.
Note that these are just simple circuits aimed at a low parts count and a lot of improvement may be obtained by just adding parts. |
|
|
| Onvinyl |
| quote: | Originally posted by AndrewT
Consider +ve half control. If emitter is connected to 2*6.8V zeners then emitter voltage is -24+13.6=-10.4V. Base is at +24V, therefore Vbe=34.4V :hot:
|
Argh, this strange arrangement makes me dizzy, I looked for Vbc instead of Vbe.
I notice you want to use 6v8 Zeners. To get these into a good operating range it is normally recommended that they carry a minimum current of more than 10% of maximum. Using low voltage Zeners forces you to use very high current to meet this requirement.
[/QUOTE]Thanks for the hint.
Rüdiger |
|
|
| richie00boy |
| Have you considered an LM317T ? That will use less parts and be smaller and easier to wire up, and probably regulate better as well. |
|
|
| Onvinyl |
| quote: | Originally posted by Eva
I'm sorry, I think you didn't understood my description. |
Obviously *loool*
Thanks for clearing,
Rüdiger |
|
|
| Eva |
| quote: | Originally posted by richie00boy
Have you considered an LM317T ? That will use less parts and be smaller and easier to wire up, and probably regulate better as well. |
It all depends on voltage and current requirements. LM317 is quiet and precise, but it has its limitations in comparison with some standard power MOSFET like a IRF540 when the requirements are just a high output current and a very low dropout voltage.
By the way, I've seen a lot of people with their brand-new transformers producing too high voltages for their amplifiers and asking for some simple high-efficiency anti-overvoltage pre-regulator. This circuit may be very useful for them.
|
|
|
| AndrewT |
Hi Eva,
is your ground symbol connected to the correct end of the load?
As drawn this is the +ve rail supply. |
|
|
| Eva |
Yes, I intentionally arranged it this way to show the possibility of grounding the tabs of the TO-220 devices and using the metal case of the project as a heatsink without mica or sil-pad insulators nor plastic washers. This requires independent floating transformer windings for each supply (not center tapped), though.
Having floating transformer secondaries also allows to employ cheaper N-channel devices for both positive and negative rails, altough the tabs of the power devices from the negative rail won't be gounded and would require insulation.
|
|
|
| Onvinyl |
Hi,
that may be a basic question as well; but how to choose a 'low impedance point' where all currents return with such a floating design?
Is it wise to have different grounds?
Rüdiger |
|
|
| Eva |
Only the transformer winding is floating in that design and this is fine as long as the same winding is not employed for anything else. The output of the regulator is not floating, it's grounded... In fact it's ground! :D:D:D
This ground is the one you should use for the rest of the circuit.
|
|
|
|
