Hey guys...
I'm at it again!!!
Here's the rough draft...
2 x 1KVA Antek AN-10440 Trafos 40VAC @ 25A in parallel + 12V 2A + 18V 2A.
Cap Bank will be 10 x 15000uF 63V Nichicon KMH's ( 5 per Rail - Bought 24 of them HERE )
This supply was designed around 2 x 1KVA Antek 10440 Toroids with 40V Secondaries tied in parallel for maximum current output...
Each Transformer will feed each board separately. As you have noticed the positive and negative side is identical besides the polarities on caps and bridges...
C1, C2, C3, C4, C5, C6, C7, C8 ( typo for C9 ) -> 0.1uF 400V Vishay Film Caps
C11, C12, C13, C14, C15 -> 15000uF 63V Nichicon KMH
C16 & C17 -> 1000uF 100V Nichicon FW
C20, C21, C22 -> 4700uF 35V Nichicon FW
C18 & C23 -> 150uF 63V Film EPCOS
C19 -> 15uF 100V Film EPCOS
Main rectifiers ( D1 ) are 2X Crydom M5060SB600
Secondary Bridges are standard 10A off the shelf units.
Driver Board will supplied separately from the output board at a higher voltage which can be selectable from SW1 and SW2 at 80VDC or 88VDC.
Also thinking C11 through C15 instead of 1 cap run 2 caps in series, but need to figure out the proper resistance to keep an equal voltage across both caps.
INPUT PLEASE !
I'm at it again!!!
Here's the rough draft...
An externally hosted image should be here but it was not working when we last tested it.
2 x 1KVA Antek AN-10440 Trafos 40VAC @ 25A in parallel + 12V 2A + 18V 2A.
Cap Bank will be 10 x 15000uF 63V Nichicon KMH's ( 5 per Rail - Bought 24 of them HERE )
This supply was designed around 2 x 1KVA Antek 10440 Toroids with 40V Secondaries tied in parallel for maximum current output...
Each Transformer will feed each board separately. As you have noticed the positive and negative side is identical besides the polarities on caps and bridges...
C1, C2, C3, C4, C5, C6, C7, C8 ( typo for C9 ) -> 0.1uF 400V Vishay Film Caps
C11, C12, C13, C14, C15 -> 15000uF 63V Nichicon KMH
C16 & C17 -> 1000uF 100V Nichicon FW
C20, C21, C22 -> 4700uF 35V Nichicon FW
C18 & C23 -> 150uF 63V Film EPCOS
C19 -> 15uF 100V Film EPCOS
Main rectifiers ( D1 ) are 2X Crydom M5060SB600
Secondary Bridges are standard 10A off the shelf units.
Driver Board will supplied separately from the output board at a higher voltage which can be selectable from SW1 and SW2 at 80VDC or 88VDC.
Also thinking C11 through C15 instead of 1 cap run 2 caps in series, but need to figure out the proper resistance to keep an equal voltage across both caps.
INPUT PLEASE !
Last edited:
Here's a better shot of the schematic...
Better ??
An externally hosted image should be here but it was not working when we last tested it.
Better ??
Last edited:
Well, it's visible now..
1. Why all the caps?
2. Why all the film caps?
3. Why four transformers? If you don't have enough VA capacity in one, I can understand that. Then,
4. Why four FWBs? You only need one for the main +/-63V rails, not two.
5. Why two extra transformers for the +/-80V supplies? How much current is required? Why not use a DC-DC converter to generate these, especially if they are low current?
6. You claim you need a "stiff" supply. Can you put a number on that? What impedance is actually required? If it's for an amplifier driving 4 ohm speakers, I'm guessing at most 10% of that, or under 0.4 ohms. Is that reasonable?
Tim
1. Why all the caps?
2. Why all the film caps?
3. Why four transformers? If you don't have enough VA capacity in one, I can understand that. Then,
4. Why four FWBs? You only need one for the main +/-63V rails, not two.
5. Why two extra transformers for the +/-80V supplies? How much current is required? Why not use a DC-DC converter to generate these, especially if they are low current?
