Ripple for PSU Solid State power amp

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
What is considered low ripple for Solid State power amp?
I am about to design a PSU for a SS amp (low watt SE Mosfet amp). It is non regulated and just filtered using Resistors, Chokes and Capacitors. The design I am thiking on simulates to a ripple at approx. 15 mV. Is that considered OK for a SS power amp?


First I was looking for a 10H / 4A choke but found out that it will be quite big so now I am looking for 0.5 H / 4A instead which seems more resonable but of course....the ripple will be higher.
 
Before I cut you the right length it might be useful to have some more details...
 

Attachments

  • DSCF6081.JPG
    DSCF6081.JPG
    286.6 KB · Views: 429
Using the PSU designer I have made a R-C-L-C filter (30 VAC in and 19 VDC / 1A out). The amp will draw approx. 1A constant (kl A). It is the AmpCamp amp (PassLab). With my chosen components I have approx. 15 mV ripple on the 19 VDC at 1 A current. When I made a PSU for my 300B tube amp I designed for less than 1 mV ripple at 535 VDC. So for me 15 mV ripple is a lot......but maybe it is ok for Solid State Apms?
What is your design goal in general for SS amps......if it is possible to say something in general?
 
Yes, I have simulated it and the ripple drops to about 1.3 mV......but so does the output voltage......drops to approx. 15 V DC so need 35 VAC on the secondary. But that is the penalty.....I guess......to get lower ripple. It is Ok.....have not purchased any components yet......
 
Disabled Account
Joined 2002
Hello,
Voltage depends a lot of DCR of the choke. My friend nagini262 uses the LL2733 from Lundahl using one coil in the positive wire and the other one in the ground wire. ASK him what AC voltage he used. I have seen and heard the amp but not looked at the transformer a few weeks ago in Vietnam.
It was a big improvement switching to choke input. The R in LCRC does not have to be big most of the filtering is done by the choke. Something like 0,22 0,33 ohm will be ok.
The choke however should be big. LL2733 is 400 mH.
Chinese newyear just started so Vietnam people will be busy....
Greetings, Eduard
 
Using the PSU designer I have made a R-C-L-C filter (30 VAC in and 19 VDC / 1A out). The amp will draw approx. 1A constant (kl A). It is the AmpCamp amp (PassLab). With my chosen components I have approx. 15 mV ripple on the 19 VDC at 1 A current. When I made a PSU for my 300B tube amp I designed for less than 1 mV ripple at 535 VDC. So for me 15 mV ripple is a lot......but maybe it is ok for Solid State Apms?
What is your design goal in general for SS amps......if it is possible to say something in general?

Ouch!

I suppose I am not Hi-Fi in as much as when I see people seeking extremely low ripple then I think they might be chasing Rainbows.

Being a knuckle dragger I would expect a Solid State amplifier to achieve 'extreme' figures for PSSR, at least at low frequencies.

The input stage is designed to do so and then feedback trounces the rest... until it runs out of steam. However you have one of these...

Amp Camp Amp #1

I wondered how it would perform in terms of PSSR so I went and stuffed it up Spice, Yeah, Yeah, Yeah, to try to see what parts of it would be susceptible to power supply ripple and...

OK. In terms of DC operating points I got a bit lucky with the components that are already available in LTSpice.

First File and Picture.

There is more to this than meets the eye but fools rush in and the first attempt is to remove the low level stuff from the main rail and give it a separate supply. That'll fix things.

Second Picture.

I seem to have Trashed the PSSR. Fine. I'll isolate the input JFET.

Third Picture.

...And it is still Trashed. OK, isolate the power side drive.

Fourth Picture.

That's better but not as good as it was to begin with so put things back to the original and.. what about the difference in those low value resistors? Change the 340mR to 240mR...

Fifth Picture.

Trashed Again.

Of course having been through that one then, being me, the first thing I guess is parasitic capacitance. Second guess says that is a silly one. Obviously RC is going to inject supply noise into things and then...

Mr Self mentions 'valve type things' in his introduction to the circuit. That might be a 'red herring' but I guess that Mr Self knows a bit about load lines or whatever they are called.

http://www.linearsystems.com/lsdata...APACITANCE_MONOLITHIC_DUAL_N-CHANNEL_JFET.pdf

https://www.vishay.com/docs/91210/91210.pdf

Device gm varies with VDS, the power supply voltage. I think Mr Self has been clever here without fully explaining what is going on.

Of course if you want to chase extremely low values of supply voltage ripple then I am not going to say the chase might be misplaced. Lower is better. However it might be worth investing some time in 'component selection' in an effort to improve the PSSR of the base circuit.

Of itself that may be chasing Rainbows but this is Hi-Fi. Perhaps Mr Pass can comment.
 

Attachments

  • pass_01.asc
    3.8 KB · Views: 36
  • Screenshot at 2018-02-18 12:20:30.png
    Screenshot at 2018-02-18 12:20:30.png
    30.1 KB · Views: 356
  • Screenshot at 2018-02-18 12:24:51.png
    Screenshot at 2018-02-18 12:24:51.png
    30 KB · Views: 345
  • Screenshot at 2018-02-18 12:30:35.png
    Screenshot at 2018-02-18 12:30:35.png
    30.4 KB · Views: 335
  • Screenshot at 2018-02-18 12:33:23.png
    Screenshot at 2018-02-18 12:33:23.png
    31.1 KB · Views: 339
  • Screenshot at 2018-02-18 12:37:50.png
    Screenshot at 2018-02-18 12:37:50.png
    30.6 KB · Views: 78
  • Screenshot at 2018-02-18 12:52:39.png
    Screenshot at 2018-02-18 12:52:39.png
    280.2 KB · Views: 77
PSSR is that Power Supply Rejection Ratio?
The pictures you show is how much the amp itself reduces the ripple from the PSU?
So around 50 dB?
From the tube world I was just surprised how much a serial choke could do to reduce ripple. In commercial amps......it is just bridge and caps and most of these seems to work fine.
 
