Hi!
I am thinking about using a SMPS for the B+ in a poweramp (for output tubes). The SMPS from here
http://www.coldamp.com/opencms/opencms/coldamp/en/productos/SPS80/index.html?idioma=en
seems to have an center tap which is isolated from the main ground. Joining the -ve rail to the circuit ground and the +ve rail will give 120V supply. However, will the switching noise gets into the circuit and affect the 1st gain stage?
How about using different wall sockets. One for the first gain stage and a different wall socket for the output stage and isolating the 2 stages electrically. Woudl this solve the problem?
I am thinking about using a SMPS for the B+ in a poweramp (for output tubes). The SMPS from here
http://www.coldamp.com/opencms/opencms/coldamp/en/productos/SPS80/index.html?idioma=en
seems to have an center tap which is isolated from the main ground. Joining the -ve rail to the circuit ground and the +ve rail will give 120V supply. However, will the switching noise gets into the circuit and affect the 1st gain stage?
How about using different wall sockets. One for the first gain stage and a different wall socket for the output stage and isolating the 2 stages electrically. Woudl this solve the problem?
If you purchase a supply and ask for it, we can supply a couple of chokes for free that we usually add in series with the supply lines when operating the SPS80 with our Class-D modules to avoid noise to interfere with them and produce "beating" or "whistles".
They help a lot in cleaning up the rails. However, you will almost for sure find them unnecessary in a conventional amplifier.
They help a lot in cleaning up the rails. However, you will almost for sure find them unnecessary in a conventional amplifier.
ackcheng said:
This is asking for trouble especially when earth circuits are joined up, a circulating earth current could create problems in supporting equipment. As I see it, if you want to use 2 smps within an amp (one for power stage and other for front end) with a common chassis, that means 2 xy input filters sharing a centre ground. This is against EMI rules because in the case of a physical earthing fault within the chassis one could see double leakage Y capacitor current. Switchmode power supplies running at different frequencies can create a host of problems esp towards common and differential mode earthing. Generally I find the noise levels not low enough.....i.e much better filtering is required for HiFi applications.
richj
More smps = more interference.
On safety grounds
Well, obviously, the more SMPS, the more EMi, but in the case of SPS80 it is so low that we have been running multiple SMPS in the same chassis with no problems at all. The audible noise level was such that you couldn't hear more noise than a _very small_ hiss inherent to the amplifiers. And that was glueing your ear to the tweeters! 
The amplifier was actually a 19" 1U rack, formed by 3 SPS80 + 6 BP4078 class D modules, all running free. It is actually working at a recording studio, I think, and no noise problemas have appeared, and there's lot of electronic stuff in that environment.
You must be careful with the cabling but... nothing more.
I think that with valve amplifiers it will be even less prone to problems, as they don't switch as Class-D amps do.
By the way, SPS80 PSUs have onboard mains filter.
Hope this helps
The amplifier was actually a 19" 1U rack, formed by 3 SPS80 + 6 BP4078 class D modules, all running free. It is actually working at a recording studio, I think, and no noise problemas have appeared, and there's lot of electronic stuff in that environment.
You must be careful with the cabling but... nothing more.
I think that with valve amplifiers it will be even less prone to problems, as they don't switch as Class-D amps do.
By the way, SPS80 PSUs have onboard mains filter.
Hope this helps
ssanmor said:
Ooop's understatement ! Some tube amp designs do have s/n ratios down in the -80 to-90db/w ballpark where one would see fine grass electron noise. That's a magnitude lower than current EMI levels for smps between 1-10MHz... however I've lost track and bit out of date of current FCC/ECC legislation levels.
Traditionally, earthing the circuit wiring of tube amps was made only at one place i.e input socket directly connected to chassis, and no other place. Put an SMPS onboard and the earthing circuit for the tube amp is fine but for the smps the tube amp earth is often too high impedance. I had this problem and the only way is to thicken up all earth wiring and even braid the power end return to the chassis. It isn't hit and miss and with stereo circuits things can get quite difficult, it's a question of providing the lowest impedance return for charge injected smps noise and not upsetting amp channel crosstalk rejection at 15KHz. Experience is often the only way and bad cabling will show this up.
This is probably the reason why smps get bad press on tube amps.....lack of EMI understanding on how to redress problems and conflicts....An RF engineer should swallow this with ease... .
Go carefully with the live sides....
richj
No, ackcheng, what I meant is that the SPS80 didn't add audible noise to the amplifier. Any amplifier has a background hiss, that in the case of BP4078 is only audible when all is absolutely silent and you put your ear next to the tweeter. (in the order of 100dB s/n)
Of course the switching frequency is well above the audible band, but sometimes with Class-D amplifiers, the amp and supply freqs. "beat" and move to the audible band. This is not the case with SPS80 and Bp4078 (and according to our customers, not with UcD180/400 or zappulse either).
Rich,
I didn't want to understate the s/n ratio of tube amps, what I was saying is that in theory you are in worst conditions using SMPS with switching amplifiers, for the reason stated above: beating frequencies that may produce "whistles" or colored hiss.
In tube amps you don't have these effects. This doesn't mean, as you pointed out, that you must be exquisite with the wiring (as you have to with other kinds of amplifiers).
In any case, a very interesting mix: I am anxious to hear what our tube customers report about using our SMPS with their amps!
I must confess to be a tube ignorant, by the way.
Sergio
Of course the switching frequency is well above the audible band, but sometimes with Class-D amplifiers, the amp and supply freqs. "beat" and move to the audible band. This is not the case with SPS80 and Bp4078 (and according to our customers, not with UcD180/400 or zappulse either).