6. You claim you need a "stiff" supply. Can you put a number on that? What impedance is actually required? If it's for an amplifier driving 4 ohm speakers, I'm guessing at most 10% of that, or under 0.4 ohms. Is that reasonable?
Tim
For a simulation there is a freeware that is designed for power supplies, I can't recall the name, but it should be easy enough to find.
The problem with doing a simulation will be getting all the paramaters for each component "correct". Raw capacitance, for example, will not tell you how one cap will perform over another - so in that case all the "extra" caps you have there will do nothing more than simulate like a single larger cap.
Also, the power supply will only show "issues" with a real world load - which is often a dynamic load. That is beyond the capabilities of LTSpice, afaik (well there may be some workarounds, but that is really complex).
If you want an exercise in simulation, start with simple, and work up. Also LTSpice afaik will not simulate a real-world transformer...
And a "stiff supply" is simply a low output Z supply WRT what it is loaded with. The same supply with a bigger load isn't quite so 'stiff' anymore??
_-_-bear
PS. what software are you using to draw the schematic btw??
The problem with doing a simulation will be getting all the paramaters for each component "correct". Raw capacitance, for example, will not tell you how one cap will perform over another - so in that case all the "extra" caps you have there will do nothing more than simulate like a single larger cap.
Also, the power supply will only show "issues" with a real world load - which is often a dynamic load. That is beyond the capabilities of LTSpice, afaik (well there may be some workarounds, but that is really complex).
If you want an exercise in simulation, start with simple, and work up. Also LTSpice afaik will not simulate a real-world transformer...
And a "stiff supply" is simply a low output Z supply WRT what it is loaded with. The same supply with a bigger load isn't quite so 'stiff' anymore??
_-_-bear
PS. what software are you using to draw the schematic btw??
1. For the most part they are there to filter out noise, and the other part is to store and dampen the large transient currents drawn by the amp.
2. Film caps are there to decouple the large electrolitics
See this THREAD for more details...
3&4. There are actually only 2 transformers but the program that i'm using only has a single secondary, hence why i used multiple. Also would like to have 25A of current on tap on each side of the "median". Since the 2 transformers are not center tapped its easier to build 2 mirrored supplies to deliver the needed voltage and current.
Also the start up will place large amounts of stress on the bridge rectifiers when the caps charge and maybe a peak of 200-300A flowing through the rectifiers upon the charge of the caps. By using 2 rectifiers the current draw will be theoretically half of what i would be drawing across 2.
5. There is no extra transformers, they are meerly a second set of secondaries of each transformer. They will be used as the supply the LME49810's and at most 1A will be needed there.
6. The amplifier itself uses 8 Complimentary pairs of NJL4302DG & NJL4281DG.
Its basicly a current dumping amplifier and hopeing that it could drive loads as difficult as 0.5Ohms ( not that i will ever use it, but why not overbuild things to be able to withstand abuse from reactive and capacitive loads )
Im using PCB Expresses - Xpress Scematic.
I like it as its very simple to use and can draw up a schematic very fast.
smaller film caps do not "decouple" the larger filter caps, they "bypass" them. The idea is that at HF the proper cap has a lower reactance, since the self inductance is low at RF and then VHF and up (depending on what you pick...) beyond a certain point there is no benefit to more and more little caps... at least in theory. I'd pay more attention to pragmatic aspects like layout and wiring, grounds especially...
_-_-bear
_-_-bear
OK.
Do you know how much capacitance is needed?
This thread seems to disagree:
A cursory reading of a few factual posts suggests I am right to be suspicious.
Ah, OK.
Then why two FWB?
The FWB are in series, not parallel. The charging current is identical.
FWB are quite robust, I've never seen one fail from charging spikes. In fact, I've never even seen one fail from blowing a breaker. And I've blown big breakers through them before.
Example:
I was testing my 10kW induction heater with a single 35A, 800V FWB. This was obviously insufficient at the power level (7kW by my estimate), causing it to overheat. One of the diodes failed short circuit; the other three survived, despite blowing the 50A, 250V circuit breaker. The other diodes must've been near 175C or above, and they survived the surge current, which must've been several thousand amperes.
