Hi Roberto,
My main point is that in class B amplifiers (and also class D) the current requirements are signal dependent (unlike class A).
With transients a class B power supply must have a very low inner resistance in order not to "compress" the transient signal, and IMO the transformer is the most important here; banks of high value capacitors will not help if the transformer is not able to supply the repetitive high transient current without compression; that will lead to what you call slow response.
It is clear that not all SMPS's behave the same, and the fact that you recognize the importance of this phenomenon, and succesfully implement it in your SMPS, is proof of your engineering capability.
At the same time however I believe that equally good behaviour can be reached by well designed low inner resistance linear supplies; there are, like you say, no mysteries.
I completely agree, big Banck capacity is consumed in the first 50% of a single shot drum. haha! precisely because of this , need a transformer high flow dendity, but this means weight and space and does not allow the creation of new compact design amp. This was the input that did invest time and money in new technologies..
I have no desire to buy 4 x 1200va transformers, a bank of very fast diodes and a world of caps for an amp not much bigger than a matchbox, ok 4 channels = 4 matchboxes. I actually think the pursuit of such a linear supply is ridiculous when there are other ways to even get close. also remember that we are dealing with a pretty heavily biased AB amp here that can actually be biased into complete 100% CLASS A. i do not think that many here will go so far against the design spec to produce something that could be called class B
last a lifetime? hmm under these pressures you are talking of i doubt your caps will and how many of us will still be using the same amp, as is; in even 10-15yrs? i dont want my amp to die in a couple of years, but i look at an SMPS like this as something a little akin to an SSD drive, unheard of performance in a small package rather efficiently and the highest quality ones will last quite a long time, but the lifetime IS limited by the technology. i accept that and if the amp finds a place in my system for longer than that i figure i'll treat it to something new; i'll probably be able to power it with a higher quality supply 1/4 the size and more efficiently by waves transmitted through the air by then.
Last edited:
Qusp, no world of caps; please no more than two good quality caps.
As long as the amp is not pure class A there is signal dependent current draw on the power supply, and the differences in power supply designs can become apparent.
Remember we are talking subtile differences here, not "bad" and "good".
The impact of the power supply is big; in an amp we are mere modulating the power supply aren't we?
it is obvious that more complexity, may be at risk. I try to do things properly and do not omit anything to chance, the experience certainly helps me.
I believe that the person's character is mixed when designing.
One curiosity: this SMPS series, have a sub protection. in case of breakage of the regulator, the SMPS switch-off permanently. one year but no one returns. or they came back and said I did not have.
as in class D, the protection circuits are very important, and some require an effort equal, if not more, of the amplifier stage. this is my way of thinking and see.
To Qusp:
eheheh..I have some measure for you (get this as optional,just i not know if OPC have VHF spectrum analizer), yust you have asked..Mhz show.
I believe that the person's character is mixed when designing.
One curiosity: this SMPS series, have a sub protection. in case of breakage of the regulator, the SMPS switch-off permanently. one year but no one returns. or they came back and said I did not have.
as in class D, the protection circuits are very important, and some require an effort equal, if not more, of the amplifier stage. this is my way of thinking and see.
To Qusp:
eheheh..I have some measure for you (get this as optional,just i not know if OPC have VHF spectrum analizer), yust you have asked..Mhz show.
Yes, a life time (not for caps, but for transformers yes).
I just sold my Yamaha B2 amp (with pain in my heart) because someone begged me; that's an over 30 years old and excellent sounding design, and it has regulated +/- 80V supplies for the driver stage (hey not quite unlike the Wire heh??, and look at other threads where the same technology is enjoying a new life again - power JFets....). What I mean to say: not so much happened in amplifier design justifying replacing an amp so often.
measured yesterday, but ... I forgot that the power unit was opened, after the changes I made, I remember that they should be -3dB lower in the first pic.
Again, this may be an essential complement to definitive those who wait by the OPC.
10 MHz band or above not have sense, as already seen, is completely flat clean.
Again, this may be an essential complement to definitive those who wait by the OPC.
10 MHz band or above not have sense, as already seen, is completely flat clean.
Attachments
Last edited:
yes, but a psu like you describe for me for eventually 6 channels 6 x 1.2kW transformers and 12 (or 24 if you mean per phase as this is a balanced amp) is HUUUGE and $$$ and caps i would trust just one with here are not cheap either, neither the diodes. I already have a box of PEH169 i can use, i trust they are acceptable.
Your scenario is as i said, crazy in an amp many like me are building with these to match digital XO's. several times the cost of the amp section and maybe 10 x the size of it...... sorry but that would be insane and for what? what advantage do you claim? only based on disagreeing with an opinion opc stated? and you propose an alternate opinion that trumps it yes?
the distances we are talking about creating here with the size of this thing and the radiated fields will make the speed of the main reservoir caps and TX a bit redundant for transient response dont you think? you seem to be striving for an ideal that is if not unattainable, not worth the lengths. yes of course the power and signal are one and the same, but we do actually have pretty decent PSRR here.