Rich,
I didn't want to understate the s/n ratio of tube amps, what I was saying is that in theory you are in worst conditions using SMPS with switching amplifiers, for the reason stated above: beating frequencies that may produce "whistles" or colored hiss.
In tube amps you don't have these effects. This doesn't mean, as you pointed out, that you must be exquisite with the wiring (as you have to with other kinds of amplifiers).
In any case, a very interesting mix: I am anxious to hear what our tube customers report about using our SMPS with their amps!
I must confess to be a tube ignorant, by the way.
Sergio
ssanmor said:
-->Tube ignorant ..... Doesn't matter......However, one can get birdies, whistles with tube amps. More than one suspects. The common high slope ECC88 RF tube is a good example. I use it in my preamp output. Badly wired and sloppy decoupled, it will peacefully oscillate at low levels around 100-400Mc/s and this will intermodulate with any audio signal passing through and one wouldn't know about it. The cathode follower is a particular configuration that can cause this. The remedy is simple.....use a screening can and better layout.
Noticing your comment.......in days gone long past it was common design practice in tube tuner mixer oscillators simply to put stub of wire from the oscillator close to the mixer tube grid pin in question.....and mixing happens.
The statement--> in tube amps doesn't has this effect.....quite wrong. If you look at a standard tube amp with global neg feedback, notice one of the output transformer sec windings returns via resistor and cap (or step filter) to the input tube cathode. The speaker wire inductance/ impedance is high relative to RF and can pass strong AM signals back into the first stage. Demodulätion can take place and one hears the radio in your speaker. This is the so-called problem of CB radio "busting" through stereo systems.
richj
One thing to keep in mind...
SMPS have gotten a bad rap..and for good reason....
Most of the textbooks on switchers are way outdated with 1980 technology...as well as many SMPS engineers are outdated and clueless as well.....
It is very possible to design high efficiency, low noise SMPS...
Especially making them low in EMI .....
The use of a properly stabilized Quasi-Resonant switcher at or above 500kHz will work out really well for audio applications..
Chris
SMPS have gotten a bad rap..and for good reason....
Most of the textbooks on switchers are way outdated with 1980 technology...as well as many SMPS engineers are outdated and clueless as well.....
It is very possible to design high efficiency, low noise SMPS...
Especially making them low in EMI .....
The use of a properly stabilized Quasi-Resonant switcher at or above 500kHz will work out really well for audio applications..
Chris
Better Way !
I had used +B SMPS for TUBE AMPLIFIER !
100KHz RT*CT, 550V for power tube, -150V for grid bias (variable), 25.2V for 2CH filaments (6.3+6.3+6.3+6.3), choke 10H, remote on function
Max power is 250W but power consumtion for 30Wrms is only 150W !
Next project is 100Wrms with PP EL509 included sgreen bias circuit
Regard
analog guy
I had used +B SMPS for TUBE AMPLIFIER !
100KHz RT*CT, 550V for power tube, -150V for grid bias (variable), 25.2V for 2CH filaments (6.3+6.3+6.3+6.3), choke 10H, remote on function
Max power is 250W but power consumtion for 30Wrms is only 150W !
Next project is 100Wrms with PP EL509 included sgreen bias circuit
Regard
analog guy
Attachments
Both Good
When I compared both Linear and SMPS I found same result about maximum power output ! but there're some different in audio term...
Linear : more deeper bass and airy : 6KGs : BIGGER
SMPS : focus and warm human voice : 0.6KGs : SMALL and SLIM
When I sold them, I'm very happy about weight shipping (20USD)
Uploaded picture is my inspiration !
Thanks
analog guy
When I compared both Linear and SMPS I found same result about maximum power output ! but there're some different in audio term...
Linear : more deeper bass and airy : 6KGs : BIGGER
SMPS : focus and warm human voice : 0.6KGs : SMALL and SLIM
When I sold them, I'm very happy about weight shipping (20USD)
Uploaded picture is my inspiration !
Thanks
analog guy
Attachments
cerrem said:
Facit-> Chris, In the mid 90's I did alot of work with the UC3895 and other resonant ic's, but it had a serious bug in it which was too risky for high power applications. This ic had ramp troubles when light load or no load conditions prevailed.
You will be glad to know I use a kW hard switching PFC in my 500W tube amp which creates no interference.
QUOTE ...> As well as many SMPS engineers are outdated and clueless as well.....true true...(I hope you aren't implying me ??.....) However, there isn't such a thing as outdating the laws of physics. the use of is another matter.
The concept of resonant switching has been around for decades but expense has often ruled it out of a particular design. A SMPS design engineer should have full RF design experience as well an allround in magnetics. Not many of us about anymore.
richj
Well, as I said I don't have experience with tube amps, but with Class-D and linear solid state amplifiers, our experience is the opposite: sound is at least as clear and bass seems deeper and more controlled with SMPS (I agree in the size/weight considerations, of course)
Let's see what our customers find when they test this curious combination.
gmarinelli said:
We need to know what the power requirement of a particular amp. Flyback topology is dirt simple at low cost, but there becomes a line in the sand when other configurations are more attractive.
For example I have a tube preamp that requires 300V B+ at 50mA so I'm using 50KHz flyback with a HV doubler. The piecemeal advantage with flyback is excellent cross regulation with other windings, so heater 6.3V+ 0.6V (keeping to whole number of turns) can be taken as the design integer (v.s primary) for calculating the HV number of turns.
The rotton business with the forward is a choke required for each secondary unless one uses the coupled inductor approach.
richj
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.