It is remarkably difficult to destroy a GBPC3508 outside of thermal overload.
Ah, then why not connect them all in series and use only two FWB:
+End
3
3 12/18V
3
+Tap
3
3 45V
3
CT
3
3 45V
3
-Tap
3
3 12/18V
3
-End
Use a big FWB between +/-Tap, and a smaller one between +/-End.
Ok, so let's figure the supply should be less than 0.05 ohms.
This doesn't even take into account the fact that PSRR is probably fairly high, or that you only need maybe 20V into that 0.5 ohm load to complete destroy the amp chip. Both of these factors would allow a much higher impedance to work effectively, so this is a conservative estimate.
Do you have the datasheets for the capacitors you're using? What is rated ESR, how does it compare to this figure?
What is the RC time constant of the capacitors with 0.5 ohm? 0.05 ohm? What about the film capacitors, what's their time constant against 0.5, 0.05 and the electrolytic ESR? How does this time constant compare to the stray inductance of the circuit? (Guess 20nH per electrolytic, 10nH per film, plus stray wiring inductance, which may be around 5nH/cm distance between capacitors.) Will the film capacitors ring in the circuit (worsening RFI) or will they be well-damped?
Tim
As I allude to above, going by the input impedance of the circuit could allow for "interesting" numbers.
For instance, a power amplifier driving 8 ohms with 80dB PSRR has an effective impedance rail-to-output of 80kohms. In other words, the output transistors' collectors have a quite high impedance. This is typical of three-terminal regulators, of course, to the point where LM317 is used quite often as a current source!
If it weren't for power supply ripple, most applications could get away with very little capacitance indeed.
Dynamic loads, of course, have much the same effect as ripple, and can have a more dramatic effect as well: a subwoofer amp driving 20Hz can suck the voltage right out of capacitors which are otherwise sufficient for 120Hz ripple. Under this condition, several line cycles pass for one cycle of load, so line and transformer regulation is the controlling factor!
Tim
( 5 x 15000uF per rail ) x 2 rails
75000uf x 2 = 150mF
Originally i was thinking 8 x 15000uF per rail ( 120mF per rail ) for a total of 240mF which i might still persue...
The transformers are 2 x ANTEK AN-10440
1KVA - 40VAC @ 25A ( secondaries in parallel ), 12VAC & 18VAC @ 2 A
Last edited:
***:*:*** FM ACOUSTICS LTD. ***:*:*** company***:*:***
"In audio electronics, especially with power amplifiers, the importance of a true low-impedance, wide-frequency power supply is generally underrated. Here every mOhm (one thousand's of one Ohm) counts. The connection between the large electrolytic capacitors (which also acts as the central ground for the amplifier's circuitry) is made with a massive bar of specially coated high-purity copper. Instead of standard flat washers, specially corrugated washers are used: these dig deep into the copper and guarantee an even lower contact resistance.
A component of major importance are electrolytic capacitors. It is widely believed that it is the capacitance value that describes a capacitors' capability. This is not correct.
More important than capacitance per se are the loss factor, the Electric Series Resistance (ESR), the maximum ripple current that the capacitor can handle and other less known details, e.g. how the internal aluminium foil is connected to the terminals."
"In audio electronics, especially with power amplifiers, the importance of a true low-impedance, wide-frequency power supply is generally underrated. Here every mOhm (one thousand's of one Ohm) counts. The connection between the large electrolytic capacitors (which also acts as the central ground for the amplifier's circuitry) is made with a massive bar of specially coated high-purity copper. Instead of standard flat washers, specially corrugated washers are used: these dig deep into the copper and guarantee an even lower contact resistance.
A component of major importance are electrolytic capacitors. It is widely believed that it is the capacitance value that describes a capacitors' capability. This is not correct.
More important than capacitance per se are the loss factor, the Electric Series Resistance (ESR), the maximum ripple current that the capacitor can handle and other less known details, e.g. how the internal aluminium foil is connected to the terminals."
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.