I expect running a psu like you describe to be literally illegal in many countries within the sort of period you are talking about.
yep all true
x 2
hey pieter t: ive seen you derail many threads with these sort of endless predictable and rather circular arguments when you arent actually building the amp or even interested, do you think you could just leave it out this time?
we are discussing this amp and the realistic power supply options for it among us the builders, there may be a few building supplies as you describe, but i bet many of them are pretty interested in the possibility of something much cheaper and much smaller that matches the performance and they like me, would probably like to just get on with discussing these alternatives. i have what i need for a linear supply and ive been down that road.
Last edited:
Qusp,
Sorry if I brutally entered your 5.1 world; I am living in a 2.0 environment (stereo so to say...) so my remarks are based merely upon left and right....
I question the craziness of my scenario; my 1200VA power supply transformer would come (potted) in a 6 inches square can; rectification and two 10000 uF quality capacitors would not add so much to the required space (it's for a + and - supply).
Besides I never said that the linear supply would dissipate 1200 watt; I just indicated to need a high power core for a low inner resistance power supply. Power supply dissipation would be determined by the load, and the load does not distinguish between the power supply being SM or linear.
But hey that's what I would do; feel free to do and think different!
I posted because I (still) think that a well designed linear supply will not perform worse than a well designed SMPS, and I understand that AP2 basically agrees.
In a three-way active set up you could perfectly use a single high quality supply, with additional small capacitance power supply decoupling for each amplifier board.
Your last remark is actually rather offensive.
First of all I still consider building the amp (I was on the GB list) as I need a power amp for my open baffle woofers.
It might be a LME49830 based one, or maybe one of Roberto's class d modules, and I even might consider an SMPS. So I not so much feel the need "to leave it out this time".
You perfectly seem to know the price difference between a quality SMPS and a quality linear supply, so please make up your own mind.
Sorry if I brutally entered your 5.1 world; I am living in a 2.0 environment (stereo so to say...) so my remarks are based merely upon left and right....
I question the craziness of my scenario; my 1200VA power supply transformer would come (potted) in a 6 inches square can; rectification and two 10000 uF quality capacitors would not add so much to the required space (it's for a + and - supply).
Besides I never said that the linear supply would dissipate 1200 watt; I just indicated to need a high power core for a low inner resistance power supply. Power supply dissipation would be determined by the load, and the load does not distinguish between the power supply being SM or linear.
But hey that's what I would do; feel free to do and think different!
I posted because I (still) think that a well designed linear supply will not perform worse than a well designed SMPS, and I understand that AP2 basically agrees.
In a three-way active set up you could perfectly use a single high quality supply, with additional small capacitance power supply decoupling for each amplifier board.
Your last remark is actually rather offensive.
First of all I still consider building the amp (I was on the GB list) as I need a power amp for my open baffle woofers.
It might be a LME49830 based one, or maybe one of Roberto's class d modules, and I even might consider an SMPS. So I not so much feel the need "to leave it out this time".
You perfectly seem to know the price difference between a quality SMPS and a quality linear supply, so please make up your own mind.
Just as a comparison:
FetZilla running at 470mAdc of bias through the 1pr output stage.
600VA transformer.
Mains current at 240Vac is 138mAac. That's a total consumption for the amplifier and PSU and transformer of 33W.
If I had a 1200VA transformer I would not expect the power consumption to rise as high as 45W. It might even be <40W
FetZilla running at 470mAdc of bias through the 1pr output stage.
600VA transformer.
Mains current at 240Vac is 138mAac. That's a total consumption for the amplifier and PSU and transformer of 33W.
If I had a 1200VA transformer I would not expect the power consumption to rise as high as 45W. It might even be <40W
Hi Guys,
Everyone needs to simmer a little I think... It's just a discussion on power supplies, and you'd think we were debating religion or something!
pieter t:
I should have qualified my statements a little more I suppose. I was referring specifically to the dynamic performance of the DPS600, and I was making the statement that you cannot achieve that level of performance with an un-regulated linear supply.
You're thinking about things the old fashion way when you refer to "Inner resistance". That means nothing when you have a regulated system with a feedback loop. That loop eliminates all the resistances with the exception of the copper between the feedback point and the load.
In your example, you have 0.8 ohms on the primary, 0.1 ohms on each secondary for a total resistance of 1 ohm in the transformer. Next step is the diodes which have a variable voltage drop with respect to current, so the "resistance" of the diode increases as current increases. This could add between a 0.4V drop or as much as a 1.2V drop depending on the current. Next up is all the wiring to get from the transformer to the diodes and then to the caps. I would be willing to bet that it would be very difficult to build a linear supply with a true output impedance of less than 2 ohms. I would also imagine the magnetization current on a toroid that size would be significant (probably 30-40 watts) and the magnetic noise generated by a transformer that size is also very significant. In addition to all that, you have higher levels of 120Hz ripple on the output of the supply, since you have no feedback loop to get rid of it.