Disabled Account
Joined 2002
Hello,
Most people will be more impressed if you tell them you use a BIG 1000VA transformer and some beer can capacitors than if you tell them you use a 300VA , 2 chokes ( preferably choke input) and a few 22000µF in parallel.
Meper will kind of use the same power supply my friend in Vietnam is using. 300VA toriodal, 2 secundairy windings, one for left the other for right channel.
LL2733 Lundahl in what Lundahl calls common mode connection. CHOKE INPUT so the first cap is connected to the diodes by two coils each 1,7 ohm DCR. To much, much to much DCR the experts say. Did they try? No!
To be honest i thought my friend did put the two coils in parallel to reduce DCR. Only after he showed me the interior i realized this was not true.
My friend also told me that after switching to choke input the differences between caps in the power supply became less obvious. He explained that with choke input it is the choke that has to work hard not the cap. The cap will probably have a longer life. The rectifier and the power transformer also can kind of relax.
Greetings, Eduard
 
PSSR is that Power Supply Rejection Ratio?
The pictures you show is how much the amp itself reduces the ripple from the PSU?
So around 50 dB?
From the tube world I was just surprised how much a serial choke could do to reduce ripple. In commercial amps......it is just bridge and caps and most of these seems to work fine.

Yes. PSSR, Power Supply Rejection Ratio.

The pictures are AC plots of VOUT with VDC, the supply voltage being subjected to a 1V AC signal so it is an indication of how much the circuit rejects that ripple.

I guess Chokes are good for High Voltage stuff. The problem you get at lower voltage might be or is indicative of possible energy storage versus voltage and current for a given component volume.

I could be making this up but it is known that for electrolytic capacitors in terms of size the CV product is more or less constant but energy goes as [0.5]CV^2. You need bigger capacitors at lower voltages to store the same energy.

There may be a similar analogy in respect of inductors but without looking harder I cannot claim that would be the case.

If things run that way then the difference would be 500^2/20^2 or 625 in volume terms.

Again Solid State, low voltage, largely relies on making the low level stuff immune to ripple and having a huge open loop gain to sort out the power stuff unless....

You are Mr Self, have more insight and put some thought into things.
 
Disabled Account
Joined 2002
Hello,
My Vietnamese friend just uses 1 33000µF on each channel at 19 volts DC so probably not to much energy stored. The bias is 1,4 A in his ACA amp.
Some people say the energy is stored in the choke and it will act as an electronic flywheel. Dont ask me to explain.
Dont ask me why if chokes can be used for solid state designs why almost nobody is using them. I cannot explain why so called coffee fanatics go to the big chains serving .....
greetings, Eduard
 
Seeking tiny supply rail ripple is a standard newbie mistake. Two reasons why it is a mistake:
1. any decent amp design is likely to reject supply rail ripple - this is PSRR
2. once the rail ripple is low enough, hum then comes from grounding mistakes

Chokes were used in old valve amps because big caps were not available. Now people use big caps instead, mostly.
 
Assume your design requires a maximum of 100mV ripple, whether it's tube or solid state. (As Morbid points out above, that's a questionable assumption because different amplifier topologies have different PSRRs). Vrip = (I*t)/C, where Vrip is in volts, I is current in amperes, t is time in seconds, and C is capacitance in farads. Assume the tube amp is drawing 100mA at 500 volts; the solid state amp draws 1A at 50 volts, i.e., both 50 watts in. As you can see, everything else being equal, the tube amp requires 1/10 C of the solid state amp to achieve the design ripple. But a 100,000uF/50 volt capacitor is less expensive than a 10,000uF/500 volt capacitor. The Vrip calculation is more complicated with an inductor, but "back of the envelope," let's assume it's the same 10:1 ratio. A 2H choke rated at 100mA for the tube amp is less expensive than a 200mH/1A choke for the solid state amp. There's plenty more going on, but that's the simplified explanation. And don't forget, in the heyday of tubes, 40 or 50 years ago, a 10,000uF/500V capacitor was science fiction.
 
When playing with PSU designer 2 it looks like a 400 mH / 2A choke will do a good job in a low voltage PSU for the AmpCAmp. What I also think it will do is to act like a good mains filter to reject transients and noise on the mains supply?
So why not use it. In a SE design there is one supply only. Not the symmetry as in a PP design. I would guess the symmetry in a PP design with +- supply will reject ripple better than in a SE design with only a + supply? .....I am only trying to use logic here....and that can be dangerous......as Niels Bohr said "no no you are not thinking you are just being logical". Some argue that it is difficult to change current fast if a choke is in the design but in a class A amp the current is quite stable and to deliver fast energy there is still a big capacitor (e.g. 47 mF) directly to the amp (the last capacitor in the filter). So should be safe also here. For bigger class A amps where you need 10A constant current a choke could be a problem. It will be very big.
 
Nothing wrong with that. Large value, high voltage caps are much more readily available these days than back when the text books were written. Hence it made sense at the time to use chokes more.

For SE amps I've found a resistive Pi filter arrangement is very effective at getting rid of 100/120Hz hum
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.