In the case of a regulated SMPS, the output impedance could be down in the milliohm range if you design the feedback loop properly. You just can't touch that with an un-regulated linear supply.
With the un-regulated linear supply, you have no way of compensating for changes in current draw. There is no closed loop, so you have to rely on brute force to try and counter the effects of dynamic changes in the load on the supply.
Now none of this applies if you start getting into regulated linear supplies. In that case, as long as you have enough pre-reg voltage to stay above the drop out of the reg, then you'd probably get as good or better performance with the linear approach. This does have the drawback of horrendous efficiency however, and the linear regs would need heatsinks nearly the size of the amp itself!
With an SPMS, the regulation doesn't cause any loss of efficiency. You can still get a supply that is 90% efficient with a fully regulated output.
See where I'm going with this?
Cheers,
Owen
Everyone needs to simmer a little I think... It's just a discussion on power supplies, and you'd think we were debating religion or something!
pieter t:
I should have qualified my statements a little more I suppose. I was referring specifically to the dynamic performance of the DPS600, and I was making the statement that you cannot achieve that level of performance with an un-regulated linear supply.
You're thinking about things the old fashion way when you refer to "Inner resistance". That means nothing when you have a regulated system with a feedback loop. That loop eliminates all the resistances with the exception of the copper between the feedback point and the load.
In your example, you have 0.8 ohms on the primary, 0.1 ohms on each secondary for a total resistance of 1 ohm in the transformer. Next step is the diodes which have a variable voltage drop with respect to current, so the "resistance" of the diode increases as current increases. This could add between a 0.4V drop or as much as a 1.2V drop depending on the current. Next up is all the wiring to get from the transformer to the diodes and then to the caps. I would be willing to bet that it would be very difficult to build a linear supply with a true output impedance of less than 2 ohms. I would also imagine the magnetization current on a toroid that size would be significant (probably 30-40 watts) and the magnetic noise generated by a transformer that size is also very significant. In addition to all that, you have higher levels of 120Hz ripple on the output of the supply, since you have no feedback loop to get rid of it.
In the case of a regulated SMPS, the output impedance could be down in the milliohm range if you design the feedback loop properly. You just can't touch that with an un-regulated linear supply.
With the un-regulated linear supply, you have no way of compensating for changes in current draw. There is no closed loop, so you have to rely on brute force to try and counter the effects of dynamic changes in the load on the supply.
Now none of this applies if you start getting into regulated linear supplies. In that case, as long as you have enough pre-reg voltage to stay above the drop out of the reg, then you'd probably get as good or better performance with the linear approach. This does have the drawback of horrendous efficiency however, and the linear regs would need heatsinks nearly the size of the amp itself!
With an SPMS, the regulation doesn't cause any loss of efficiency. You can still get a supply that is 90% efficient with a fully regulated output.
See where I'm going with this?
Cheers,
Owen
I thought you found that even your regulated linear supply was bettered in dynamic performance? (though it was possibly quieter?) given that was the other option i was considering as ideal, this comparison was of particular interest. i would have had to limit its size and the output of the amp to a realistic max to deal with the size and expense, something i wouldnt have to do with the DPS600. which would mean the system can scale with my needs.
Most linear unregulated supplies tend to just throw a heap of power away in the name of lowering ripple with resistors, coils etc. so coupled with the massive overkill needed for the transformer that allows for the dynamic overhead, it is a part of poweramp building that doesnt fill me with a warm glow. bare in mind i dont drive a big car, or run a stupid size television either, so saying things like oh but its nothing compared to the massive amount of power i'm wasting with X dont really apply to me
i'm done bickering though, i just wanted to make my point and also mention that i did not want to see this turned into another N-Core thread.......... it seems any high performance amplifier that utilizes new technology brings these old arguments up again. i dont quite understand stubbornness myself and i wouldnt have thought this thread was the place for it, being that it uses new techniques and new parts to create great performance.
myself, if i can do the same or more with less everything, then i'm a happy diyer....
Most linear unregulated supplies tend to just throw a heap of power away in the name of lowering ripple with resistors, coils etc. so coupled with the massive overkill needed for the transformer that allows for the dynamic overhead, it is a part of poweramp building that doesnt fill me with a warm glow. bare in mind i dont drive a big car, or run a stupid size television either, so saying things like oh but its nothing compared to the massive amount of power i'm wasting with X dont really apply to me
i'm done bickering though, i just wanted to make my point and also mention that i did not want to see this turned into another N-Core thread.......... it seems any high performance amplifier that utilizes new technology brings these old arguments up again. i dont quite understand stubbornness myself and i wouldnt have thought this thread was the place for it, being that it uses new techniques and new parts to create great performance.
myself, if i can do the same or more with less everything, then i'm a happy diyer....
I thought this was a reasonably valid question, got lost in the commotion a few pages back.
My understanding is that it is *possible* to run 2 amps from a single DPS-400 but not from from the DPS-600.
I'd be interested to hear from AP2/opc on viability of this, sure its less than ideal, but how much less?